List of contributors

D.G. Amaral, Department of Psychiatry and Behavioral Sciences, The M.I.N.D. Institute and the
California National Primate Research Center, UC Davis, 2825 50th Street, Sacramento, CA 95817, USA
L. Acsády, Institute of Experimental Medicine, Hungarian Academy of Sciences, PO Box 67, 1450
Budapest, Hungary
C.A. Barnes, Arizona Research Laboratories Division of Neural Systems, Memory & Aging, University of
Arizona, Tucson, AZ and Evelyn F. McKnight Brain Institute, University of Arizona, Tucson, AZ and
Departments of Psychology and Neurology, University of Arizona, Tucson, AZ, USA
I. Bechmann, Institute of Clinical Neuroanatomy, J.W. Goethe-University, Theodor-Stern-Kai 7, D-60590
Frankfurt/Main, Germany
D.K. Binder, Department of Neurological Surgery, University of California, Irvine, CA 92868-3298, USA
M. Blaabjerg, Anatomy and Neurobiology, Institute of Medical Biology, University of Southern Denmark,
Winslowparken 21, DK-5000 Odense C, Denmark
C.R. Bramham, Department of Biomedicine and Mental Health Research Center, University of Bergen,
Jonas Lies vei 91, N-5009 Bergen, Norway
G. Buzsáki, Center for Molecular and Behavioral Neuroscience, Rutgers, The State University of
New Jersey, 197 University Avenue, University Heights, Newark, NJ 07102, USA
G.C. Carlson, The Children’s Hospital of Philadelphia, Abramson Pediatric Research Center, Room 410D,
3516 Civic Center Boulevard, Philadelphia, PA 19104-4318, USA
C. Chavkin, Department of Pharmacology, University of Washington, Seattle, WA 98195, USA
M.K. Chawla, Arizona Research Laboratories Division of Neural Systems, Memory & Aging, University
of Arizona, Tucson, AZ and Evelyn F. McKnight Brain Institute, University of Arizona, Tucson, AZ,
USA
B.J. Claiborne, Department of Biology, University of Texas at San Antonio, One UTSA Circle,
San Antonio, TX 78249, USA
W.F. Colmers, Department of Pharmacology, University of Alberta, 9-36 Medical Sciences Building,
Edmonton, AL T6G 2H7, Canada
D.A. Coulter, The Children’s Hospital of Philadelphia, Abramson Pediatric Research Center, Room 410D,
3516 Civic Center Boulevard, Philadelphia, PA 19104-4318, USA
D. Del Turco, Institute of Clinical Neuroanatomy, J.W. Goethe-University, Theodor-Stern-Kai 7, D-60590
Frankfurt/Main, Germany
T. Deller, Institute of Clinical Neuroanatomy, J.W. Goethe-University, Theodor-Stern-Kai 7, D-60590
Frankfurt/Main, Germany
B.E. Derrick, Department of Biology, The Cajal Neuroscience Research Institute, The University of Texas
at San Antonio, 6900 N. Loop 1604 West, San Antonio, TX 78249-0662, USA
L. DeToledo-Morrell, Department of Neurological Sciences, Rush University Medical Center, Chicago,
IL 60612, USA
C.T. Drake, Division of Neurobiology, Department of Neurology and Neuroscience, Weill-Cornell
Medical College, 411 East 69th Street, New York, NY 10021, USA

v
vi

M.R. Drew, Department of Neuroscience and Psychiatry, Division of Integrative Neuroscience, Columbia
University, New York, NY 10032, USA
F.E. Dudek, Department of Physiology, University of Utah, Salt Lake City, UT, USA
E. Förster, Institute of Anatomy and Cell Biology, University of Freiburg, Albertstr. 17, D-79104 Freiburg,
Germany
C.J. Frazier, Department of Pharmacodynamics and Department of Neuroscience, University of Florida,
College of Medicine, JHMHC 100487, 1600 S.W. Arder Road, Gainesville, FL 32610, USA
M. Frotscher, Institute of Anatomy and Cell Biology, University of Freiburg, Albertstr. 17, D-79104
Freiburg, Germany
R. Gutiérrez, Department of Physiology, Biophysics and Neurosciences, Center for Research and
Advanced Studies, Mexico City Apartado Postal 14-740, Mexico D.F. 07000, Mexico
T. Hajszan, Department of Obstetrics, Gynecology, and Reproductive Sciences, Yale University School
of Medicine, 333 Cedar Street, FMB 313, New Haven, CT 06520, USA
T. Hamilton, Department of Pharmacology, University of Alberta, 9-36 Medical Sciences Building,
Edmonton, AL T6G 2H7, Canada
C.W. Harley, Department of Psychology, Memorial University of Newfoundland, St. John’s, NL
A1B 3X9, Canada
R. Hen, Departments of Neuroscience, Psychiatry, and Pharmocology, Division of Integrative
Neuroscience, Columbia University, New York, NY 10032, USA
D.A. Henze, Neuroscience Drug Discovery Merck Research Laboratories, 770 Sunney town Pike, P.O.B. 4
WP 26A-2000, West Point, PA 19486, USA
C.R. Houser, Department of Neurobiology, David Geffen School of Medicine at UCLA, 73-235 CHS,
10833 Le-Conte Avenue, Los Angeles, CA 90095, USA
D. Hsu, Department of Neurology, University of Wisconsin, 600 Highland Avenue, H6/526, Madison,
WI 53792, USA
D.B. Jaffe, Department of Biology, University of Texas at San Antonio, One UTSA Circle, San Antonio,
TX 78249, USA
M. Joëls, Swammerdam Institute of Life Sciences, Center for NeuroScience (SILS-CNS), University of
Amsterdam, Kruislaan 320, 1098 SM Amsterdam, The Netherlands
S. Káli, Institute of Experimental Medicine, Hungarian Academy of Sciences, PO Box 67, 1450 Budapest,
Hungary
R.P. Kesner, Department of Psychology, University of Utah, 380 S. 1530 E., Room 502, Salt Lake City,
UT 84121, USA
P. Lavenex, Department of Medicine, Unit of Physiology, University of Fribourg, 1700 Fribourg,
Switzerland
C. Leranth, Department of Obstetrics, Gynecology, and Reproductive Sciences, Yale University, School
of Medicine, 333 Cedar Street, FMB 312, New Haven, CT 06520, USA
G. Li, Department of Neurology, Programs in Neuroscience, Developmental Biology, University of
California, San Francisco, 1550 4th Street, Rock Hall Room 448C, San Francisco, CA 94158, USA
J.E. Lisman, Department of Biology and Volen Center for Complex Systems, Brandeis University,
Waltham, MA 02454, USA
T.A. Milner, Department of Neurology and Neuroscience, Division of Neurobiology, Weill-Cornell
Medical College, 411 East 69th Street, New York, NY 10021, USA
R.J. Morgan, Department of Anatomy and Neurobiology, 193 Irvine Hall, University of California, Irvine,
CA 92697, USA
J.J. O’Connor, UCD School of Biomolecular and Biomedical Science, Conway Institute of Biomolecular
and Biomedical Research, University College Dublin, Belfield, Dublin 4, Ireland
 vii

T.G. Ohm, Institute of Integrative Neuroanatomy, Department of Clinical Cell and Neurobiology, Charité
CCM, 10098 Berlin, Germany
J.M. Parent, Department of Neurology, University of Michigan Medical Center, 109 Zina Pitcher Place,
5021 BSRB, Ann Arbor, MI 48109-2200, USA
P.R. Patrylo, Department of Physiology, Southern Illinois University School of Medicine, Carbondale,
IL 62901, USA
M. Pickering, UCD School of Biomolecular and Biomedical Science, Conway Institute of Biomolecular
and Biomedical Research, University College Dublin, Belfield, Dublin 4, Ireland
S.J. Pleasure, Department of Neurology, Programs in Neuroscience, Developmental Biology, University of
California, San Francisco, 1550 4th Street, Rock Hall Room 448C, San Francisco, CA 94158, USA
B. Pöschel, Department of Cell Biology and Anatomy, New York Medical College, Valhalla, NY 10595,
USA
O. Rahimi, Department of Cellular and Structural Biology, University of Texas Health Science Center at
San Antonio, San Antonio, TX 78229, USA
A. Rappert, Institute of Clinical Neuroanatomy, J.W. Goethe-University, Theodor-Stern-Kai 7, D-60590
Frankfurt/Main, Germany
C.E. Ribak, Department of Anatomy and Neurobiology, University of California at Irvine, Irvine,
CA 92697-1275, USA
A Sahay, Departments of Neuroscience and Psychiatry, Division of Integrative Neuroscience, Columbia
University, New York, NY 10032, USA
V. Santhakumar, Department of Neurology, David Geffen School of Medicine, University of California,
Los Angeles, CA, USA
H.E. Scharfman, Departments of Pharmacology and Neurology, Columbia University, College of
Physicians and Surgeons, New York, NY, USA and Center for Neural Recovery and Rehabilitation
Research, Helen Hayes Hospital, New York State Department of Health, Rte 9W, West Haverstraw,
NY 10993-1195, USA Present address: Nathan Kline Institute for Psychiatric Research, 140 Old
Orangeburg Rd., Bldg. 35, Orangeburg, NY 10962, USA
L. Seress, Central Electron Microscopic Laboratory, Faculty of Medicine, University of Pécs, Szigeti str.
12, 7624 Pécs, Hungary
L.A. Shapiro, Department of Anatomy and Neurobiology, University of California at Irvine, Irvine,
CA 92697-1275, USA
I. Soltesz, Department of Anatomy and Neurobiology, 193 Irvine Hall, University of California, Irvine,
CA 92697, USA
G. Sperk, Department of Pharmacology, Medical University Innsbruck, Peter-Mayr-Str. 1a, 6020
Innsbruck, Austria
P.K. Stanton, Department of Cell Biology and Anatomy, New York Medical College, Valhalla, NY 10595,
USA
T.R. Stoub, Department of Neurological Sciences, Rush University Medical Center, Chicago, IL 60612,
USA
T.P. Sutula, Department of Neurology H6/570 CSC, University of Wisconsin, 600 Highland Avenue,
Madison, WI 53792, USA
M.K. Tallent, Department of Pharmacology and Physiology, Drexel University College of Medicine,
245 N. 15 Street, Philadelphia, PA 19102, USA
C. Wang, Department of Neurological Sciences, Rush University Medical Center, Chicago, IL 60612,
USA
A. Williamson, Department of Neurosurgery, Yale University School of Medicine, New Haven, CT 06518,
USA
viii

M.P. Witter, Department of Anatomy and Neurosciences, VUMC, MF-G102C, P.O.Box 7057, 1007 MB
Amsterdam, The Netherlands
S. Zhao, Institute of Anatomy and Cell Biology, University of Freiburg, Albertstr. 17, D-79104 Freiburg,
Germany
J. Zimmer, Anatomy and Neurobiology, Institute of Medical Biology, University of Southern Denmark,
Winslowparken 21, DK-5000 Odense C, Denmark
 Preface

Part I. Why is a guide to the dentate gyrus necessary?

As more knowledge is acquired about the nervous system, one sometimes is lost in a sea of literature. The
current generation surveys the literature primarily from PubMed, which provide a vast resource for data,
but often there is little context for it, and even less of a guide to seminal but pre-1990 papers. Presumably
this context, and a historical overview, can be gained by courses, lectures, and from reviews, but this is often
difficult. One example of this problem is our understanding of the dentate gyrus. This part of the brain has
always been a topic of intense interest, and pioneering studies from the distant and recent past attest to it.
But where in a textbook or review can one find a comprehensive overview some basic information about the
dentate gyrus? The primary goal of the book is to address this problem by bringing together in an organized
manner a series of chapters that address, in an approachable format, the fundamental concepts about
dentate gyrus structure, function, and their implications. Some have more detail than others, some provide
distinct perspectives on a similar topic, but to some extent this is welcome because each reader may desire
different degrees of detail, and some may want to compare viewpoints of different experts in the field.
Often a region of the brain is best understood by first outlining its components, so the first section of the
book deals with the fundamental structure of the dentate gyrus. In the first chapter, David Amaral, Helen
Scharfman, and Pierre Lavenex describe the neuronal organization and intrinsic circuitry. László Seress
discusses this further by comparing the dentate gyrus of different species, emphasizing those that are
primarily used in research (mouse, rat, monkey). Menno Witter outlines the topography of arguably the
most important afferent system, the input from entorhinal cortex (the perforant path). Csaba Leranth and
Tibor Hajszan discuss other extrinsic afferents, such as those from septum, mammillary bodies, and other
areas that are poorly understood compared to the perforant path, particularly from the physiological
perspective. Morten Blaabjerg and Jens Zimmer discuss in detail one of the most impressive and complex
aspects of the dentate gyrus, the mossy fibers axons of the dentate gyrus granule cells. The remarkable
structural complexity of the mossy fiber system, including the unique specializations that characterize their
terminal arbors, is like no other in the hippocampus. Following this overview is a complementary
discussion of the expression, plasticity, and physiology of the mossy fiber system, by David Jaffe and Rafael
Gutierrez. This section then ends with two overviews about the development of the dentate gyrus, the first
by Michael Frotscher, Shanting Zhao, and Eckart Förster, and the second by Guangnan Li and Sam
Pleasure. The emphasis of both is on the unique set of signals that orchestrate the development of the
normal laminar organization and cytoarchitecture of the adult dentate gyrus.
The second section of the book addresses the major neuronal cell types in the dentate gyrus, mostly in the
rodent. Charles Ribak and Lee Shapiro begin this section with an overview of the structural characteristics
of the neuronal cell types, emphasizing their unique ultrastructural ‘‘signatures.’’ Omid Rahimi and Brenda
Claiborne discuss the granule cell from a developmental and morphological perspective. Anne Williamson
and Peter Patrylo review characteristics of granule cells from an electrophysiological viewpoint, both in
rodent and in human tissue studies. The non-granule cells are then addressed in two chapters that cover the
two primary subtypes of non-granule cells: the hilar mossy cells and the GABAergic interneurons. Darrell

ix
x

Henze and Gyuri Buzsáki provide an overview about the mossy cells. Carolyn Houser discusses the
GABAergic neurons, which are heterogeneous in anatomy, expression, and functional organization.
Section three attempts to cover the numerous transmitters and neuromodulators that influence the
dentate gyrus. Glutamatergic inputs are discussed in other parts of the book (the perforant path in Section
1, extrinsic inputs in Section 1, the CA3 input in Section 5), so this leaves the first chapter to consider
GABA (Doug Coulter and Greg Carlson), followed by a series of reviews about non-classical
neurotransmitters. Carrie Drake, Charlie Chavkin, and Teri Milner review the expression and organization
of the opioid system, a complex topic because of the numerous types of opioid peptides, receptors, and
actions. Melanie Tallent addresses the neuropeptide somatostatin, which is normally co-localized with
GABA in a subset of interneurons. Gunther + Sperk, Trevor Hamilton, and William Colmers review
neuropeptide Y and its receptors, functions, and plasticity. Carolyn Harley provides an overview of
norepinephrine, a neuromodulator that has been known to have a robust influence in the dentate gyrus for
decades. Jason Frazier addresses a topic that has been appreciated relatively recently, endocannabinoids in
the dentate gyrus. Mark Pickering and John O’Connor discuss the pro-inflammatory cytokines,
neuromodulators that have not been widely acknowledged until recent years. Marian Joëls reviews the
glucocorticoids, and Devin Binder discusses the neurotrophins, a family of ‘‘growth factors’’ that are
expressed in high concentration in the dentate gyrus, and modulate function in a profound manner. The
reproductive steroids, estrogen, progesterone, and androgen, are addressed by Tibor Hajszan, Teri Milner,
and Csaba Leranth.
One of the more remarkable characteristics of the dentate gyrus is its plasticity, and this is the focus of
the 4th part of the volume. Brian Derrick begins with a perspective that ties together studies of in vitro and
in vivo synaptic plasticity, circuitry, and the behaving animal. Clive Bramham reviews long-term
potentiation (LTP) and its candidate mechanisms. Beatrice Pöschel and Patric Stanton address long-term
depression (LTD). Other examples of plasticity follow, such as lesion-induced plasticity. One of the most
widely studied examples involves the response to a lesion of the entorhinal cortex, and this large literature is
reviewed by Thomas Deller, Domenico Del Turco, Angelika Rappert, and Ingo Bechmann. They also
provide a modern context by updating the field with what is currently understood from studies of mice. On
a separate subject, Jack Parent provides a review of one of the most exciting developments in the history of
dentate gyrus research, and could not illustrate plasticity any better: that neurogenesis occurs in the dentate
gyrus throughout the lifespan of mammals. Finally, another remarkable type of plasticity, is discussed by
Tom Sutula and Ed Dudek: mossy fiber sprouting. This reorganization of the mossy fiber pathway has
captured the attention of many, because it involves dramatic anatomical and physiological changes, and
can occur in response to many types of insults or injury.
In the 5th part of the volume, network considerations are addressed. This is a rich and diverse area of
research that is reflected by many perspectives, each suggesting distinct roles of the dentate gyrus as a
structure. Ray Kesner begins, followed by László Acsády and Szabolcs Káli. David Hsu discusses the
concept that the dentate gyrus is a ‘‘gate’’ or ‘‘filter.’’ John Lisman turns to the question of the perforant
path input, and specifically how the medial vs. lateral perforant path provide distinct roles. Helen
Scharfman discusses the evidence that CA3 is a major input to the dentate gyrus, by virtue of axon
collaterals that innervate diverse dentate neurons. This section ends with a comprehensive model of the
dentate gyrus network, presented by Robert Morgan, Vijy Santhakumar, and Ivan Soltesz.
In the last section of the book, an effort is made to review some of the literature that addresses how
abnormalities within the dentate gyrus contribute to disease, or aging. Monica Chawla and Carol Barnes
discuss how granule cell function changes with age, drawing upon the wealth of information from
electrophysiology, imaging, and other approaches. Peter Patrylo and Anne Williamson also provide a
review of this large field, but with a different perspective. The potential role of dentate gyrus neurogenesis in
depression is reviewed by Amar Sahay, Michael Drew, and Rene Hen, whose lab was one of the first to
provide evidence that deficits in neurogenesis in the dentate gyrus might be a reason for depression. Thomas
 xi

Ohm discusses the diverse anatomical changes in the dentate gyrus that occur in Alzheimer’s disease. Leyla
DeToledo-Morrell, Travis Stoub, and Changsheng Wang review their studies of Alzheimer’s disease,
providing a persuasive argument that a major contributing factor is pathology that develops in the
entorhinal area. Finally, Ed Dudek and Tom Sutula review a large and active area of research:
epileptogenesis in the dentate gyrus.
Although a book such as this can never cover all the topics and all information, and certainly cannot do
justice to the multitude of contributors both from the past and present, its purpose is primarily as a guide,
and readers will be led to references that can address topics that are not covered. Particularly for those who
are not well versed in ‘‘hippocampology,’’ it is hoped that this volume can help – and in doing so, provide
new impetus to resolve the many outstanding questions that remain about the dentate gyrus.
Helen E. Scharfman
Columbia University and Helen Hayes Hospital
Part II. A guide to terminology and orientation
The dentate gyrus is a complex, three-dimensionally curved structure that is coupled to another similarly
curved structure, the hippocampus. As such, it is often difficult to understand its orientation in sections
taken from any one plane (e.g., coronal, horizontal, etc.). Moreover, some of the terms used to describe the
components of the dentate gyrus are confusing, and not completely defined in any textbook, to our
knowledge. Therefore, here we provide a general guide to the three-dimensional organization of the dentate
gyrus, and to terminology.
The dentate gyrus in rodents and in primates is a composed of a compact layer containing densely packed
granule cells. Their dendrites mainly lie on one side of the cell layer, and their axons on the other side. The
dendritic layer, called the molecular layer, is bounded by the hippocampal fissure and the granule cell layer.
The layer containing granule cell axons is a polymorphic zone, called the hilus. It includes two main classes
of non-granule cells, GABAergic interneurons, and glutamatergic mossy cells (Fig. 1A). The borders of the
hilus in the rat are shown in Fig. 1. In primates, the border between the hilus and CA3c is difficult to define,
because the hilus is relatively small, and the large mossy cells of the hilus are close to the modified
pyramidal cells of CA3.
The tri-laminar dentate gyrus is folded such that in a cross-section it looks like a V- or C-shaped
structure, having two ‘‘blades’’ and an area where these two blades meet (Fig. 1B). Moreover, it also curves
as a whole, along a dorsal-ventral and medio-lateral axis (Figs. 1C, D). In the rat, the dorsal dentate gyrus
is medial and anterior, and as it extends posteriorly, it moves to a more ventral and lateral position,
eventually curving anteriorly again in its most ventrolateral position (Fig. 1D). Because the result is a
structure that is curved in three-dimensions, it is suggested here that the best way to discuss the ‘‘blades’’ of
the dentate gyrus cell layer is not the way that is most common. As shown in Table 1, the portion of the
dentate gyrus that is closest to CA1 is often called the ‘‘upper’’ or ‘‘dorsal’’ blade. In actuality, this same
part of the granule cell layer becomes ventral and inferior as the dentate gyrus curves from dorsal to ventral
along the longitudinal axis. Therefore, it is more consistent to use the term ‘‘enclosed’’ or ‘‘hidden’’ for this
portion of the cell layer. In other words, it is always enclosed in the CA1–CA3 pyramidal cell regions of the
hippocampal formation. Similarly, the opposite part of the granule cell layer, typically referred to as a
‘‘lower’’ or ‘‘ventral’’ blade, always lies ‘‘exposed’’ or ‘‘open’’ because it is not embedded in the pyramidal
cell layers of the hippocampal formation. Therefore, the terminology that applies to all parts of the dentate
gyrus is ‘‘exposed’’ vs. ‘‘enclosed.’’
In the rodent, the curve of the tri-laminar dentate gyrus blades is slightly different in dorsal and ventral
levels. This is reflected in the fact that sections cut in the horizontal plane through the middle-third of the
dentate gyrus, show a cell layer that is folded into a ‘‘C’’ shape, yet coronal sections in the dorsal area show
a more ‘‘V’’ like structure (Fig. 1B). At the point that is most ventral and anterior (also referred to as
xii

Fig. 1. The three-dimensional organization of the dentate gyrus. (A) The fundamental organization of the neurons in the dentate gyrus
is shown, with the main categories of neurons (granule cells and non-granule cells, the latter including GABAergic interneurons and
glutamatergic mossy cells). (B) The two-dimensional perspective of the dentate gyrus changes with septotemporal position and plane of
section. In the dorsal, coronal plane, the layers are folded similar to a ‘‘V’’, with the apex or crest as the location of the point of the V.
In the middle of the hippocampus cut in horizontal section, the granule cell layer is curved. At extreme septal locations, dorsal sections
demonstrate an arc to the ‘‘V’’; at extreme temporal sites, the exposed blade is relatively small. The hilus of the rodent is illustrated
schematically for two orientations. (C) Coronal sections from rostral to caudal levels illustrate the changes in the blades and illustrate
the lack of consistency of terms such as ‘‘dorsal blade’’ across the entire length of the hippocampus. (D) Comparative views of the
rodent and primate hippocampus illustrate differences in relative locations schematically.

temporal pole), the dentate appears as if almost no exposed blade is present. The most dorsal and anterior
region (also referred to as septal pole) is distinct again, because the curve is flattened and the crest twisted
upward, toward the dorsal aspect of the brain. Some of these changes impact the way the layers are
discriminated, particularly the separation between the hilus and CA3. In the dorsal dentate gyrus, the CA3c
 xiii

Table 1

Formal term Synonyms

A. Common nomenclature
Stratum moleculare Molecular layer
Stratum granulosum Granule cell layer
Polymorphic layer (zone) Hilus, hilar region, CA4, zone 4 of Amaral (1978)
Subgranular layer (zone) 50–100 mm of the hilus, next to the granule cell layer
Granule cell Granular cell
GABAergic neuron Interneuron
Glutamatergic hilar neuron Mossy cell
Granule cell axons Mossy fibers
Enclosed blade Upper, dorsal, inner, hidden, superior blade
Exposed blade Lower, ventral, outer, open, inferior blade

Rodents Primates

B. Species-specific nomenclature
Dorsal, septal Dorsal, posterior
Ventral, temporal Ventral, temporal
Rostral, anterior, dorsal pole Rostral, anterior, temporal pole

pyramidal cell layer extends far into the dentate gyrus, almost reaching the point where the blades meet (the
apex or crest). At temporal levels, the CA3c pyramidal cell layer curves toward the enclosed blade. These
issues stress the importance of clear descriptive criteria to differentiate between subfields and layers based
on Nissl-stained sections, in conjunction with other staining methods, as needed.
Another confusing set of terms describes the parts of the longitudinal axis of the dentate gyrus (and the
entire hippocampal formation, for that matter). In rodents, this axis is most commonly referred to as the
septal-to-temporal axis or dorsal-to-ventral axis (Fig. 1). In primates, the longitudinal axis is generally
referred to as the posterior-to-anterior axis, where posterior is comparable to dorsal in rodents. The
confusion lies in the fact that the dentate is positioned differently in the primate, with the more anterior tip
located in the ventral part of the brain (Fig. 1).
Helen Scharfman
Columbia University and Helen Hayes Hospital
Menno Witter
Vrije Universiteit, Amsterdam and Norwegian University of Science and Technology, Trondheim
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 1

The dentate gyrus: fundamental neuroanatomical

organization (dentate gyrus for dummies)

David G. Amaral1,, Helen E. Scharfman2 and Pierre Lavenex3

1
Department of Psychiatry and Behavioral Sciences, The M.I.N.D. Institute and the California National Primate Research
Center, UC Davis, Davis, CA, USA
2
Departments of Pharmacology and Neurology, Columbia University, New York, NY 10032 and the Center for Neural
Recovery and Rehabilitation Research, Helen Hayes Hospital, New York State Department of Health, Rte 9W, West
Haverstraw, NY 10993-1195, USA
3
Department of Medicine, Unit of Physiology, University of Fribourg, 1700 Fribourg, Switzerland

Abstract: The dentate gyrus is a simple cortical region that is an integral portion of the larger functional
brain system called the hippocampal formation. In this review, the fundamental neuroanatomical organ-
ization of the dentate gyrus is described, including principal cell types and their connectivity, and a
summary of the major extrinsic inputs of the dentate gyrus is provided. Together, this information provides
essential information that can serve as an introduction to the dentate gyrus — a ‘‘dentate gyrus for
dummies.’’

Keywords: neuroanatomy; circuits; connections; cell types

Introduction One of the unique features of the hippocampal

The dentate gyrus is a simple cortical region that is
an integral portion of the larger functional brain
system called the hippocampal formation (Fig. 1)
(Amaral and Lavenex, 2007). The regular organ-
ization of its principal cell layers coupled with the
highly ordered laminar distribution of many of its
inputs has encouraged its use as a model system
for many facets of modern neurobiology. In this
chapter, we present the fundamental principles of
neuroanatomical organization of the dentate
gyrus, its principal cell types and their connectiv-
ity, and a summary of the major extrinsic inputs of
the dentate gyrus.
Corresponding author. Tel.: +1 916 703 0225;
Fax: +1 916 703 0287; E-mail: dgamaral@ucdavis.edu
formation is that many of its connections are uni-
directional (Fig. 1). Thus, the entorhinal cortex
provides the major input to the dentate gyrus via
fibers called the perforant path (see Chapter 3).
However, the dentate gyrus does not return a pro-
jection to the entorhinal cortex. Since the ent-
orhinal cortex is the source of much of the cortical
sensory information that the hippocampal forma-
tion uses to carry out its functions, and since the
dentate gyrus is the major termination of projec-
tions from the entorhinal cortex, it is reasonable to
consider the dentate gyrus as the first step in the
processing of information processing that ulti-
mately leads to the production of episodic mem-
ories. Moreover, the unique neuroanatomy of the
dentate gyrus predicts that it carries out a specific
information-processing task with the information

DOI: 10.1016/S0079-6123(07)63001-5 3
4
 5

that it receives from the entorhinal cortex and
ultimately conveys to the CA3 field of the hippo-
campus.
In this chapter, we will first give a general over-
view of the neuroanatomical organization of the
dentate gyrus. We will then put the spotlight on
three of its most important neurons: the dentate
granule cell, the dentate pyramidal basket cell and
the mossy cell. We will summarize the major
structural features of these neurons as well as the
connections they make within and beyond the
dentate gyrus. We will then very briefly discuss
some of the other neurons located in the dentate
gyrus. Finally, we briefly summarize the origin and
termination of the major extrinsic inputs to the
dentate gyrus. Many of these topics will be dis-
cussed in greater detail in other chapters of this
book. But, this chapter provides a synoptic view in
order to serve as an introduction to the dentate
gyrus — this is dentate gyrus for dummies.
Overall organization
The dentate gyrus has three layers (Figs. 1–2).
There is a relatively cell-free layer called the mo-
lecular layer which, in the rat, is approximately
250 mm thick and is occupied by, among other
things, the dendrites of the dentate granule cells.
The other major occupants of the molecular layer
are the fibers of the perforant path that originate
in the entorhinal cortex. There are also a small
number of interneurons that reside in the molec-
ular layer and fibers from a variety of extrinsic
inputs terminate there. The principal cell layer, the
granule cell layer, is made up largely of densely
packed granule cells. The thickness of the granule
cell layer ranges from 4 to 8 neurons or
60 mm.
While the granule cell layer is mainly made up of
granule cells, there are some other neurons that are
located at the boundary of the granule and pol-
ymorphic layers. The cell body of the dentate py-
ramidal basket cell, for example, is often located
just within the granule cell layer at its border with
the polymorphic layer. The granule cell layer en-
closes a cellular region, the polymorphic cell layer,
which constitutes the third layer of the dentate
gyrus (Fig. 1). A number of cell types are located
in the polymorphic layer but the most prominent is
the mossy cell that we will describe below.
The dentate gyrus along with many other fields
of the hippocampal formation create a banana-
shaped structure that, in the rat, runs from the
septal nuclei rostrally to the temporal cortex cau-
dally. Thus, the long axis is typically called the
septotemporal axis. The axis at right angles to the
septotemporal axis is typically called the transverse
axis. The dentate gyrus has a relatively similar
structure at all septotemporal levels of the hippo-
campal formation and is not generally divided into
subregions. However, it does tend to have more of
a ‘‘V’’ shape septally and more of a ‘‘U’’ shape
temporally (Fig. 3). In discussing features of the
dentate gyrus, it is often useful to refer to a

Fig. 1. The rat hippocampal formation. (A) Nissl-stained horizontal section through the hippocampal formation of the rat. The major
fields are indicated. Projections (1) originate from layer II of the entorhinal cortex (EC) and terminate in the molecular layer of the
dentate gyrus (DG) and in the stratum lacunosum-moleculare of the CA3 field of the hippocampus. An additional component of the
perforant path originates in layer III and terminates in the CA1 field of the hippocampus and the subiculum. Granule cells of the DG
give rise to the mossy fibers (2) that terminate both within the polymorphic layer of the DG and within stratum lucidum of the CA3
field of the hippocampus. The CA3 field, in turn, gives rise to the Schaffer collaterals (3) that innervate the CA1 field of the
hippocampus. Pyramidal cells in CA1 project to the subiculum and to the deep layers of the EC. The subiculum also gives rise to
projections to the deep layers of the EC. (B) Line drawing to illustrate the major regional and laminar organization of the DG. The DG
is divided into a molecular layer (ml) a granule cell layer (gcl) and a polymorphic layer (pl). The molecular layer is divided into three
sublayers based on the laminar organization of inputs. The hippocampus is divided into CA3, CA2 and CA1 subfields. Within CA3, a
number of layers are defined. The main cell layer is the pyramidal cell layer (pcl). Deep to the pyramidal cell layer is the stratum oriens
(so); deep to this is the white matter of the alveus (al). Superficial to the pyramidal cell layer is stratum lucidum (sl), stratum radiatum
(sr) and stratum lacunosum-moleculare (sl-m). Fields CA2 and CA1 have the same layers as CA3 except for stratum lucidum. The
remainder of the hippocampal formation is made up of the subiculum (Sub), presubiculum (Pre), parasubiculum (Para) and EC. Layers
of these latter structures are indicated with roman numerals. Additional abbreviations: ab, angular bundle; fi, fimbria; hf, hippocampal
fissure. (C) Schematic illustration of DG and hippocampus to illustrate position of suprapyramidal blade, infrapyramidal blade and
crest of the DG. This model is used in subsequent illustrations to demonstrate the major cell types and connections of the DG. (See
Color Plate 1.1 in color plate section.)
6

Fig. 2. Comparative Nissl and Timm’s staining of the rat and monkey dentate gyrus. Approximately the same portions of the rat (A &
B) and monkey (C & D) hippocampal formation are illustrated. The magnification of the paired images are different. The increased
complexity of the hilar region in the monkey compared to the rat is obvious in these photomicrographs. In the monkey, the CA3 field
inserts more deeply into the dentate gyrus and accounts for a much greater area of the hilar region. While the darkly stained
polymorphic layer is pronounced in the rat, it is much thinner in the monkey brain. The condensed layer of mossy fibers is also much
more pronounced in the rat compared to the monkey. Also note that the strict lamination of the molecular layer into thirds in the rat is
not as sharp as in the monkey. This corresponds to a more gradient-like termination of the medial and lateral perforant paths in the
monkey compared to the rat. Calibration, 250 mm (A, B); 350 mm (C, D).

particular transverse portion of the ‘‘V’’- or ‘‘U’’-
shaped structure. The portion of the granule cell
layer that is located between the CA3 field and the
CA1 field (separated by the hippocampal fissure) is
called the suprapyramidal (above CA3) blade and
the portion opposite to this, the infrapyramidal
(below CA3) blade. The region bridging the two
blades (at the apex of the ‘‘V’’ or ‘‘U’’) is called the
crest (Fig. 3).
Comparative neuroanatomy of the dentate gyrus
The comparative neuroanatomy of the dentate
gyrus is covered in detail in Chapter 2. However,
we would like to raise a few issues for which there
remains substantial confusion in the literature.
First, the basic trilaminar structure of the dentate
gyrus is common across all species studied. And,
relative to other hippocampal structures such as
the CA1 field of the hippocampus or the ent-
orhinal cortex, there has not been substantial
phylogenetic modification. For example, in the rat,
there are approximately one million granule cells
in the dentate gyrus. There are approximately ten
times more granule cells in the monkey than in the
rat, a ratio that parallels the overall volume differ-
ences. But there are only 15 times more dentate
granule cells in humans when compared to rats,
while the volume of the dentate gyrus plus the
hippocampus is
100 times larger in humans than
in rats (reviewed in more detail in Amaral and
Lavenex, 2007).
Despite the similarities, there are also some
obvious differences. In the rat, the CA3 field of the
hippocampus inserts into the dentate gyrus and
abuts the cells of the polymorphic layer. By using
stains such as the Timm’s stain that identifies
 7

Fig. 3. Horizontal sections through the rat hippocampal formation. This figure illustrates a more dorsally situated (A) and a more
ventrally situated (B) horizontal section through the rat hippocampal formation. The approximate level of the section is illustrated on a
3D reconstruction of a magnetic resonance image series of the rat brain. Subtle differences in the cytoarchitectonic organization are
seen throughout the hippocampal formation. The dentate gyrus, takes on a more ‘‘V’’ shape dorsally and a more ‘‘U’’ shape ventrally.
Calibration bar ¼ 250 mm. (See Color Plate 1.3 in color plate section.)

depositions of metals such as zinc (Fig. 2), the
heavily labeled polymorphic layer can be easily
differentiated from the CA3 field. This differenti-
ation is not so clear in the nonhuman primate or in
the human dentate gyrus. This is because the CA3
field is much more prominent within the confines
of the dentate gyrus and the polymorphic layer has
thinned to a very narrow subgranular region
(Fig. 2). In the early nomenclature espoused by
Lorente de Nó, he used a term ‘‘CA4’’ that was
vaguely defined but appears to have been intended
for the part of CA3 that inserts within the dent-
ate gyrus. However, sometimes Lorente de Nó ap-
pears to have used the CA4 term for the
polymorphic layer. We have recommended that
this term be abandoned and that even the part of
CA3 that enters the limbs of the dentate gyrus
be called CA3.
There are a variety of other differences in the
rat, monkey and human dentate gyrus. The gran-
ule cells, which we will describe in much more de-
tail shortly, only have apical dendrites in the rat.
But in the monkey and human, many granule cells
also have basal dendrites (Seress and Mrzljak,
1987). Similarly, the mossy cells of the rat have
very rare dendrites that enter the molecular layer.
Yet, in the macaque monkey, many of the mossy
cells give rise to numerous dendrites that enter the
molecular layer (Buckmaster and Amaral, 2001).
The implication of this is that rat mossy cells get
little if any perforant path input whereas monkey
mossy cells could get a quite substantial perforant
path input.
Unfortunately, many of the connections of the
monkey and human dentate gyrus have yet to be
investigated. One example of a projection that has
been studied in both species is the commissural
connection. While cells of the polymorphic layer
(mainly the mossy cells) give rise to a very robust
commissural projection to the molecular layer of
the contralateral dentate gyrus in the rat, this pro-
jection does not exist in the nonhuman primate
brain or in the human brain. As connections of the
primate and human dentate gyrus are better stud-
ied, it is undoubtedly the case that there will be
other fundamental differences that will affect not
only normal function but also pathological proc-
esses. This raises the caveat that while the rodent
8

dentate gyrus is an extremely valuable tool for
evaluating function and potential mechanisms of
pathology, caution must be exercised when these
results are extrapolated to the human dentate
gyrus. For the remainder of this chapter, we will
focus on neuroanatomical findings from the rat
brain.
Major cell types of the dentate gyrus — and their
connections
The dentate granule cell
The principal cell type of the dentate gyrus is the
granule cell (Figs. 4 and 5). The dentate granule
cell has an elliptical cell body with a width of ap-
proximately 10 mm and a height of 18 mm (Clai-
borne et al., 1990). The granule cell bodies are
tightly packed together and, in most cases, there is
no glial sheath interposed between cells. The gran-
ule cell has a characteristic cone-shaped tree of
spiny apical dendrites. The branches extend
throughout the molecular layer and the distal tips
of the dendritic tree end just at the hippocampal
fissure or at the ventricular surface. The total
length of the dendritic trees of granule cells lo-
cated in the suprapyramidal blade are, on average,
larger than those of cells in the infrapyramidal
blade (3500 mm vs. 2800 mm, respectively). Dend-
rites of cells in the suprapyramidal blade have
1.6 spines/mm, whereas dendrites in the infrapy-
ramidal blade have 1.3 spines/mm (Desmond and
Levy, 1985). Thus, an estimate for the number of
spines on the average suprapyramidal granule cell
would be around 5600 and for an infrapyramidal
cell 3640. Since virtually all of the excitatory inputs
to granule cells are on these dendritic spines, these
numbers indicate the approximate number of ex-
citatory synapses that dentate granule cells receive
from all sources. As noted earlier, the total
number of granule cells in one dentate gyrus of
the rat is
1.2  106 (West et al., 1991; Rapp
and Gallagher, 1996). Although neurogenesis in

Fig. 4. The dentate granule cell. The characteristic features of the dentate granule cell are illustrated, including its axonal arbor. A
collateral plexus gives rise to numerous (
200) typical synapses on cells located within the polymorphic layer. Most of these synapses
are onto the dendrites of inhibitory interneurons. Some of the large mossy fiber expansions are also distributed in the polymorphic
layer. Many of these terminate on the proximal dendrites of mossy cells. The mossy fiber axons ultimately enter the CA3 field where
they travel through the full transverse extent of the field. On their course, they terminate with mossy fiber expansions on a small
number (15–20) of CA3 pyramidal cells. Additional abbreviations: gc, granule cell; pc, pyramidal cell. (See Color Plate 1.4 in color
plate section.)

Fig. 5. The dentate granule cell. A photomicrograph (A) and line drawing (B) of a prototypical granule cell that was filled with Lucifer
yellow in a hippocampal slice. The dendrites arise primarily from the apical surface of the cell body, and the axon emerges from the
basal surface. The spiny dendrites extend into the molecular layer until the hippocampal fissure, and the axon collateralizes profusely
within the polymorphic layer. Calibration bar ¼ 25 mm (A), 20 mm (B).
the dentate gyrus persists into adulthood, and
appears to be under environmental control, mod-
ern stereological studies have shown that the total
number of granule cells does not vary in adult
animals (Rapp and Gallagher, 1996). This implies
that there is a steady state turnover of granule cells
rather than a continuous accretion.
The packing number of granule cells and their
ratio to CA3 pyramidal cells varies along the
septotemporal axis (Gaarskjaer, 1978b); the pack-
ing density is higher septally than temporally.
Since the packing density of CA3 pyramidal cells
follows an inverse gradient, the net result is that at
septal levels of the hippocampal formation, the
ratio of granule cells to CA3 pyramidal cells is
something on the order of 12:1, whereas at the
temporal pole the ratio drops to 2:3. Since the CA3
pyramidal cells are the major recipients of granule
cell innervation, and the number of mossy fiber
synapses is roughly the same along the septotem-
poral axis, contact probability is much lower
septally than temporally.
The mossy fibers — projections to the polymorphic
layer
The granule cells give rise to distinctive unmyelin-
ated axons which Ramón y Cajal called mossy
9
fibers. The mossy fibers have unusually large
boutons that form en passant synapses with the
mossy cells of the polymorphic layer and with the
CA3 pyramidal cells of the hippocampus. Less
appreciated is the fact that the mossy fiber axons
give rise to a distinctive set of collaterals that
heavily innervate cells within the polymorphic
layer of the dentate gyrus. Each principal mossy
fiber (which is on the order of 0.2–0.5 mm in di-
ameter) gives rise to about seven thinner collater-
als within the polymorphic layer before entering
the CA3 field of the hippocampus. As much as
2300 mm of collateral axonal plexus is generated by
a single mossy fiber in the polymorphic layer
(Claiborne et al., 1986). Within the polymorphic
layer, the mossy fiber collaterals branch exten-
sively and the daughter branches bear two types of
synaptic varicosities. Numerous small (approxi-
mately 2 mm) spherical synaptic varicosities are
distributed unevenly along these collaterals. There
are
160–200 of these varicosities distributed
throughout the axonal collateral plexus of a
single granule cell, and these form contacts on
dendrites located in the polymorphic layer. At the
end of many of the collateral branches there are
also larger (3–5 mm diameter), irregularly shaped
varicosities that resemble, although are smaller
than, the mossy fiber boutons found in CA3. The
mossy fiber terminals in the polymorphic layer
10
establish contacts with the proximal dendrites of
the mossy cells (Ribak et al., 1985), the basal
dendrites of the pyramidal basket cells as well as
with other, unidentified cells. Acsady et al. (1998)
reported that the vast majority of mossy fiber col-
laterals in the polymorphic cell layer terminate on
GABAergic interneurons. Since there are 160–200
such varicosities (compared with
20 of the larger
thorny excrescences), mossy fiber axons synapse
on a larger number of interneurons than mossy
cells or CA3 pyramidal cells. Mossy fiber collat-
erals occasionally enter the granule cell layer, but
they rarely enter the molecular layer under normal
conditions. The collaterals that enter the granule
cell layer appear to terminate preferentially on the
apical dendritic shafts of pyramidal basket cells.
The lack of mossy fiber innervation of the molec-
ular layer changes dramatically in pathological
conditions and sprouting of mossy fibers is one of
the major hallmarks of temporal lobe epilepsy (see
Chapter 29).
The mossy fibers — projections to CA3
The rat dentate gyrus does not project to any brain
region other than the CA3 field of the hippocam-
pus (but see Chapter by Zimmer for exceptions
in other species). The mossy fibers terminate in a
relatively narrow zone mainly located just above
the CA3 pyramidal cell layer (Blackstad et al.,
1970; Swanson et al., 1978; Gaarskjaer, 1978a;
Claiborne et al., 1986). In the proximal portion of
CA3, mossy fibers are also located below and
within the pyramidal cell layer. The layer of mossy
fiber termination located just above the pyramidal
cell layer is called stratum lucidum. There is no
indication that dentate neurons other than the
granule cells project to CA3. In particular, cells
in the polymorphic layer do not project to the
hippocampus, at least in the rodent. The dentate
projection to CA3 stops near the border of CA3
with CA2, and the lack of granule cell input is one
of the main features that distinguishes CA3 from
CA2 pyramidal cells.
As far as can be determined, all dentate granule
cells appear to project to CA3 and the axon tra-
jectory is partially correlated with the position of
the parent cell body. In the proximal portion of
CA3 (closer to the dentate gyrus), mossy fibers
are distributed below, within and above the py-
ramidal cell layer. The fibers located below the
layer, i.e., those that are in the area occupied pri-
marily by basal dendrites, are generally called the
infrapyramidal bundle (Fig. 2). The fibers located
within the pyramidal cell layer are called the in-
trapyramidal bundle and those located above the
pyramidal cell layer (in the area occupied mainly
by proximal apical dendrites) the suprapyramidal
bundle. The suprapyramidal bundle occupies the
stratum lucidum. At mid and distal portions of
CA3, the intra- and infrapyramidal bundles are
largely eliminated and those fibers that were in
these regions cross the pyramidal cell layer and
join the other mossy fibers within stratum lucid-
um.
Granule cells at all transverse positions within
the granule cell layer generate mossy fibers that
extend for the full transverse extent of CA3
(Gaarskjaer, 1981). Cells located in the infrapy-
ramidal blade of the granule cell layer have axons
that tend to enter CA3 in the infrapyramidal
bundle, but ultimately cross the pyramidal cell
layer to enter the deep portion of stratum lucidum.
The axons of granule cells located in the crest of
the dentate gyrus tend to enter CA3 in the intra-
pyramidal bundle and also ultimately ascend into
stratum lucidum. Cells located in the suprapyram-
idal blade of the dentate gyrus give rise to axons
that enter CA3 in the stratum lucidum and
continue within the most superficial portion of
stratum lucidum (Claiborne et al., 1986).
Early Golgi anatomists indicated that the
mossy fiber axons were mainly oriented perpen-
dicular to the long axis of the hippocampus.
Blackstad and colleagues confirmed the largely
transverse trajectory of the mossy fibers using de-
generation track tracing methods. Mossy fibers
originating from any particular septotemporal
level travel through the full transverse extent of
CA3 in a very narrow septotemporal zone of

1 mm, which has been described as a lamella.
There is one peculiarity of the mossy fiber pro-
jection, however, that has eluded explanation.
Lorente de Nó (1934) and McLardy (McLardy
and Kilmer, 1970) indicated that the mossy fibers

actually do change their course and take a lon-
gitudinal direction, but not until they reach the
distal portion of CA3. The extent of this distal
longitudinal projection was further clarified by
Swanson and colleagues, using the autoradio-
graphic method of neural tract tracing (Swanson
et al., 1978). They showed that granule cells
located at septal levels give rise to mossy fibers
that travel throughout most of the transverse
extent of CA3 at the same septotemporal level.
But just at the CA3/CA2 border, they abruptly
change course and travel toward the temporal
pole for nearly 2 mm. The extent of this longitu-
dinal component, however, appears to depend on
the septotemporal location of the cells of origin.
Granule cells in the mid to temporal portions of
the dentate gyrus have mossy fibers that exhibit
only a slight temporal inclination at their distal
extremity. And mossy fibers that originate at the
extreme temporal pole of the dentate gyrus barely
extend to the CA3/CA2 border and have little or
no longitudinal component. On the face of it, this
indicates that some CA3 pyramidal cells that are
located very close to the border with CA2 may be
contacted by mossy fiber axons from granule cells
spread out over a much broader septotemporal
extent of the dentate gyrus. They might, therefore,
form a special class of integrator CA3 cells. An
elegant neuroanatomical verification of this has
come from the work of Acsady et al. (1998)
who stained dentate granule cells by using in vivo
intracellular techniques. Not only do mossy
fibers travel temporally for 2–3 mm, but they also
demonstrate typical mossy fiber expansions along
this portion of their trajectory.
There is substantial evidence indicating that the
granule cells use glutamate as their primary trans-
mitter, and the asymmetric contacts made between
the mossy fiber expansions and the thorny excres-
cences would tend to confirm this notion. The
mossy fibers are, nonetheless, also immunoreactive
for several other neuroactive substances. At least
some of the mossy fibers demonstrate immunore-
activity for the opioid peptide dynorphin and they
are also immunoreactive for GABA (Walker et al.,
2002). The chemical neuroanatomy of the dentate
granule cell will be dealt with in more detail in
Chapter 6.
11
The dentate pyramidal basket cell
As with most other brain regions, some neurons of
the dentate gyrus, such as the granule cells, are
excitatory whereas others are inhibitory. The most
intensively studied type of inhibitory interneuron
in the dentate gyrus is the pyramidal basket cell
(Figs. 6 and 7; Ribak et al., 1978; Ribak and
Seress, 1983). These cells are generally located
along the interface between the granule cell layer
and the polymorphic layer. They have pyramidal-
shaped cell bodies that are substantially larger
(25–35 mm in diameter) than the granule cells
(10–18 mm). Ramón y Cajal first described the
pyramidal basket cells as having a single, principal
aspiny apical dendrite directed into the molecular
layer that divides into several aspiny branches,
and several basal dendrites that divide and extend
into the polymorphic cell layer. The number of
basket cells is not constant throughout either the
transverse or septotemporal extents of the dentate
gyrus (Seress and Pokorny, 1981). At septal levels,
the ratio of basket cells to granule cells is 1:100
in the suprapyramidal blade and 1:180 in the
infrapyramidal blade. At temporal levels, the
number is 1:150 for the suprapyramidal blade
and 1:300 for the infrapyramidal blade.
The basket portion of the name refers to the
fact that the axon of these cells forms pericellular
plexuses, like the covering of a chianti bottle, that
form encompassing synapses with the cell bodies
of granule cells. The dentate pyramidal basket
cell along with other basket cells located just be-
low the granule cell layer contribute to a very
dense terminal plexus that is confined to the
granule cell layer (Struble et al., 1978; Sik et al.,
1997). The terminals in this basket plexus are
GABAergic and form symmetric, inhibitory con-
tacts, located primarily on the cell bodies and
proximal dendritic shafts of the apical dendrites
of the granule cells. Analysis of Golgi-stained
axonal plexuses from single basket cells in the rat
indicates that they extend for distances greater
than 900 mm in the transverse axis and
1.5 mm
in the septotemporal axis. This widely distributed
axonal plexus would allow a single basket cell to
influence as many as 10,000 (1%) of the granule
cells.
12

Fig. 6. The pyramidal basket cell. The cell body of the pyramidal basket cell is located at the interface between the granule cell layer
and the polymorphic cell layer. The axon (arrow) emerges from the apical dendrite. Collaterals of this axon form a curtain of terminals
that synapse with the granule cell bodies. Additional abbreviations: pbc, pyramidal basket cell. (See Color Plate 1.6 in color plate
section.)

Fig. 7. The pyramidal basket cell. A prototypical pyramidal basket cell is shown after intracellular injection of Neurobiotin. The
montage that was created after visualization of the cell (A) and line drawing (B) illustrate the characteristics of this cell type. An arrow
points to the axon. Calibration bar ¼ 25 mm (A), 40 mm (B).
 13

Fig. 8. The mossy cell. A line drawing of a mossy cell (mc) in the polymorphic layer. The axon (arrow) develops a plexus within the
polymorphic layer, and also an ipsilateral projection to the inner molecular layer, known as the associational pathway. The main axon
also projects contralaterally to the inner molecular layer, forming the commissural pathway. The ipsilateral projection increases in
density with distance from the cell body of origin. (See Color Plate 1.8 in color plate section.)

Fig. 9. The mossy cell. A montage of several focal planes (A), and line drawing (B) of a classic mossy cell, illustrating the characteristic
thorny excrescences and dendritic tree of this cell type. Thorny excrescences are present proximal to the soma. The dendrites of the cell
extend throughout nearly the entire polymorphic region, but few enter either the granule cell or molecular layers. Calibration
bar ¼ 25 mm (A), 50 mm (B).

The mossy cell located in his ‘‘subzone of fusiform cells’’, and is

The polymorphic layer harbors a variety of neu-
ronal cell types. The most common, and certainly
the most impressive, is the mossy cell (Figs. 8 and
9) This cell type is probably what Ramón y Cajal
referred to as the ‘‘stellate or triangular’’ cells
undoubtedly what Lorente de Nó referred to as
‘‘modified pyramids’’. The mossy cell received its
current name from Amaral (1978) who studied this
and other neurons of the polymorphic layer using
the Golgi method. The cell bodies of the mossy
cells are large (25–35 mm) and are often triangular
14
or multipolar in shape. Three or more thick dend-
rites originate from the cell body and extend
for long distances within the polymorphic layer.
Each principal dendrite bifurcates once or twice
and generally gives rise to a few side branches.
While most of the dendritic branches remain
within the polymorphic layer, an occasional dend-
rite pierces the granule cell layer and enters the
molecular layer. The mossy cell dendrites virtually
never enter the adjacent CA3 field in the rat.
The most distinctive feature of the mossy cell is
that all of its proximal dendrites are covered by
very large and complex spines, the so-called thorny
excrescences (Fig. 9). These are the distinctive sites
of termination of the mossy fiber axons. While
thorny excrescences are also observed on the prox-
imal dendrites of CA3 pyramidal cells, they are
never as dense or as complex as the ones on the
mossy cells. The distal dendrites of the mossy cell
have typical pedunculate spines that appear to be
less dense than those on the distal dendrites of the
hippocampal pyramidal cells.
Mossy cell projections
The inner third of the molecular layer (Fig. 2) re-
ceives a projection that originates exclusively from
neurons in the polymorphic layer (Laurberg and
Sorensen, 1981; Frotscher et al., 1991; Buckmaster
et al., 1992, 1996). Since, in the rat, this projection
originates both in the ipsilateral and contralateral
sides of the dentate gyrus, it has been called the
associational/commissural projection. As noted
above, the commissural portion of this projection
does not exist in the primate brain. There is sub-
stantial evidence from neural track tracing studies
that this projection arises almost exclusively from
cells in the polymorphic layer and mainly from the
mossy cells. The fact that the mossy cells are
immunoreactive for glutamate (Soriano and
Frotscher, 1994) adds credence to the notion that
the associational/commissural projection is excita-
tory. A direct electrophysiological demonstration
of the excitatory nature of the mossy cell synapse
on the granule cell has been provided by Scharf-
man (1994). Scharfman (1995) has also demon-
strated that mossy cells innervate both granule
cells and GABAergic interneurons. It should also
be noted for the sake of completeness that a small
component of the commissural connection in the
rat does appear to arise from GABAergic neurons
(Ribak et al., 1986).
There are a number of interesting features about
the ‘‘feedback’’ projection from the mossy cells to
the granule cells. First, the projection from mossy
cells located at any particular level of the dentate
gyrus is distributed widely along the longitudinal
axis, both septally and temporally from the point
of origin. Axons from any particular septotempo-
ral point in the dentate gyrus may innervate as
much as 75% of the long axis of the dentate gyrus
(Amaral and Witter, 1989). Second, the projection
to the molecular layer at the septotemporal level of
origin is very weak, but gets increasingly stronger
at levels that are progressively more distant from
the cells of origin. Remembering that mossy cells
are the recipients of massive innervation from the
granule cells at their same level (via the mossy fiber
collaterals into the polymorphic layer), it would
appear that the mossy cells pass on the collective
output of granule cells from one septotemporal
level to granule cells located at distant levels of the
dentate gyrus.
Other neurons of the dentate gyrus
There has been an explosion in the number of in-
terneurons identified in the rat hippocampal for-
mation. A detailed overview of the characteristics
of the various hippocampal interneurons has been
published by Freund and Buzsaki (1996) and is
also covered in Chapter 13 of this book. Many of
the cell types can be distinguished on the basis of
the distribution of their axonal plexus. Some have
axons that terminate on cell bodies, whereas others
have axons that terminate exclusively on the initial
segments of other axons. Interneurons have also
been distinguished on the basis of their inputs.
Some are preferentially innervated, for example,
by the serotonergic fibers originating from the
raphe nuclei. Interneurons can also be differenti-
ated from principal cells on the basis of their
electrophysiological characteristics. At least some
interneurons have high rates of spontaneous

activity and fire in relation to the theta rhythm.
For this reason, interneurons are often called theta
cells (Sik et al., 1997).
Within the same subgranular region occupied by
the cell bodies and dendrites of the pyramidal
basket cells are several other cell types with dis-
tinctly different somal shapes, as well as different
dendritic and axonal configurations. Some of these
cells are multipolar with several aspiny dendrites
entering the molecular and polymorphic layers,
while others tend to be more fusiform-shaped with
a similar dendritic distribution. As Ribak and
colleagues have pointed out, many of these cells
share fine structural characteristics such as infold-
ed nuclei, extensive perikaryal cytoplasm with
large Nissl bodies and intranuclear rods. More-
over, it appears that all of these cells give rise to
axons that contribute to the basket plexus in the
granule cell layer. Most of these neurons are
immunoreactive for GABA, form symmetrical
synaptic contacts with the cell bodies, proximal
dendrites and occasionally with axon initial seg-
ments of granule cells, and therefore function as
inhibitory interneurons. These cells are not neuro-
chemically homogeneous, however, since subsets
appear to colocalize distinct categories of other
neuroactive substances (Ribak, 1992).
Neurons of the molecular layer
The molecular layer is occupied primarily by
dendrites of the granule cells, pyramidal basket
cells and polymorphic layer cells, as well as axons
and terminal axonal arbors from the entorhinal
cortex and other sources. At least two neuron
types are also present in the molecular layer. The
first is located deep in the molecular layer, has a
multipolar or triangular cell body and gives rise to
an axon that produces a substantial terminal
plexus largely limited to the outer two thirds of
the molecular layer. This neuron has aspiny dend-
rites that remain mainly within the molecular layer
and has been called the MOPP cell (molecular
layer perforant path-associated cell). This termi-
nology was proposed by Han et al. (1993) to bring
some order to naming interneurons in the hippo-
campal formation. The lettering system refers to
15
the location of the cell body and to the region
where the axon is distributed. Unfortunately, this
terminology has not been embraced and more
widely applied to neurons of the hippocampal
formation.
Frotscher and colleagues have described a
second type of neuron in the molecular layer that
resembles the chandelier or axo-axonic cell origi-
nally found in the neocortex (Soriano and Frotsc-
her, 1989). These cells are generally located
immediately adjacent or even within the superfi-
cial portion of the granule cell layer. The axo-
axonic cell is named for the fact that its axon
descends from the molecular layer into the granule
cell layer, collateralizes profusely and then termi-
nates, with symmetric synaptic contacts, exclu-
sively on the axon initial segments of granule cells.
Thus, their shape resembles that of a chandelier.
Each axo-axonic cell may innervate the axon in-
itial segments of as many as 1000 granule cells.
Since these cells are immunoreactive for markers
of GABAergic neurons and make symmetrical
synapses, it is likely that they provide an addi-
tional means of inhibitory control of granule cell
output.
Neurons of the polymorphic cell layer
Besides the mossy cell, there are a number of fusi-
form cells in the polymorphic layer. The main
difference between the fusiform cell types is
whether they have spines or not and the charac-
teristic shapes and sizes of the spines. One type,
the long-spined multipolar cell first described by
Amaral (1978) has recently been called the HIPP
cell (hilar perforant path-associated cell) (Fig. 10;
Han et al., 1993). The conspicuous feature of this
cell is the distribution of copious, long and often
branched spines over its cell body and dendrites.
Intracellular staining techniques demonstrate that
these cells have axons that ascend into the outer
two-thirds of the molecular layer (i.e., the perfo-
rant path zone) and terminate with symmetrical
and presumably inhibitory synapses on the dend-
rites of granule cells. An amazing feature of these
neurons is that their axonal plexus can extend for
as much as 3.5 mm along the septotemporal axis of
16

Fig. 10. The long-spined cell. A montage (A) and line drawing (B) of a long-spined cell in the polymorphic layer. The extremely long
spines that characterize this cell type are marked by arrowheads, and can be proximal as well as distal to the cell body, although in this
example they were primarily located along distal dendrites. The axon of this cell collateralized in the molecular layer. Some of the long-
spined cells correspond to the GABAergic interneurons that colocalize somatostatin, so-called HIPP cells. Calibration bar ¼ 25 mm
(A), 50 mm (B).

the dentate gyrus (the entire length of the dentate
gyrus in the rat is only
10 mm) and may generate
as many as 100,000 synaptic terminals. Since in-
hibitory interneurons typically have aspiny den-
drites and relatively local axonal plexuses, this
long spined multipolar/HIPP cell is a very atypical
interneuron! At least some of these HIPP cells
appear to correspond to the somatostatin/GABA
cells that give rise to the somatostatin innervation
of the outer portion of the molecular layer.
Antibodies directed against the peptide som-
atostatin have revealed that neurons such as the
HIPP cells scattered throughout the polymorphic
layer are immunoreactive for this peptide, and ac-
count for approximately 16% of the GABAergic
cells in the dentate gyrus (Morrison et al., 1982;
Bakst et al., 1986; Freund and Buzsaki, 1996;
Sik et al., 1997; Boyett and Buckmaster, 2001).
The somatostatin-positive cells all colocalize with
GABA, and are the source of the somatostatin
immunoreactive fibers and terminals in the outer
two-thirds of the molecular layer. This system of
fibers, which forms contacts on the distal dendrites
of the granule cells, provides a third means for in-
hibitory control of granule cell activity, in addition
to the GABAergic basket cell plexus and axo-ax-
onic terminals of the chandelier cells. Since elec-
tron microscopic studies have demonstrated that
the somatostatin cells are contacted by mossy fiber
terminals, the projection to the outer molecular
layer thus constitutes a local feedback inhibitory
circuit.
 17

Fig. 11. Neurons of the polymorphic region. The summary figure reprinted from Amaral (1978) illustrates the diversity of cells types
within the dentate gyrus and proximal CA3 region of the rat. Based on Golgi staining and cameral lucida drawings of individual
neurons from multiple preparations. See original paper for description of cell types.

Interestingly, unlike the mossy cell associational in the polymorphic layer of the dentate gyrus
projection that terminates more heavily at distant whose axonal plexus have not yet been well
levels of the dentate gyrus, the GABA/somatosta- described. A summary of the neurons in the pol-
tin projection terminates most heavily around the ymorphic region, to some extent still incompletely
level of the cells of origin, and termination rapidly understood, is shown in Fig. 11.
decreases within approximately 1.5 mm septally
and temporally to the cells of origin. Thus, the
Extrinsic afferents
mossy cell projection and the somatostatin/GABA
cell projection have terminal fields that are spa-
Afferent projections
tially complementary in both radial and septotem-
poral axes. The distribution suggests that the two
Although this topic is covered in more details in
cell types mediate distal excitation and local inhi-
other chapters (see Chapters 3 and 4), we will
bition, respectively.
provide a brief general synopsis of the afferent
There are also multipolar or triangular cells in
projections of the dentate gyrus.
the polymorphic layer with thin, aspiny dendrites
that extend both within the hilus and within the
molecular layer. The axons of these HICAP cells Entorhinal cortex projection to the dentate gyrus
(hilar commissural-associational pathway related
cells) extend through the granule cell layer and The dentate gyrus receives its major input from the
branch profusely in the inner third of the molec- entorhinal cortex, via the so-called perforant path-
ular layer. There is a variety of other neuron types way (Ramón y Cajal, 1893). The projection to the
18
dentate gyrus arises mainly from cells located in
layer II of the entorhinal cortex (Fig. 1), although
a minor component of the projection also comes
from layers V and VI (Steward and Scoville, 1976;
Deller et al., 1996). The entorhinal terminals are
strictly confined to the outer two-thirds of the
molecular layer where they form asymmetric
synapses that account for nearly 85% of the total
axospinous terminations (Nafstad, 1967; Hjorth-
Simonsen and Jeune, 1972). These contacts occur
primarily on the dendritic spines of granule cells,
although a small number of perforant path fibers
also form asymmetric synapses on the shafts of
GABA-positive interneurons.
The perforant pathway can be divided into two
parts based on the region of origin, pattern of ter-
mination and appearance in histochemical and
immunohistochemical preparations. In the rat, the
two divisions have been called the lateral and me-
dial perforant paths because they originate from
the lateral and medial entorhinal areas, respec-
tively. Interestingly, while the cells of origin for
these projections look very similar and the light
microscopic appearance of their projections is in-
distinguishable, they do demonstrate substantial
histochemical and chemical neuroanatomical
differences which remain largely unexplored.
Perforant path fibers originating in the lateral ent-
orhinal area terminate in the most superficial third
of the molecular layer, whereas the fibers origi-
nating from the medial entorhinal area terminate
in the middle third of the molecular layer (see
Chapter 3 for more detail on the perforant path
projection). These terminal zones are readily dis-
tinguished by the classical Timm’s stain method
for visualization of heavy metals which demon-
strates dense staining in the outer third of the mo-
lecular layer, a near absence of staining in the
middle third and a dark staining in the inner third
that is associated with the commissural/associat-
ional connection (Fig. 2). Projections from both
areas of the entorhinal cortex innervate the entire
transverse extent of the molecular layer. The thin
axon branches (0.1 mm) in the molecular layer of
the dentate gyrus show periodic varicosities with a
thickness of 0.5–1.0 mm. Most of entorhinal cortex
layer II spiny stellate cells project up to 2 mm in
the septotemporal direction forming a sheet-like
axon arbor in the molecular layer (Tamamaki and
Nojyo, 1993).
While it is often assumed that the perforant path
fibers from the entorhinal cortex are the only hip-
pocampal input reaching the dentate gyrus, it is
now clear that at least minor projections also arise
in the presubiculum and parasubiculum (Kohler,
1985). These fibers enter the molecular layer of the
dentate gyrus and ramify in a zone that is inter-
spersed between the lateral and medial perforant
path projections. The presubicular axons tend to
be thicker than those from the entorhinal cor-
tex and give rise to collaterals that take a radial
course in the molecular layer. Virtually nothing is
currently known about which cells these fibers in-
nervate or what type of transmitter they use. Since
the presubiculum receives the only direct input
from the anterior thalamic nucleus, these fibers
provide a potential link by which thalamic infor-
mation could reach the dentate gyrus.
Basal forebrain inputs: projections from the septal
nuclei
The dentate gyrus receives relatively few inputs
from subcortical structures. The most robust is the
projection from the septal nuclei (Mosko et al.,
1973; Swanson, 1978; Amaral and Kurz, 1985).
The septal projection arises from cells of the me-
dial septal nucleus and the nucleus of the diagonal
band of Broca. Septal fibers heavily innervate cells
of the polymorphic layer, particularly in a narrow
region just subjacent to the granule cell layer.
Septal fibers are lightly distributed throughout the
molecular layer.
A major portion of the fibers of the septal pro-
jection to the dentate gyrus are cholinergic. Thirty
to fifty percent of the cells in the medial septal
nucleus and 50–75% of the cells in the nucleus of
the diagonal band that project to the hippocampal
formation are cholinergic. Many of the other sep-
tal cells that project to the dentate gyrus are GAB-
Aergic. The most interesting facet of this hetero-
geneous septal projection is that the cholinergic
and GABAergic components target different cell
types. Fibers of the septal GABAergic projections
terminate preferentially on other GABAergic

nonpyramidal cells, such as the basket pyramidal
cells of the dentate gyrus, and they form symmet-
rical, presumably inhibitory contacts. The heaviest
GABAergic septal termination is on interneurons
located in the polymorphic layer. The cholinergic
septal projection to the dentate gyrus, in contrast,
terminates mainly on granule cells, making asym-
metric, presumably excitatory contacts on dendri-
tic spines, chiefly in the inner third of the
molecular layer. Only 5–10% of the septal
cholinergic synapses are on dentate interneurons.
The large mossy cells are innervated by cholinergic
fibers (Lübke et al., 1997).
Supramammillary and other hypothalamic inputs
The major hypothalamic projection to the dentate
gyrus arises from a population of large cells in the
supramammillary area that caps and partially sur-
rounds the medial mammillary nuclei ( Wyss et al.,
1979; Dent et al., 1983; Vertes, 1993; Magloczky et
al., 1994). The supramammillary projection mainly
terminates in a narrow zone located just superficial
to the granule cell layer and lightly in the poly-
morphic layer or the remaining portion of the
molecular layer. The vast majority of the sup-
ramammillary fibers terminate on the proximal
dendrites of granule cells. This projection is exci-
tatory and is likely using glutamate as a primary
neurotransmitter (Kiss et al., 2000). Most, but not
all, of the glutamatergic supramammillary neurons
that project to the dentate gyrus also colocalize
calretinin; some of these cells also colocalize sub-
stance P (Borhegyi and Leranth, 1997).
Brainstem inputs
The dentate gyrus receives a particularly promi-
nent noradrenergic input from the nucleus locus
coeruleus (Pickel et al., 1974; Swanson and Hart-
man, 1975; Loughlin et al., 1986). The nor-
adrenergic fibers terminate mainly in the
polymorphic layer of the dentate gyrus and ex-
tend into the stratum lucidum of CA3, as if pref-
erentially terminating in the zones occupied by
mossy fibers.
19
The dentate gyrus receives a minor and diffusely
distributed dopaminergic projection that arises
mainly from cells located in the ventral tegmental
area. The dopaminergic fibers terminate mainly in
the polymorphic layer.
The serotonergic projection that originates from
median and dorsal divisions of the raphe nuclei
also terminates most heavily in the polymorphic
layer in an immediately subgranular portion of the
layer (Conrad et al., 1974; Moore and Halaris,
1975; Köhler and Steinbusch, 1982; Vertes et al.,
1999). A number of GABAergic interneurons
appear to be preferentially innervated by the se-
rotonergic fibers. The targets are often the pyram-
idal basket cells. Fusiform neurons in the region,
particularly those that are stained for the calcium
binding protein calbindin, are also very heavily
innervated. As with the cholinergic projection
from the septum, many of the cells in the raphe
nuclei that project to the hippocampal formation
appear to be nonserotonergic, but their transmitter
is not known.
Conclusions
In the remainder of this book, the reader will
receive far greater detail on many of the neuroan-
atomical features of the dentate gyrus that were
only briefly touched on in this chapter. Every
effort will also be made to correlate the neuro-
anatomy of the dentate gyrus with both its elect-
rophysiological and functional attributes. We will
close with a short reflection on the position of
the dentate gyrus both within the hippocampal
formation and within the brain at large.
It has been fashionable at times to highlight the
simple neuroanatomy of the dentate gyrus and to
claim that it provides a heuristic for studying and
understanding the much more complicated neo-
cortex. This is probably a misguided strategy.
There are numerous features of the dentate gyrus
that make it absolutely unique from a neuroana-
tomical point of view and thus presumably from a
functional point of view as well. The largely uni-
directional nature of its inputs and outputs is one
distinctive feature. The distinctive mossy fibers
that give rise to massive synaptic endings that have
20
as many as 40 active sites onto postsynaptic neu-
rons is another. Our view is that the distinctive
neuroanatomy of the dentate gyrus has been
sculpted by evolution to play a precise and unique
role in the hippocampal information processing
that ultimately leads to the production of declar-
ative memories. Whatever computations the dent-
ate gyrus executes will undoubtedly be different
from, and will work in concert with, the compu-
tations carried out by field CA3 and other portions
of the hippocampal formation. The challenge for
the future will be to understand enough about the
input/output patterns of the dentate gyrus to iden-
tify the specific contribution it makes to memory
formation.
There are other unique features of the dentate
gyrus. Why, for example, is there the conspicuous
production during adult life of new granule cells
and how do these support the unique function of
the dentate gyrus? Does the formation of these
new neurons have significance for the broader
questions of stem cell recovery of function or does
it reflect again the unique attributes of one partic-
ular brain region? Despite the ‘‘simplicity’’ of the
dentate gyrus, it will undoubtedly remain the topic
of extensive research for decades to come. And,
the chapters that constitute the remainder of this
book will provide the reader both with a valuable
summary of the state of dentate gyrus science and
guidance for research into the future.
References
Acsady, L., Kamondi, A., Sik, A., Freund, T. and Buzsaki, G.
(1998) GABAergic cells are the major postsynaptic targets of
mossy fibers in the rat hippocampus. J. Neurosci., 18:
3386–3403.
Amaral, D.G. and Kurz, J. (1985) An analysis of the origins of
the cholinergic and noncholinergic septal projections to the
hippocampal formation of the rat. J. Comp. Neurol., 240:
37–59.
Amaral, D.G. (1978) A Golgi study of cell types in the hilar
region of the hippocampus in the rat. J. Comp. Neurol., 182:
851–914.
Amaral, D.G. and Lavenex, P. (2007) Hippocampal neuro-
anatomy. In: Andersen P., Morris R., Amaral D., Bliss T.
and O’Keefe J. (Eds.), The Hippocampus Book. Oxford
University Press, New York, p. 872.
Amaral, D.G. and Witter, M.P. (1989) The three dimensional
organization of the hippocampal formation: a review of
anatomical data. Neuroscience, 31: 571–591.
Bakst, I., Avendaño, C., Morrison, J.H. and Amaral, D.G.
(1986) An experimental analysis of the origins of the
somatostatin immunoreactive fibers in the dentate gyrus of
the rat. J. Neurosci., 6: 1452–1462.
Blackstad, T.W., Brink, K., Hem, J. and Jeune, B. (1970) Dis-
tribution of hippocampal mossy fibers in the rat. An exper-
imental study with silver impregnation methods. J. Comp.
Neurol., 138: 433–450.
Borhegyi, Z. and Leranth, C. (1997) Distinct substance P- and
calretinin-containing projections from the supramammillary
area to the hippocampus in rats — a species difference be-
tween rats and monkeys. Exp. Brain Res., 115: 369–374.
Boyett, J.M. and Buckmaster, P.S. (2001) Somatostatin-
immunoreactive interneurons contribute to lateral inhibitory
circuits in the dentate gyrus of control and epileptic rats.
Hippocampus, 11: 418–422.
Buckmaster, P.S. and Amaral, D.G. (2001) Intracellular re-
cording and labeling of mossy cells and proximal CA3 py-
ramidal cells in macaque monkeys. J. Comp. Neurol., 430:
264–281.
Buckmaster, P.S., Strowbridge, B.W., Kunkel, D.D., Schmiege,
D.L. and Schwartzkroin, P.A. (1992) Mossy cell axonal pro-
jections to the dentate gyrus molecular layer in the rat hip-
pocampal slice. Hippocampus, 2: 349–362.
Buckmaster, P.S., Wenzel, H.J., Kunkel, D.D. and Schwa-
rtzkroin, P.A. (1996) Axon arbors and synaptic connections
of hippocampal mossy cells in the rat in vivo. J. Comp.
Neurol., 366: 271–292.
Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1986) A light
and electron microscopic analysis of the mossy fibers of the
rat dentate gyrus. J. Comp. Neurol., 246: 435–458.
Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1990) A
quantitative three-dimensional analysis of granule cell
dendrites in the rat dentate gyrus. J. Comp. Neurol., 302:
206–219.
Conrad, L.C.A., Leonard, C.M. and Pfaff, D.W. (1974) Con-
nections of the median and dorsal raphe nuclei in the rat: an
autoradiographic and degeneration study. J. Comp. Neurol.,
156: 179–206.
Deller, T., Martinez, A., Nitsch, R. and Frotscher, M. (1996) A
novel entorhinal projection to the rat dentate gyrus — direct
innervation of proximal dendrites and cell bodies of granule
cells and gabaergic neurons. J. Neurosci., 16: 3322–3333.
Dent, J.A., Galvin, N.J., Stanfield, B.B. and Cowan, W.M.
(1983) The mode of termination of the hypothalamic projec-
tion to the dentate gyrus: An EM autoradiographic study.
Brain Res., 258: 1–10.
Desmond, N.L. and Levy, W.B. (1985) Granule cell dendritic
spine density in the rat hippocampus varies with spine shape
and location. Neurosci. Lett., 54: 219–224.
Freund, T.F. and Buzsaki, G. (1996) Interneurons of the hip-
pocampus [Review]. Hippocampus, 6: 347–470.
Frotscher, M., Seress, L., Schwerdtfeger, W.K. and Buhl, E.
(1991) The mossy cells of the fascia dentata: a comparative

study of their fine structure and synaptic connections in ro-
dents and primates. J. Comp. Neurol., 312: 145–163.
Gaarskjaer, F.B. (1978a) Organization of the mossy fiber sys-
tem of the rat studied in extended hippocampi. II. Experi-
mental analysis of fiber distribution with silver impregnation
methods. J. Comp. Neurol., 178: 73–88.
Gaarskjaer, F.B. (1978b) Organization of the mossy fiber sys-
tem of the rat studied in extended hippocampi. I. Terminal
area related to number of granule and pyramidal cells. J.
Comp. Neurol., 178: 49–72.
Gaarskjaer, F.B. (1981) The hippocampal mossy fiber system of
the rat studied with retrograde tracing techniques. Correla-
tion between topographic organization and neurogenetic
gradients. J. Comp. Neurol., 203: 717–735.
Han, Z.S., Buhl, E.H., Lorinczi, Z. and Somogyi, P. (1993) A
high degree of spatial selectivity in the axonal and dendritic
domains of physiologically identified local-circuit neurons in
the dentate gyrus of the rat hippocampus. Eur. J. Neurosci.,
5: 395–410.
Hjorth-Simonsen, A. and Jeune, B. (1972) Origin and
termination of the hippocampal perforant path in the rat
studied by silver impregnation. J. Comp. Neurol., 144:
215–232.
Kiss, J., Csaki, A., Bokor, H., Shanabrough, M. and Leranth,
C. (2000) The supramammillo-hippocampal and supramam-
millo-septal glutamatergic/aspartatergic projections in the
rat: a combined. Neuroscience, 97: 657–669.
Kohler, C. (1985) Intrinsic projections of the retrohippocampal
region in the rat brain. I. The subicular complex. J. Comp.
Neurol., 236: 504–522.
Köhler, C. and Steinbusch, H. (1982) Identification of serotonin
and non-serotonin-containing neurons of the mid-brain rap-
he projecting to the entorhinal area and the hippocampal
formation. A combined immunohistochemical and fluores-
cent retrograde tracing study in the rat brain. Neuroscience,
7: 951–975.
Laurberg, S. and Sorensen, K.E. (1981) Associational and
commissural collaterals of neurons in the hippocampal for-
mation (hilus fasciae dentate and subfield CA3). Brain Res.,
212: 287–300.
Lorente, de Nó R. (1934) Studies on the structure of the cer-
ebral cortex. II. Continuation of the study of the ammonic
system. J. Psychol. Neurol., 46: 113–177.
Loughlin, S.E., Foote, S.L. and Bloom, F.E. (1986) Efferent
projections of nucleus locus coeruleus: topographic organi-
zation of cells of origin demonstrated by three-dimensional
reconstruction. Neuroscience, 18: 291–306.
Lübke, J., Deller, T. and Frotscher, M. (1997) Septal innerva-
tion of mossy cells in the hilus of the rat dentate gyrus — an
anterograde tracing and intracellular labeling study. Exp.
Brain Res., 114: 423–432.
Magloczky, Z., Acsady, L. and Freund, T.F. (1994) Principal
cells are the postsynaptic targets of supramammillary affer-
ents in the hippocampus of the rat. Hippocampus, 4:
322–334.
McLardy, T. and Kilmer, W.L. (1970) Hippocampal circuitry.
Am. Psychol., 25: 563–566.
21
Moore, R.Y. and Halaris, A.E. (1975) Hippocampal innerva-
tion by serotonin neurons of the midbrain raphe in the rat. J.
Comp. Neurol., 164: 171–184.
Morrison, J.H., Benoit, R., Magistretti, P.J., Ling, N. and
Bloom, F.E. (1982) Immunohistochemical distribution of
prosomatostatin-related peptides in hippocampus. Neurosci.
Lett., 34: 137–142.
Mosko, S., Lynch, G. and Cotman, C.W. (1973) The distribu-
tion of septal projections to the hippocampus of the rat. J.
Comp. Neurol., 152: 163–174.
Nafstad, P.H.J. (1967) An electron microscope study on the
termination of the perforant path fibres in the hippocampus
and the fascia dentata. Zeitsch. Zellforsch. Mikrosk. Anat.,
76: 532–542.
Pickel, V.M., Segal, M. and Bloom, F.E. (1974) A radioauto-
graphic study of the efferent pathways of the nucleus locus
coeruleus. J. Comp. Neurol., 155: 15–42.
Ramón y Cajal, S. (1893) Estructura del asta de Ammon y
fascia dentata. Ann. Soc. Esp. Hist. Nat., 22.
Rapp, P.R. and Gallagher, M. (1996) Preserved neuron number
in the hippocampus of aged rats with spatial learning deficits.
Proc. Natl. Acad. Sci. U.S.A., 93: 9926–9930.
Ribak, C.E. (1992) Local circuitry of GABAergic basket cells in
the dentate gyrus. Epilepsy Res. Suppl., 7: 29–47.
Ribak, C.E. and Seress, L. (1983) Five types of basket cell in the
hippocampal dentate gyrus: a combined Golgi and electron
microscopic study. J. Neurocytol., 12: 577–597.
Ribak, C.E., Seress, L. and Amaral, D.G. (1985) The
development, ultrastructure and synaptic connections of
the mossy cells of the dentate gyrus. J. Neurocytol., 14:
835–857.
Ribak, C.E., Seress, L., Peterson, G.M., Seroogy, K.B., Fallon,
J.H. and Schmued, L.C. (1986) A GABAergic inhibitory
component within the hippocampal commissural pathway. J.
Neurosci., 6: 3492–3498.
Ribak, C.E., Vaughn, J.E. and Saito, K. (1978) Immunocyto-
chemical localization of glutamic acid decarboxylase in neu-
ronal somata following colchicine inhibition of axonal
transport. Brain Res., 140: 315–332.
Scharfman, H.E. (1994) Evidence from simultaneous intracel-
lular recordings in rat hippocampal slices that area CA3
pyramidal cells innervate dentate hilar mossy cells. J. Ne-
urophysiol., 72: 2167–2180.
Scharfman, H.E. (1995) Electrophysiological evidence that
dentate hilar mossy cells are excitatory and innervate
both granule cells and interneurons. J. Neurophysiol., 74:
179–194.
Seress, L. and Mrzljak, L. (1987) Basal dendrites of granule
cells are normal features of the fetal and adult dentate gyrus
of both monkey and human hippocampal formations. Brain
Res., 405: 169–174.
Seress, L. and Pokorny, J. (1981) Structure of the granular layer
of the rat dentate gyrus. A light microscopic and Golgi study.
J. Anat., 133: 181–195.
Sik, A., Penttonen, M. and Buzsaki, G. (1997) Interneurons in
the hippocampal dentate gyrus — an in vivo intracellular
study. Eur. J. Neurosci., 9: 573–588.
22
Soriano, E. and Frotscher, M. (1989) A GABAergic axo-axonic
cell in the fascia dentata controls the main excitatory hippo-
campal pathway. Brain Res., 503: 170–174.
Soriano, E. and Frotscher, M. (1994) Mossy cells of the rat
fascia dentata are glutamate-immunoreactive. Hippocampus,
4: 65–69.
Steward, O. and Scoville, S.A. (1976) Cells of origin of ent-
orhinal cortical afferents to the hippocampus and fascia den-
tata of the rat. J. Comp. Neurol., 169: 347–370.
Struble, R.G., Desmond, N.L. and Levy, W.B. (1978) Ana-
tomical evidence for interlamellar inhibition in the fascia
dentata. Brain Res., 152: 580–585.
Swanson, L.W. (1978) The anatomical organization of septo-
hippocampal projections. In: Gray A. (Ed.), Functions of the
Septo-Hippocampal System. Elsevier North Holland, Am-
sterdam, pp. 25–48.
Swanson, L.W. and Hartman, B.K. (1975) The central ad-
renergic system. An immunofluorescence study of the loca-
tion of cell bodies and their efferent connections in the rat
utilizing dopamine-beta-hydroxylase as a marker. J. Comp.
Neurol., 163: 467–506.
Swanson, L.W., Wyss, J.M. and Cowan, W.M. (1978) An au-
toradiographic study of the organization of intrahippocampal
association pathways in the rat. J. Comp. Neurol., 181:
681–716.
Tamamaki, N. and Nojyo, Y. (1993) Projection of the
entorhinal layer-II neurons in the rat as revealed by intra-
cellular pressure-injection of neurobiotin. Hippocampus, 3:
471–480.
Vertes, R.P. (1993) PHA-L analysis of projections from the
supramammillary nucleus in the rat. J. Comp. Neurol., 326:
595–622.
Vertes, R.P., Fortin, W.J. and Crane, A.M. (1999) Projections
of the median raphe nucleus in the rat [Review]. J. Comp.
Neurol., 407: 555–582.
Walker, M.C., Ruiz, A. and Kullmann, D.M. (2002) Do mossy
fibers release GABA? Epilepsia, 43: 196–202.
West, M.J., Slomianka, L. and Gundersen, H.J.G. (1991) Un-
biased stereological estimation of the total number of neu-
rons in the subdivisions of the rat hippocampus using the
optical fractionator. Anat. Rec., 231: 482–497.
Wyss, J.M., Swanson, L.W. and Cowan, W.M. (1979) Evi-
dence for an input to the molecular layer and the stra-
tum granulosum of the dentate gyrus from the
supramammillary region of the hypothalamus. Anat. Em-
bryol., 156: 165–176.
Plate 1.1. The rat hippocampal formation. (A) Nissl-stained horizontal section through the hippocampal formation of the rat. The
major fields are indicated. Projections (1) originate from layer II of the entorhinal cortex (EC) and terminate in the molecular layer of
the dentate gyrus (DG) and in the stratum lacunosum-moleculare of the CA3 field of the hippocampus. An additional component of
the perforant path originates in layer III and terminates in the CA1 field of the hippocampus and the subiculum. Granule cells
of the DG give rise to the mossy fibers (2) that terminate both within the polymorphic layer of the DG and within stratum lucidum of
the CA3 field of the hippocampus. The CA3 field, in turn, gives rise to the Schaffer collaterals (3) that innervate the CA1 field of the
hippocampus. Pyramidal cells in CA1 project to the subiculum and to the deep layers of the EC. The subiculum also gives rise to
projections to the deep layers of the EC. (B) Line drawing to illustrate the major regional and laminar organization of the DG. The DG
is divided into a molecular layer (ml) a granule cell layer (gcl) and a polymorphic layer (pl). The molecular layer is divided into three
sublayers based on the laminar organization of inputs. The hippocampus is divided into CA3, CA2 and CA1 subfields. Within CA3, a
number of layers are defined. The main cell layer is the pyramidal cell layer (pcl). Deep to the pyramidal cell layer is the stratum oriens
(so); deep to this is the white matter of the alveus (al). Superficial to the pyramidal cell layer is stratum lucidum (sl), stratum radiatum
(sr) and stratum lacunosum-moleculare (sl-m). Fields CA2 and CA1 have the same layers as CA3 except for stratum lucidum. The
remainder of the hippocampal formation is made up of the subiculum (Sub), presubiculum (Pre), parasubiculum (Para) and EC. Layers
of these latter structures are indicated with roman numerals. Additional abbreviations: ab, angular bundle; fi, fimbria; hf, hippocampal
fissure. (C) Schematic illustration of DG and hippocampus to illustrate position of suprapyramidal blade, infrapyramidal blade and
crest of the DG. This model is used in subsequent illustrations to demonstrate the major cell types and connections of the DG. (For
B/W version, see page 4 in the volume.)
Plate 1.3. Horizontal sections through the rat hippocampal formation. This figure illustrates a more dorsally situated (A) and a more
ventrally situated (B) horizontal section through the rat hippocampal formation. The approximate level of the section is illustrated on a
3D reconstruction of a magnetic resonance image series of the rat brain. Subtle differences in the cytoarchitectonic organization are
seen throughout the hippocampal formation. The dentate gyrus, takes on a more ‘‘V’’ shape dorsally and a more ‘‘U’’ shape ventrally.
Calibration bar ¼ 250 mm. (For B/W version, see page 7 in the volume.)

Plate 1.4. The dentate granule cell. The characteristic features of the dentate granule cell are illustrated, including its axonal arbor. A
collateral plexus gives rise to numerous (
200) typical synapses on cells located within the polymorphic layer. Most of these synapses
are onto the dendrites of inhibitory interneurons. Some of the large mossy fiber expansions are also distributed in the polymorphic
layer. Many of these terminate on the proximal dendrites of mossy cells. The mossy fiber axons ultimately enter the CA3 field where
they travel through the full transverse extent of the field. On their course, they terminate with mossy fiber expansions on a small
number (15–20) of CA3 pyramidal cells. Additional abbreviations: gc, granule cell; pc, pyramidal cell. (For B/W version, see page 8 in
the volume.)
Plate 1.6. The pyramidal basket cell. The cell body of the pyramidal basket cell is located at the interface between the granule cell layer
and the polymorphic cell layer. The axon (arrow) emerges from the apical dendrite. Collaterals of this axon form a curtain of terminals
that synapse with the granule cell bodies. Additional abbreviations: pbc, pyramidal basket cell. (For B/W version, see page 12 in the
volume.)

Plate 1.8. The mossy cell. A line drawing of a mossy cell (mc) in the polymorphic layer. The axon (arrow) develops a plexus within the
polymorphic layer, and also an ipsilateral projection to the inner molecular layer, known as the associational pathway. The main axon
also projects contralaterally to the inner molecular layer, forming the commissural pathway. The ipsilateral projection increases in
density with distance from the cell body of origin. (For B/W version, see page 13 in the volume.)
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 3

The perforant path: projections from the entorhinal

cortex to the dentate gyrus

Menno P. Witter1,2,

1
Institute for Clinical and Experimental Neurosciences, Department of Anatomy & Neurosciences, VU University Medical
Center, MF-G102C, P.O. Box 7057, 1007 MB Amsterdam, The Netherlands
2
Centre for the Biology of Memory, Norwegian University of Science and Technology, Trondheim, Norway

Abstract: This paper provides a comprehensive description of the organization of projections from the
entorhinal cortex to the dentate gyrus, which together with projections to other subfields of the hippo-
campal formation form the so-called perforant pathway. To this end, data that are primarily from an-
atomical studies in the rat will be summarized, complimented with comparative data from other species.
The analysis of the organization of any of the connections of the hippocampus, including that of the
entorhinal cortex to the dentate gyrus, is severely hampered because of the complex three-dimensional
shape of the hippocampus. In particular in rodents, but to a lesser extent also in primates, all traditional
planes of sectioning will result in sections that at some point or another do not cut through the hippo-
campus at an angle that is perpendicular to its long axis. To amend this, we will describe own unpublished
tracing data obtained in the rat with the use of the so-called extended preparation. A number of issues will
be addressed. First, data will be summarized which will clarify the laminar origin of the perforant pathway
within the entorhinal cortex. Second, we will discuss whether or not a radial organization, along the
proximo-distal dendritic axis of granule cells, characterizes the entorhinal-dentate projection. Third, we will
discuss whether this projection is governed by any transverse organization, and fourth, we will focus on the
organization along the longitudinal axis. Finally, the synaptic organization and the contralateral ent-
orhinal-dentate projection will be described briefly. Taken together, the available data suggest that the
projection from the entorhinal cortex to the dentate gyrus is a fairly well conserved connection, present in
all species studied, exhibiting a grossly similar organization.

Keywords: hippocampus; anatomy; parahippocampal region; topographical organization; subiculum

Introduction dentate gyrus with its major cortical input. Sec-

A description of the projection from the ent-
orhinal cortex to the dentate gyrus is an indispen-
sable part of any work on the dentate gyrus for at
least two reasons. First, this pathway, generally
referred to as the perforant pathway, provides the
Corresponding author. Tel.:+31 20 4448048;
Fax:+31 20 4448054; E-mail: mp.witter@vumc.nl
ond, the perforant pathway is among the path-
ways in the brain that have been most actively
investigated, beginning with the earliest pioneers
in neuroscience who described its remarkable
structure and organization. These studies were
followed by others, which clarified the central im-
portance of the perforant pathway to hippocam-
pal function and plasticity, studies which continue
to this day.

DOI: 10.1016/S0079-6123(07)63003-9 43
44
The elaborate Golgi studies of Ramón y Cajal
(1911) and Lorente de Nó (1933), first demonstrated
that the entorhinal cortex is the origin of an im-
mensely strong projection to the dentate gyrus.
These observations were subsequently corroborated
and extended in a seemingly continuous stream of
tracing studies using axonal degeneration techniques
(Blackstad, 1958; Hjorth-Simonsen, 1972; Hjorth-
Simonsen and Jeune, 1972), followed by transport
of radioactively labeled amino acids and horse-
radish peroxidase (Van Hoesen and Pandya, 1975;
Steward, 1976; Wyss, 1981; Witter and Groenewe-
gen, 1984; Witter, 1989; Witter et al., 1989b) and
finally the more recently introduced sensitive tracing
with lectines and dextran-amines (Köhler, 1985;
Witter, 1989; Tamamaki and Nojyo, 1993; Deller
et al., 1996; Deller, 1998). In the same period,
a number of retrograde tracing studies, again using
a variety of different tracers, have further contrib-
uted to our current understanding of the entorhinal-
dentate projection (Ruth et al., 1982, 1988; Witter
et al., 1989b; Dolorfo and Amaral, 1998). Although
the dentate gyrus is the ‘‘traditional’’ target of the
entorhinal-hippocampal fibers, there is ample evi-
dence that the entorhinal cortex also projects to the
hippocampal fields CA1–CA3, and to the subiculum
(Steward, 1976; Steward and Scoville, 1976; Witter
et al., 1989a; Desmond et al., 1994; Naber et al.,
2001; Baks-Te-Bulte et al., 2005).
In all species, the hippocampus and entorhinal
cortex show a complex three-dimensional orienta-
tion and relationship. In rats in particular, this has
lead to a general tendency to restrict hippocampal
and entorhinal studies to those parts that are most
easily accessible, such as dorsal hippocampus and
central parts of entorhinal cortex. Initially this re-
sulted in rather coarse descriptions of the overall
functionality and connectivity of entorhinal-dent-
ate relationships, and it was only after analyses
started to include the full extent of both entorhinal
cortex and dentate gyrus that we became aware of
the complex topographical organization of this
connection. It is now more widely appreciated that
the hippocampus and entorhinal cortex are not as
homogeneous in their organization as initially
conceived, (cf. Witter and Groenewegen, 1990).
A major reason for this new perspective was the
use of tools to circumvent a major problem with
hippocampal studies in rodents, the fact that the
hippocampal formation and the entorhinal cortex
are curved structures. Cutting through such curved
structures in any plane will result in sections that
at some point or another do not cut through the
hippocampus and entorhinal cortex at an angle
that is perpendicular to the respective long axes.
Initially described by Gaarskjaer (1978), and
advocated by Ishizuka (2001), and likewise by
Amaral and Witter (1989), this problem can
be amended by using the so-called extended prep-
aration, and in a number of studies this approach
has resulted in improved understanding of the
connectional organization of the system (Amaral
and Witter, 1989; Ishizuka et al., 1990; Ishizuka,
2001).
Many of the disputes about functions of the
entorhinal-dentate projection, and entorhinal-hip-
pocampal connectivity more generally, are likely
to be related to this spatial problem, at least in
part. A proper understanding of the complex
three-dimensional organization of the system is
essential to evaluate data from functional studies
which suggest heterogeneity, as has been suggested
for the dorsal versus the ventral hippocampus, for
example (Witter et al., 1989a; Moser et al., 1993).
Another, more recent example deals with the
rather confusing data addressing the functional
relevance of the entorhinal cortex. Here, again,
studies have benefited significantly from taking
into account the complex three-dimensional or-
ganization of cortico-entorhinal-hippocampal con-
nections (Fyhn et al., 2004; Hafting et al., 2005;
Hargreaves et al., 2005; Steffenach et al., 2005;
Sargolini et al., 2006). Over the years, one of our
aims has been to further our understanding of the
system through systematic descriptions of its com-
plex anatomical organization. For the present pa-
per, the focus will be on the entorhinal-dentate
projection; in a recently published parallel paper,
the focus was on entorhinal projections to the
subiculum (Witter, 2006). The data support a cur-
rently prevailing view that there are two differen-
tially organized components within the entorhinal-
dentate projection, originating from the lateral
and medial entorhinal cortex respectively.

Nomenclature
Dentate gyrus
In all species, the dentate gyrus represents part of
the cortex situated closest to the free rim of the
original cortical anlage, but the final position and
overall structural organization of the dentate gyrus
and of the hippocampus as a whole may be quite
different among species (Stephan, 1975; Voogd
et al., 1998). To give a striking example, one just
has to compare the position of the hippocampus in
the rat with that in primates. In the rat the struc-
ture extends from a dorsomedial position, in close
proximity to the most caudodorsal part of the
septal complex to a ventrolateral position in close
proximity to the most caudomedial parts of the
amygdaloid complex. In contrast, in primates, this
same structure starts from a dorsocaudal position,
close to the splenium of the corpus callosum, ex-
tending ventrorostrally all the way again to appo-
sition in close proximity to the amygdaloid
complex. Moreover, in the rat, the dorsal portion
of the dentate gyrus, plus parts of the hippocam-
pus proper, seem to fold backwards, continuing
into a structure generally referred to as the gyrus
fasciolaris (fasciolaris cinereum), whereas in pri-
mates such a folding seems to be the main char-
acteristic of the most ventral and anterior part,
referred to as genu and uncal portions, spatially
associated with the amygdaloid complex. A second
example refers to the overall connotation that the
principal cell layer of dentate gyrus contains gran-
ule cells, i.e. cells without basal dendrites but with
an extensive apical tuft. In contrast, in the human
and monkey dentate, a sizeable proportion, in hu-
mans up to 30%, of the granule cells do have basal
dendrites (Seress and Mrzljak, 1987; Lim et al.,
1997).
Irrespective of such differences, it is useful to
develop a coherent and generally applicable no-
menclature, for example to be able to distinguish
one portion of the granule cell layer from another.
For this paper, I will call the portion of the granule
cell layer that is adjacent to CA1, the enclosed
blade (which is synonymous with suprapyramidal,
dorsal, or inner blade/limb), and the opposite
45
portion of the granule cell layer will be called the
exposed blade (which is synonymous with infra-
pyramidal, ventral, outer, or free blade/limb). The
region of the ‘‘V’’ or ‘‘U’’ that unites the two
blades will be referred to as the crest, also called
apex or vertex.
Aside from the longitudinal or dorsoventral
axis, a description of the organization of the dent-
ate gyrus is strongly facilitated by defining two
additional axes, a transverse and a radial axis. The
transverse axis is oriented perpendicular to the
long axis, similar to the term ‘lamellar axis’ be-
cause a section through the transverse axis would
reveal the classic lamellar arrangement of the hip-
pocampus. The radial axis refers to an axis that is
essentially perpendicular to the transverse axis. It
can be defined as a perpendicular to the granule
cell layer, extending from molecular layer through
the granule cell layer to the hilus. It therefore rep-
resents the orientation of the granule cell dendrites
from their most distal apical extent to their origin
at the granule cell soma. For the present descrip-
tion we will refer to the three layers of the dentate
gyrus as molecular layer, granule layer, and the
hilus of the dentate gyrus as the third, inner, or
polymorph layer. The molecular layer will be
further subdivided into inner, middle, and outer
one-thirds.
Entorhinal cortex
In view of the classical notion that the entorhinal
cortex is the source of the projection to the dentate
gyrus (Ramón y Cajal, 1911), we and others have
used this to delineate the entorhinal cortex of the
rat and the monkey (Insausti et al., 1997). The
entorhinal cortex comprises six layers, including
the two cell-sparse layers I and IV, respectively
called the molecular layer and lamina dissecans.
There are several reasons to use this nomenclature
for entorhinal cortical layers instead of the initially
proposed terms defined by Lorente de Nó (1933),
who considered the lamina dissecans as the deep,
sparsely populated part of layer III, and used the
designation IV for the layer of large pyramidal
cells which is layer Va in the currently employed
46
terminology (Witter et al., 1989a; Witter and
Amaral, 2004).
In general, the entorhinal cortex can be subdi-
vided into two components generally referred to as
lateral and medial entorhinal cortex or Brodman’s
areas 28a and 28b respectively (for a more detailed
description and comparison of different nomen-
clatures used in the rat and in different species the
reader is referred to a number of reviews (Witter
et al., 1989a). In the rat, and likewise in the mouse,
the entorhinal cortex has been further subdivided
into dorsolateral (DLE), dorsal-intermediate
(DIE), ventral-intermediate (VIE), caudal (CE),
and medial (ME) subdivisions (Insausti et al.,
1997; van Groen et al., 2003). Initially, we included
a sixth domain called amygdalo-entorhinal cortex
(AE) as part of the entorhinal cortex because our
own tracing data indicated that this region also
contributed to the projections to the dentate gyrus
(Insausti et al., 1997). In a recent and more de-
tailed analysis of the efferent connectivity of this
region, it was shown that it is quite unlikely that
this so-called area AE indeed contributes projec-
tions to the dentate gyrus (Kemppainen et al.,
2002; Majak and Pitkanen, 2003). Therefore, we
exclude this region in the present description.
In monkeys, humans, and in other species in
which the entorhinal cortex was described, such as
cat, dog, and guinea pig (Krettek and Price, 1977;
Amaral et al., 1987; Witter et al., 1989a; Insausti et
al., 1995; Uva et al., 2004; Woznicka et al., 2006)
comparable partitioning schemes into multiple
subdivisions of the entorhinal cortex have been
proposed. However, in those species there is also a
tendency to consider the entorhinal cortex as com-
posed of two primary components, most likely re-
flecting functional differences.
For the present description we will therefore
maintain a subdivision of the entorhinal cortex
into two main subdivisions, the lateral (LEC) and
medial (MEC) entorhinal cortices respectively.
Studies of the perforant pathway projection to
the dentate gyrus are most numerous and most
detailed in the rat. Therefore, the descriptions
will be initially provided for material gathered
from experiments in rats. When relevant, data
from other species will be added in a comparative
fashion.
Fiber pathways
After leaving the entorhinal cortex, perforant path
fibers enter the underlying white matter and the an-
gular bundle. They then traverse the pyramidal cell
layer of the subiculum and cross the hippocampal
fissure to enter the dentate gyrus, or distribute to the
molecular layer of the subiculum and the hippocam-
pus. The entorhinal cortex fibers also take alternative
routes such as projecting through the alveus before
entering the hippocampus, or by traversing the mo-
lecular layers of the entorhinal cortex, pre and par-
asubiculum. The general orientation of entorhinal
fibers within hippocampal subfields is parallel to the
principal cell layers, and it has been described that
entorhinal fibers in the molecular layer of the CA-
fields continue around the tip of the enclosed blade
of the dentate gyrus to enter the dentate molecular
layer (Ramón y Cajal, 1911; Witter, 1989).
Layers of origin in the entorhinal cortex
In rat, mouse, or monkey, the ipsilateral projection
to the dentate gyrus appears to arise mainly, if not
exclusively from layer II of the entorhinal cortex
(Steward and Scoville, 1976; Schwartz and
Coleman, 1981; Ruth et al., 1982, 1988; Witter
et al., 1989b; van Groen et al., 2003; Chrobak and
Amaral, 2006). In humans, fetal material indicates a
similar origin (Hevner and Kinney, 1996). It is of
interest to note that in the brain of Alzheimer pa-
tients, the entorhinal cortex layer II neurons are
among the ones preferentially implicated in the dis-
ease, such that up to 50% of those neurons appar-
ently disappear. This rather selective loss of layer II
neurons has been associated with changes in the
outer part of the molecular layer, such as a decrease
in free glutamate and a decrease in markers asso-
ciated with the perforant path, suggesting that in
humans also the origin, and obviously termination,
of the entorhinal-dentate projection is similar to that
reported in other species (Hyman et al., 1986, 1987,
1988; Morys et al., 1994). There is evidence that a
minor component of the projection to the dentate
gyrus also comes from the deep layers (IV–VI) of
the entorhinal cortex (Köhler, 1985; Witter and
Amaral, 1991; van Groen et al., 2003). In rat and

monkey, neurons in layer II not only project to
the dentate gyrus but also to CA3 (Steward and
Scoville, 1976; Witter and Amaral, 1991). Intra-
cellular tracing in the rat revealed that a single layer
II cell may innervate not only the dentate gyrus and
CA3 but also the subiculum (Tamamaki and Nojyo,
1993), but whether this holds true for the monkey as
well is currently unknown. In contrast, in the
mouse, at least in one of the strains (C57BL/6J),
layer II cells appear to project only to the dentate
gyrus and does not extend collaterals to other hip-
pocampal subfields (van Groen et al., 2003). Al-
though this appears a striking species difference,
emphasized by several authors, it is currently not
known whether this is a general phenomenon in all
mouse strains or typical for this one strain. More-
over, our own material has indicated that in the rat
layer III cells contribute to the projection to CA3.
As illustrated in Fig. 1A, a ventrally positioned in-
jection in layers II and III of CE (case 88246) results
in strong labeling of the dentate gyrus, all CA fields
and the subiculum. In contrast, a similarly posi-
tioned injection confined to layer III (case 89339)
does not result in any labeling in the dentate gyrus,
as expected on the basis of the selective origin of the
dentate component of the perforant path in layer II.
However, in this case we did find some labeling in
CA3, as well as strong labeling in CA1 and the
subiculum (Fig. 1B, D). Likewise, this can be ob-
served in case of an injection in layer III of VIE
(case 88178; Fig. 1C, E).
On the basis of these data, it thus seems safe to
conclude that in all species, layer II is the pre-
dominant if not exclusive source of the entorhinal
projections to the outer two-thirds of the molec-
ular layer of the ipsilateral dentate gyrus (cf.
Kunzle, 2002). In addition, these data indicate that
in the rat, the projection to CA3 originates pre-
dominantly in layer II with a minor component
arising from layer III. The layer III projection to
area CA3 is apparently more prominent in mice.
Radial or layered terminal organization in the
dentate gyrus
One of the more convincing arguments to differ-
entiate between the lateral and medial subdivisions
47
of the entorhinal cortex, aside from overall cyto-
architectonic differences, was based on observa-
tions that fibers originating in the LEC terminate
in the outer one-third of the molecular layer and
fibers from the MEC terminate in the middle one-
third of the molecular layer (Hjorth-Simonsen and
Jeune, 1972; Hjorth-Simonsen, 1972). This initial
idea was strengthened further by the striking la-
minar staining pattern that was revealed by Timm
stain for heavy metals, (Haug, 1976; Stanfield and
Cowan, 1979) and by immunocytochemistry using
antibodies against enkephalin and cholecystokinin
(Fredens et al., 1984).
Although most consider a non-overlapping dis-
tribution in the dentate of LEC and MEC fibers, it
is important to note that there may be a contin-
uum in these fiber systems. This was originally
suggested on the basis of experiments with antero-
grade transport of tritiated amino acids by Steward
(1976), who suggested that the perforant path
should actually be subdivided into lateral, interme-
diate, and medial components. The intermediate
pathway was thought to originate in the inter-
mediate entorhinal area, an area most likely
including VIE and ME, and to terminate in the
molecular layer between the projections from LEC
and MEC. In an elegant series of anterograde
tracing experiments using small injections of triti-
ated amino acids in rats, Wyss (1981) reported that
there is a striking radial gradient in the terminal
distribution of entorhinal fibers depending on the
position of the source in the entorhinal cortex.
Lateral portions of LEC project close to the pial
surface and successively more medial portions
project closer to the granule cells. However, the
data from this study also supported an intermedi-
ate perforant pathway, distributing to the border
region between the typical lateral and medial per-
forant path terminal zones in the outer and middle
one-thirds of the molecular layer, respectively. A
comparable conclusion was reached on the basis of
a study using very small injections of the ante-
rogradely transported tracer DiI in portions of the
entorhinal cortex close to the rhinal fissure, i.e.
projecting to the dorsal part of the dentate gyrus
(Tamamaki, 1997). When looking at a summary
figure taken from this particular study (Fig. 2),
two conclusions are apparent. First, the radial
48

Fig. 1. Layer II and to a lesser extent layer III contribute to the projections to CA3. Sections taken from unfolded preparations
(Gaarskjaer, 1978), stained for the presence of BDA and counterstained for Nissl-substances with cresyl violet. (A) Section from a case
with an injection centered in layers II and III of MEC (Fig. 1, case 89246). Strong labeling is present in the dentate gyrus, as well as in
CA3 (and CA1 and subiculum). (B) Section from a case with an injection of which the location corresponds to that illustrated in A, but
no tracer has been taken up by layer II neurons (case 88339). Note the absence of labeling in the dentate gyrus, strong labeling in CA1
and subiculum but also weak labeling in CA3 (boxed area shown in D). (C) Section from a case with an injection in layer III of LEC
(Fig. 2, case 88178). Similar to the case shown in B, labeling is present in CA1 and subiculum, absent in dentate gyrus, but weak
labeling can be seen in CA3 (boxed area shown in E). (D) Higher magnification of boxed area marked in B, showing the light terminal
plexus in CA3. (E) Higher magnification of boxed area marked in C, showing the light terminal plexus in CA3. Scale bar in A equals
1 mm; also holds for B and C. Scale bars in D and E equal 100 mm.

extent of the labeled terminal field in the dentate
gyrus almost never covers one-third of the molec-
ular layer: one-fifth to one-sixth would be a more
appropriate description. Second, the terminal
fields show a gradual transition from a very distal
to a more proximal position along the apical
dendrites of dentate granule cells. These observa-
tions led to the proposition that the entorhinal-
dentate projection is a continuum rather than
subdivided into two or three discrete components.
Interestingly, analysis of intracellularly labeled
cells in layer II of the medial entorhinal cortex
which projected to the dorsal part of the dentate
gyrus revealed yet another pattern, where the axon
is confined to the middle one-third in the enclosed
blade but extends into the outer one-third of the
exposed blade, i.e. the terminal zone becomes
thicker in the exposed blade (Tamamaki and
Nojyo, 1993). This finding was supported by the
results of anterograde tracing using DiI. Injection
of DiI into lateral entorhinal cortex led to a thick
band of terminal labeling in the enclosed blade but
 49

Fig. 2. Confocal microphotographs showing anterograde labeling in the enclosed blade of the dentate gyrus following a series of
anterograde tracer injections (DiI) along the rostro-caudal extent of the rhinal fissure. A is taken from the most anterior injection,
which is in LEC, F is taken from the most posterior position, which is in MEC. Cases D and E are most likely along the border between
LEC and MEC, although this has not been assessed in the original paper. Arrows demarcate laminar boundaries. Scale bar equals
100 mm. Adapted with permission from Tamamaki (1997).

comparatively thin labeling in the exposed blade.
Following DiI injections in the medial entorhinal
cortex, the reverse pattern was observed. Our own
data, based on visualizing the projections with
PHA-L and BDA (Witter, 1989, 1990), to a large
extent corroborates the observations reported by
Tamamaki and Wyss (Wyss, 1981; Tamamaki,
1997). Whereas large injections generally label the
full width of the outer or middle one-third of the
dentate molecular layer, small injections never re-
sult in consistent labeling throughout the entire
outer or middle one-third. In these cases, the labe-
led domain is closer to one-fifth or one-sixth of the
entire width of the molecular layer, similar to the
observations of Tamamaki (1997). There is a weak
topographical arrangement emerging from these
experiments. Injections in the lateral part of LEA
or dorsolateral part of MEA more densely target
the more distal portions of the dendrite in their
respective terminal zone, whereas more ventrally
and medially positioned injections tend to distrib-
ute fibers preferentially to more proximal portions
of their corresponding one-third of the molecular
layer.
What remains to be discussed is whether or not
the perforant path is organized as a continuum, as
proposed by Tamamaki (1997), whether it com-
prises two non-overlapping systems — lateral and
medial, or whether there are three components —
lateral, intermediate and medial. To address this
point, we compared the radial distribution of an-
terogradely labeled fibers following two closely
spaced injections, centered around the border be-
tween VIE and ME in the rat (Figs. 3 and 4). The
injection in ME (case 88246), which is part of
MEC, clearly produces a narrow terminal field in a
narrow distal part of the middle one-third of the
molecular layer (Fig. 3A). In contrast, a spatially
close injection, but in VIE (case 88334), which is
part of LEC, labels a narrow terminal field in a
proximal part of the outer one-third of the molec-
ular layer (Fig. 3B). One can easily imagine that a
slightly larger injection around the border region
between VIE and ME would result in a terminal
50

Fig. 3. Radial distribution of the entorhinal-dentate projections. (A) High power photomicrograph of a small portion of the enclosed
blade of the dentate gyrus, taken from a case with an injection in MEC. Indicated is the border between the molecular layer and
lacunosum-moleculare of CA1 (white line). Arrows indicate the measures of the width of the molecular layer (a), expressed as 100%
(see also the Y-axis in D), the position of the outer border or the terminal field (b) and that of the inner border (c), as measured from
the outer border of the molecular layer (see also C). (B) High power photomicrograph of a small portion of the enclosed blade of the
dentate gyrus, taken from a case with an injection in LEC. The picture is size-matched and merged with A to illustrate the striking
difference in position of the two terminal fields (see also D). (C) Section with terminal labeling in the dentate gyrus, illustrating the
measuring protocol applied in this study. The center of the line has been derived, and equally distanced lines, perpendicular to this
center-axis, have been generated to measure the position of the terminal field as indicated in A. This procedure is carried out in all
sections with a labeled terminal field. (D) Graphic representation of four representative cases with terminal fields confined to either the
outer 35% (cases 88183R and 88334; see also Fig. 2), or the middle portion between 35 and 66.5%. Exposed and enclosed blades have
been analyzed separately in view of the overall fluctuations in density and position of the terminal fields between the two blades. Note
that two closely positioned injections on either side of the border between LEC and MEC (cases 88334 and 89246 respectively) show
completely non-overlapping, though adjacent, terminal fields. Scale bar in C, D equals 100 mm.
 51

Fig. 4. Longitudinal and transverse distribution of the entorhinal-dentate projections. (A) Unfolded representations of the dorso-
ventral extent of the dentate gyrus and terminal labeling resulting from injections in LEC (upper row) and MEC (lower row). (B)
Unfolded representation of the entorhinal cortex with injections in lateral and medial subdivisions. Case 88178 (light grey) is an
injection in layer III, illustrated in Fig. 1C. Abbreviations: AE, amygdalo-entorhinal area; LEC, lateral entorhinal cortex; MEC,
medial entorhinal cortex; PaS, parasubiculum.
52
pattern that is best described as an intermediate
pathway (see for example Fig. 2D, E). A problem
with all such types of analyses is that the data are
taken from a single level of a highly divergent
projection. Note that most studies indicated differ-
ences between exposed and enclosed blades (see
above) or between different longitudinal levels.
Therefore, we aimed to assess our terminal distri-
butions on the basis of all available data collected
throughout the entire terminal domain. As illus-
trated schematically in Fig. 3C, we calculated the
centerline of the terminal field for multiple posi-
tions and subsequently measured the radial thick-
ness of the molecular layer, taken perpendicular to
this centerline (measure a). We also measured the
outer (measure b) and inner border (measure c) of
the terminal field at all these radial points through-
out all sections of the extended dentate gyrus that
showed anterogradely labeled perforant path fib-
ers (Figs. 4 and 5). For both illustrated experi-
ments, as well for all other experiments illustrated
throughout this paper, we collected such a popu-
lation of outer and inner border measures. We
subsequently established that the outer border
measures for the medial perforant pathway where
significantly different from the inner border meas-
ures for the lateral perforant pathway (Fig. 3D;
Po0.05).
At this point, one may wonder what we know
about this issue in other species. In the mouse, no
detailed studies with respect to the radial differ-
ences in the distribution of the entorhinal projec-
tions to the dentate are available. The only
detailed account in mice seems to support a sep-
aration between fibers originating in LEC and
MEC (van Groen et al., 2003). Irrespective of their
origin in either DLE, DIE or VIE, LEC fibers ter-
minate throughout the extent of the outer one-
third of the molecular layer. Fibers arising from
MEC preferentially terminate throughout the mid-
dle one-third, although in this pathway, thinner
terminal zones have been noted, similar to what
has been observed in the rat. In contrast, the sit-
uation in the macaque monkey is in support of a
more continuous organization of the entorhinal-
dentate projection (Witter et al., 1989b; Witter and
Amaral, 1991). Irrespective of the origin in EC, at
all levels of the dentate gyrus innervated, labeling
was present throughout the extent of the outer
two-thirds of the molecular layer. Subtle indica-
tions for differences between lateral and medial
perforant pathways were observed however. Ros-
tral parts of EC, most likely homologous to LEC
of the rat, project most densely to the outer one-
third, whereas projections originating in the most
caudal parts of the monkey EC, most likely
an area comparable to MEC of the rat, showed
a preference for the middle one-third. Interest-
ingly, also the typical differentiation between the
terminal zones of the two perforant path compo-
nents, as seen with the Timm stain, is largely ab-
sent in the monkey dentate, with the exception of
the anterior genu and uncal portion (Witter and
Amaral, 1991).
Taken together, the data do not allow for a
conclusive statement whether or not the radial or-
ganization of the entorhinal-dentate projections
supports the differentiation into two discrete lat-
eral and medial components. One question is
whether such distinctions influence function. We
do know that lateral and medial entorhinal cortex
most likely convey different types of information
to the dentate molecular layer, and the most strik-
ing difference seem to relate to spatial modulation,
with the cells in the medial entorhinal cortex being
more strongly spatially modulated relative to those
in the lateral entorhinal cortex (Hafting et al.,
2005; Hargreaves et al., 2005). We also know that
these pathways are differently modulated and that
the overall postsynaptic effects are different (see
the section on synaptic organization below). More
important, however, is the conclusion that in
all species, information from functionally differ-
ent entorhinal domains converges onto a single
population of dentate granule cells (see section on
synaptic organization below).
Transverse organization in the dentate gyrus
There are conflicting papers on the transverse dis-
tribution of the perforant path projection. In ear-
lier studies no differences were reported (Hjorth-
Simonsen, 1971, 1972; Hjorth-Simonsen and
Jeune, 1972; Steward, 1976). Wyss (1981) reported
that the lateral perforant pathway preferentially
 53

Fig. 5. Representative examples of longitudinal, transverse, and radial terminal distributions of the entorhinal-dentate projections.
(A–C) Injections in LEC. (D–F) Injections in MEC. For details see text.
54
projects to the enclosed blade of the dentate gyrus,
whereas the medial component either does not
show a preference or predominantly targets the
exposed blade. It should be stressed, however, that
dimensions of injection sites may lead to different
results, similar to what was described with respect
to the radial organization. The findings of Wyss
were supported by the observation that the termi-
nal zone of LEC fibers is wider in the enclosed
blade, covering almost the entire outer one-third
of the enclosed blade, whereas the terminal field in
the exposed blade is much thinner (Tamamaki,
1997). For the MEC, the reverse appears to be the
case. Importantly, individual neurons show exten-
sive variation in their target projection, demon-
strated by the axon arbors of layer II cells that
were intracellularly labeled in vivo. An individual
layer II cell may send axon collaterals along the
entire transverse extent of the dentate gyrus, or
project preferentially to either one of the two
blades. These data are important because they
suggest that the EC projections as a whole may
provide homogeneous innervation of the dentate,
but individual components could have specific and
selective functional influence. Indeed, using the
extended preparations and subsequent unfolding,
we noticed no systematic organization in a total of
21 experiments. In most cases, when in the core of
the projection, labeling was present in both blades,
although it was unequal in density in some in-
stances (Figs. 4 and 5). In case of injections in
LEC, we generally observed, in the dorsal and
ventral domains of the terminal field, a clear pref-
erence for the enclosed blade, whereas even in the
core, labeling was denser in the enclosed blade
(Fig. 5A–C). In particular, more laterally placed
injections in LEC (cases 90170, 881813R, 89156)
showed extensive labeling in the dorsal part of the
enclosed blade, exclusively (Fig. 4, top row), in line
with the observations published by Wyss (1981).
Regarding the projections originating from MEC
(Fig. 4, lower row; Fig. 5D–F) no striking prefer-
ence for either blade was apparent although differ-
ences in density do occur between experiments
(compare Fig. 5D, E and F) as well as within ex-
periments (Fig. 5E).
In the mouse and the monkey, no transverse
organization has been described. Irrespective of
the origin of the projection, when labeling is
present in the dentate molecular layer, it appears
to be distributed equally along both blades (Witter
et al., 1989b; van Groen et al., 2003).
In conclusion, the entorhinal-dentate projection
appears to distribute along the transverse axis of
the dentate gyrus in a homogeneous fashion. Al-
though a subtle topographical arrangement cannot
be excluded, and individual cells may differ in their
particular target projection, there is no obvious
pattern. It is worth to emphasize though that con-
vergent evidence points to functional differences
between the enclosed and exposed blades of the
dentate gyrus. For example in guinea pigs, cells in
the enclosed blade as well as the associated hilar
mossy cells show morphological sex differences
whereas such differences have not been observed
in the exposed blade and associated hilar mossy
cells (Bartesaghi et al., 2003; Guidi et al., 2006).
Differences between the two blades have further
been reported with respect to sensitivity for hypo-
xia (Hara et al., 1990), neurogenesis (Choi et al.,
2003), efficacy in activating hippocampal circuits
(Scharfman et al., 2002), and loss of granule cells
following adrenalectomy (Jaarsma et al., 1992);
(see also Chawla et al., this volume).
Longitudinal organization of entorhinal-dentate
projections
Entorhinal projections show a striking organiza-
tion along the longitudinal axis of the dentate
gyrus. Originally described in the cat (Witter and
Groenewegen, 1984), it was later discovered to be
a general governing principle in all species studied
(Ruth et al., 1982, 1988; Witter et al., 1989b;
Dolorfo and Amaral, 1998; van Groen et al.,
2003). In the rat, cells located laterally in the ent-
orhinal cortex project to dorsal parts of the dent-
ate gyrus while cells located progressively more
medially project to more ventral levels of the dent-
ate gyrus (Fig. 4). This organization leads to a
pattern of connections such that the dorsal part of
the dentate gyrus receive inputs from lateral parts
of the LEC and lateral and caudal parts of MEC,
whereas the ventral portions of the dentate gyrus
receive input from more medial portions of both

LEC and MEC (Ruth et al., 1982, 1988; Witter
et al., 1989a, 2000). This topographical organization
has been convincingly demonstrated in a series of
retrograde tracing experiments, in which discrete
injections in dorsal, mid-dorsoventral, and ventral
levels of the dentate gyrus resulted in labeled pop-
ulations of entorhinal neurons, in both LEC and
MEC, with a different lateral to medial position
(Dolorfo and Amaral, 1998). Although the do-
mains of the entorhinal cells projecting to these
different longitudinal levels do not show much
overlap, one should take into account that a single
entorhinal layer II neuron, as shown with intra-
cellular filling, may distribute an axon along as
much as 2 mm (20–25%) of the dorsoventral ex-
tent (Tamamaki and Nojyo, 1993) and that re-
stricted injections in EC in almost all cases result
in labeling extending over at least one-half to up to
two-thirds of the long axis (Witter et al., 1989a;
Tamamaki, 1997), such that overlapping terminal
zones in the dentate gyrus of these populations of
layer II neurons are likely to exist. This widespread
distribution along the long axis demonstrated by
anterograde tracing seems at odds with the rather
restrictive labeling in the entorhinal cortex result-
ing from retrograde tracing methods. One may
therefore pose the question whether this longitu-
dinal topographical organization actually holds
true (Tamamaki, 1997). However, it is essential to
take into account the effect of injection size, which
quite often is not easily determined in anatomical
tracing experiments. Most data support the idea
that small foci of entorhinal layer II cells adhere to
the overall 25% longitudinal axonal distribution,
as described on the basis of single cell axonal dis-
tributions.
The unfolded approach unmasks another fea-
ture of this projection that has not been appreci-
ated before. Irrespective of the size of the injection,
the position of the origin in the entorhinal cortex,
and thus the overall terminal distribution along
the longitudinal axis, appears critical. In case of an
injection in the entorhinal cortex, which strongly
labels the extreme dorsal or ventral tips of the
dentate gyrus (Fig. 4, cases 90170, 88334, 89361,
and 88400), the terminal fields do not extend over
more than about 20–25% of the long axis. In con-
trast, following injections that produce strong
55
labeling in the intermediate levels of the hippo-
campus, the longitudinal extent is generally larger
(Fig. 4, cases 88183R, 89156, 89247, and 89246).
The longitudinal organization of the entorhinal-
dentate projection in the mouse, cat, and monkey
is strikingly comparable. In these species again,
lateral and caudal parts of the entorhinal cortex,
encompassing a band that parallels the lateral
border between entorhinal and perirhinal/parahip-
pocampal cortex and thus including parts of both
the lateral and medial entorhinal cortex preferen-
tially project to the dorsal (posterior) dentate
gyrus. Progressively more medial portions of EC,
again encompassing portions of both LEC and
MEC project to increasingly more ventral (ante-
rior) portions of the dentate gyrus (Witter et al.,
1989a, b; van Groen et al., 2003).
We may thus conclude that of all the principles
governing the organization of the perforant path-
way projection to the dentate gyrus, the longitu-
dinal one is most reliable throughout the animal
kingdom. Whether this also holds true for the hu-
man is currently unknown but it seems plausible
given the number of studies supporting functional
differences along the long axis that would be likely
to reflect this longitudinal topography. It is clear,
however, that further research is needed to fully
understand and assess the potential functional im-
plications of these findings. Complicating factors
are the strong longitudinal associational pathways
present within several of the hippocampal subfields
(Witter and Amaral, 2004). Since we do not yet
appreciate the functional relevance of those strong
longitudinal associational connections, functional
differences along the long axis may be hard to as-
sess.
Contralateral projections
In the rat, the entorhinal cortex projects to the
ipsilateral dentate gyrus, and also gives rise to a
crossed projection to the contralateral dentate
gyrus, as well as the contralateral CA3 and CA1
subfields, and the subiculum. The crossed ent-
orhinal projection is most prominent to the more
dorsal portions of the hippocampal subfields and
rapidly diminishes in density at more temporal
56
levels (Goldowitz et al., 1975; Steward, 1976).
With respect to the laminar origin of the crossed
projections, Steward and Scoville (Steward and
Scoville, 1976) reported that it matches the ipsi-
lateral origin. The crossed dentate projection, that
thus originates from layer II cells, mainly takes the
more common perforant path trajectory, i.e. cross-
ing the midline through the ventral hippocampal
commissure. In the mouse, almost no crossed pro-
jection to the dentate gyrus have been observed, in
contrast to the rat (van Groen et al., 2003). In
rabbit and cat, a distinct projection to the cont-
ralateral DG, mainly limited to the dorsal part,
has been described, similar to that in the rat
(Goldowitz et al., 1975; Hjorth-Simonsen and
Zimmer, 1975; Steward and Scoville, 1976; Wyss,
1981). In the monkey, this projection is very mod-
est, limited to the uncal/genu portion only (Witter
and Amaral, 1991).
Synaptic organization of the entorhinal-dentate
projection
In the molecular layer of the dentate gyrus in the
rat, the terminals of the perforant path fibers in the
outer two-thirds of the molecular layer make up at
least 85% of the total synaptic population (Nafstad,
1967; Matthews et al., 1976). Entorhinal fibers form
mainly, if not exclusively, asymmetric synapses
(Nafstad, 1967; Matthews et al., 1976; Deller and
Leranth, 1990; Leranth et al., 1990). These occur
most frequently on the dendritic spines of dentate
granule cells, although a small proportion of perfo-
rant path fibers terminate on non-spiny dendrites of
presumed interneurons. These include parvalbumin/
GABA-immunoreactive (Zipp et al., 1989), som-
atostatin-positive, and NPY-positive neurons with
cell bodies located in the hilus and apical dendrites
which extend into the outer portions of the molec-
ular layer (Scharfman, 1991; Soriano and Frotscher,
1993). Since at least half of the NPY-positive neu-
rons also stain for somatostatin (Swanson and
Köhler, 1986) and some co-localization of NPY and
parvalbumin has been reported (Deller and Leranth,
1990), it cannot be excluded that what may seem
as three different population of interneuronal tar-
gets, may in fact be the population of interneurons
known to distribute local axonal plexi to the per-
forant path terminal zone (Bakst et al., 1985, 1986;
Halasy and Somogyi, 1993; Boyett and Buckmaster,
2001). It is most likely that the so-called molecular
layer perforant path-associated cells (MOPP) as well
as the hilar commissural-associational pathway re-
lated cells (HICAP)(Han et al., 1993) are also
among the postsynaptic targets of entorhinal axons,
but this remains to be established. Note that in
monkeys, NPY-positive cells in the hilus do not
extend dendrites into the molecular layer (Nitsch
and Leranth, 1991). Whether this implies a species
difference in connectivity or in the expression of
certain proteins is not clear.
In addition to the main innervation arising from
layer II cells in the entorhinal cortex, a projection
originating from deep layers has been described
above (Köhler, 1985). This projection preferen-
tially distributes to the inner portion of the mo-
lecular layer, the granule cell layer, as well as the
subgranular zone, where it establishes asymmetri-
cal synapses onto granule cell dendrites as well as
on their somata and onto spine-free dendrites in
the subgranular zone. The latter most likely rep-
resent dendrites of interneurons (Deller et al.,
1996).
The observation that most of the entorhinal
synapses in the dentate gyrus are asymmetric
strongly suggests that the pathway is largely exci-
tatory, most likely using glutamate as primary
transmitter (White et al., 1977). Although not
studied in detail, prominent differences in these
respects between the lateral and medial perforant
pathway are unlikely (Nafstad, 1967; Matthews et
al., 1976), similar to observations recently reported
for the lateral and medial entorhinal projections to
the subiculum (Baks-Te-Bulte et al., 2005). How-
ever, regarding the chemical characteristics, differ-
ences between the two pathways have been
described. Terminals of the lateral perforant path-
way are also enkephalin immunoreactive, whereas
those of the medial pathway are immunoreactive
for CCK and dynorphin (Fredens et al., 1984; van
Abeelen, 1989). Medial perforant path fibers are
immunoreactive for the metabotropic glutamate
receptor, mGLUR 2/3, whereas the lateral fibers
are not. There is also convincing evidence that
neurons in LEC and MEC that project to the

dentate gyrus are markedly different with respect
to their electrophysiological properties (van Der
Linden and Lopes da Silva, 1998; Wang and Lam-
bert, 2003). Surprisingly, there is, as yet, no dis-
tinctive marker at the level of the entorhinal cortex
for cells that give rise to the lateral and medial
perforant path projections.
In view of the extensive focus on whether or not
the two pathways predominantly target different
segments of granule cell dendrites (see section on
radial organization above), there has been surpris-
ingly less interest in the question whether or not it
is a general feature for lateral and medial perforant
path fibers to converge onto a single granule cell.
Although this seems likely in view of the overall
‘‘en passant’’ type of termination of both pathways
and the fact that both indeed distribute along the
entire transverse extend of the dentate gyrus, thus
increasing the likelihood that each dentate granule
cells receives convergent input from multiple ent-
orhinal layer II cells, there is no strong anatomical
data to support the idea of convergence. In view of
the surprising observation of an apparent lack of
convergence in similarly organized pathways in the
subiculum (Cappaert et al., 2005), this issue of
convergence at the single cell level remains to be
resolved.
Summary and functional comments
The main origin of the projection from the ent-
orhinal cortex to the dentate gyrus is throughout
layer II cells in EC. Entorhinal axons of layer II
cells have their terminals predominantly, if not
exclusively, in the outer two-thirds of the molec-
ular layer of the dentate gyrus. An additional pro-
jection arises from cells in deeper layers of the
entorhinal cortex, mainly deep V and VI. In the
rat, and likely in the mouse as well, these deeply
located cells appear to be the source of most if not
all of the perforant path fibers that distribute out-
side the ‘‘traditional’’ terminal zone in the outer
molecular layer, showing a more dispersed termi-
nal distribution in the hilar region, inner molecular
and granular layers of the DG. However, these
minor projections have received little emphasis,
and remain largely unexplored, and most would
57
suggest that they play a minor role in the ent-
orhinal input to the dentate gyrus compared to the
primary projection from layer II neurons.
Although there is converging evidence that the
entorhinal-dentate projection can be subdivided
into two, functionally different systems, the precise
spatial relationship between those two systems and
the radial terminal distribution in the molecular
layer of the dentate gyrus is not entirely clear at
the present time. Although generally accepted in
rodents, and quite often used as a model system to
understand mechanisms for lamina-specific axon
outgrowth (see chapter by M. Frotscher in this
volume), this radial differentiation is less apparent
in the monkey. Taken together with the other spe-
cies differences in the two pathways discussed
above, the question can be raised whether these
species differences in the topographic organization
of the perforant path impart functional differences
to distinct species. Physiological differences would
be expected if the organization of the projections is
not the same, given the evidence that the lateral
and medial pathways differentially affect dentate
activity upon stimulation. Moreover, the effects of
either pathway on dentate activity can be manip-
ulated with a number of pharmacological tools
that appear selective for only one or the other (see
chapter by Bramham in this volume).
A critical question that seems yet to be ad-
dressed is whether or not the position of a single
entorhinal synapse along the proximo-distal extent
of the apical dendrite of a granule cells is a nec-
essary element to explain the function of that in-
put. Furthermore, it is not clear whether fibers
from LEC and MEC must target individual gran-
ule cells in specific convergent patterns in order to
transfer information from entorhinal cortex nor-
mally. In view of the primate organization, this
seems an unlikely proposition. Surprisingly, there
appears no solid anatomical evidence for the gen-
erally accepted view that all granule cells receive
convergent inputs from both pathways.
No clear conclusions can be derived with respect
to a transverse organization of this projection. From
this overview, the generally accepted view that there
is no transverse organization appears acceptable. In
contrast, the entorhinal-dentate system shows a
striking organization along the longitudinal axis of
58
the dentate gyrus such that lateral and caudal por-
tions of the entorhinal cortex project preferentially
to the dorsal (posterior in primates) part of the
dentate gyrus, whereas more medial and anterior
portions of the entorhinal cortex projects preferen-
tially to more ventral (anterior in primates) portions
of the dentate gyrus. It is of interest that the pro-
jections to the dorsal (posterior) part generally show
a much more dispersed terminal distribution, up to
almost 60% of the entire length, whereas those to
the ventral (anterior) dentate appear more restricted
to maximally 30% of the long axis. This appears to
be in correspondence to the differential longitudinal
spread of the intrinsic CA3 associative system. This
clear cut organization along the long axis has trig-
gered numerous experiments trying to relate differ-
ences in inputs to the different lateral-to-medial
bands in entorhinal cortex to a functional differen-
tiation along the long axis of the hippocampus, as
initially proposed (Witter et al., 1989a). It has been
established in rats that dorsal hippocampal lesions
selectively interfere with certain forms of spatial
learning and memory (Moser and Moser, 1998).
Likewise, lesions of the related entorhinal input
zone produces spatial deficits (Steffenach et al.,
2005). In contrast, ventral hippocampal lesions do
not result in clear spatial deficits but lead to more
motivational deficits (Kjelstrup et al., 2002) and
again comparable behavioral effects have been re-
ported to result from lesions of the corresponding
entorhinal input zone (Steffenach et al., 2005). In
non-human primates this issue has not been ad-
dressed in any detail, but human functional imaging
experiments indeed did report functional differences
along the longitudinal axis. However, these exper-
imental data mainly support differences in encoding
and retrieval processes and as such are not easily
comparable to the rodent data. In order to reconcile
this current apparent mismatch between animal and
human data, we most likely have to take the strong
intrinsic wiring of both the hippocampus and the
entorhinal cortex into account (Witter and Amaral,
2004; Witter and Moser, 2006). In particular the
intrinsic interactions of the hippocampus, both
along the transverse and the long axis may compli-
cate our understanding of this issue. In order to be
able to value the relevance of the organizational and
topographical issues described here, we have to un-
derstand the unique, and most likely complemen-
tary, contributions of each subfield of the
hippocampus and the entorhinal cortex, as well as
the roles of the longitudinal associational connec-
tions.
Acknowledgments
This paper is based on large quantities of work
carried out by many colleagues over the years. I am
greatly indebted to them and their well-prepared
publications on this subject. Moreover, for the ex-
perimental data described, I owe my former tech-
nician, Mrs. B. Jorritsma-Byham, who performed
most of the analyses, assisted by a sizable group of
under-graduate students. Dr. G. Doctor developed
the statistical routines and performed the measure-
ments on the radial position of terminal fields.
Finally, I want to thank Helen Scharfman for her
valuable discussions and contributions to the man-
uscript. This work has been supported in part by
grants from HFSPO and the Dutch Organization
for Scientific Research (NWO).
References
van Abeelen, J.H. (1989) Genetic control of hippocampal
cholinergic and dynorphinergic mechanisms regulating nov-
elty-induced exploratory behavior in house mice. Experientia,
45(9): 839–845.
Amaral, D.G., Insausti, R. and Cowan, W.M. (1987) The ent-
orhinal cortex of the monkey. I. Cytoarchitectonic organiza-
tion. J. Comp. Neurol., 264(3): 326–355.
Amaral, D.G. and Witter, M.P. (1989) The three-dimensional
organization of the hippocampal formation: a review of an-
atomical data. Neuroscience, 31(3): 571–591.
Bakst, I., Avendano, C., Morrison, J.H. and Amaral, D.G.
(1986) An experimental analysis of the origins of somatosta-
tin-like immunoreactivity in the dentate gyrus of the rat. J.
Neurosci., 6(5): 1452–1462.
Bakst, I., Morrison, J.H. and Amaral, D.G. (1985) The distri-
bution of somatostatin-like immunoreactivity in the monkey
hippocampal formation. J. Comp. Neurol., 236(4): 423–442.
Baks-Te-Bulte, L., Wouterlood, F.G., Vinkenoog, M. and Wit-
ter, M.P. (2005) Entorhinal projections terminate onto prin-
cipal neurons and interneurons in the subiculum: a
quantitative electron microscopical analysis in the rat.
Neuroscience, 136(3): 729–739.

Bartesaghi, R., Guidi, S., Severi, S., Contestabile, A. and Ciani,
E. (2003) Sex differences in the hippocampal dentate gyrus of
the guinea-pig before puberty. Neuroscience, 121(2): 327–339.
Blackstad, T.W. (1958) On the termination of some afferents to
the hippocampus and fascia dentata. An experimental study
in the rat. Acta Anat., 35: 202–214.
Boyett, J.M. and Buckmaster, P.S. (2001) Somatostatin-
immunoreactive interneurons contribute to lateral inhibitory
circuits in the dentate gyrus of control and epileptic rats.
Hippocampus, 11(4): 418–422.
Cappaert, N.L.M., Wadman, W.J. and Witter, M.P. (2005)
Spatiotemporal analyses of interactions between entorhinal
and CA1 projections to the subiculum of the rat. Soc. Ne-
urosci. Abstr., 73.13.
Choi, Y.S., Lee, M.Y., Sung, K.W., Jeong, S.W., Choi, J.S.,
Park, H.J., Kim, O.N., Lee, S.B. and Kim, S.Y. (2003) Re-
gional differences in enhanced neurogenesis in the dentate
gyrus of adult rats after transient forebrain ischemia. Mol.
Cells, 16(2): 232–238.
Chrobak, J.J. and Amaral, D.G. (2006) The entorhinal cortex
of the monkey. VII. Intrinsic connections. J. Comp. Neurol.,
500(4): 612–633.
Deller, T. (1998) The anatomical organization of the rat fascia
dentate — new aspects of laminar organization as revealed by
anterograde tracing with Phaseolus vulgaris-leucoagglutinin
(Pha-L). Anat. Embryol., 197(2): 89–103.
Deller, T. and Leranth, C. (1990) Synaptic connections of ne-
uropeptide Y (NPY) immunoreactive neurons in the hilar
area of the rat hippocampus. J. Comp. Neurol., 300(3):
433–447.
Deller, T., Martinez, A., Nitsch, R. and Frotscher, M. (1996) A
novel entorhinal projection to the rat dentate gyrus: direct
innervation of proximal dendrites and cell bodies of granule
cells and GABAergic neurons. J. Neurosci., 16(10):
3322–3333.
van Der Linden, S. and Lopes da Silva, F.H. (1998) Compar-
ison of the electrophysiology and morphology of layers III
and II neurons of the rat medial entorhinal cortex in vitro.
Eur. J. Neurosci., 10(4): 1479–1489.
Desmond, N.L., Scott, C.A., Jane, J.A. and Levy, W.B. (1994)
Ultrastructural identification of entorhinal cortical synapses
in CA1 stratum lacunosum-moleculare of the rat. Hippo-
campus, 4(5): 594–600.
Dolorfo, C.L. and Amaral, D.G. (1998) Entorhinal cortex of
the rat: topographic organization of the cells of origin of the
perforant path projection to the dentate gyrus. J. Comp.
Neurol., 398(1): 25–48.
Fredens, K., Stengaard-Pedersen, K. and Larsson, L.I. (1984)
Localization of enkephalin and cholecystokinin immunore-
activities in the perforant path terminal fields of the rat hip-
pocampal formation. Brain Res., 304(2): 255–263.
Fyhn, M., Molden, S., Witter, M.P., Moser, E.I. and Moser,
M.B. (2004) Spatial representation in the entorhinal cortex.
Science, 305(5688): 1258–1264.
Gaarskjaer, F.B. (1978) Organization of the mossy fiber system
of the rat studied in extended hippocampi. I. Terminal area
59
related to number of granule and pyramidal cells. J. Comp.
Neurol., 178(1): 49–72.
Goldowitz, D., White, W.F., Steward, O., Lynch, G. and Cot-
man, C. (1975) Anatomical evidence for a projection from the
entorhinal cortex to the contralateral dentate gyrus of the rat.
Exp. Neurol., 47(3): 433–441.
van Groen, T., Miettinen, P. and Kadish, I. (2003) The ent-
orhinal cortex of the mouse: organization of the projection to
the hippocampal formation. Hippocampus, 13(1): 133–149.
Guidi, S., Severi, S., Ciani, E. and Bartesaghi, R. (2006) Sex
differences in the hilar mossy cells of the guinea-pig before
puberty. Neuroscience, 139(2): 565–576.
Hafting, T., Fyhn, M., Molden, S., Moser, M.B. and Moser,
E.I. (2005) Microstructure of a spatial map in the entorhinal
cortex. Nature, 436(7052): 801–806.
Halasy, K. and Somogyi, P. (1993) Subdivisions in the multiple
GABAergic innervation of granule cells in the dentate gyrus
of the rat hippocampus. Eur. J. Neurosci., 5(5): 411–429.
Han, Z.S., Buhl, E.H., Lorinczi, Z. and Somogyi, P. (1993) A
high degree of spatial selectivity in the axonal and dendritic
domains of physiologically identified local-circuit neurons in
the dentate gyrus of the rat hippocampus. Eur. J. Neurosci.,
5(5): 395–410.
Hara, H., Onodera, H., Kogure, K. and Akaike, N. (1990) The
regional difference of neuronal susceptibility in the dentate
gyrus to hypoxia. Neurosci. Lett., 115(2–3): 189–194.
Hargreaves, E.L., Rao, G., Lee, I. and Knierim, J.J. (2005)
Major dissociation between medial and lateral entorhinal in-
put to dorsal hippocampus. Science, 308(5729): 1792–1794.
Haug, F.M. (1976) Sulphide silver pattern and cytoarchitec-
tonics of parahippocampal areas in the rat. Special reference
to the subdivision of area entorhinalis (area 28) and its de-
marcation from the pyriform cortex. Adv. Anat. Embryol.
Cell Biol., 52(4): 3–73.
Hevner, R.F. and Kinney, H.C. (1996) Reciprocal entorhinal-
hippocampal connections established by human fetal mid-
gestation. J. Comp. Neurol., 372(3): 384–394.
Hjorth-Simonsen, A. (1971) Hippocampal efferents to the ipsi-
lateral entorhinal area: an experimental study in the rat. J.
Comp. Neurol., 142(4): 417–437.
Hjorth-Simonsen, A. (1972) Projection of the lateral part of the
entorhinal area to the hippocampus and fascia dentata. J.
Comp. Neurol., 146(2): 219–232.
Hjorth-Simonsen, A. and Jeune, B. (1972) Origin and termina-
tion of the hippocampal perforant path in the rat studied by
silver impregnation. J. Comp. Neurol., 144(2): 215–232.
Hjorth-Simonsen, A. and Zimmer, J. (1975) Crossed pathways
from the entorhinal area to the fascia dentata. I. Normal in
rabbits. J. Comp. Neurol., 161(1): 57–70.
Hyman, B.T., Kromer, L.J. and Van Hoesen, G.W. (1988) A
direct demonstration of the perforant pathway terminal zone
in Alzheimer’s disease using the monoclonal antibody Alz-50.
Brain Res., 450(1–2): 392–397.
Hyman, B.T., Van Hoesen, G.W. and Damasio, A.R. (1987)
Alzheimer’s disease: glutamate depletion in the hippocampal
perforant pathway zone. Ann. Neurol., 22(1): 37–40.
60
Hyman, B.T., Van Hoesen, G.W., Kromer, L.J. and Damasio,
A.R. (1986) Perforant pathway changes and the memory
impairment of Alzheimer’s disease. Ann. Neurol., 20(4):
472–481.
Insausti, R., Herrero, M.T. and Witter, M.P. (1997) Entorhinal
cortex of the rat: cytoarchitectonic subdivisions and the or-
igin and distribution of cortical efferents. Hippocampus, 7(2):
146–183.
Insausti, R., Tunon, T., Sobreviela, T., Insausti, A.M. and
Gonzalo, L.M. (1995) The human entorhinal cortex: a cyto-
architectonic analysis. J. Comp. Neurol., 355(2): 171–198.
Ishizuka, N. (2001) Laminar organization of the pyramidal cell
layer of the subiculum in the rat. J. Comp. Neurol., 435(1):
89–110.
Ishizuka, N., Weber, J. and Amaral, D.G. (1990) Organization
of intrahippocampal projections originating from CA3 py-
ramidal cells in the rat. J. Comp. Neurol., 295(4): 580–623.
Jaarsma, D., Postema, F. and Korf, J. (1992) Time course and
distribution of neuronal degeneration in the dentate gyrus
of rat after adrenalectomy: A silver impregnation study.
Hippocampus, 2(2): 143–150.
Kemppainen, S., Jolkkonen, E. and Pitkanen, A. (2002) Pro-
jections from the posterior cortical nucleus of the amygdala
to the hippocampal formation and parahippocampal region
in rat. Hippocampus, 12(6): 735–755.
Kjelstrup, K.G., Tuvnes, F.A., Steffenach, H.A., Murison, R.,
Moser, E.I. and Moser, M.B. (2002) Reduced fear expression
after lesions of the ventral hippocampus. Proc. Natl. Acad.
Sci. U.S.A., 99(16): 10825–10830.
Köhler, C. (1985) A projection from the deep layers of the
entorhinal area to the hippocampal formation in the rat
brain. Neurosci. Lett., 56(1): 13–19.
Krettek, J.E. and Price, J.L. (1977) Projections from the am-
ygdaloid complex and adjacent olfactory structures to the
entorhinal cortex and to the subiculum in the rat and cat. J.
Comp. Neurol., 172(4): 723–752.
Kunzle, H. (2002) Distribution of perihippocampo-hippocampal
projection neurons in the lesser hedgehog tenrec. Neurosci.
Res., 44(4): 405–419.
Leranth, C., Malcolm, A.J. and Frotscher, M. (1990) Afferent
and efferent synaptic connections of somatostatin-immuno-
reactive neurons in the rat fascia dentata. J. Comp. Neurol.,
295(1): 111–122.
Lim, C., Blume, H.W., Madsen, J.R. and Saper, C.B. (1997)
Connections of the hippocampal formation in humans. I. The
mossy fiber pathway. J. Comp. Neurol., 385(3): 325–351.
Lorente de Nó, R. (1933) Studies on the structure of the cer-
ebral cortex. J. Psychol. Neurol., 45(6): 26–438.
Majak, K. and Pitkanen, A. (2003) Projections from the per-
iamygdaloid cortex to the amygdaloid complex, the hippo-
campal formation, and the parahippocampal region: a PHA-
L study in the rat. Hippocampus, 13(8): 922–942.
Matthews, D.A., Cotman, C. and Lynch, G. (1976) An electron
microscopic study of lesion-induced synaptogenesis in the
dentate gyrus of the adult rat. I. Magnitude and time course
of degeneration. Brain Res., 115(1): 1–21.
Morys, J., Sadowski, M., Barcikowska, M., Maciejewska, B.
and Narkiewicz, O. (1994) The second layer neurones of
the entorhinal cortex and the perforant path in physiologi-
cal ageing and Alzheimer’s disease. Acta Neurobiol. Exp.
(Wars.), 54(1): 47–53.
Moser, E., Moser, M.B. and Andersen, P. (1993) Spatial learn-
ing impairment parallels the magnitude of dorsal hippocam-
pal lesions, but is hardly present following ventral lesions. J.
Neurosci., 13(9): 3916–3925.
Moser, M.B. and Moser, E.I. (1998) Functional differentiation
in the hippocampus. Hippocampus, 8(6): 608–619.
Naber, P.A., Lopes da Silva, F.H. and Witter, M.P. (2001)
Reciprocal connections between the entorhinal cortex and
hippocampal fields CA1 and the subiculum are in register
with the projections from CA1 to the subiculum. Hippocam-
pus, 11(2): 99–104.
Nafstad, P.H. (1967) An electron microscope study on the ter-
mination of the perforant path fibres in the hippocampus and
the fascia dentata. Z. Zellforsch. Mikrosk. Anat., 76(4):
532–542.
Nitsch, R. and Leranth, C. (1991) Neuropeptide Y (NPY)-
immunoreactive neurons in the primate fascia dentata; occa-
sional coexistence with calcium-binding proteins: a light
and electron microscopic study. J. Comp. Neurol., 309(4):
430–444.
Ramón y Cajal, S. (1911) Histologie du SystemeNerveux de
l’Homme et des Vertebres. Maloine, Paris.
Ruth, R.E., Collier, T.J. and Routtenberg, A. (1982) Topog-
raphy between the entorhinal cortex and the dentate septo-
temporal axis in rats. I. Medial and intermediate entorhinal
projecting cells. J. Comp. Neurol., 209(1): 69–78.
Ruth, R.E., Collier, T.J. and Routtenberg, A. (1988) Topo-
graphical relationship between the entorhinal cortex and the
septotemporal axis of the dentate gyrus in rats. II. Cells
projecting from lateral entorhinal subdivisions. J. Comp.
Neurol., 270(4): 506–516.
Sargolini, F., Fyhn, M., Hafting, T., McNaughton, B.L., Wit-
ter, M.P., Moser, M.B. and Moser, E.I. (2006) Conjunctive
representation of position, direction, and velocity in ent-
orhinal cortex. Science, 312(5774): 758–762.
Scharfman, H.E. (1991) Dentate hilar cells with dendrites in the
molecular layer have lower thresholds for synaptic activation
by perforant path than granule cells. J. Neurosci., 11(6):
1660–1673.
Scharfman, H.E., Sollas, A.L., Smith, K.L., Jackson, M.B. and
Goodman, J.H. (2002) Structural and functional asymmetry
in the normal and epileptic rat dentate gyrus. J. Comp.
Neurol., 454(4): 424–439.
Schwartz, S.P. and Coleman, P.D. (1981) Neurons of origin of
the perforant path. Exp. Neurol., 74(1): 305–312.
Seress, L. and Mrzljak, L. (1987) Basal dendrites of granule
cells are normal features of the fetal and adult dentate gyrus
of both monkey and human hippocampal formations. Brain
Res., 405(1): 169–174.
Soriano, E. and Frotscher, M. (1993) GABAergic innervation
of the rat fascia dentata: a novel type of interneuron in the

granule cell layer with extensive axonal arborization in the
molecular layer. J. Comp. Neurol., 334(3): 385–396.
Stanfield, B.B. and Cowan, W.M. (1979) The morphology of
the hippocampus and dentate gyrus in normal and reeler
mice. J. Comp. Neurol., 185(3): 393–422.
Steffenach, H.A., Witter, M., Moser, M.B. and Moser, E.I.
(2005) Spatial memory in the rat requires the dorsolateral
band of the entorhinal cortex. Neuron, 45(2): 301–313.
Stephan, H. (1975) Allocortex. In: Borgman W. (Ed.), Handbuch
der Mikroskopischen Anatomie des Menschen. Springer-
Verlag, Berlin.
Steward, O. (1976) Topographic organization of the projections
from the entorhinal area to the hippocampal formation of the
rat. J. Comp. Neurol., 167(3): 285–314.
Steward, O. and Scoville, S.A. (1976) Cells of origin of ent-
orhinal cortical afferents to the hippocampus and fascia den-
tata of the rat. J. Comp. Neurol., 169(3): 347–370.
Swanson, L.W. and Köhler, C. (1986) Anatomical evidence for
direct projections from the entorhinal area to the entire cor-
tical mantle in the rat. J. Neurosci., 6(10): 3010–3023.
Tamamaki, N. (1997) Organization of the entorhinal projection
to the rat dentate gyrus revealed by Dil anterograde labeling.
Exp. Brain Res., 116(2): 250–258.
Tamamaki, N. and Nojyo, Y. (1993) Projection of the entorhinal
layer II neurons in the rat as revealed by intracellular pressure-
injection of neurobiotin. Hippocampus, 3(4): 471–480.
Uva, L., Gruschke, S., Biella, G., de Curtis, M. and Witter,
M.P. (2004) Cytoarchitectonic characterization of the para-
hippocampal region of the guinea pig. J. Comp. Neurol.,
474(2): 289–303.
Van Hoesen, G.W. and Pandya, D.N. (1975) Some connections
of the entorhinal (area 28) and perirhinal (area 35) cortices of
the rhesus monkey. III. Efferent connections. Brain Res.,
95(1): 39–59.
Voogd, J., Nieuwenhuis, R., van Dongen, P.A.M. and Ten
Donkelaar, H.J. (1998) Mammals. In: Dubbeldam J.L., van
Dongen P.A.M. and Voogd J. (Eds.), The Central Nervous
System of Vertebrates, Vol. 3. Springer-Verlag, Berlin,
pp. 1637–2098.
Wang, X. and Lambert, N.A. (2003) Membrane properties of
identified lateral and medial perforant pathway projection
neurons. Neuroscience, 117(2): 485–492.
White, W.F., Nadler, J.V., Hamberger, A., Cotman, C.W. and
Cummins, J.T. (1977) Glutamate as transmitter of hippo-
campal perforant path. Nature, 270(5635): 356–357.
61
Witter, M.P. (1989) Connectivity of the rat hippocampus. In:
Chan-Palay V. and Köhler C. (Eds.), The Hippocampus —
New Vistas. Allen R. Liss, New York, pp. 53–69.
Witter, M.P. (1990) Organization of the entorhinal projections
to the hippocampal formation of the rat. Soc. Neurosci.
Abstr., 16: 124.
Witter, M.P. (2006) Connections of the subiculum of the rat:
topography in relation to columnar and laminar organiza-
tion. Behav. Brain Res., 174(2): 251–264.
Witter, M.P. and Amaral, D.G. (1991) Entorhinal cortex of the
monkey. V. Projections to the dentate gyrus, hippocampus,
and subicular complex. J. Comp. Neurol., 307(3): 437–459.
Witter, M.P. and Amaral, D.G. (2004) Hippocampal forma-
tion. In: Paxinos G. (Ed.), The Rat Nervous System (3rd
Ed.). Elsevier Academic Press, San Diego, CA, pp. 635–704.
Witter, M.P. and Groenewegen, H.J. (1984) Laminar origin and
septotemporal distribution of entorhinal and perirhinal pro-
jections to the hippocampus in the cat. J. Comp. Neurol.,
224(3): 371–385.
Witter, M.P. and Groenewegen, H.J. (1990) The subiculum:
cytoarchitectonically a simple structure, but hodologically
complex. Prog. Brain Res., 83: 47–58.
Witter, M.P., Groenewegen, H.J., Lopes da Silva, F.H. and
Lohman, A.H. (1989a) Functional organization of the ex-
trinsic and intrinsic circuitry of the parahippocampal region.
Prog. Neurobiol., 33(3): 161–253.
Witter, M.P. and Moser, E.I. (2006) Spatial representation and
the architecture of the entorhinal cortex. Trends Neurosci.,
29(12): 671–678.
Witter, M.P., Van Hoesen, G.W. and Amaral, D.G. (1989b)
Topographical organization of the entorhinal projection to
the dentate gyrus of the monkey. J. Neurosci., 9(1): 216–228.
Witter, M.P., Wouterlood, F.G., Naber, P.A. and van Haeften,
T. (2000) Anatomical organization of the parahippocampal-
hippocampal network. Ann. N.Y. Acad. Sci., 911: 1–24.
Woznicka, A., Malinowska, M. and Kosmal, A. (2006) Cyto-
architectonic organization of the entorhinal cortex of the ca-
nine brain. Brain Res. Rev., 52(2): 346–367.
Wyss, J.M. (1981) An autoradiographic study of the efferent
connections of the entorhinal cortex in the rat. J. Comp.
Neurol., 199(4): 495–512.
Zipp, F., Nitsch, R., Soriano, E. and Frotscher, M. (1989)
Entorhinal fibers form synaptic contacts on parvalbumin-
immunoreactive neurons in the rat fascia dentata. Brain Res.,
495(1): 161–166.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 4

Extrinsic afferent systems to the dentate gyrus

Csaba Leranth1,2, and Tibor Hajszan1,3

1
Department of Obstetrics, Gynecology, and Reproductive Sciences, Yale University School of Medicine,
333 Cedar Street, FMB 312, New Haven, CT 06520, USA
2
Department of Neurobiology, Yale University School of Medicine, New Haven, CT 06510, USA
3
Department of Biophysics, Biological Research Center, Hungarian Academy of Sciences, H-6726 Szeged, Hungary

Abstract: The dentate gyrus is the first stage of the intrahippocampal, excitatory, trisynaptic loop, and a
primary target of the majority of entorhinal afferents that terminate in a laminar fashion on granule cell
dendrites and carry sensory information of multiple modalities about the external world. The electric
activity of the trisynaptic pathway is controlled mainly by different types of local, GABAergic interneu-
rons, and subcortical and commissural afferents. In this chapter we will outline the origin and postsynaptic
targets in the dentate gyrus of chemically identified subcortical inputs. These systems are afferents orig-
inating from the medial septum/diagonal band of Broca GABAergic and cholinergic neurons, neuro-
chemically distinct types of neurons located in the supramammillary area, serotonergic fibers from the
median raphe, noradrenergic afferents from the pontine nucleus, locus ceruleus, dopamine axons origi-
nating in the ventral tegmental area, and the commissural projection system. Because of the physiological
implications, these afferents are discussed in the context of the glutamatergic innervation of the dentate
gyrus. One common feature of the extrinsic dentate afferent systems is that they originate from a relatively
small number of neurons. However, the majority of these afferents are able to exert a powerful control over
the electrical activity of the hippocampus. This strong influence is due to the fact that the majority of the
extrinsic afferents terminate on a relatively small, but specific, populations of neurons that are able to
control large areas of the hippocampal formation.

Keywords: septum; supramammillary area; median raphe; commissural connection; acetylcholine;

dopamine; noradrenaline; serotonin

Introduction and anatomical overview
It has been known for more than 100 years that
neurons present in any cortical area are far from
being uniform regarding their morphology and in-
trinsic and extrinsic connections (Ramon y Cajal,
1893, 1911). This suggests that they are able to
interact with each other in a complex fashion.
Corresponding author. Tel.: +1 203 785 4748;
E-mail: csaba.leranth@yale.edu
Even if only a fraction of these connections are
active at the same time, a network with virtually
endless numbers of possible active patterns emerge
(Freund and Buzsaki, 1996). Thus, even if we will
be able to understand the complete molecular and
biophysical properties of a single hippocampal
neuron, in order to understand the function of the
hippocampal formation, including the dentate
gyrus, as a whole, it is indispensable to elucidate
both the intrinsic and extrinsic connections of
these cells.

DOI: 10.1016/S0079-6123(07)63004-0 63
64
The major cell types and the majority of intrin-
sic and extrinsic connections of the hippocampal
formation, which comprises the cornu ammonis,
divided into three subfields, CA1–CA3 (Lorente de
No, 1934) and the dentate gyrus (fascia dentata
and hilus), have been well known since the studies
of Ramon y Cajal (1893) and Lorente de No
(1934) and have been reviewed extensively (e.g.,
Amaral and Witter, 1989, 1995; Lopes de Silva
et al., 1990). It should be noted, however, that the
term hippocampal formation sometimes is used
to include the subicular complex and entorhinal
cortex (Amaral and Witter, 1995).
The three dimensional position of the hippo-
campal formation in the brain is rather complex.
In the majority of species, the hippocampal for-
mation is an elongated structure with its long axis
extending in a C-shaped fashion from the septal
nuclei of the basal forebrain rostrodorsally, over
and behind the diencephalon, to the incipient
temporal lobe caudoventrally (see more details in
Amaral and Witter, 1995).
Although there is still a debate about the classi-
fication of hippocampal neurons, the simple
terms of principal and non-principal neurons are
accepted (Freund and Buzsaki, 1996). The hippo-
campal principal neurons are the pyramidal and
granule cells located in the CA1–CA3 subfields of
the ammon’s horn and dentate gyrus, respectively.
In addition, the mossy cells of the dentate gyrus can
be referred to as principal cells, as discussed further
below. The granule cells and pyramidal neurons,
with their intrahippocampal connections, are the
major components of the so-called trisynaptic cir-
cuit of the hippocampus, which is unidirectional
and glutamatergic. Layer 2 pyramidal cells of the
entorhinal cortex project to granule cells (and
to some extent, to CA3 pyramidal cells) via the
perforant pathway. Granule cells, via their mossy
fiber projection, organized in a laminar fashion,
terminate on CA3 pyramidal neurons, which send
their Schaffer collaterals to the CA1 pyramidal
cells. CA1 pyramidal neurons, in turn, project back
to the entorhinal cortex, via the subiculum.
As mentioned above, the dentate gyrus is the
first stage of the intrahippocampal, excitatory,
trisynaptic loop, being the primary target of the
majority of entorhinal afferents that terminate in a
laminar fashion on granule cell dendrites
(see Chapter 1) and carry sensory information of
multiple modalities about the external world. The
principal cells of the dentate gyrus are the granule
cells,
1 million in rats and 5 million in non-
human primates (Claiborne et al., 1986; Seress,
1988), and the mossy cells (Amaral, 1978). The
small cell bodies of the granule cells (8–12 mm)
form the granule cell layer. Granule cells have two
main, radially oriented, spiny dendrites emitting
several fine branches, which reach the hippocam-
pal fissure. The axons of the granule cells, the
mossy fiber axons, originate from the opposite
pole of the soma relative to the dendrites, and
enter the dentate hilus, where they give rise to
several collaterals (Claiborne et al., 1986). Recur-
rent collaterals periodically enter the granule cell
layer, climb along the cell bodies and dendrites of
presumed basket cells, and form synapses (Ribak
and Peterson, 1991). The main axonal projection
of the granule cells leaves the hilar region and
courses through the stratum lucidum of the CA3
subfield, where it forms the giant mossy terminals
synapsing on the proximal dendrites of pyramidal
cells.
The dentate hilus is located subjacent to the
granule cell layer and extends to the border of the
dendritic layer of CA3 that is interposed between
the upper (suprapyramidal) and lower (infrapy-
ramidal) blades of the dentate gyrus. The principal
and most numerous cell type in the hilus is the
mossy cell. These neurons are characterized by
their densely spiny dendrites and several thorny
excrescences on both the cell body and proximal
dendritic shafts and their dendrites are mostly
confined to the hilus (Amaral, 1978). However,
single mossy dendrites were observed penetrating
the stratum moleculare (Scharfman, 1991; Scharf-
man, 1995; Soltesz and Mody, 1994). Axons of
mossy cells innervate the inner third of the dentate
molecular layer of both the ipsi- and contralateral
dentate gyrus and also have collaterals in the hilus
(Amaral, 1978; Laurberg and Sorensen, 1981;
Ribak et al., 1985; Buckmaster et al., 1996). These
intrahilar collaterals of mossy cells terminate on
unidentified dendrites in the hilus and on dend-
rites of hilar interneurons (Frotscher and Zimmer,
1983a, b; Frotscher et al., 1984; Ribak et al., 1985;

Buckmaster et al., 1996). Because both the CA3
pyramidal cells and mossy axons form asymmetric,
glutamatergic, excitatory synapses, many authors
consider the mossy cells as modified CA3 pyram-
idal neurons (Soriano and Frotscher, 1994; Scharf-
man, 1995, 1999). Therefore, mossy cells are
considered as excitatory principal cells that pro-
vide long-range ipsilateral and commissural pro-
jections into the dentate gyrus (Amaral, 1978).
The electric activity of the aforementioned ex-
citatory signal loop is controlled mainly by differ-
ent types of local, GABAergic interneurons (for
review, see Freund and Buzsaki, 1996) and sub-
cortical and commissural afferents. In this chapter
we will outline our recent knowledge regarding the
origin and postsynaptic targets in the dentate
gyrus of chemically identified subcortical inputs,
including afferents originating from the medial
septum/diagonal band of Broca (MSDB) GAB-
Aergic and cholinergic neurons, the dentate pro-
jection of neurochemically different types of
neurons located in the supramammillary area
(SUM), serotonergic fibers from the median rap-
he (MR), noradrenergic afferents from the pontine
nucleus, locus ceruleus, dopamine axons originat-
ing in the ventral tegmental area, and the com-
missural system.
Septo-hippocampal connections
In the 1970s, it had been shown that anterograd-
ely-transported radiolabeled amino acids, injected
into the MSDB, could be detected in hippocampal
neurons (Rose and Schubert, 1977). Studies in the
seventies and eighties, using the combination of
retrograde tracer technique and immunostaining
for choline acetyltransferase (ChAT) and gluta-
mate decarboxylase (GAD) showed that there
were two major populations of MSDB neurons
projecting to the hippocampus (for review, see
Leranth and Frotscher, 1989). Later, experiments
using anterograde labeling with the lectin PHAL
revealed two types of septohippocampal fibers.
One has large boutons that occur in clusters,
whereas the other has small boutons and arborizes
more diffusely (Nyakas et al., 1987). Type 1 axons
are immunoreactive for GABA and are processes
65
of GABAergic, parvalbumin (PA)-containing
(Freund, 1989) neurons located in the midline of
the MSDB (Kiss et al., 1990). The type 2 axons are
GABA-negative (Freund and Antal, 1988), but are
immunoreactive for ChAT, the synthesizing en-
zyme of acetylcholine (Frotscher and Leranth,
1985, 1986; Leranth and Frotscher, 1987). The
parent neurons of these axons are also in the
MSDB and are positioned in a way so that they
surround the septo-hippocampal GABAergic cells
(Kiss et al., 1990).
MSDB cholinergic innervation of the dentate gyrus
Basic and clinical studies have long recognized the
importance of cholinergic mechanisms in cognitive
function (Givens and Sarter, 1997), and drugs
which increase synaptic acetylcholine levels are
currently the most common for the treatment of
cognitive deficits associated with disorders such as
Alzheimer’s disease, albeit, with limited effective-
ness (Benzi and Moretti, 1998). The septo-hippo-
campal pathway (SH), which originates in the
MSDB and shows progressive degeneration in
Alzheimer’s disease (Whitehouse et al., 1982), has
specifically been implicated in cognitive function.
Lesions of the fimbria-fornix, which carry SH
cholinergic (Lewis and Shute, 1967) fibers to the
hippocampal formation, including the dentate
gyrus, interfere both with learning and memory
tasks and with generation of the theta rhythm in
rats (Brito and Brito, 1990). These deficits can be
attenuated by grafting acetylcholine-producing
cells to the hippocampus (Dunnett et al., 1982;
Dickinson-Anson et al., 1998).
Electrophysiological experiments have demon-
strated that excitatory, inhibitory, and disinhibi-
tory responses can be recorded as a result of the
iontophoretic application of acetylcholine in the
dentate gyrus (Wheal and Miller, 1980; Fricke and
Prince, 1984). It is not clear, however, from these
experiments whether acetylcholine exerts these
effects at the level of the primary dendrites,
directly onto the perikaryon of principal cells or
indirectly, via interneurons (Hounsgaard, 1978).
Histochemical staining procedures to visualize
putative cholinergic neurons and fibers applying
66
acetylcholinesterase (ACHE) reaction were first
used in the dentate gyrus in the early 1960’s
(Storm-Mathisen and Blackstad, 1964). These au-
thors described a laminar pattern of ACHE fiber
staining in the dentate gyrus; the densest band oc-
curring at the interface between the granule cell
layer and the molecular layer. Shute and Lewis
(1966) modified this ACHE histochemical tech-
nique to permit examination of the dentate gyrus,
stained for ACHE, using the electron microscope.
Their study revealed several neurons histochemi-
cally stained for ACHE, numerous axonal fibers
and some synaptic boutons in contact with pre-
dominantly dendritic shafts. Biochemical measure-
ments of the levels of ChAT in the dentate gyrus
also indicate a supragranular band of intense
cholinergic expression (Fonnum, 1970).
The first detailed light and electron microscopic
studies using immunostaining for ChAT on the
cholinergic innervation of the dentate gyrus were
performed in the middle 1980s (Frotscher and
Leranth, 1985, 1986; Wainer et al., 1984; Clarke,
1985; Leranth and Frotscher, 1987). The light mi-
croscopic analyses revealed a dense plexus of
ChAT-immunoreactive fibers in the dentate gyrus
forming a supragranular band at the interface be-
tween the granule cell layer and molecular layer.
Very little staining was in the granule cell layer and
only a few fibers were located in the hilar region.
In addition, ChAT-immunoreactive neurons could
also be observed in the dentate hilus (Clarke, 1985;
Frotscher et al., 1986). These ChAT-positive cells
are rare and non-principal neurons. They are rel-
atively small with round or ovoid perikarya, which
give rise to thin spine-free dendrites and are very
similar to ChAT-immunoreactive cells in the neo-
cortex of the same animals but were quite different
from cholinergic neurons in the basal forebrain,
medial septal nucleus, and neostriatum, which
were larger and more intensely immunostained.
Electron-microscopic analysis of ChAT-contain-
ing cells in the hippocampus and fascia dentata
revealed that their afferent synaptic contacts are
mainly the asymmetric type, and are located on
their cell bodies and smooth proximal dendrites.
The nuclei of the immunoreactive cells exhibited
deep indentations, which are a characteristic of
non-pyramidal neurons (Frotscher et al., 1986).
At the electron microscopic level, the vast
majority of ChAT-positive synaptic boutons con-
tacts dendritic shafts and forms predominantly
symmetric synaptic contacts. In addition, some
asymmetric synapses,
11% of the total number
of ChAT synapses (Clarke, 1985), could also be
observed on dendritic spines. There is a possibility
that axons forming symmetric and asymmetric
contacts originate from different population of
cholinergic neurons; intrinsic and extrinsic. How-
ever, it seems unlikely since, lesions of the septo-
hippocampal pathway cause an almost complete
removal of cholinergic markers in the dentate
gyrus (Mellgren and Srebro, 1973).
In the dentate gyrus, the majority of the postsy-
naptic targets of cholinergic boutons are the gran-
ule cells and the ChAT-immunoreactive boutons
form both axodendritic and axospinous synapses
with these neurons. All of the axospinous synapses
observed are asymmetric (Fig. 1). In addition,

5–10% of all postsynaptic elements of choliner-
gic axons show ultrastructural features of inter-
neurons. A combined electron microscopic, Golgi
impregnation and ChAT immunostaining study
has shown symmetric synaptic contacts between
ChAT-containing axon terminals and non-spiny,
smooth dendritic shafts, characteristic for inter-
neurons (Frotscher and Leranth, 1986; Fig. 2).
This 5–10% appears to correspond to the propor-
tion of interneurons in the neuropil, suggesting
that the interneurons are contacted in a ‘‘quasi-
random’’ fashion. Direct evidence that interneu-
rons are indeed innervated by cholinergic fibers is
provided by an electron microscopic double-
immunostaining study that showed ChAT-posi-
tive axon terminals forming symmetric synaptic
contacts with hilar GAD (glutamic acid dec-
arboxylase)- and SOM (somatostatin)-containing
neurons (Leranth and Frotscher, 1987).
In addition to the cholinergic innervation of
dentate granule cells and interneurons, MSDB
cholinergic cells innervate mossy cells densely. A
correlated light and electron microscopic double-
immunostaining study has demonstrated numerous
axosomatic synaptic contacts between ChAT-
containing axon terminals and calcitonin gene-
related peptide (CGRP)-immunoreactive neurons
in the dentate hilar area (Deller et al., 1999). CGRP
 67

Fig. 1. Low (Panel a) and high power (Panel b) electron micrographs (received from Dr. Michael Frotscher) show the result of a
combined Golgi impregnation and ChAT immunostaining experiment. On Panel a, Golgi-impregnated (gold-toned) spiny granule cell
dendrites are seen. Arrow on the same panel points at a ChAT immunoreactive bouton contacting the spine head of the dendrite (D).
Panel b shows the asymmetric synaptic contact (arrow) between the two profiles. Bar scales ¼ 1 mm.

of these neurons could modulate granule cell excit-

ability throughout large portions of the dentate
gyrus.

MSDB GABAergic innervation of the dentate gyrus

It has been shown that when the perforant path is

Fig. 2. Electron micrograph taken from the molecular layer of
the dentate gyrus immunostained for ChAT. A ChAT-immuno-
reactive axon terminal forms asymmetric synaptic contact with
a non-spiny dendritic shaft (D). The lack of dendritic spines
indicates that this dendrite is the process of an interneuron.
Asterisks label axon terminals forming asymmetric synapses
with the same dendrite. Bar scale ¼ 1 mm.
is an accepted marker for mossy cells (Freund et al.,
1997). Since mossy cells project for a long distance
along the longitudinal axis of the hippocampus
(Ribak et al., 1985), the cholinergic innervation
stimulated with a brief test pulse, it evokes exci-
tatory postsynaptic potentials (EPSPs) in the gran-
ule cell dendrites, which could be recorded
collectively in the dentate hilus as a positive-going
field potential. At greater stimulus intensities, a
negative-going component is superimposed on the
positive-going field potential, as the result of
the synchronous activation of many granule cells.
This component is referred to as a population
spike (PS). When a conditioning pulse is applied to
the medial septum just prior to a perforant path
test pulse, the EPSP remains unchanged, but the
PS is greater than that expected from the sum of
the potentials evoked by either stimulus alone
(Alvarez-Leefmans and Gardner-Medwin, 1975;
68
Fantie and Goddard, 1982). Thus, the septohip-
pocampal input appears to facilitate the discharge
of granule cells. An interesting and unexpected
finding of these studies was that stimulation of the
medial septum alone, at a site that produced the
facilitation of the PS in the dentate gyrus, did not
necessarily evoke a field potential in the dentate
gyrus. This finding, together with the failure of
medial septal stimulation to affect the perforant
path-evoked field potential, upon which the PS is
superimposed suggests that the facilitation does
not result from summation of excitatory input to
the granule cells. Two mechanisms that have been
proposed by Buzsaki (1984) to account for these
effects are: (1) medial septal input acts directly on
granule cells to facilitate activation (Robinson and
Racine, 1984), and (2) the medial septum acts in-
directly, via the inhibitory interneurons responsi-
ble for feed-forward and feed-back inhibition
(Kandel et al., 1961; Anderson and Eccles, 1962;
Mosko et al., 1973; Buzsaki and Eidelburg, 1982).
Anatomical studies of the septo-dentate termi-
nation patterns show that the innervation of the
supragranular and molecular layers are consider-
ably less dense than the innervation of the dentate
hilus and subgranular layer (Lynch et al., 1978;
Chandler and Crutcher, 1983). Although MSDB
fibers in the supragranular and molecular layers
are likely to terminate on granule cells, fibers in the
subgranular and hilar region appear to terminate
on cells having the morphological characteristics
of interneurons (Rose and Schubert, 1977;
Chandler and Crutcher, 1983). If SH fibers form
excitatory connections onto inhibitory interneu-
rons, their activation might synchronize granule
cell responses, resulting in a larger population
spike from a subsequent perforant path input
(Buzsaki, 1984). Alternatively, if they form inhib-
itory connections onto the interneurons, septal ac-
tivation might reduce both tonic and feed-forward
inhibition of the granule cells, allowing a larger
number of these neurons to be activated by the
perforant path input, again resulting in a larger
population spike. To determine the validity of
these hypotheses, Bilkey and Goddard (1985) per-
formed a very elegant study. They have shown that
a conditioning pulse to the medial septum, al-
though eliciting no field potential of its own,
facilitated the granule cell population spike evoked
by perforant path stimulation, and infusion into
the dentate hilus of a GABAA receptor antagonist,
picrotoxin, blocked the facilitation. This observa-
tion suggests that the facilitatory effect of MSDB
stimulation is mediated through an inhibitory
connection from the MSDB onto inhibitory inter-
neurons in the dentate gyrus, and that this con-
nection may utilize the neurotransmitter GABA.
Indeed, it has been demonstrated in a series
of concomitant morphological studies that SH (in-
cluding the septo-dentate) GABAergic terminals
always terminate on GABAergic interneurons, in
the rat (Freund and Antal, 1988; Gulyas et al.,
1990) and monkey hippocampus (Gulyas et al.,
1991). Experiments using double immunostaining
for PHAL and markers for different subsets of in-
terneurons have revealed that all examined sub-
populations of hippocampal GABAergic
interneurons, including those co-expressing parv-
albumin, calbindin, SOM, neuropeptide Y, chole-
cystokinin, and vasoactive intestinal polypeptide,
receive input from GABAergic septohippocampal
afferents (Freund and Antal, 1988; Gulyas et al.,
1990; Miettinen and Freund, 1992a, b; Acsady
et al., 1993). The typical innervation pattern is
multiple, ‘‘climbing fiber-like’’ contacts on the
soma, and proximal and distal dendrites of the
postsynaptic interneurons (Fig. 3). All these con-
tacts have proved to be symmetrical synapses. The
area where most of the interneurons appear to re-
ceive septal GABAergic input is the CA3 subfield,
particularly strata oriens and pyramidale (Rose
and Schubert, 1977; Nyakas et al., 1987; Freund
and Antal, 1988; Gaykema et al., 1990). The dent-
ate hilus is also heavily innervated, but in the CA1
region, a relatively smaller proportion of the inter-
neurons receive multiple synaptic inputs. No sys-
tematic differences have been observed in the
termination pattern of septal afferents between
the dorsal and ventral hippocampus.
Interactions between the septohippocampal
cholinergic and GABAergic systems
In order to better understand the function of the
septohippocampal projection system that contains

Fig. 3. Light micrograph (kindly provided by Dr. Attila
Gulyas) shows the result of a combined anterograde tracing
and calretinin immunostaining study, in the dentate gyrus. The
anterograde tracer, biotinylated dextran amine (BDA) was in-
jected into the medial septum diagonal band. Large boutons of
BDA-containing axons form multiple, basket-like, putative
synaptic contacts with the soma of calretinin-immunoreactive
(labeled with a brown diaminobenzidine reaction product) neu-
rons. One of these cells (arrows) located in the supragranular
layer (SgL) the other is at the border between the granule cell
layer (GcL) and dentate hilar area (H). Bar scale ¼ 50 mm.
(See Color Plate 4.3 in color plate section.)
a cholinergic and a GABAergic component, the
anatomical and functional interaction between
these two systems must be briefly discussed, at
the level of the MSDB.
It has to be noted that intrinsic cholinergic
mechanisms operating within the MSDB are also
critical for learning and memory. Infusion of
muscarinic agonists directly into the MSDB elicit
continuous hippocampal theta rhythm (Monmaur
and Breton, 1991; Lawson and Bland, 1993) and
facilitate learning and memory-related behaviors
both in young (Givens and Olton, 1990) and aged
rats (Markowska et al., 1995). Also, intraseptal
infusions of muscarinic agonists can alleviate
systemic scopolamine-induced amnesia, suggesting
that the MSDB is a critical locus for the mnemonic
effects of muscarinic drugs (Givens and Olton,
1995). In general, it is assumed that improvements
in MSDB-related learning and memory tasks
occur as a result of an increase in hippocampal
69
acetylcholine (ACh) release (Monmaur and
Breton, 1991; Givens and Olton, 1994, 1995;
Apartis et al., 1998; Bland and Oddie, 1998;
Dickinson-Anson et al., 1998). As such, it has
been presumed that the memory-enhancing effects
of intraseptal administration of muscarinic ago-
nists occurs due to increased firing of MSDB
cholinergic neurons (Markowska et al., 1995;
Givens and Sarter, 1997). These assumptions have
been based on earlier studies that reported a mu-
scarinic receptor-mediated increase in firing of
MSDB neurons (Dutar et al., 1983; Lamour et al.,
1984), which lead to the hypothesis that Ach, via
muscarinic receptors, has a positive feed-back
effect on MSDB cholinergic neurons (Dutar
et al., 1983; Lamour et al., 1984; see review in,
Wu et al., 2004). However, a recent combined
electrophysiological and morphological study
(Wu et al., 2004), using a novel fluorescent labe-
ling technique to selectively visualize live septo-
hippocampal cholinergic neurons has demonstrated
that administration of muscarinic agonists to the
MSDB do not excite septo-hippocampal choliner-
gic neurons, instead they inhibit a subpopula-
tion of them. In contrast, septo-hippocampal
GABAergic neurons (visualized by retrograde
tracing and parvalbumin immunostaining tech-
niques) are profoundly excited by muscarine
administration into the MSDB. Thus the cogni-
tion enhancing effects of muscarinic drugs in the
MSDB cannot be attributed to an increase in hip-
pocampal ACh release. Instead, disinhibitory
mechanisms (Freund and Antal, 1988), due to in-
creased impulse flow in the septo-hippocampal
GABAergic pathway, may underlie the cognition-
enhancing effects of muscarinic agonists. In sup-
port of this view is the morphological observation
that axon collaterals of the septo-hippocampal
cholinergic neurons heavily innervate MSDB se-
pto-hippocampal GABAergic neurons (Leranth
and Frotscher, 1989; Fig. 4). These connections
could be involved in the aforementioned process.
Supramamillo-dentate connections
It has been shown that prestimulation of sup-
ramammillary (SUM) neurons significantly
70
Fig. 4. Electron micrograph shows the result of a double
immunostaining experiment for choline acetyltransferase
(ChAT) and glutamic acid decarboxylase (GAD), in the rat
medial septum diagonal band of Broca. Immunoreactivity for
ChAT and GAD was visualized with two contrasting immuno-
marker, diaminobenzidine reaction and ferritin labeling, re-
spectively. A ChAT-immunoreactive bouton forms asymmetric
synaptic contact (arrowheads) with a GAD immunoreactive
dendrite. Bar scale ¼ 1 mm.
enhances perforant path-elicited population spikes
in the fascia dentata (Mizumori et al., 1989), an
effect that could be mimicked by glutamate injec-
tions to the lateral supramammillary area (Carre
and Harley, 1991). Moreover, Dahl and Winson
(1986) speculated about a neuronal control mech-
anism, a ‘‘gate,’’ mediated by supramammillary
afferents that would facilitate information flow in
the rat dentate gyrus, in a behavior-dependent
manner.
Since the mid-1970s, it has been known that
a projection exists between the SUM and
hippocampus in the rat (Segal and Landis, 1974).
Later, Amaral and Cowan (1980) showed that
horseradish peroxidase (HRP) injected into the
monkey hippocampus also results in labeled cells
in the SUM. Furthermore, it was noted that more
cells ipsilateral to the site of injection were labeled
than cells in the contralateral SUM and a large
number of SUM efferents terminate in the dentate
gyrus since, local application of the retrograde
tracer, Evans blue to the upper blade of the dent-
ate gyrus produced labeled cell bodies in the SUM
(Harley et al., 1983). Labeling was observed
throughout the rostrocaudal aspect of the SUM.
Therefore, Amaral and Cowan (1980) and Harley
et al. (1983) postulated that afferents from the
SUM to the hippocampus involve at least as many
cells in the SUM as the septal neurons, which give
rise to the septohippocampal projection. The spe-
cific pathway containing SUM efferents to the
hippocampus remains debatable. The medial fore-
brain bundle (Haglund et al., 1984) and fornix
(Veazey et al., 1982) have been suggested as likely
candidates. More specifically, Veazey et al. (1982)
have postulated that the medial forebrain bundle
carries SUM efferents to septal nuclei, whereas the
fornix carries projection from the SUM to the
hippocampus. At the level of dorsal hippocampus,
labeled fibers have been described in the fimbria
(Haglund et al., 1984) or the subcallosal fornix
(Wiss et al., 1979).
Regarding the neurochemical nature of the
SUM-hippocampal pathway, studies were able to
demonstrate that calretinin-immunoreactive axon
terminals in the inner molecular layer of the dent-
ate gyrus (and in the pyramidal layer of CA2)
originate in the SUM. Colocalization studies pro-
vided evidence that these large projecting neurons
contained both calretinin and substance P (SP) but
lacked GABA (Nitsch and Leranth, 1993). This
observation indicates that the nature of the hypo-
thalamo-hippocampal projection to the dentate is
likely to be excitatory. In fact, SP-containing, long
projection systems have been shown to exert ex-
citatory actions (Nicoll et al., 1980). Conversely, in
the rat, the functional properties of the hypotha-
lamo-hippocampal afferent system have been re-
ported to be at least partially inhibitory (Segal,
1979). The study by Segal (1979) raises a question
 71

of how can the physiological action on hippocam-

pal principal neurons of a putatively excitatory
pathway become inhibitory. A possible explana-
tion for this is that the supramammillo-hippocam-
pal afferents, in addition to terminating on
principal neurons (Magloczky et al., 1994), inner-
vate hippocampal GABAergic interneurons.
Indeed, in a study dealing with the innervation
of the primate hippocampal formation, thick and
varicose SP-immunoreactive axons, forming bas-
ket-like structures were identified adjacent to the
granule cell layer and in the hilar area of the dent-
ate gyrus and in the molecular layer of the middle
portion of CA3 (Nitsch and Leranth, 1994). The
location of these basket-like structures outside the
principal cell layers indicates that they contact
hippocampal non-principal neurons. Furthermore,
these SP-containing axons disappear after fimbria-
fornix transection, indicating that they are of
extrinsic origin (Nitsch and Leranth, 1994). A
subsequent study on non-human primates (Lera-
nth and Nitsch, 1994), applying correlated light
and electron microscopic immunocytochemical,
double-labeling technique for SP and parvalbu-
min and SP and calbindin and subsequent post-
embedding GABA-immunostaining revealed that
Fig. 5. Light micrograph taken from a double immunostained
this supramammillo-hippocampal afferent system vibratome section of the monkey dentate gyrus. Immunoreac-
establishes multiple, exclusively asymmetric tivity for substance P was labeled with a dark-blue Ni-di-
synapses with three specific subpopulations of aminobenzidine reaction, while immunostaining for
non-pyramidal cells: (1) a small portion of parv- parvalbumin was visualized by a brown diaminobenzidine re-
albumin-containing basket cells located in or action. The soma and dendrites of the parvalbumin-containing
cell embedded into the granule cell layer (GcL) is contacted by
adjacent to the granule cell layer of the dentate several substance P-immunoreactive axon terminals (arrows).
gyrus (Fig. 5), which, therefore, inhibit only a Bar scale ¼ 10 mm. (See Color Plate 4.5 in color plate section.)
subpopulation of granule cells; (2) some of the
calbindin-immunoreactive neurons located in the will transfer excitatory signals differently than
hilar area and in the granule cell layer (Fig. 6); and those that are controlled by a feed-forward inhib-
(3) calbindin-positive cells occurring exclusively in itory mechanism initiated by these fibers (see
the stratum moleculare of the middle portion of Fig. 10, in Leranth and Nitsch, 1994).
the CA3 subfield. Postembedding immunostaining
for GABA revealed that the aforementioned
calbindin-containing cells in area CA3 are Catecholaminergic brainstem-dentate connections
GABAergic inhibitory neurons. The results of this
study indicate that supramammillary afferents, a Serotonergic afferents
portion of which is glutamatergic (Kiss et al.,
2000) can effectively filter the information flow at The serotonergic raphe-hippocampal pathway has
different points along the trisynaptic circuit in the a powerful effect on hippocampal electric activity,
monkey hippocampal formation. Dentate granule depression-associated synaptic plasticity, and
cells, which are only stimulated by SUM afferents, cognitive behavior. Electrophysiological and
72

Fig. 6. Light (Panel a) and electron micrographs (Panels b, c, d) depicted from the monkey dentate gyrus demonstrate the result of a
correlated light and electron microscopic double immunostaining for substance P (labeled by a dark-blue Ni-diaminobenzidine
reaction) and calbindin (brown diaminobenzidine chromogen). Panel a shows a calbindin-immunoreactive neuron embedded into the
granule cell layer forming putative synaptic contacts (A–D) with substance P-containing axon terminals. Electron microscopic analysis
of ultrathin sections cut from the same area shows that boutons A and D form robust asymmetric synaptic contacts (arrowheads on
panels c and d) with the soma of this calbindin-containing cell (CB on panel b). Gc-granule cell. Bar scales ¼ panel a, 10 mm; panels
b–c, 1 mm.

pharmacological studies have shown that the typ-
ical effect of serotonin on hippocampal neurons is
a hyperpolarization evoked by an increase in K+
conductance, but depolarization and reduction of
afterhyperpolarization were also reported (for
review, see Freund et al., 1990).
The serotonergic innervation of the hippocam-
pus originates largely from the MR and a less
numerous projection arises from the dorsal raphe
(for review, see Tork, 1990) and includes two types
of fibers (Kosofsky and Molliver, 1987). In the
dentate gyrus, the most numerous are the thin ax-
ons with small evenly distributed varicosities that
are present in the subgranular area. The other fiber
type has larger boutons that form clusters along
the hilar border of the stratum granulosum and
contact secondary dendritic branches of two
cell types: GABAergic, pyramidal-shaped neurons
located subjacent to the granule cell layer, and
fusiform cells in the hilus. Both types of these
GABAergic cells seem to contain calbindin, but
none of them are immunoreactive for parvalbumin
(Freund et al., 1990; Halasy et al., 1992).
Based on these morphological data, it has been
suggested that the two types of serotonergic axons
have different mechanisms of action in the hippo-
campus. Axons with small varicosities (originating
in the dorsal raphe; Kosofsky and Molliver, 1987)
release serotonin at non-synaptic sites, diffusely,
and target cells having 5-HT1-2 receptors, to exert
a slow, tonic, G-protein-mediated action. The
other type of serotonergic axons with large
boutons (originating in the median raphe; Kosof-
sky and Molliver, 1987) always form synaptic
contacts with GABAergic interneurons that have
5-HT3 receptors (Halasy et al., 1992). Thus, stim-
ulation of median raphe serotonin neurons could
result in fast excitation of these GABAergic cells
and an enhanced GABAA receptor-mediated inhi-
bition of granule cells, a mechanism described in
the CA1 area (Roppert and Guy, 1991).
An important question is whether there is
convergence of the MSDB and MR hippocampal
inputs on the same hippocampal interneurons.
A very elegant morphological study of Miettinen
and Freund (1992a) has demonstrated that parv-
albumin-containing interneurons are innervated
by MSDB GABAergic afferents, but are avoided
73
by axons originating in the median raphe. On
the other hand, calbindin- and, to a smaller
extent, cholecystokinin-containing interneurons
are targets for both pathways. In some cases
the same individual calbindin- or cholecystokinin-
containing neurons received multiple contacts
from afferents of both MSDB and median raphe
origin. Thus, these observations indicate that
different subcortical nuclei modulate largely differ-
ent inhibitory circuits. However, considering
the occasional convergence of the two subcortical
nuclei not only onto the same type, but also onto
the same individual interneurons, the authors pro-
posed that a particular inhibitory function, most
probably feed-forward inhibition in the distal
dendritic region, is under the control of both
pathways.
It has to be noted that the MR serotonin system
could also effect the hippocampus via an indirect
route. The MR serotonergic neurons heavily in-
nervate the MSDB (Leranth and Vertes, 1999) and
exert a robust stimulatory effect on parvalbumin-
containing septohippocampal GABAergic neurons
(Alreja, 1996). These GABAergic septal cells, in
turn, selectively innervate hippocampal basket and
chandelier cells (Freund and Antal, 1988), which
are known to have powerful inhibitory effects on
the output sector (soma and axon hillock) of prin-
cipal neurons (Freund and Antal, 1988). Thus, se-
rotonergic stimulation of the septohippocampal
GABA system results in a disinhibition of princi-
pal cells. The situation is more complex, because a
population of MR serotonergic neurons project to
both the hippocampus and MSDB (e.g., McKenna
and Vertes, 2001).
Noradrenergic and dopaminergic afferents to the
dentate gyrus
Noradrenergic fibers in the hippocampus originate
from the locus ceruleus. Noradrenaline in the
dentate gyrus promotes and permits long-term
perforant path potentiation. Furthermore, phasic
locus ceruleus activation produces a delayed pro-
tein synthesis-dependent long-term potentiation of
synaptic plasticity, suggesting a selective role in
long-term memory and increases in the perforant
74
path-evoked population spike (see Chapter 10).
Potentiation of the perforant path population
spike by noradrenaline could be the result of in-
creased granule cell excitability or reduced inhibi-
tion from interneurons of granule cells, or both.
However, unit recording in the dentate gyrus
shows that exogenous noradrenaline inhibits gran-
ule cells and excites presumed inhibitory interneu-
rons (Brown et al., 2005). On the other hand,
activation of a2- or b-adrenoceptors excites both
the interneurons and granule cells (Brown et al.,
2005). Increased inhibitory interneuron activity
and inhibition of granule cells is inconsistent with
the enhanced granule cell responsiveness and en-
hanced plasticity in the dentate gyrus reported
with population recording. Thus, a critical ques-
tion concerns the action of noradrenaline on the
dentate gyrus interneurons.
In the hippocampal formation, noradrenergic
innervation is particularly dense in areas receiving
mossy fiber inputs, including the hilus of the dent-
ate gyrus and stratum lucidum of the CA3 (e.g.,
Moudy et al., 1993). In these two areas, the ma-
jority of noradrenergic varicosities do not make
conventional synaptic contacts. However, those
that form synapses terminate on dendritic shafts
and somata of GABAergic interneurons forming
symmetric membrane specializations (Frotscher
and Leranth, 1988; Milner and Bacon, 1989). It
should be noted, however, that asymmetric synap-
tic contacts between tyrosine hydroxylase-contain-
ing (presumably noradrenergic) axon terminals
and dendritic spines were also observed, in the
stratum lucidum of CA3 (Frotscher and Leranth,
1988). Symmetric (presumably inhibitory)
synapses support earlier physiological studies sug-
gesting that noradrenaline disinhibits hippocampal
pyramidal neurons by decreasing the excitability
of GABAergic interneurons (e.g., Madison and
Nicoll, 1988).
In contrast to the very dense noradrenergic in-
nervation of the dentate gyrus, this structure re-
ceives only a minor and diffusely distributed
dopaminergic projection that arises mainly from
the ventral tegmental area (Swanson, 1982). In
spite of the well-established effects of dopamine in
hippocampus-related mnemonic functions, little is
known about the mechanisms of these effects,
which are likely to reside within the hippocampus.
Most of our knowledge derives from receptor
studies, characterizing the role of D1–D5 recep-
tors. The results of these studies indicate that in
the dentate gyrus, via the aforementioned
dopamine receptor subtypes, dopamine is in-
volved in depotentiation and serves to maintain
synaptic facilitation in recently potentiated path-
ways. The net effect helps consolidate information
storage (e.g., in Manahan-Voughan and Kulla,
2003).
Commissural connections of the dentate gyrus
The majority of commissural fibers occupy the
inner molecular layer of the dentate gyrus (e.g.,
Blackstad, 1956). However, some commissural fib-
ers were described that do not follow the ‘‘classi-
cal’’ pattern of fiber lamination and terminate in
the outer molecular layer (Deller et al., 1995,
1996a, b; Deller, 1998). The main postsynaptic
targets of the commissural projection are the gran-
ule cell dendrites (e.g., Frotscher and Zimmer,
1983b; Seress and Ribak 1984), but they also
innervate interneurons (Frotscher and Zimmer,
1983a; Seress and Ribak, 1984), which were shown
to contain GAD (Frotscher et al., 1984). These
interneurons do not represent a homogeneous cell
population since, they could be distinguished by
their different neuropeptide content, e.g., vasoac-
tive intestinal polypeptide (Leranth and Frotscher,
1983), neuropeptide Y (Deller and Leranth, 1990),
and parvalbumin (Deller et al., 1994). This ar-
rangement, that commissural fibers terminate on
both principal and inhibitory interneurons in the
dentate gyrus, represents the anatomical basis of
the feed-forward inhibitory mechanism elicited by
stimulation of the commissural system (Buzsaki
and Eidelburg, 1981; Buzsaki, 1984).
Retrograde tracing experiments revealed that
the majority of the cells of origin of the commis-
sural system located in the hilus of the fascia den-
tata (e.g., Hjorth-Simonsen and Laurberg, 1977;
Berger et al., 1980) and several groups of neurons
have been identified that contribute to this fiber
system. The majority of commissural fibers are
axon collaterals of the glutamate-containing

mossy cells that are also considered as associatio-
nal/commissural neurons (Ribak et al., 1985;
Scharfman and Schwartzkroin, 1988; Scharfman,
1992, 1995; Soriano and Frotscher, 1994). The
second major population of cells projecting to the
contralateral hippocampus includes different types
of GABAergic interneurons. The first hint of a
possible contribution of GABAergic interneurons
to the commissural projection was published by
Seress and Ribak (1983). They showed that in the
hilar area, more than 60% of neurons were GAD
positive, whereas
80% of hilar neurons project
commissurally, suggesting that at least some of
the hilar projection must be GABAergic. Later,
this proposition was verified by experiments
using combination of retrograde tracing and
GAD immunochemistry, as well as anterograde
labeling and degeneration (Ribak et al., 1986).
The GABAergic commissural projection is not
homologous. Different populations of commissu-
rally projecting GABAergic cells co-express SOM
(Zimmer et al., 1983; Leranth and Frotscher, 1987),
neuropeptide Y (Deller and Leranth, 1990), parv-
albumin (Goodman and Sloviter, 1992), or other
markers (Leranth and Frotscher, 1987; Sloviter
and Nilaver, 1987). However, not all neurons of a
given cell type have a commissural collateral.
While most, if not all mossy cells appear to project
bilaterally (e.g., Frotscher et al., 1991; Scharfman,
1992), only 4–5% of the SOM- (Zimmer et al.,
1983) and only 2% of the NPY-containing
neurons (Deller and Leranth, 1990) project to the
contralateral dentate gyrus. In addition to the
aforementioned neurons, commissurally projecting
CA3c pyramidal neurons were also described
(Gottlieb and Cowan, 1973; Laurberg, 1979;
Voneida et al., 1981). These CA3c pyramids
appear to have commissural collaterals, which
arborize in the contralateral hilus, similar to their
ipsilateral collaterals (Ishizuka et al., 1990;
Li et al., 1994).
Based on the existence of a GABAergic com-
missural projection, Freund and Buzsaki (1996)
have suggested that there is a component of direct
inhibition in the feed-forward inhibitory response
evoked in the dentate gyrus by commissural stim-
ulation (Buzsaki and Eidelburg, 1981; Buzsaki,
1984).
75
Glutamatergic innervation of the dentate gyrus
During the last several decades, visualizing neu-
rons that utilize glutamate as a neurotransmitter
has proven to be a difficulty due to the lack of
specific glutamatergic markers. As a result, one of
the most influential and significant neuronal
systems has remained grossly under investigated.
At the turn of the millennium, two independent
research groups discovered that a protein that
has previously been suggested to mediate the Na-
dependent uptake of inorganic phosphate across
the plasma membrane, also transports glutamate
into synaptic vesicles (Bellocchio et al., 2000;
Takamori et al., 2000). This protein, originally
called brain-specific Na+-dependent inorganic
phosphate contransporter (BNPI) (Ni et al.,
1994), became the first of what are now called ve-
sicular glutamate transporters (VGLUT1), and
revolutionized research about the central glut-
amatergic system. Shortly after this groundbreak-
ing discovery, a second (VGLUT2) and a third
(VGLUT3) vesicular glutamate transporter has
also been identified and cloned (Aihara et al.,
2000; Bai et al., 2001; Fremeau et al., 2002; Gras
et al., 2002; Schafer et al., 2002; Takamori et al.,
2002). Since then, a vast amount of data has been
collected that has led to a better understanding of
the brain glutamatergic circuitry, including that in
the dentate gyrus.
BNPI/VGLUT1
Distribution of BNPI/VGLUT1 containing fibers
and terminals in the rat dentate gyrus has been
investigated by several groups (Bellocchio et al.,
1998; Fremeau et al., 2001; Kaneko and Fujiyama,
2002; Kaneko et al., 2002). VGLUT1 fiber density
is moderate to intense in most regions of the dent-
ate gyrus except the granule cell layer (Kaneko
et al., 2002). A more detailed description has come
from Bellocchio and colleagues (Bellocchio et al.,
1998). The outer two-thirds of the molecular layer
label more strongly for BNPI than does the inner
one-third. Further examination of the CA3 region
under higher magnification revealed coarse gran-
ular labeling at the periphery of the pyramidal cell
76
layer, strongly suggestive of mossy fiber synapses.
Electron microscopic, immunoperoxidase labeling
in stratum lucidum of CA3 showed prominent re-
action product in large axon terminals having the
morphological characteristics of mossy fiber bou-
tons. In the hilar area of the dentate gyrus, where
mossy fiber collaterals also terminate, a few large
terminals similar to those in the CA3 region
contain BNPI immunoreactivity. Many smaller
terminals in the dentate gyrus that form asymmet-
ric (Gray type I) synapses with dendritic spines
are also labeled for BNPI. Symmetric synapses do
not contain detectable BNPI, supporting a specific
role for the protein in excitatory transmission. In
addition, many terminals forming asymmetric
synapses do not contain detectable BNPI, indicat-
ing expression only in a subset of excitatory
synapses (Bellocchio et al., 1998).
The distribution of BNPI containing boutons in
the stratum moleculare suggests that they repre-
sent mainly perforant path inputs from the ent-
orhinal cortex. Indeed, hybridization with the
BNPI probe resulted in a strong hybridization sig-
nal in neuron-enriched regions of the entorhinal
cortex (Ni et al., 1995). On the other hand, the
BNPI positivity of mossy fiber-like terminals in the
hilus and CA3 suggests that they originate from
granule cells. Similar to cell bodies elsewhere in the
brain, the somata of dentate gyrus granule
cells show no detectable BNPI immunoreactivity
(Bellocchio et al., 1998). By contrast, a strong
BNPI hybridization signal has been observed in
the pyramidal neurons of the hippocampus and
granule cells of the dentate gyrus, suggesting that
these cells synthesize the protein and then trans-
port it to their terminals (Ni et al., 1994, 1995;
Fremeau et al., 2001; Herzog et al., 2001). In con-
clusion, BNPI/VGLUT1 appears to be the pri-
mary glutamatergic marker of principal neurons,
including dentate granule cells (Varoqui et al.,
2002).
DNPI/VGLUT2
Prior to its identification as VGLUT2 (Aihara
et al., 2000), this protein has been known as differ-
entiation-associated Na+-dependent inorganic
phosphate cotransporter (DNPI), having similar
functions as BNPI (Hisano et al., 2000). In a
rather comprehensive study, VGLUT2-positive
fibers and terminals in the rat hippocampus have
been characterized, and their possible sources of
origin defined (Halasy et al., 2004). The highest
density of VGLUT2-immunoreactive boutons has
been observed in the inner molecular layer, while
only scattered VGLUT2 positive fibers have been
detected in the other areas of the dentate gyrus,
such as the stratum moleculare and the hilus. In
general, VGLUT2-immunoreactive boutons estab-
lish exclusively asymmetric synapses. Analysis of
VGLUT2-positive synaptic boutons revealed that
the majority of postsynaptic targets in the dentate
gyrus are dendritic spines, followed by dendritic
shafts and granule cell somata (Halasy et al.,
2004). These findings are consistent with the re-
sults of previous studies in the rat (Fremeau et al.,
2001; Kaneko and Fujiyama, 2002; Kaneko et al.,
2002; Varoqui et al., 2002).
Because only low levels of VGLUT2 mRNA
expression have been observed within the hippo-
campus of rats (Hisano et al., 2000; Fremeau et al.,
2001; Herzog et al., 2001), and the granule cell
layer of the dentate gyrus shows no signal for
DNPI by in situ hybridization (Fremeau et al.,
2001), dentate VGLUT2 positive boutons are
likely to be of extrahippocampal origin. It has to
be noted, however, that VGLUT2 mRNA in mice
is concentrated in pyramidal neurons of the hip-
pocampus (Bai et al., 2001), indicating the poten-
tial for considerable species differences in the
hippocampal expression of VGLUT2. In rats,
Fremeau and colleagues suggest a hypothalamic
origin for the supragranular VGLUT2-positive
fiber cluster because cells in the hypothalamus that
project to this layer strongly express DNPI mRNA
(Fremeau et al., 2001). In addition to the hypo-
thalamus, VGLUT2 has been found in most septal
neurons (Hajszan et al., 2004), and several of these
neurons are also positive for GABAergic or
cholinergic markers (Danik et al., 2005). How-
ever, deafferentation studies (fimbria-fornix tran-
section and entorhinal cortex ablation) led to no
significant differences in either the density or dis-
tribution pattern of VGLUT2-positive boutons in
the dentate gyrus (Halasy et al., 2004), suggesting

that the majority of dental VGLUT2 boutons are
of intrahippocampal origin. Indeed, light micro-
scopic observation of the hippocampus of col-
chicine-treated rats revealed a large number of
VGLUT2-immunoreactive cell bodies. In the dent-
ate gyrus, the hilar mossy cells represent the most
heavily labeled population, while a small propor-
tion of granule cells in the subgranular zone are
also VGLUT2-positive (Halasy et al., 2004). It is
well known that mossy cells in the dentate gyrus
project to the inner molecular layer (Frotscher
et al., 1991). Thus, VGLUT2-containing hilar
mossy cells may be the source of the VGLUT2
fiber network in the dentate gyrus.
VGLUT3
VGLUT3 is the only known vesicular glutamate
transporter that began its ‘‘career’’ without previ-
ous history as inorganic phosphate cotransporter.
It has been identified in a direct search for addi-
tional VGLUT molecules, and cloned (Fremeau
et al., 2002; Schafer et al., 2002; Takamori et al.,
2002). VGLUT3-immunoreactive terminals sur-
round the granule cell layer of the rat dentate
gyrus. VGLUT3 labeling is intense in the molec-
ular layer, corresponding to the proximal part of
the granule cell dendrites, and along the hilar bor-
der with the granule cell layer, corresponding to
the zone where granule cell axons emerge. A dense
VGLUT3 network also surrounds the soma of
granule cells (Fremeau et al., 2002; Gras et al.,
2002; Herzog et al., 2004). A similar distribution
pattern has been observed in mice (Schafer et al.,
2002). In rats, immunoparticles for VGLUT3 ac-
cumulate over vesicle clusters in terminals making
classical asymmetrical synapses in the hippocam-
pus (Gras et al., 2002). However, a large number
of VGLUT3-positive terminals also form symmet-
rical synapses (Fremeau et al., 2002; Gras et al.,
2002). Further, the labeled terminals make contact
with the shaft of proximal dendrites, a location
characteristic of inhibitory synapses. However,
only a subset of symmetric synapses in the hippo-
campus appears to stain for VGLUT3 (Fremeau
et al., 2002).
77
Regarding the source of dentate VGLUT3 fib-
ers, one possibility is an intrahippocampal origin.
Although moderate levels of VGLUT3 mRNA
expression have been observed in the principal
cells of pyramidal and dentate granule cell layers
in the rat (Fremeau et al., 2002), the distribution
of VGLUT3-immunoreactive fibers resembles
that observed with markers of GABA terminals.
Indeed, the VGLUT3 gene is expressed in scat-
tered interneurons in the hilus of the dentate gyrus
(Fremeau et al., 2002; Gras et al., 2002; Herzog et
al., 2004). A similar distribution pattern of
VGLUT3 mRNA has been demonstrated in mice
(Schafer et al., 2002). Immunoreactivity for
VGLUT3 is undetectable in pyramidal and dent-
ate granule cells (Fremeau et al., 2002; Somogyi
et al., 2004). More importantly, a partial colocal-
ization of VGLUT3 and the GABA marker glut-
amic acid decarboxylase (GAD) has been observed
in the hilar interneurons of the dentate gyrus by
means of both in situ hybridization (Herzog et al.,
2004) and immunohistochemistry (Fremeau et al.,
2002). In addition, colocalization of VGLUT3
with vesicular inhibitory amino acid transporter,
a marker of GABAergic neurons and boutons
revealed numerous double-labeled nerve endings
in the perisomatic terminals that contact the soma
of granule cells (Herzog et al., 2004). More spe-
cifically, Somogyi and colleagues (Somogyi et al.,
2004) have shown that all VGLUT3-positive so-
mata are immunoreactive for cholecystokinin, a
marker of a subpopulation of basket cells, and
none express markers for other interneuron types
in the hippocampus. Boutons expressing
VGLUT3, cholecystokinin and GAD are most
abundant in the cell layers of the hippocampus
(Somogyi et al., 2004).
Another source of dentate VGLUT3 fibers may
be the subcortical neurons, such as the septohip-
pocampal neurons, the mesolimbic dopaminergic,
and raphe serotonergic cells. However, the data
obtained so far are controversial. In the rat, the
substantia nigra pars compacta and ventral teg-
mental area contain moderate levels of VGLUT3
mRNA, suggesting expression by dopamine neu-
rons (Fremeau et al., 2002). By contrast, no
VGLUT3 mRNA expression has been detected
in the substantia nigra by Herzog and colleagues
78
(Herzog et al., 2004), and lack of VGLUT3 ex-
pressing neurons has also been observed in the
septum and the vertical limb of the diagonal band
in the same study. Furthermore, a very high den-
sity of VGLUT3 mRNA has been found in the
dorsal raphe and MR nuclei (Fremeau et al., 2002;
Herzog et al., 2004). All of the serotonin trans-
porter-positive neurons also express VGLUT3 in
the dorsal raphe and MR. Interestingly, numerous
neurons from the dorsal raphe express VGLUT3
but no serotonin transporter (Gras et al., 2002).
However, double-staining for VGLUT3 and the
catecholamine marker tyrosine hydroxylase or the
plasma membrane serotonin transporter revealed
no colocalization in the axons of the dentate
gyrus (Fremeau et al., 2002). On the other hand,
Somogyi and colleagues (Somogyi et al., 2004)
have found a population of VGLUT3 boutons in
the hippocampus that are negative for GAD but
are labeled for vesicular monoamine transporter
type 2 (VMAT2), plasmalemmal serotonin trans-
porter or serotonin, but no colocalization has been
found in terminals containing VGLUT3 and ve-
sicular acetylcholine transporter. In mice, the
highest VGLUT3 mRNA expression levels and
highest density of VGLUT3 mRNA-expressing
cells have been observed in the raphe nuclei. In
contrast to its mRNA, VGLUT3 immunoreactiv-
ity is undetectable in the cell bodies of raphe nuclei
in vivo (Schafer et al., 2002). In the hippocampus,
VGLUT3 immunoreactivity is absent from choli-
nergic synapses. Confocal double-immunofluores-
cence revealed the presence of VGLUT3 in a
subpopulation of both thick and thin VMAT2-
positive varicose fibers in the hippocampus, as well
as the absence of VGLUT3 from tyrosine hydro-
xylase positive terminals. Thus, the hippocampal
VGLUT3/VMAT2 system most likely stems from
the dorsal raphe and MR nuclei and represents a
novel neuronal projection subsystem encoding
both glutamatergic and serotonergic neurotrans-
mission in the mouse (Schafer et al., 2002).
Conclusion
One common feature of the extrinsic dentate affer-
ent systems is that they originate from a relatively
small number of neurons. However, the majority of
these afferents are able to exert a powerful control
over the electric activity of the hippocampus. In
some of the afferent systems, this efficacy is due to
the fact that the majority of the extrinsic afferents
terminate on a relatively small, but specific popu-
lations of neurons. These target cells include differ-
ent types of GABAergic interneurons, which, in
turn have the ability to control a large number of
principal cells (Buzsaki, 1984; Freund and Buzsaki,
1996). For example, The MSDB GABAergic
neurons seem to innervate only the GABAergic
chandelier and basket cells that have a great influ-
ence (disinhibition) on the output sector, soma and
axon hillock, of granule cells (Freund and Antal,
1988). Serotonergic fibers originating in the MR,
also selectively innervate a specific population of
calbindin-containing GABA cells that terminate
on the dendritic shafts of principal cells. Thus, in
contrast to the disinhibitory function of the MSDB-
hippocampal GABA system, activation of the
MR-hippocampal serotonergic pathway could re-
sult in inhibition of the input of principal neurons
(Freund et al., 1990). The termination pattern of
SUM afferents also seems to be very specific. The
overwhelming majority of fibers originating in the
SUM form robust asymmetric synaptic contacts
with the primary dendrites of granule cells (Leranth
and Nitsch, 1994; Magloczky et al., 1994; Nitsch
and Leranth, 1996) and, in addition, terminate on
specifically positioned GABA interneurons, in a
way that they could effectively filter/contrast the
signal flow in the trisynaptic circuit (Leranth and
Nitsch, 1994).
The functional importance of these three major
extrahippocampal afferent systems is highlighted
by recent observations related to the mnemonic
functions of the hippocampus that are associated
with synaptoplastic effects of gonadal hormones.
Local estrogen administration into the MSDB
(Lâm and Leranth, 2003), MR (Prange-Kiel et al.,
2004), and SUM (Leranth and Shanabrough, 2001)
of ovariectomized rats all resulted in a robust in-
crease in the density of spine synapses in the hip-
pocampus. In contrast, transection of the fimbria
fornix, which contains the hippocampal projection
of these subcortical structures, prevents (Leranth
et al., 2000) the known effects (Gould et al., 1990).

Acknowledgments
This work was supported by NIH grants
MH060858 and NS042644 to C.L. and
MH074021 to T.H., as well as by a Hungarian
National Office for Research and Technology
grant RET-08/04.
References
Acsady, L., Halasy, K. and Freund, T.F. (1993) Calretinin is
present in non-pyramidal cells of the rat hippocampus–III.
Their inputs from the median raphe and medial septal nuclei.
Neuroscience, 52: 829–841.
Aihara, Y., Mashima, H., Onda, H., Hisano, S., Kasuya, H.,
Hori, T., Yamada, S., Tomura, H., Yamada, Y., Inoue, I.,
Kojima, I. and Takeda, J. (2000) Molecular cloning of a
novel brain-type Na(+)-dependent inorganic phosphate co-
transporter. J. Neurochem., 74(6): 2622–2625.
Alreja, M. (1996) Excitatory actions of serotonin on GAB-
Aergic neurons of the medial septum diagonal band of Broca.
Synapse, 22: 15–27.
Alvarez-Leefmans, F.J. and Gardner-Medwin, A.R. (1975) In-
fluences of septum on the hippocampal dental area which are
unaccompanied by field potentials. J. Physiol. (Lond.), 249:
14–16.
Amaral, D.G. (1978) A Golgi study of cell types in the hilar
region of the hippocampus in the rat. J. Comp. Neurol., 182:
851–914.
Amaral, D.G. and Cowan, W.M. (1980) Subcortical afferents
to the hippocampal formation in the monkey. J. Comp.
Neurol., 189: 573–591.
Amaral, D.G. and Witter, M.P. (1989) The three-dimensional
organization of the hippocampal formation: a review of
anatomical data. Neuroscience, 31: 571–591.
Amaral, D.G. and Witter, M.P. (1995) Hippocampal forma-
tion. In: Paxinos G. (Ed.), The rat nervous system (2nd ed.).
Academic Press, New York, pp. 443–494.
Anderson, P. and Eccles, J.C. (1962) Inhibitory phasing of
neuronal discharge. Nature (London), 196: 645–647.
Apartis, E., Poindessous-Jazat, F.R., Lamour, Y.A. and Bass-
ant, M.H. (1998) Loss of rhythmically bursting neurons in rat
medial septum following selective lesion of septohippocampal
cholinergic system. J. Neurophysiol., 79: 1633–1642.
Bai, L., Xu, H., Collins, J.F. and Ghishan, F.K. (2001)
Molecular and functional analysis of a novel neuronal vesic-
ular glutamate transporter. J. Biol. Chem., 276(39):
36764–36769.
Bellocchio, E.E., Hu, H., Pohorille, A., Chan, J., Pickel, V.M.
and Edwards, R.H. (1998) The localization of the brain-spe-
cific inorganic phosphate transporter suggests a specific pre-
synaptic role in glutamatergic transmission. J. Neurosci.,
18(21): 8648–8659.
Bellocchio, E.E., Reimer, R.J., Fremeau, R.T.J. and Edwards,
R.H. (2000) Uptake of glutamate into synaptic vesicles by an
79
inorganic phosphate transporter. Science, 289(5481):
957–960.
Benzi, G. and Moretti, A. (1998) Is there a rationale for the
use of acetylcholinesterase inhibitors in the therapy of
Alzheimer’s disease? Eur. J. Pharmacol., 346: 1–13.
Berger, T.W., Semple-Rowland, S. and Basset, J. (1980) Hip-
pocampal polymorph neurons are the cells of origin for ipsi-
lateral association and commissural afferents to the dentate
gyrus. Brain Res., 215: 329–336.
Bilkey, D.K. and Goddard, G.V. (1985) Medial septal facili-
tation of hippocampal granule cell activity is mediated by
inhibition of inhibitory interneurons. Brain Res., 361:
99–106.
Blackstad, T.W. (1956) Commissural connections of the hip-
pocampal region of the rat, with special reference to their
mode of termination. J. Comp. Neurol., 105: 417–537.
Bland, B.H. and Oddie, S.D. (1998) Anatomical, electrophys-
iological and pharmacological studies of ascending brainstem
hippocampal synchronizing pathways. Neurosci. Biobehav.
Rev., 22: 259–273.
Brito, G.N. and Brito, L.S. (1990) Septohippocampal system
and the prelimbic sector of frontal cortex: a neuropsycho-
logical battery analysis in the rat. Behav. Brain Res., 36:
127–146.
Brown, R.A., Walling, S.G., Milway, J.S. and Harley, C.W.
(2005) Locus ceruleus activation suppresses feedforward in-
terneurons and reduces beta-gamma electroencephalogram
frequencies while it enhances theta frequencies in rat dentate
gyrus. J. Neurosci., 25: 1985–1991.
Buckmaster, P.S., Wentzel, H.J., Kunkel, D.D. and Schwa-
rtzkroin, P.A. (1996) Axon arbors and synaptic connections
of hippocampal mossy cells in the rat in vivo. J. Comp.
Neurol., 366: 270–292.
Buzsaki, G. (1984) Feed-forward inhibition in the hippocampal
formation. Progr. Neurobiol., 22: 131–153.
Buzsaki, G. and Eidelburg, E. (1981) Commissural projection
to the dentate gyrus of the rat: evidence for feed-forward
inhibition. Brain Res., 230: 346–350.
Buzsaki, G. and Eidelburg, E. (1982) Direct afferent excitation
and long-term potentiation of hippocampal interneurons. J.
Neurophysiol., 48: 597–607.
Carre, G.P. and Harley, C.W. (1991) Population spike facili-
tation in the dentate gyrus following glutamate administra-
tion to the lateral supramammillary nucleus. Brain Res., 568:
307–310.
Chandler, J.P. and Crutcher, K.A. (1983) The septohippocam-
pal projection in the rat: an electron microscopic horseradish
peroxidase study. Neuroscience, 10: 685–696.
Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1986) A light
and electron microscopic analysis of the mossy fibers of the
rat dentate gyrus. J. Comp. Neurol., 246: 435–458.
Clarke, D.J. (1985) Cholinergic innervation of the rat dentate
gyrus: an immunocytochemical and electron microscopical
study. Brain Res., 360: 349–354.
Dahl, D. and Winson, J. (1986) Influence of neurons of the
parafascicular region on neuronal transmission from perfo-
rant path through dentate gyrus. Brain Res., 377: 211–219.
80
Danik, M., Cassoly, E., Manseau, F., Sotty, F., Mouginot, D.
and Williams, S. (2005) Frequent coexpression of the vesic-
ular glutamate transporter 1 and 2 genes, as well as coex-
pression with genes for choline acetyltransferase or glutamic
acid decarboxylase in neurons of rat brain. J. Neurosci. Res.,
81(4): 506–521.
Deller, T. (1998) The anatomical organization of the rat fascia
dentata: new aspects of laminar organization as revealed by
anterograde tracing with phaseolus vulgaris-leucoagglutinin.
Anat. Embryol., 197: 89–103.
Deller, T., Katona, I., Cozzari, C., Frotscher, M. and Freund,
T.F. (1999) Cholinergic innervation of mossy cells in the rat
fascia dentata. Hippocampus, 9: 314–320.
Deller, T. and Leranth, C. (1990) Synaptic connections of NPY-
immunoreactive neurons in the rat hilar area. J. Comp.
Neurol., 300: 433–447.
Deller, T., Martinez, A., Nitsch, R. and Frotscher, M. (1996a)
A novel entorhinal projection to the rat dentate gyrus: direct
innervation of proximal dendrites and granule cell bodies of
granule cells and GABAergic neurons. J. Neurosci., 16:
3322–3333.
Deller, T., Nitsch, R. and Frotscher, M. (1994) Associational
and commissural afferents of parvalbumin-immunoreactive
neurons in the rat hippocampus: a combined immunocyto-
chemical and PHAL study. J. Comp. Neurol., 350: 612–622.
Deller, T., Nitsch, R. and Frotscher, M. (1995) Phaseolus vul-
garis- Leucoagglutinin (PHAL) tracing of commissural fibers
to the rat fascia dentata: evidence for a previously unknown
commissural projection to the outer molecular layer. J.
Comp. Neurol., 352: 55–68.
Deller, T., Nitsch, R. and Frotscher, M. (1996b) Heterogeneity
of the commissural projection to the rat dentate gyrus: a
Phaseolus vulgaris- Leucoagglutinin tracing study. Neurosci-
ence, 75: 111–121.
Dickinson-Anson, H., Aubert, I., Gage, F.H. and Fisher, L.J.
(1998) Hippocampal grafts of acetylcholine-producing cells
are sufficient to improve behavioral performance following a
unilateral fimbria-fornix lesion. Neuroscience, 84: 771–781.
Dunnett, S.B., Low, W.C., Iversen, S.D., Stenevi, U. and
Bjorklund, A. (1982) Septal transplants restore maze learning
in rats with fornix-fimbria lesions. Brain Res., 251: 335–348.
Dutar, P., Lamour, Y. and Jobert, A. (1983) Acetylcholine
excites identified septo-hippocampal neurons in the rat.
Neurosci. Lett., 43: 43–47.
Fantie, B.D. and Goddard, G.V. (1982) Septal modulation of
the population spike in the fascia dentata produced by
perforant path stimulation in the rat. Brain Res., 252:
227–237.
Fonnum, F. (1970) Topographical and subcellular localization
of choline acetyltransferase in rat hippocampal formation. J.
Neurochem., 17: 1029–1037.
Fremeau, R.T., Burman, J., Qureshi, T., Tran, C.H., Proctor,
J., Johnson, J., Zhang, H., Sulzer, D., Copenhagen, D.R.,
Storm-Mathisen, J., Reimer, R.J., Chaudhry, F.A. and Ed-
wards, R.H. (2002) The identification of vesicular glutamate
transporter 3 suggests novel modes of signaling by glutamate.
Proc. Natl. Acad. Sci. U.S.A., 99: 14488–14493.
Fremeau, R.T., Troyer, M.D., Pahner, I., Nygaard, G.O.,
Tran, C.H., Reimer, R.J., Bellocchio, E.E., Fortin, D.,
Storm-Mathisen, J. and Edwards, R.H. (2001) The expres-
sion of vesicular glutamate transporters defines two classes of
excitatory synapse. Neuron, 31: 247–260.
Freund, T.F. (1989) GABAergic septohippocampal neurons
contain parvalbumin. Brain Res., 478: 375–381.
Freund, T.F. and Antal, M. (1988) GABA-containing neurons
in the septum control inhibitory interneurons in the hippo-
campus. Nature, 336: 170–173.
Freund, T.F. and Buzsaki, G. (1996) Interneurons of the hip-
pocampus. Hippocampus, 6: 347–470.
Freund, T.F., Gulyas, A.I., Acsady, L., Gorcs, T. and Toth, K.
(1990) Serotonergic control of the hippocampus via local in-
hibitory interneurons. Proc. Natl. Acad. Sci. U.S.A., 87:
8501–8505.
Freund, T.F., Hajos, N., Acsady, L., Gorcs, T.J. and Katona, I.
(1997) Mossy cells of the rat dentate gyrus are immunore-
active for calcitonin gene-related peptide (CGRP). Eur. J.
Neurosci., 9: 1815–1830.
Fricke, R.A. and Prince, D.A. (1984) Electrophysiology of
dentate gyrus granule cells. J. Neurophysiol., 51: 195–209.
Frotscher, M. and Leranth, C. (1985) Cholinergic innerva-
tion of the rat hippocampus as revealed by choline ace-
tyltransferase immunocytochemistry: A combined light and
electron microscopic study. J. Comp. Neurol., 239:
237–246.
Frotscher, M. and Leranth, C. (1986) The cholinergic innerva-
tion of the rat fascia dentata: identification of target struc-
tures on granule cells by combining choline acetyltransferase
immunocytochemistry and Golgi impregnation. J. Comp.
Neurol., 243: 58–70.
Frotscher, M. and Leranth, C. (1988) Catecholaminergic in-
nervation of pyramidal and GABAergic nonpyramidal neu-
rons in the rat hippocampus. Double label immunostaining
with antibodies against tyrosine hydroxylase and glutamate
decarboxylase. Histochemistry, 88: 313–319.
Frotscher, M., Leranth, C., Lubbers, K. and Oertel, W.H.
(1984) Commissural afferents innervate glutamate dec-
arboxylase immunoreactive nonpyramidal neurons in the
guinea pig hippocampus. Neurosci. Lett., 46: 137–143.
Frotscher, M., Schlander, M. and Leranth, C. (1986) Choli-
nergic neurons in the hippocampus. A combined light- and
electron-microscopic immunocytochemical study in the rat.
Cell Tissue Res., 246: 293–301.
Frotscher, M., Seress, L., Schwerdtfeger, W.K. and Buhl, E.
(1991) The mossy cells of the fascia dentata: a comparative
study of their fine structure and synaptic connections in
rodents and primates. J. Comp. Neurol., 312: 145–163.
Frotscher, M. and Zimmer, J. (1983a) Commissural fibers ter-
minate on nonpyramidal neurons in the guinea pig hippo-
campus-a combined Golgi/EM degeneration study. Brain
Res., 265: 289–293.
Frotscher, M. and Zimmer, J. (1983b) Lesion-induced
mossy fibers to the inner molecular layer of the rat fascia
dentata: identification of postsynaptic granule cells by the
Golgi-EM technique. J. Comp. Neurol., 336: 170–173.

Gaykema, R.P., Luiten, P.G., Nyakas, C. and Traber, J. (1990)
Cortical projection patterns of the medial septum-diagonal
band complex. J. Comp. Neurol., 293: 103–124.
Givens, B.S. and Olton, D.S. (1990) Cholinergic and
GABAergic modulation of medial septal area: effect on
working memory. Behav. Neurosci., 104: 849–855.
Givens, B.S. and Olton, D.S. (1994) Local modulation of basal
forebrain: effects on working and reference memory. J.
Neurosci., 14: 3578–3587.
Givens, B.S. and Olton, D.S. (1995) Bidirectional modulation
of scopolamine-induced working memory impairments by
muscarinic activation of the medial septal area. Neurobiol.
Learn. Mem., 63: 269–276.
Givens, B. and Sarter, M. (1997) Modulation of cognitive
processes by transsynaptic activation of the basal forebrain.
Behav. Brain Res., 84: 1–22.
Goodman, J.H. and Sloviter, R.S. (1992) Evidence for com-
missurally projecting parvalbumin-immunoreactive basket
cells in the dentate gyrus of the rat. Hippocampus, 2: 13–22.
Gottlieb, D.I. and Cowan, W.M. (1973) Autoradiographic
studies of the commissural and ipsilateral association con-
nections of the hippocampus and dentate gyrus of the rat. I.
The commissural connections. J. Comp. Neurol., 149:
393–422.
Gould, E., Woolley, C.S., Frankfurt, M. and McEwen, B.S.
(1990) Gonadal steroids regulate dendritic spine density in
hippocampal pyramidal cells in adulthood. J. Neurosci., 10:
1286–1291.
Gras, C., Herzog, E., Bellenchi, G.C., Bernard, V., Ravassard,
P., Pohl, M., Gasnier, B., Giros, B. and El Mestikawy, S.
(2002) A third vesicular glutamate transporter expressed by
cholinergic and serotoninergic neurons. J. Neurosci., 22:
5442–5451.
Gulyas, A.I., Gorcs, T.J. and Freund, T.F. (1990) Innervation
of different peptide-containing neurons in the hippocampus
by GABAergic septal afferents. Neuroscience, 37: 31–44.
Gulyas, A.I., Seress, L., Toth, K., Acsady, L., Antal, M. and
Freund, T.F. (1991) Septal GABAergic neurons innervate
inhibitory interneurons in the hippocampus of the macaque
monkey. Neuroscience, 41: 381–390.
Haglund, L., Swanson, L.W. and Kohler, C. (1984) The pro-
jection of the supramammillary nucleus to the hippocampal
formation: an immunohistochemical and anterograde trans-
port study with the lectin PHA-L in the rat. J. Comp.
Neurol., 229: 171–185.
Hajszan, T., Alreja, M. and Leranth, C. (2004) Intrinsic vesic-
ular glutamate transporter 2-immunoreactive input to septo-
hippocampal parvalbumin-containing neurons: novel
glutamatergic local circuit cells. Hippocampus, 14: 499–509.
Halasy, K., Hajszan, T., Kovacs, E.G., Lam, T.T. and Leranth,
C. (2004) Distribution and origin of vesicular glutamate
transporter 2-immunoreactive fibers in the rat hippocampus.
Hippocampus, 14: 908–918.
Haglund, L., Swanson, L.W. and Kohler, C. (1984) The pro-
jection of the supramammillary nucleus to the hippocampal
formation: an immunohistochemical and anterograde
81
transport study with the lectin PHA-L in the rat. J. Comp.
Neurol., 229: 171–185.
Harley, C.W., Lacaille, J.-C. and Galway, M. (1983) Hypo-
thalamic afferents to the dorsal dentate gyrus contain ace-
tylcholinesterase. Brain Res., 270: 335–339.
Herzog, E., Bellenchi, G.C., Gras, C., Bernard, V., Ravassard,
P., Bedet, C., Gasnier, B., Giros, B. and El Mestikawy, S.
(2001) The existence of a second vesicular glutamate trans-
porter specifies subpopulations of glutamatergic neurons. J.
Neurosci., 21: 1–6.
Herzog, E., Gilchrist, J., Gras, C., Muzerelle, A., Ravassard, P.,
Giros, B., Gaspar, P. and El Mestikawy, S. (2004) Locali-
zation of VGLUT3, the vesicular glutamate transporter type
3, in the rat brain. Neuroscience, 123: 983–1002.
Hisano, S., Hoshi, K., Ikeda, Y., Maruyama, D., Kanemoto,
M., Ichijo, H., Kojima, I., Takeda, J. and Nogami, H. (2000)
Regional expression of a gene encoding a neuron-specific
Na(+)-dependent inorganic phosphate cotransporter
(DNPI) in the rat forebrain. Brain Res. Mol. Brain Res.,
83: 34–43.
Hjorth-Simonsen, A. and Laurberg, S. (1977) Commissural
connections of the fascia dentata in the rat. J. Comp. Neurol.,
174: 591–606.
Hounsgaard, J. (1978) Presynaptic inhibitory action of acetyl-
choline in area CA1 of the hippocampus. Exp. Neurol., 62:
787–797.
Ishizuka, N., Weber, J. and Amaral, D.G. (1990) Organization
of intrahippocampal projections originating from CA3 py-
ramidal cells in the rat. J. Comp. Neurol., 295: 580–623.
Kandel, E.R., Spencer, W.A. and Brinley, F.J. (1961) Electro-
physiology of hippocampal neurons. I. Sequential invasion
and synaptic organization. J. Neurophysiol., 24: 225–242.
Kaneko, T. and Fujiyama, F. (2002) Complementary distribu-
tion of vesicular glutamate transporters in the central nervous
system. Neurosci. Res., 42: 243–250.
Kaneko, T., Fujiyama, F. and Hioki, H. (2002) Immunohisto-
chemical localization of candidates for vesicular glutamate
transporters in the rat brain. J. Comp. Neurol., 444(1):
39–62.
Kiss, J., Csaki, A., Bokor, H., Shanabrough, M. and Leranth,
C. (2000) The supramammillo-hippocampal and supramam-
millo-septal glutamatergic/aspartatergic projection in the rat:
a combined [3H]D-aspartate autoradiographic and immuno-
histochemical study. Neuroscience, 97: 657–669.
Kiss, J., Patel, A.J. and Freund, T.F. (1990) Distribution of
septohippocampal neurons containing parvalbumin or cho-
line acetyltransferase in the rat brain. J. Comp. Neurol., 298:
362–372.
Kosofsky, B.E. and Molliver, M.E. (1987) The serotonergic
innervation of cerebral cortex: different classes of axon ter-
minals arise from dorsal and medial raphe nuclei. Synapse, 1:
153–168.
Lâm, T.T. and Leranth, C. (2003) Role of the medial septum
diagonal band of Broca cholinergic neurons in estrogen-
induced spine synapse formation on hippocampal CA1 py-
ramidal cells of female rats. Eur. J. Neurosci., 17: 1997–2005.
82
Lamour, Y., Dutar, P. and Jobert, A. (1984) Septo-hippocam-
pal and other medial septum diagonal band neurons: elect-
rophysiological and pharmacological properties. Brain Res.,
309: 227–239.
Laurberg, S. (1979) Commissural and intrinsic connections of
the rat hippocampus. J. Comp. Neurol., 184: 685–708.
Laurberg, S. and Sorensen, K.E. (1981) Associational and
commissural collaterals of neurons in the hippocampal for-
mation (hilus fasciae dentatae and subfield CA3). Brain Res.,
212: 287–300.
Lawson, V.H. and Bland, B.H. (1993) The role of the septo-
hippocampal pathway in the regulation of hippocampal field
activity and behavior: analysis by the intraseptal microinfu-
sion of carbachol, atropine, and procaine. Exp. Neurol., 120:
132–144.
Leranth, C. and Frotscher, M. (1983) Commissural afferents to
the rat hippocampus terminate on vasoactive intestinal po-
lypeptide-like immunoreactive non-pyramidal neurons: An
EM immunocytochemical degeneration study. Brain Res.,
276: 357–361.
Leranth, C. and Frotscher, M. (1987) Cholinergic innervation
of hippocampal GAD- and somatostatin-immunoreactive
commissural neurons. J. Comp. Neurol., 261: 33–47.
Leranth, C. and Frotscher, M. (1989) The organization of the
septal region in the rat brain: cholinergic-GABAergic inter-
connections and the termination of hippocampo-septal fibers.
J. Comp. Neurol., 289: 304–314.
Leranth, C. and Nitsch, R. (1994) Morphological evidence that
hypothalamic substance P-containing afferents are capable of
filtering the signal flow in the monkey hippocampal forma-
tion. J. Neurosci., 14: 4079–4094.
Leranth, C. and Shanabrough, M. (2001) Supramammillary
area mediates subcortical estrogenic action on hippocampal
synaptic plasticity. Exp. Neurol., 167: 445–450.
Leranth, C., Shanabrough, M. and Horvath, T.L. (2000) Hor-
monal regulation of hippocampal spine synapse density in-
volves subcortical mediation. Neuroscience, 101: 349–356.
Leranth, C. and Vertes, R.P. (1999) Median raphe serotonergic
innervation of medial septum/diagonal band of Broca
(MSDB) parvalbumin-containing neurons: possible involve-
ment of the MSDB in the desynchronization of hippocampal
EEG. J. Comp. Neurol., 410: 586–598.
Lewis, P.R. and Shute, C.C.D. (1967) The cholinergic limbic
system: projections to hippocampal formation, medial cortex
nuclei of the ascending cholinergic reticular system and the
subfornical organ and supra-optic crest. Brain, 90: 521–537.
Li, X.-G., Somogyi, P., Ylinen, A. and Buzsaki, G. (1994) The
hippocampal CA3 network: an in vivo intracellular labeling
study. J. Comp. Neurol., 339: 181–208.
Lopes de Silva, F.H., Witter, M.P., Boeijinga, P.H. and
Lohman, A.H.M. (1990) Anatomic organization and phys-
iology of the limbic cortex. Physiol. Rev., 70: 453–511.
Lorente de No, R. (1934) Studies on the structure of the cer-
ebral cortex II. Continuation of the study of the ammonic
system. J. Psychol. Neurol., 46: 113–177.
Lynch, G., Rose, G. and Gall, C. (1978) Anatomical and
functional aspects of the septo-hippocampal projections.
In: Elliot K. and Whelan J. (Eds.), Functions of the Hippo-
campal System. Elsevier, Amsterdam, pp. 5–24.
Madison, D.V. and Nicoll, R.A. (1988) Norepinephrine de-
creases synaptic inhibition in the rat hippocampus. Brain
Res., 442: 131–138.
Magloczky, Z., Acsady, L. and Freund, T.F. (1994) Principal
cells are the postsynaptic targets of supramammillary affer-
ents in the hippocampus of the rat. Hippocampus, 4:
322–334.
Manahan-Voughan, D. and Kulla, A. (2003) Regulation of de-
potentiation and long-term potentiation in the dentate gyrus
of freely moving rats by dopamine D2-like receptors. Cereb.
Cortex, 13: 123–135.
Markowska, A.L., Olton, D.S. and Givens, B. (1995) Choli-
nergic manipulations in the medial septal area: age-related
effects on working memory and hippocampal electrophysi-
ology. J. Neurosci., 15: 2063–2073.
McKenna, J.T. and Vertes, R.P. (2001) Collateral projections
from the median raphe nucleus to the medial septum and
hippocampus. Brain Res. Bull., 54: 619–630.
Mellgren, S.I. and Srebro, B. (1973) Changes in acetylcholine-
sterase and distribution of degenerating fibres in the hippo-
campal region after septal lesion in the rat. Brain Res., 52:
19–35.
Miettinen, R. and Freund, T.F. (1992a) Convergence and seg-
regation of septal and median raphe inputs onto different
subsets of hippocampal inhibitory interneurons. Brain Res.,
594: 263–272.
Miettinen, R. and Freund, T.F. (1992b) Neuropeptide-Y-con-
taining interneurons in the hippocampus receive synaptic in-
put from median raphe and GABAergic septal afferents.
Neuropeptides, 22: 185–193.
Milner, T.A. and Bacon, C.E. (1989) GABAergic neurons in the
rat hippocampal formation: ultrastructure and synaptic re-
lationships with catecholaminergic terminals. J. Neurosci.,
10: 3410–3427.
Mizumori, S.J.Y., McNaughton, B.L. and Barnes, C.A. (1989)
A comparison of supramammillary and medial septal influ-
ences on hippocampal field potentials and single-unit activity.
J. Neurophysiol., 61: 15–31.
Monmaur, P. and Breton, P. (1991) Elicitation of hippocampal
theta by intraseptal carbachol injection in freely moving rats.
Brain Res., 544: 150–155.
Mosko, S., Lynch, G. and Cotman, C.W. (1973) The distribu-
tion of septal projections to the hippocampus in the rat.
J. Comp. Neurol., 52: 163–174.
Moudy, A.M., Kunkel, D.D. and Schwartzkroin, P.A. (1993)
Development of dopamine-beta-hydroxylase-positive fber in-
nervation of the rat hippocampus. Synapse, 15: 307–318.
Ni, B., Rosteck, P.R.J., Nadi, N.S. and Paul, S.M. (1994)
Cloning and expression, of a cDNA encoding a brain-specific
Na(+)-dependent inorganic phosphate cotransporter. Proc.
Natl. Acad. Sci. U.S.A., 91(12): 5607–5611.
Ni, B., Wu, X., Yan, G.M., Wang, J. and Paul, S.M. (1995)
Regional expression and cellular localization of the Na(+)-
dependent inorganic phosphate cotransporter of rat brain.
J. Neurosci., 15(8): 5789–5799.

Nicoll, R.A., Schenker, C. and Leeman, S.E. (1980) Substance
P as a transmitter candidate. Ann. Rev. Neurosci., 2:
227–268.
Nitsch, R. and Leranth, C. (1993) Calretinin immunoreactivity
in the monkey hippocampal formation. II: Intrinsic GAB-
Aergic and hypothalamic non-GABAergic systems. An ex-
perimental tracing and coexistence study. Neuroscience, 55:
797–812.
Nitsch, R. and Leranth, C. (1994) Substance P-containing
hypothalamic afferents to the monkey hippocampus. An
immunocytochemical, tract tracing, and coexistence study.
Exp. Brain Res., 101: 231–240.
Nitsch, R. and Leranth, C. (1996) GABAergic neurons in the
rat dentate gyrus are innervated by subcortical calretinin-
containing afferents. J. Comp. Neurol., 364: 425–438.
Nyakas, C., Luiten, P.G.M., Spencer, D.G. and Traber, J.
(1987) Detailed projection patterns of septa1 and diagonal
band efferents to the hippocampus in the rat with emphasis
on innervation of CAI and dentate gyrus. Brain Res. Bull.,
18: 533–545.
Prange-Kiel, J., Rune, G.M. and Leranth, C. (2004) Median
raphe mediates estrogenic effects to the hippocampus in fe-
male rats. Eur. J. Neurosci., 19: 309–317.
Ramon y Cajal, S. (1893) Estructura del asfa de Ammon y
fascia dentata. Ann. SOC Esp. Hist. Nat., 22.
Ramon y Cajal, S. (1911) Histologie de systeme nerveux de
I’Homme et des vertebres tomme II. Maloine, Paris.
Ribak, C.E. and Peterson, G.M. (1991) Intragranular mossy
fibers in rats and gerbils form synapses with the somata
and proximal dendrites of basket cells in the dentate gyrus.
Hippocampus, 1: 355–364.
Ribak, C.E., Seress, L. and Amaral, D.G. (1985) The
development, ultrastructure and synaptic connections of
the mossy cells of the dentate gyrus. J. Neurocytol., 14:
835–857.
Ribak, C.E., Seress, L., Peterson, G.M., Serooggy, K.B., Fall-
on, J.H. and Schmued, L.C. (1986) A GABAergic inhibitory
component within the hippocampal commissural pathway. J.
Neurosci., 6: 3492–3498.
Robinson, G.B. and Racine, R.J. (1984) Interactions between
septodentate and perforant-path inputs to the dentate gyrus.
Soc. Neurosci. Abstr., 10: 79.
Roppert, N. and Guy, N. (1991) Serotonin facilitates
GABAergic transmission in the CA1 region of the rat hip-
pocampus in vitro. J. Physiol. (London), 441: 121–136.
Rose, G. and Schubert, P. (1977) Release and transfer of
[3H]adenosine derivatives in the cholinergic septa1 system.
Brain Res., 121: 353–357.
Schafer, M.K., Varoqui, H., Defamie, N., Weihe, E. and Erick-
son, J.D. (2002) Molecular cloning and functional identifi-
cation of mouse vesicular glutamate transporter 3 and its
expression in subsets of novel excitatory neurons. J. Biol.
Chem., 277(52): 50734–50748.
Scharfman, H.E. (1991) Dentate hilar cells with dendrites in the
molecular layer have lower thresholds for synaptic activation
by perforant path than granule cells. J. Neurosci., 11:
1660–1673.
83
Scharfman, H.E. (1992) Differentiation of rat dentate neurons
by morphology and electrophysiology in hippocampal slices:
granule cells, spiny hilar cells and aspiny ‘fast-spiking’ cells.
In: Ribak, C.E., Gall, C.M. and Mody, I. (Eds.), The Dentate
Gyrus and its Role in Seizures: Epilepsy Research Supple-
ments. Elsevier, Amsterdam, Vol. 7, pp. 93–109.
Scharfman, H.E. (1995) Electrophysiological evidence that
dentate hilar mossy cells are excitatory and innervate both
granule cells and interneurons. J. Neurophysiol., 74: 179–194.
Scharfman, H.E. (1999) The role of nonprincipal cells in dent-
ate gyrus excitability and its relevance to animal models of
epilepsy and temporal lobe epilepsy. Adv. Neurol., 79:
805–820.
Scharfman, H.E. and Schwartzkroin, P.A. (1988) Electrophys-
iology of morphologically identified mossy cells of the dent-
ate hilus recorded in guinea pig hippocampal slices.
J. Neurosci., 8: 3812–3821.
Segal, M. (1979) A potent inhibitory monosynaptic hypotha-
lamo-hippocampal connection. Brain Res., 162: 137–141.
Segal, M. and Landis, S. (1974) Afferents to the hippocampus
of the rat studied with the method of retrograde transport of
horseradish peroxidase. Brain Res., 78: 1–15.
Seress, L. (1988) Interspecies comparison of the hippocampal
formation shows increased emphasis on the regio superior in
the Ammon’s horn of the human brain. J. Hirnforsch., 29:
335–340.
Seress, L. and Ribak, C.E. (1983) GABAergic cells in the dent-
ate gyrus appear to be local circuit and projection neurons.
Exp. Brain Res., 50: 173–182.
Seress, L. and Ribak, C.E. (1984) Direct commissural connec-
tions to basket cells of the hippocampal dentate gyrus: an-
atomical evidence for feed-forward inhibition. J. Neurocytol.,
13: 215–225.
Shute, C.C.D. and Lewis, P.R. (1966) Electron microscopy of
cholinergic terminals and acetylcholinesterase-containing
neurons in the hippocampal formation of the rat.
Z. Zellforsch. Mikrosk. Anat., 69: 334–343.
Sloviter, R.S. and Nilaver, G. (1987) Immunocytochemical lo-
calization of GABA-, cholecystokinin, vasoactive intestinal
polypeptide, and somatostatin-like immunoreactivity in the
area dentata and hippocampus of the rat. J. Comp. Neurol.,
256: 42–60.
Soltesz, I. and Mody, I. (1994) Patch-clamp recordings reveal
powerful GABAergic inhibition in dentate hilar neurons.
J. Neurosci., 14: 2365–2376.
Somogyi, J., Baude, A., Omori, Y., Shimizu, H., El Mestikawy,
S., Fukaya, M., Shigemoto, R., Watanabe, M. and Somogyi,
P. (2004) GABAergic basket cells expressing cholecystokinin
contain vesicular glutamate transporter type 3 (VGLUT3) in
their synaptic terminals in hippocampus and isocortex of the
rat. Eur. J. Neurosci., 19: 552–569.
Soriano, E. and Frotscher, M. (1994) Mossy cells of the rat
fascia dentata are glutamate-immunoreactive. Hippocampus,
4: 65–70.
Storm-Mathisen, J. and Blackstad, T.W. (1964) Cholinesterase
in the hippocampal region. Distribution and relation to
architectonics and afferent systems. Acta Anat., 56: 216–253.
84
Swanson, L.W. (1982) The projection of the ventral tegmental
area and adjacent regions: a combined fluorescent and ret-
rograde tracer and immunofluorescence study in the rat.
Brain Res. Bull., 9: 321–353.
Takamori, S., Malherbe, P., Broger, C. and Jahn, R. (2002)
Molecular cloning and functional characterization of human
vesicular glutamate transporter 3. EMBO Rep., 3: 798–803.
Takamori, S., Rhee, J.S., Rosenmund, C. and Jahn, R. (2000)
Identification of a vesicular glutamate transporter that de-
fines a glutamatergic phenotype in neurons. Nature, 407:
189–194.
Tork, I. (1990) Anatomy of the serotonergic system. Ann. N.Y.
Acad. Sci. U.S.A., 600: 9–34.
Varoqui, H., Schafer, M.K., Zhu, H., Weihe, E. and Erickson,
J.D. (2002) Identification of the differentiation-associated
Na+/PI transporter as a novel vesicular glutamate trans-
porter expressed in a distinct set of glutamatergic synapses.
J. Neurosci., 22: 142–155.
Veazey, R.B., Amaral, D.G. and Cowan, W.M. (1982) The
morphology and connections of the posterior hypothalamus
in the cynomolgus monkey (Macaca fascicularis). II. Efferent
connections. J. Comp. Neurol., 207: 135–156.
Voneida, T.J., Vardaris, R.M., Fish, S.E. and Reiheld, C.T.
(1981) The origin of the hippocampal commissure in the rat.
Anat. Rec., 201: 91–103.
Wainer, B.H., Bolam, J.P., Freund, T.F., Henderson, Z., Tot-
terdell, S. and Smith, A.D. (1984) Cholinergic synapses in the
rat brain: a correlated light and electron microscopic
immunohistochemical study employing a monoclonal anti-
body against choline acetyltransferase. Brain Res., 308:
69–76.
Wheal, H.V. and Miller, J.J. (1980) Pharmacological identifi-
cation of acetylcholine and glutamate excitatory systems in
the dentate gyrus of the rat. Brain Res., 182: 145–155.
Whitehouse, P.J., Price, D.L., Struble, R.G., Clark, A.W.,
Coyle, J.T. and Delon, M.R. (1982) Alzheimer’s disease
and senile dementia: loss of neurons in the basal forebrain.
Science, 215: 1237–1239.
Wiss, J.M., Swanson, L.W. and Cowan, W.M. (1979) Evidence
for an input to the molecular layer and the stratum granulo-
sum of the dentate gyrus from the supramammillary region
of the hypothalamus. Anat. Embryol., 156: 165–176.
Wu, M., Hajszan, T., Xu, C., Leranth, C. and Alreja, M. (2004)
Group I metabotropic glutamate receptor activation pro-
duces a direct excitation of identified septohippocampal
cholinergic neurons. J. Neurophysiol., 92: 1216–1225.
Zimmer, J., Laurberg, S. and Sunde, N. (1983) Neuroanatom-
ical aspects of normal and transplanted hippocampal tissue.
In: Seifert W. (Ed.), Neurobiology of the Hippocampus.
Academic Press, New York, pp. 39–64.
Plate 4.3. Light micrograph (kindly provided by Dr. Attila Gulyas) shows the result of a combined anterograde tracing and calretinin
immunostaining study, in the dentate gyrus. The anterograde tracer, biotinylated dextran amine (BDA) was injected into the medial
septum diagonal band. Large boutons of BDA-containing axons form multiple, basket-like, putative synaptic contacts with the soma
of calretinin-immunoreactive (labeled with a brown diaminobenzidine reaction product) neurons. One of these cells (arrows) located in
the supragranular layer (SgL) the other is at the border between the granule cell layer (GcL) and dentate hilar area (H). Bar
scale ¼ 50 mm. (For B/W version, see page 69 in the volume.)

Plate 4.5. Light micrograph taken from a double immunostained vibratome section of the monkey dentate gyrus. Immunoreactivity
for substance P was labeled with a dark-blue Ni-diaminobenzidine reaction, while immunostaining for parvalbumin was visualized by a
brown diaminobenzidine reaction. The soma and dendrites of the parvalbumin-containing cell embedded into the granule cell layer
(GcL) is contacted by several substance P-immunoreactive axon terminals (arrows). Bar scale ¼ 10 mm. (For B/W version, see page 71
in the volume.)
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 5

The dentate mossy fibers: structural organization,

development and plasticity

Morten Blaabjerg and Jens Zimmer

Anatomy and Neurobiology, Institute of Medical Biology, University of Southern Denmark, Winslowparken 21,
DK-5000 Odense C, Denmark

Abstract: Hippocampal mossy fibers are the axons of the dentate granule cells and project to hippocampal
CA3 pyramidal cells and mossy cells of the dentate hilus (CA4) as well as a number of interneurons in the
two areas. Besides their role in hippocampal function, studies of which are still evolving and taking
interesting turns, the mossy fibers display a number of unique features with regard to axonal projections,
terminal structures and synaptic contacts, development and variations among species and strains, as well as
to normal occurring and lesion-induced plasticity and neural transplantation. These features are the topic
of this review, which will use the mossy fiber system of the rat as basis and reference in its aim to provide an
up-to-date, yet historically based guide to students in the field.

Keywords: granule cells; hippocampus; CA3; CA1; dentate hilus; lesion-induced sprouting; neural
transplantation

Introduction to the dentate mossy fiber system
The dentate mossy fibers (MFs) are generally
known to be the second link in the classical trisy-
naptic pathway from the entorhinal cortex to the
hippocampal subfield CA1, including: (1) the ento-
rhinal perforant path projection to dentate granule
cells; (2) the dentate MF projection to CA3 py-
ramidal cells; and (3) the CA3 Schaffer collateral
projection to CA1 (Andersen et al., 1969). In line
with this, MFs arise from dentate granule cells,
located in the tightly packed cell layer of the fascia
dentata or dentate gyrus, and project their main
axons through the dentate hilus (CA4) and to the
proximal parts of the apical (and basal) dendrites
of CA3 pyramidal cells. The extensive network of
Corresponding author. Tel.: (+45) 6550 3801;
Fax: (+45) 6550 6321; E-mail: jzimmer@health.sdu.dk
collaterals of the MFs in the dentate hilus also
contact several types of neurons there, deep in the
hilus as well as immediately below the granule cell
layer. The fine, filopodial processes, originally
identified only to extend from the
classical, large MF terminals, have recently been
shown to make separate and specific synaptic con-
tacts with interneurons, both within the dentate
hilus and in CA3, thereby adding new functional
dimensions to the dentate MF system and its role
in hippocampal function (Frotscher et al., 1994,
2006; Acsády et al., 1998; Henze et al., 2000).
The dentate MFs were first described by Golgi
(1886) and Sala (1891), and were named mossy
fibers by Ramón y Cajal (1893, 1911), based on the
similarity of their large, irregular terminals with
the moss-like terminal varicosities of the cere-
bellar MFs, which he had described earlier. Addi-
tional contributions, based on Golgi-impregnated

DOI: 10.1016/S0079-6123(07)63005-2 85
86
material, were made by Schaffer (1882) and
Koelliker (1896) and later Lorente de Nó (1934),
a student of Ramón y Cajal, who added further
details about the synaptic nature of contacts be-
tween the giant MF terminals and the complex
spines emerging from the proximal dendrites of
large pyramidal cells of regio inferior (CA3) and
the so-called mossy cells of the dentate hilus
(CA4).
The dentate MFs display various differences in
number, trajectory and termination among species
and strains and individuals within a given species.
Here we present a guide to the structural organ-
ization of the MF system. For the sake of sim-
plicity, we will use the common Wistar and
Sprague-Dawley laboratory rats as the standard
to which other species are compared.
Visualization of dentate mossy fibers and terminals
The axonal projection and termination of dentate
MF have been visualized by several techniques,
including classical Golgi-impregnation (Schaffer,
1882; Golgi, 1886; Sala, 1891; Ramón y Cajal,
1893, 1911; Koelliker, 1896; Lorente de Nó, 1934;
Amaral, 1978, 1979; Duffy and Rakic, 1983;
Soriano et al., 1983); axonal tract tracing meth-
ods, like silver staining of lesion-induced antero-
grade axonal degeneration (Blackstad et al., 1970;
Gaarskjaer, 1978a) and axonal transport of extra-
and intracellularly injected tracers (Lynch et al.,
1973; Swanson et al., 1978; Gaarskjaer, 1981; West
et al., 1982; Claiborne et al., 1986; Acsády et al.,
1998; Vida and Frotscher, 2000); histochemical
and immunocytochemical staining for proteins like
calbindin (Lim et al., 1997; Keuker et al., 2003)
and neuropeptides like enkephalin, dynorphin,
cholecystokinin (CCK), neuropeptide Y (NPY)
and substance P (Fitzpatrick and Johnson, 1981;
Stengaard-Pedersen et al., 1983; Gall, 1984, 1988;
Zimmer et al., 1988a, b; Holm et al., 1992, 1993);
and ultrastructural analysis by transmission elec-
tron microscopy alone (Fig. 1) (Blackstad and
Kjaerheim, 1961; Amaral, 1979) or in combination
with the other methods mentioned above (see
Amaral, 1979; Frotscher and Zimmer, 1983;
Frotscher, 1985, 1989; Frotscher and Seress,
1991; Frotscher et al., 1994; Acsády et al., 1998).
One rather unique and widely used method for
‘‘bulk visualization’’ of MF terminals, and thereby
the MF terminal projection, is the histochemical
Timm sulfide silver method (Timm, 1958; Haug,
1967; Danscher, 1981), which takes advantage of
a very high concentration of chelatable zinc in the
MF terminals (see Danscher et al., 1985). After
precipitation by sulfide, the chelatable zinc can be
visualized by physical development and the stained
elements analyzed by both light and electron mi-
croscopy (Danscher and Zimmer, 1978; Laurberg
and Zimmer, 1981). After Timm staining, the MF
projection and its individual terminals stand out
distinctly black, in contrast to the other unstained
to lightly-stained (yellowish to brown) laminar-
specific terminal fields in the hippocampus and
fascia dentata (Figs. 2–4).
Trajectory and termination of mossy fibers in
CA3 (regio inferior)
Within the hippocampal region, there are about
one million granule cells in the rat fascia dentata
(West et al., 1991), about 10 million granule cells
in the domestic pig (Holm and West, 1994) and
about 15 million granule cells in man (West and
Gundersen, 1990). Each cell gives rise to a thin,
unmyelinated MF (diameter 0.1–0.7 mm) (Gaarsk-
jaer, 1978a; Amaral, 1979; Claiborne et al., 1986;
Acsády et al., 1998), which projects into the sub-
jacent dentate hilus. During its passage through
the hilus, each MF gives off a number of branch-
ing collaterals, which contact various hilar and
immediately subgranularly positioned neurons,
whereafter the main axons, grouped in bundles
without intervening glia (Fig. 1) (Blackstad and
Kjaerheim, 1961), continue toward the pyramidal
cell layer of regio inferior (CA3). Within CA3, the
main MFs do not branch or give off collaterals as
they traject in a laminar fashion along the pyram-
idal cell layer across the transverse axis of the
hippocampus, before bending to take a more
longitudinal course in temporal direction when
approaching CA1 (Gaarskjaer, 1978b, 1986;
Acsády et al., 1998).
 87

Fig. 1. Electron microscopy (EM) (A, B) of large MF terminals (mft) from the suprapyramidal MF layer in CA3 of the adult rat
hippocampus. The large terminals and terminal profiles form asymmetric synaptic contacts with a number of dendritic spines (sp).
Slender extension from these (spinules, s) invade the terminals. Micrographs were kindly provided by M. Frotscher. Abbreviations: a,
unmyelinated MFs; d, dendrites; mft, MF terminal; s, small dendritis spine extensions, spinules, invading MF terminals; sp, dendritic
spines. Scale bars: 2 mm (A), 1 mm (B).

Supra- and infrapyramidal bands of hippocampal
mossy fibers and terminals
As they approach the part of the hippocampal
pyramidal cell layer is closest to the hilus (CA3c),
the MFs form, in classical terms, a ‘‘suprapyram-
idal band’’, passing along the most proximal parts
of the apical dendrites of CA3 pyramidal cells, and
an ‘‘infrapyramidal band’’, contacting correspond-
ing proximal parts of the basal dendrites
(Figs. 2–4A). Due to its homogeneous transpar-
ency in unstained sections, the suprapyramidal
MF layer is also referred to as stratum lucidum
(Ganser, 1882). The infrapyramidal MF projection
is especially subject to variation among species
and strains (Fig. 2 and below). By studying the
MF system in so-called ‘‘extended hippocampi’’ of
Wistar rats, i.e., after having stretched out the
longitudinal septotemporal bend of dissected
hippocampi, Gaarskjaer (1978b) found that the
suprapyramidal MF terminals in CA3c actually
were located within the rather dispersed pyramidal
cell layer, and that only a few infrapyramidal MF
terminals were present in the corresponding
area of CA3 basal dendrites. After horseradish
peroxidase-labeling of MFs following intra- and
extracellular injections in different parts of the
granule cell layer of acute hippocampal slices from
young adult Sprague-Dawley rats, Claiborne et al.
(1986) found that MFs from granule cells located
at different supra- to infrapyramidal locations
along the transverse axis of the dentate granule cell
layer projected into MF bundles both above and
within the pyramidal cell layer in CA3c. MF
bundles within the deep part and below the CA3c
pyramidal cell layer originated from granule cells
located in the infrapyramidal (free or outer) blade
of the dentate. Granule cells from different
88

Fig. 2. Timm stained hippocampal sections from rat (A), guinea pig (B), cat (C), European hedgehog (D), dog (E) and man (F),
illustrating the distribution of intensely stained, black MF terminals in the different species, as well as other terminal fields and general
organization. Note species specific characteristics, like the ‘‘end bulb’’ at the terminal part of the MF layer in guinea pig (black asterisk,
B) and the layering of the dentate hilus with MF terminal-free, intermediate or plexiforme sublayer in the same species (white asterisk,
B), and MF projections into CA1 in some cats (arrows, C) and European hedgehog (D). Except for the black MF terminal staining, the
neuropil of the human hippocampus is virtually unstained, due to the use of special Timm staining procedures for non-perfused tissue
(Danscher and Zimmer, 1978). Abbreviations: CA1, hippocampal regio superior or subfield CA1; CA3, hippocampal regio inferior or
subfield CA3; FD, fascia dentata (or denate gyrus); g, dentate granule cell layer; h, dentate hilus or CA4; m, dentate molecular layer;
mf, MF layer; Sub, subiculum. Scale bars: 500 mm (See Color Plate 5.2 in color plate section.)
 89

Fig. 3. Timm stained mouse hippocampal sections, illustrating strain differences in MFs terminal projections between BALB7C (A)
and C57B mice (B) and Reeler mutants (C). BALB/c mice have no infrapyramidal MFs, but a short intrapyramidal bundle (black
asterisk, A), compared to long infrapyramidal bundles in C57B mice (black asterisk, B). In Reeler mutant mice, granule and pyramidal
cell layers are disorganized, resulting in a corresponding distribution of MF terminals in the confluent granule cell layer/hilar area and
CA3. However, MFs still avoid CA1, similar to normal mice. Abbreviations: g, dentate granule cell layer; h, dentate hilus (CA4); m,
dentate molecular layer. Scale bar: 500 mm (See Color Plate 5.3 in color plate section.)

locations along the transverse axis of the granule
cell layer all sent MFs to the distal parts of the
suprapyramidal MF layer in CA3a, adjacent to
CA2/CA1 (Lorente de Nó, 1934). Granule cells
located near the tip of the suprapyramidal (hidden
or inner) part of the granule cell layer projected
directly into the suprapyramidal MF layer, distal
to the hilar border, after passing through the in-
tervening part of startum radiatum, as also shown
in Ramón y Cajal’s figure 478 from a Golgi-stained
guinea pig hippocampus (Ramón y Cajal, 1911).
Having considered the presence and composi-
tion of supra-, intra- and infrapyramidal MF bun-
dles in primarily Wistar and Sprague-Dawley rats,
it should be noted that considerable differences
exist between various strains of rats (Dimitrieva
et al., 1993) and mice. BALB/cJ mice have no
infrapyramidal MF bundles, but a single intrapy-
ramidal bundle (Fig. 3A), whereas SM/J mice
have a distinct infrapyramidal, but no intrapy-
ramidal MF bundle (Barber et al., 1974). These
mouse strain differences have been linked to
different patterns of pyramidal cell generation
(SM/J ¼ inside-out; BALB/cJ ¼ outside-in) by
Vaughn et al. (1977), but might also be induced
by strain differences in thyroid hormone levels,
which affect the formation of dentate granule cells
and the size of the MF projection (Lauder and
Mugnaini, 1980; Fredens, 1981; Lipp et al., 1984;
Madeira et al., 1988), or differences in the ratio
between the number of granule cells and target
CA3 pyramidal cells (Gaarskjaer, 1978a). The
difference in size of the granule cell layer between
the BALB/cJ and C57 mice is seen by comparing
Fig. 3A and B. Systemic application of the cell
proliferation inhibitor methylazoxymethanol-
acetate (MAM) to pregnant rats during the late
fetal period, when the majority of hippocampal py-
ramidal cells are formed, induces larger than normal
infrapyramidal MF projections from the virtually
normal sized fascia dentata into a smaller than
normal CA3. In contrast, infrapyramidal MFs are
almost absent, when MAM is administered to early
postnatal pups during the time when most dentate
granule cells develop. This is likely to be due to the
fact that MAM reduces the number of the granule
cells relative to the virtually normal numbers of
CA3 pyramidal cells (unpublished observations).
90

Fig. 4. Timm stained MF terminals (black) in stratum lucidum of CA3 (mf, A) and dentate hilus (B) in semi thin plastic section from
adult rat. (A) In CA3, the densely stained MF terminals in the suprapyramidal MF layer (mf) are usually large and contact CA3
pyramidal cell dendrites, but there are more homogeneous, lighter stained MF terminals in similar locations. A small group of MF
terminals (imf) are present below the pyramidal cell layer (p). Lighter, Timm stained elements in stratum oriens (o) and stratum
radiatum (r) are terminals of the commissural/associational projections to these layers. (B) In the hilus, a large mossy cell body (mc) is
seen in contact with MF terminals, which are of variable size and also contacting other dendritic structures. Scale bars: 25 mm (A);
15 mm (B).

Both the width of the suprapyramidal layer and
the extent of the infrapyramidal MF layer in CA3
vary along the longitudinal (septotemporal) axis
of the rat hippocampus (Gaarskjaer, 1978a, b). At
septal levels, MF terminals form an infrapyrami-
dal layer, which extends along the deep side of the
CA3 pyramidal cell layer to the CA2-CA1 transi-
tion as an almost equal counterpart to the supra-
pyramidal MF layer. Many MF terminals are also
located within the pyramidal cell layer itself, which
at these levels is wide and loosely packed. Moving
temporally, the intrapyramidal and infrapyramidal
MF terminals disappear first from the distal
(CA3a) and intermediate (CA3b) parts of CA3.
In parallel, the suprapyramidal terminal zone
becomes relatively thin. The reduction in the size
of MF terminal fields corresponded to a reduction
in the ratio between the number of granule cells
and CA3 pyramidal cells moving from septal to
more posterior and temporal levels (Gaarskjaer,
1978a). Toward the temporal end of the hippo-
campus, the border between the suprapyramidal
MF and CA3 pyramidal cell layer becomes
‘‘fuzzy,’’ and a small portion of intrapyramidal
MFs reappear distally (Gaarskjaer, 1978a, b).
Transverse and longitudinal trajectories of mossy
fibers in CA3
While projecting along the more proximal parts of
CA3 (CA3c and adjacent parts of CA3b), rat MFs
deviate only slightly in temporal direction relative

to the level of origin (Blackstad et al., 1970;
Gaarskjaer 1978b, 1981; Claiborne et al., 1986;
Amaral and Witter, 1989; Tamamaki and Nojyo,
1991), remaining within a septotemporal span or
lamina of 1.1 mm (Acsády et al., 1998) or less
(Gaarskjaer, 1981). As the MFs enter more distal
parts of CA3 (distal CA3b and CA3a) facing CA1,
the fibers do, however, bend in the temporal di-
rection to project for additional 1–2 mm along the
longitudinal axis of the hippocampus (Gaarskjaer,
1978, 1981; Amaral and Witter, 1989). MFs orig-
inating at septal levels and from the tip of the
suprapyramidal granule cell layer at other levels
display the longest temporal descent (Gaarskjaer,
1978b, 1981; Swanson et al., 1978; Tamamaki
and Nojyo, 1991; Acsády et al., 1998). After 3-
dimensional reconstruction of seven biocytin-
injected granule cells from dorsal levels of the
fascia dentata in adult Sprague-Dawley rats,
Acsády et al. (1998) measured the MF axon in
CA3 to be in avarage 3.24 mm long, of which
0.80 mm was in CA3c, 1.05 mm in CA3b and
1.39 mm in CA3a.
In guinea pigs, a rather abrupt turn of MFs
followed by a steep, longitudinal descent in tem-
poral direction takes place just close to and within
the so-called ‘‘end bulb’’, which is a characteristic
bulge on the very distal part of the suprapyramidal
MF layer in this species, as demonstrated very
clearly by Timm staining (Fig. 2B).
Mossy fibers projections to CA1 (regio superior)
Based on Golgi-impregnated brain sections from
rabbit, guinea pig, macaque and rodents, Ramón y
Cajal (1893) and Lorente de Nó (1934) described
the large, irregular MF terminals as ‘‘boutons en
passant’’ which contacted the proximal dendrites
of the large pyramidal cells of regio inferior (CA3),
and characterized them by the presence of ‘‘thorny
excrescences’’ or complex spines. In particular,
Lorente de Nó (1934) noted that MFs did not
proceed to contact the smaller pyramidal cells in
regio superior (CA1). Located between the two
subfields is, in most species, a narrow transitional
area or ‘‘Mischzone’’ (Doinikow, 1908), which
Lorente de Nó (1934) defined as a separate
91
subfield, named CA2, and characterized by a
mixed content of small CA1-like pyramidal cells
and larger CA3-like pyramidal cells, which, how-
ever, were devoid of complex spines and MF input
(see also Tamamaki et al., 1988; Barthesaghi and
Ravasi, 1999). The lack of MF projection to CA1
was confirmed by lesion-induced anterograde ax-
onal tracing (rat, Blackstad et al., 1970), antero-
grade axonal transport of injected tracers, like
horseradish peroxidase (rat, Lynch et al., 1973;
West et al., 1982; Claiborne et al., 1986), isotope
labelled amino acids (rat, Swanson et al., 1978)
and biocytin (man, Lim et al., 1997; rat, Acsády
et al., 1998). The same conclusion was drawn from
visualization of the MF terminal projection by
Timm staining in rat (Fig. 2A) (Haug, 1974), sev-
eral strains of mice (Fig. 3) (Barber et al., 1974;
Fredens, 1981), guinea pig (Fig. 2B) (Blackstad,
1963; Geneser-Jensen et al., 1974), rabbit (Hjorth-
Simonsen, 1977), dog (Fig. 2E), domestic pig
(Holm and Geneser., 1991b) and man (Fig. 2F)
(Danscher and Zimmer, 1978; Cassell and Brown,
1984). The continuation of tufts of Timm stained,
thread-like elements with minor terminal-like
swellings from the suprapyramidal MF layer of
CA3 into CA2 in several species both confirms
that CA2 acts as a transitional zone and that MFs
do not reach into CA1.
A dogmatic generalization of the absence of an
MF projection to CA1 to all species, was, how-
ever, proven to be wrong by the discovery that
MFs occasionally project into CA1 in some do-
mestic cats (Fig. 2C) (Laurberg and Zimmer, 1980)
and is normal in European hedgehogs (Fig. 2D)
(Gaarskjaer et al., 1982; West et al., 1984). In the
occasional cat with MFs in CA1, the projection
was variable, being most extensive or only present
at septodorsal levels (Fig. 2C). The MF projection
to CA1 in European hedgehogs was present at
midposterior (Fig. 2D) to temporal levels only,
or along the entire septotemporal extent of the
hippocampus.
By demonstrating an overlap of MF terminal-
like, Timm stained elements and pyramidal cells
with small cell nuclei corresponding to CA1
pyramidal cells, Gaarskjaer (1986) later provided
evidence that the most temporal levels of the rat
CA1 also normally receive a MF projection, a
92

phenomenon which might therefore appear to be

rather common among species.
Distinct guidance or stop cues for MF innerva-
tion of CA1 have not been identified so far. Cer-
tainly an increased number of granule cells relative
to target CA3 pyramidal cells at septal levels in
rodents does not by itself cause MFs to enter CA1,
even that this is paralleled by a more extensive
infrapyramidal termination in CA3 (see above),
and also may be a causal factor behind the more
extensive MF projections into CA1 in some cats
at septodorsal levels (Fig. 2C) (Laurberg and
Zimmer, 1980). Occasional observations of MF
projections to CA1 in rats with early postnatal
lesions of the CA3 to CA2/CA1 transition area
suggest that cells in this area act as a barrier
(Fig. 5; see also Cook and Crutcher, 1985). Clearly
cells of the CA2 subfield have been shown to differ
from the surrounding CA3 and CA1 subfields
both with regard to neuronal birthdate (Angevine,
1965; Bayer, 1980) and gene expression profile
(Lein et al., 2005).

Mossy fiber connections in the dentate hilus (CA4)

The cytoarchitectural organization of the dentate

hilus displays significant differences among spe-
cies, ranging from a laminar organization in the
guinea pigs (Geneser-Jensen et al., 1974) and pigs
(Sus scrofu domesticus) (Holm and Geneser, 1991)
to an almost confluent mixture of cell types in ro-
dents (Blackstad, 1956; Amaral, 1978). Observa-
tions of the hilar distribution of MF collaterals
and terminals in one species may accordingly not
apply to another species, as exemplified by laminar
distribution of large and small Timm stained MF
terminals in the dentate hilus of guinea pig
(Fig. 2B) compared to the confluent distribution
in the dentate hilus of rat and European hedgehog
(Fig. 2A and D). Another obstacle to comparisons
of dentate hilar cell and neuropil layers between
species is the use of different nomenclatures.
As one of the first illustrations of the exten-
sive MF projections within the dentate hilus,
(Koelliker, 1896, figure 786; also reproduced by
Gaarskjaer, 1986) depicted two Golgi-impregnated
granule cells from a three-day-old cat. The main
Fig. 5. Timm stained section from dorsoposterior level of
young adult, rat hippocampus. When newborn, the rat received
a mechanical lesion of the CA3–CA1 transition (les) as well a
transection of the entorhinal perforant path to the fascia den-
tata. The CA3–CA1 lesion induced an aberrant ingrowth of
MFs into CA1, illustrated by presence of small, Timm stained
MF terminals (asterisk) along the CA1 pyramidal cell layer.
The removal of the perforant path projection to the outer mo-
lecular layer (m) induced spread of associational-commissural
projections from inner part of the layer, and induction of sup-
ragranular MF terminals (arrow) (see Zimmer, 1973, 1974).
Abbreviations: g, granule cell layer. Scale bar: 500 mm (See Color
Plate 5.5 in color plate section.)
MF axons from the two cells give off several ex-
tensively branching collaterals, starting at the level
of the layer of polymorphic cells and further along
their trajectory toward the hippocampal pyrami-
dal cell layer. Primarily based on observations in
Golgi-stained sections of one-week- to one-month-
old rabbits, guinea pigs and mice, Ramón y Cajal
(1893, 1911) described and illustrated the emission

of 4–8 collaterals from main MF axons just below
the granule cell layer in what he called the plexi-
form layer (or outer part of the polymorphic cell
layer), as well as deeper into the hilus close to the
hippocampal pyramidal layer from where they re-
turned back into the hilus. Referring primarily to
Ramón y Cajal’s well-known schematic diagram
of hippocampal connections (Figure 479 in
Ramón y Cajal, 1911), the MF collateral projec-
tions in the (rodent) hilus were, however, long
taken to be of limited extension, until Soriano
et al. (1983), also using Golgi-impregnation,
described extensive MF collateral projections en-
gaging major parts of the rat dentate hilus. This
spurred Claiborne et al. (1986) to trace MFs and
collaterals from single and small group of granule
cells by intracellular and small extracellular injec-
tion of horseradish peroxidase in acute hippocam-
pal slices from young adult Sprague-Dawley rats.
Later Acsády et al. (1998) combined in vivo intra-
cellular biocytin injections of single granule cells in
the dorsal fascia dentata of adult Sprague-Dawley
rats with EM immunocytochemistry and 3D re-
consruction to provide the most comprehensive
and detailed single study of the trajectory and
synaptic connections of single MF main axons and
collaterals so far, including description of post-
synaptic target cells.
According to the combined observations of
Claiborne et al. (1986) and Acsády et al. (1998)
in the rat, each MF main axon emits between 5
and 12 collaterals, less than 0.2 mm thick, at var-
ious distances along its course through the hilus.
Including secondary and tertiary branches, the to-
tal length of the collateral plexus formed by one
MF amounts to approx. 2.3 mm (Claiborne et al.,
1986). In the transverse plane, collaterals from a
single MF cover up to one third of the total hilar
extent nearest the location of the parent granule
cell, with a total transverse extent of 0.4–0.7 mm,
depending on the location of the parent granule
cell. Along the longitudinal axis of the dentate hi-
lus, the spread of collaterals never seems to exceed
0.6 mm, with 90% of the collateral network being
within a 0.4 mm high lamina (Acsády et al., 1998).
Regarding relations to the granule cell layer, in-
dividual collaterals running close to the deep bor-
der of the granule cell layer have been described in
93
all species examined, presumably contacting the
pyramidal-shaped basket cells in the limiting sub-
zone (cf. Ribak and Peterson, 1991). Among the
studied collateral networks only one collateral was
reported to enter the granule cell layer and none to
reach the dentate molecular layer. This contrasts
to the observations of several such MF collateral
terminals in Timm staining in normal rats
(Fig. 6A) and guinea pigs (Laurberg and Zimmer,
1981; Sloviter et al., 2006), as dealt with below.
Timm staining and species differences, including
dentate hilus
‘‘Bulk’’ visualization of the zinc-rich MF terminals
by the Timm method, which intensely stains both
the characteristic large terminals of the main fibers
in CA3 and the corresponding large terminals as
well as the smaller MF collateral terminals in the
dentate hilus, have been performed for several
species, including rat (Figs. 2A and 4A, B) (Haug,
1974; Gaarskjaer, 1978b; Laurberg and Zimmer
1981; Dimitrieva et al., 1993), several strains of
mice (Fig. 3) (Barber et al., 1974; Fredens, 1981,
including C57/BL/6J, DBA/2J, NMRI, BALB/c/
A, C3H/Tif, A/J, AKR/A mouse strains), guinea
pig (Fig. 2B) (Blackstad, 1963; Geneser-Jensen et
al., 1974), rabbit (Hjorth-Simonsen, 1977), Euro-
pean hedgehog (Fig. 2D) (Gaarskjaer et al., 1982;
West et al., 1984), cat (Fig. 2C) (Laurberg and
Zimmer, 1980), dog (Fig. 2E), domestic pig (Holm
and Geneser, 1991b), and man (Fig. 2F) (Danscher
and Zimmer, 1978; Cassell and Brown, 1984). At
low magnification, the Timm staining pattern re-
veals both the general similarities and the differ-
ences between the species in the both the
hippocampal and the hilar MF projection.
Within the dentate hilus one difference among
species is a more distinct layering of hilar neurons
(not illustrated) and Timm stained MF terminals
in guinea pig (Fig. 2B), pig (Holm and Geneser,
1991b) and in part man (Fig. 2F), as compared to
other species like rat (Fig. 2A). The mentioned
species are thus characterized by having a virtually
MF terminal free zone (corresponding to Ramón
y Cajal’s (1911) intermediate or plexiforme sub-
layer), which is located between a narrow irregular
94

Fig. 6. A: Timm stained MF (collateral) terminals, found in normal adult rats to accompany dendrites (arrows), arising from neurons
in the limiting subzone just below the dentate granule cell layer or in the dentate hilus, and extending into and sometimes through the
dentate granule cell layer (g) into the, commissural-associational terminal zone (c/a). B: Timm stained, dense supragranular MF
terminal projection (sgr) found normally at temporal levels of the cat fascia dentata as well as in other species. h, dentate hilus.
Scale bar: 50 mm (A); 100 mm (B). (See Color Plate 5.6 in color plate section.)

band of MF collateral terminal staining just below
the granule cell layer (Ramón y Cajal’s limiting
subzone) and a more deeply located, densely
stained layer containing large and small MF ter-
minals (Ramón y Cajal’s deep polymorphic cell
layer).
Intra- and supragranular mossy fiber collaterals —
a normal feature and subject to plasticity
Focusing on the limiting subzone, it is possible in
all species at high magnification to observe small,
presumably MF collateral terminals extending
along dendrites from the hilus into the granule
cell layer and sometimes as far as the molecular
layer. In the rat, these intra- and supragranular
Timm-positive terminals are most frequent at the
dentate crest (Fig. 6A) and toward the tip of the
infrapyramidal blade (Haug, 1974; Laurberg and
Zimmer, 1981; Sloviter et al., 2006), but become
more widespread and numerous at temporal levels
(Gaarskjaer, 1978a). Intra- and supragranular MF
terminals can be very prominent at temporal
levels, as shown in Fig. 6B for the cat.
Aberrant growth of supragranular MF termi-
nals into the commissural-associational terminal
zone, which normally cover the inner approxi-
mately one third of the dentate molecular layer,
can be induced: (a) by sufficiently strong denerva-
tion of the dentate molecular layer in developing
and adult animals (Fig. 5) (Zimmer, 1973, 1974;
Laurberg and Zimmer, 1981; Frotscher and
Zimmer, 1983), including excitotoxic loss of hilar
mossy cells (see below, Sloviter et al., 2006); (b) in
intracerebral rat dentate allografts and mouse
dentate xenografts (Sunde and Zimmer, 1981;
Jensen et al., 1984; Frotscher and Zimmer, 1986;
 95

Fig. 7. (A) Timm stained section from midposterior level of 5-day-old rat, demonstrating a distinct Timm stained MF layer in CA3
and the hilus at this age (arrows). (B, C) Parallel sections of hippocampal slice culture, derived from 7-day-old rat and grown for
3 weeks. Timm stain reveals the distribution of aberrant supragranular and normal CA3 MFs (arrow and mf, respectively, in B) and
thionine cell staining to visualize general cellular organization (C). Abbreviations: g, granule cell layer. Scale bars: 500 mm (See Color
Plate 5.7 in color plate section.)

Zimmer et al., 1988a, b), depending on the density
of afferent host brain fiber innervation; (c) in
developing hippocampal tissue slices grown as
organotypic explant cultures (Fig. 7B) (Zimmer
and Gahwiler, 1984); and (d) in animals with
pilocarpine- or kainic acid-induced seizures
(Sutula et al., 1988; Mello et al., 1993; Frotscher
et al., 2006; Sloviter et al., 2006) and in epileptic
patients (Sutula et al., 1989; Blümcke et al., 2000).
This appears to be a response to loss of dentate
hilar mossy cells normally projecting to the dentate
molecular layer. Transection of the main MF pro-
jection to CA3 will not by itself induce excessive
supragranular MF collateral projection, excluding
an axonal pruning or compensatory sprouting
effect (Laurberg and Zimmer, 1980, 1981).
Structure of synaptic mossy fiber connections in the
dentate hilus and CA3
As outlined by Ramón y Cajal (1911), refined by
Lorente de Nó (1934) and demonstrated by EM
(Blackstad and Kjaerheim, 1961), typical MF
synapses are made by the large boutons en pass-
ant in contact with the complex spines (thorny ex-
crescences) found on the proximal dendrites of the
CA3 pyramidal cells (Fig. 1AB) and hilar mossy
cells, as well as by MF collateral terminals con-
tacting different types of neurons in the dentate
hilus. Claiborne et al. (1986) and Acsády et al.
(1998) extended the historical perspective by show-
ing that each MF makes approx. 120–150 excita-
tory synaptic connections with predominantly
inhibitory interneurons via axon collaterals in the
dentate hilus, 7–12 excitatory, large terminal
synaptic contacts with hilar mossy cells and
11–18 corresponding contacts with proximal py-
ramidal cell dendrites during the transverse MF
trajectory in CA3, plus an additional number of
terminal contacts made by the more distal and
longitudinally running parts of the MFs. These
two studies and a number of related ones (see
Frotscher et al., 2006 and references therein) have
changed the classical view of dentate-CA3 MF
connectivity from primarily being a direct synaptic
interaction between the excitatory, large MF ter-
minals and the CA3 pyramidal cells and hilar
mossy cells, into a pathway with a prominent
indirect inhibitory effect not only in the dentate
hilus, but also in CA3, apparently acting to refine
the activation of CA3 pyramidal cells.
The findings were not entirely surprising, how-
ever, because short, thick diverging processes and
the thin, up to 30 mm-long filamentous or filopo-
dial processes had been noted by Ramón y Cajal
(1911) as extending from the angles of the large,
irregularly shaped MF terminals in the dentate
hilus and CA3. These extensions appear to con-
stitute a separate, major presynaptic MF element,
contacting defined types of inhibitory GABAergic
and peptidergic interneurons in the dentate hilus
96
and the CA3 (Amaral, 1979; Claiborne et al., 1986;
Frotscher et al., 1994; Acsády et al., 1998; Henze
et al., 2000; Vida and Frotscher, 2000, and refer-
ences therein). Based on the data provided by
Acsády et al., 1998, the synaptic contacts made
by a single MF can accordingly be grouped into
contacts with hilar mossy cells (made by as few as
7–12 large terminals), CA3 pyramidal cells (11–18
large terminals), hilar interneurons (as many as
120–150 small terminal contacts) and interneurons
in CA3 (40–50 filopodial contacts).
Regarding the ultrastructural appearance of the
MF synapses, the well known, classical type is
made by the 4–10 mm large boutons en passant, or
in several cases attached to the main MF axon by a
short thin extension (Fig. 1) (Ramón y Cajal, 1911;
Blackstad and Kjaerheim, 1961; Amaral and Dent,
1981). In the hilus these terminals contact the hilar
mossy cells. In CA3 they occur at intervals of
approx. 135 mm along the main MF, which has
been estimated to be every 6th or 7th of the CA3
pyramidal cell that it meets along its trajectory
(Claiborne et al., 1986).
The contact between the giant MF terminals
and the ‘‘thorny excresences’’ on CA3 pyramidal
cells and hilar mossy cells are structurally intimate
and complex, with spinules projecting from
the excresences into the terminals (sp, Fig. 1)
(Blackstad and Kjaerheim, 1961; Amaral and
Dent, 1981; Ramón y Cajal, 1911). The terminals
themselves are anchored to the dendritic shaft of
CA3 pyramidal cells and to each other by puncta
adhaerentia.
Within the giant MF terminals (Fig. 1), three
different types of vesicles have been described. The
majority are small clear vesicles (
40 nm) contain-
ing glutamate (Storm-Mathisen, 1981), but the
boutons also contain large dense-core vesicles
containing neuropeptides such as dynorphin
(McGinty et al., 1983), enkephalin (Commons
and Milner, 1996), cholecystokinin (Chandy et al.,
1995), neuropeptide Y (Gall et al., 1990) and
neurokinin-B (Schwarzer et al., 1995) and large
clear vesicles (
200 nm; Amaral and Dent, 1981;
Chicurel and Harris, 1992; Henze et al., 2000).
Besides glutamate and neuropeptides, MF bou-
tons also contain the neuromodulator ATP/
adenosine (Terrian et al., 1989) and zinc (see
section about Timm staining, above). Interest-
ingly, dentate granule cells and MFs also express
both glutamic acid decarboxylase 67 and GABA,
particularly after seizures (Sandler and Smith,
1991; Sloviter et al., 1996).
Each large terminal synaptic complex includes
up to 35 synaptic membrane specializations or ac-
tive zones (Fig. 1) as reported by Chicurel and
Harris (1992), which together with the high density
and composition of presynaptic sodium and cal-
cium ion channels provides each terminal with a
large excitatory drive on the postsynaptic CA3
pyramidal cells (see review by Bischofberger et al.,
2006).
While the large MF terminals normally appear
to exclusively make synaptic contact with CA3
pyramidal cells and hilar mossy cells (Acsády
et al., 1998), other presynaptic MF terminals con-
tact interneurons in hilus and CA3. These presy-
naptic elements are smaller, do not have multiple
active zones, and are located en passant or at the
end of short or long thin filopodia extending from
the giant MF boutons (Acsády et al., 1998), often
up to nine per giant bouton (Amaral, 1979;
Amaral and Dent, 1981; Frotscher, 1985). In de-
veloping rats, some filopodia may extend as far
as 30 mm from the giant boutons in the hilus,
but shorten to approximately 12 mm in adult rats
(Amaral, 1979).
New emerging structural plasticity of individual
mossy fiber terminals
Based on the structural data on individual MF
axonal and terminal organization provided by the
studies cited above, and advanced imaging tech-
niques employed on in vivo hippocampal tissue and
organotypic hippocampal slice cultures derived
from transgenic mice expressing membrane tar-
geted green fluorescent protein (GFP) in few neu-
rons (Thy1-mGFPs ), Caroni and coworkers have
engaged in detailed studies on activity dependent
plasticity of individual MF termination in CA3
(De Paola et al., 2006; Galimberti et al., 2006).
Results obtained so far both in vivo and in long
term slice cultures have shown that one large MF
terminal per MF (and only one) typically is larger

than the others, and that this one, besides of hav-
ing elongated along its postsynaptic CA3 pyram-
idal cell dendrite in complex structural contact
with complex spines, also can have side branches
with other large, satellite MF terminals, contacting
the same and neighboring pyramidal cells, thereby
forming an local and interconnected, large MF
terminal cluster or complex. The formation of in-
dividual larger MF terminals and new satellite
large MF terminals was found to be activity-
dependent, regulated by age and housing in enriched
environments. With increasing age single MF ter-
minals typically attained larger size by elongating
along the postsynaptic CA3 pyramidal cell dend-
rite, at the same time expressing more presynaptic
release sites and engaging in more complex struc-
tural relations with the increasingly complex
postsynaptic spines developing during the same
period. Placing both young adult and older adult
mice in enriched environments for a period of time
on the other hand made the individual large MF
terminals respond by formation of large MF ter-
minal satellites. The mechanisms behind the MF
plasticity elicited by select large MF terminals in a
focal manner, as demonstrated by Caroni and co-
workers, and its role in dentate-hippocampal func-
tional interaction are not known at present. It is,
however, tempting to believe that the focal
expression and strengthening of synaptic contacts
include a postsynaptic feedback from those CA3
pyramidal cells, which at the time of induction are
activated by a net excitatory effect of the direct
stimulatory and indirect inhibitory MF input.
Development of the dentate mossy fiber system
Dentate neurogenesis in brief
The dentate MF system has a unique developmen-
tal profile relative to other central nervous projec-
tions, due to the late start and thereafter extended
neurogenesis of the parent dentate granule popu-
lation, lasting the entire lifetime (Altman and Das,
1965; Angevine, 1965; Bayer, 1980; Eckenhoff and
Rakic, 1988; Kuhn et al., 1996; Eriksson et al.,
1998; Gould et al., 1998; Dawirs et al., 2000; Gage,
2002; Guidi et al., 2005). In rodents, two third or
97
more of the granule cells form after birth, and
predominantly during the first two postnatal
weeks, with ongoing, although declining forma-
tion of new granule cells into adulthood and old
age (Bayer, 1982; Cameron et al., 1993; Kuhn
et al., 1996). Neurogenesis is regulated by genetic
(Kempermann et al., 1997, 2006) and epigenetic
factors, including normal activity dependent ones
and seizures (Gould et al., 1998; Kempermann
et al., 1998; Derrick et al., 2000; Gage, 2000;
Blümcke et al., 2001; Song et al., 2002; Lehmann
et al., 2005; Thom et al., 2005; McCloskey et al.,
2006, and references therein). Before dealing with
the ontogenic development of the MF projection
to the dentate hilus and CA3, it should be kept in
mind that late-forming dentate granule cells also
have been shown to extend MFs into the hilus and
CA3 (Stanfield and Trice, 1988; Hastings and
Gould, 1999; van Praag et al., 2002) as well as
becoming functionally integrated, as exemplified
by the studies of Jessberger and Kempermann
(2003), Santarelli et al. (2003), Bruel-Jungerman
et al. (2005) and Raineteau et al. (2006).
Development of MF axonal and synaptic
connections
The developmental growth of MFs into the dent-
ate hilus and CA3 with accompanying terminal
and synaptic maturation have been studied in
Wistar and Long Evans rats, mice and rabbits by
Timm staining or modifications thereof (Zimmer
and Haug, 1978; Gaarskjaer, 1985; Sanchez-
Andres et al., 1997; Slomianka and Geneser,
1997), supplemented by retrograde axonal tracing
(Gaarskjaer, 1985), Golgi impregnation and EM
(Amaral and Dent, 1981) or just by EM (Stirling
and Bliss, 1978). In rats, the first sign of MF de-
velopment by Timm staining appears at the day of
birth, where they appear in the hilus subjacent to
the early developed parts of the suprapyramidal
granule cell layer, and just above the proximal,
CA3c pyramidal cell layer (at temporal and mid
septo-temporal levels). Development at septal
levels seems to be relatively delayed. Within the
next 3–5 days, a full-length Timm-stained supra-
pyramidal MF layer develops, corresponding to
98
the formation of more granule cells and a widening
and infrapyramidal (medial) extension of cell layer
(Fig. 7A). At this time, Golgi staining revealed
dentate granule cells with a relatively mature ap-
pearance, especially in the suprapyramidal blade,
with MF axons extending into the hilus and send-
ing collaterals to form an infragranular plexus
(Amaral and Dent, 1981). By postnatal day 3,
bundles of MFs were present above the CA3 py-
ramidal cell layer. By retrograde axonal tracing
following microinjection of the fluorescent dye
True Blue in CA3, covering different proximo-
distal distances from the hilus, Gaarskjaer (1985)
demonstrated a rapid elongation of MF along the
transverse axis of CA3 during the very first post-
natal week, with some retrograde labeling of gran-
ule cells in the suprapyramidal layer from the most
distal CA1 close part of CA3 (as early as postnatal
day 1). The rate of MF outgrowth during the first
days was estimated to be at least 0.2 mm at day 1,
and nearly 1 mm at day 4. Retrograde tracing
showed a continued growth along the septotem-
poral axis with a septotemporal descent relative to
the parent cell location of about 0.5 mm at day 1,
reaching nearly 1 mm at day 4 and adult length
of 1.6 mm during the next 3–4 weeks. After 10–
12 days postnatal growth appeared to slow.
Ultrastructurally, MF boutons identified at day 1
contained both clear and dense core vesicles and
formed both symmetrical and asymmetrical synap-
tic contacts with the apical dendrites of CA3
pyramidal cells, despite the fact that spines were
not yet present (Amaral and Dent, 1981). During
the following week, the terminals developed a
more mature and complex morphology, size and
synaptic vesicle content and density. Spines on the
proximal CA3 pyramidal cell dendrites started to
develop about one week postnatal, but a mature,
classic, MF terminal and its intimate relations with
‘‘thorny excrescences’’ developed at three weeks
postnatal. Thereafter, both the size of the MF ter-
minals, and the density of Timm staining, contin-
ued to increase. It should be noted that the MF
system is not mature in two-week-old rats, as
demonstrated by Amaral and Dent (1981), who
found the mean area and perimeter of the MF
terminal profiles to increase substantially during
this period, together with a three-fold increase in
the average number of symmetrical membrane
junctions (punctae adhaerentia) to seven per ter-
minal profile at day 21. Asymmetrical synapses
at the same time reached the adult level of an
average number of nine per terminal profile,
counted in two dimensions only. It should also
be recognized that, after the average number of
invaginating spinules per terminal profile increased
from two at day 21 to five in more than one-year-
old rats.
As illustrated above, the rat MF system is sub-
ject to considerable development during the first
three postnatal weeks and even thereafter, and this
should of course be taken into consideration in
experimental and other studies of this system.
Specific points to note may be that MF synapse
formation precedes the development of thorny
excrecenses on CA3 pyramidal cells. Moreover,
asymmetrical synapses seem to develop at a steady
pace, unrelated to the dramatic structural consol-
idation expressed by a prominent increase in the
non-synaptic puncta adhaerentia from day 14 to
day 21.
Experimental studies on structural and connective
plasticity of dentate mossy fibers, including
dentate transplants
The parent dentate granule cells, the presynaptic
MFs and their terminals and the postsynaptic tar-
get cells in the dentate hilus and the hippocampal
CA3 subfield are each and interdependently
affected by genetic and epigenetic factors and con-
ditions, the effects of which are expressed as direct
structural and connective abnormalities, devia-
tions from normal or natural variations among
species, strains and individuals. New observations
on activity dependent plasticity of individual MF
terminals and their postsynaptic CA3 pyramidal
cells have already been dealt with above, just as
species (and mouse strain) differences have been
presented. This section is therefore devoted to
other experimental manipulations, ranging from
breeding studies and effects of hormonal admin-
istration to studies of lesion-induced effects and
neural transplantation.

Genetics and breeding experiments
Several studies have investigated the MF system in
mice and rats, selectively bred on the basis of be-
havioral criteria such as low and high cognitive
performance levels in defined tests. These studies
have demonstrated significant differences in the
organization and size of the MF projection to CA3
(see Bernasconi-Guastalla et al., 1994; Lipp and
Wolfer, 1998; and references therein). Using such a
heritable difference between two mouse strains
(CH3 and CPB-K) as a starting point, Nek et al.
(1993) demonstrated that MF terminals in the
strain with fewer MFs to CA3 were larger and
established more spine synapses per terminal than
in the strain with more MFs. Accordingly, the
fiber density seemed to have a regulatory effect on
the morphology and synaptic contacts of the
terminals.
Numerous mutants and gene-modified mice
with neurological functional and structural defects
exist, and have been and are currently screened for
effects on cognitive hippocampus-related function
and developmental changes in cellular and con-
nective organization. Only the mutant Reeler
mouse will be mentioned here, because the muta-
tion clearly affects the organization of the dentate
and hippocampal cell layers and as such has
involved experimental studies of the hippocampus
(Förster et al., 2006). As seen in Fig. 3C, the MF
system is present, but appears disorganized in
Reeler mice. This is, however, not due to change in
cellular specificity in termination, but a conse-
quence of the disordered location of cells.
Hormonal effects on MF development and
cognitive function
Lauder and Mugnaini (1980) and Lipp et al. (1984)
early on demonstrated the sensitivity of the rodent
MF system to early postnatal hyperthyreoidism,
inducing an increase in particular infrapyramidal
MFs. Interestingly treatment of mice from a strain
(DBA/2) virtually devoid of infrapyramidal MFs
(correlating with poor radial-maze learning) with
daily injections of 2 mg L-thyroxine early postna-
tally, both induced a prominent infrapyramidal
99
MF projection and a significant improvement in
radial-maze learning (Schwegler et al., 1991).
Stress and glucocorticoid stress hormones also
effect the MF system (McEwen, 1999), and most
interestingly such induced changes seems to be re-
versible by behavioral training (Sandi et al., 2003).
Lesion-induced MF sprouting and rerouting
Several studies referred to above have included
interference with the development or integrity of
dentate granule cells, hilar cells and CA3 pyram-
idal cells resulting in changes in the size and dis-
tribution of MFs, like for example induction of
supragranular MFs following seizures and dener-
vation of the dentate molecular layer (Fig. 5),
changes in the size of infrapyramidal projections
to CA3 or ingrowths into CA1 (Fig. 5). Only a few
studies have addressed the reaction of MFs to
transection during development or in the mature
brain. In two studies on rat MFs by Laurberg and
Zimmer (1980, 1981), supplemented by a subse-
quent EM study (Frotscher and Zimmer, 1983)
and observations in organotypic slice cultures
(Zimmer and Gahwiler, 1984, 1987), it was dem-
onstrated that: (1) the main MF projection to
CA3, with an decrease in effect over the first 3
postnatal weeks, can be rerouted along the septo-
temporal axis in response to lesion-induced re-
moval or blockade of access to CA3 pyramidal
cells at normal levels of trajectory; (2) the rerouted
MFs terminate in expanded suprapyramidal and
infrapyramidal terminal zones in accessible adja-
cent levels; (3) loss of CA3 target cells not by itself
elicits an axonal pruning effect with compensatory
sprouting of MFs or collaterals into the dentate
molecular layer; and (4) lesion-induced sprouting
of MF terminals into the dentate molecular layer
could be induced in both postnatal developing and
young adult rats in a graded manner depending on
the density of denervation of the dentate molecular
layer. As possible regrowth of severed or newly
formed MFs across lesions in CA3 were difficult to
prove in vivo, given the possibility for fibers to
circumvent the lesion and then invade the normal
target area, transections of the MF layer in hip-
pocampal slice cultures and coculturing of dentate
100

and CA3 slice cultures derived from 7-day-old rats
were employed. Under both conditions regrowth
of MFs into distal CA3 required an intimate con-
tact between the proximal transected MF layer
and the distal, normal target zone, contrasting a
vivid growth of MFs onto the substratum of glial
cells surrounding the cultures.
Neural grafting, testing MF growth and plasticity
Intracerebral grafting of tissue blocks of develop-
ing fascia dentata from newborn (or late fetal) rats
and mice have been used to identify factors reg-
ulating the growth and specificity in the formation
of nerve connections in the brain in general and to
and from the dentate gyrus specifically (Sunde and
Zimmer, 1981, 1983; Frotscher and Zimmer, 1986;
Zimmer et al., 1988a, b; Tønder et al., 1990). In
contrast to extensive growth of monoaminergic
and cholinergic graft fibers into host brain and vice
versa at all recipient brain ages, glutamatergic
projections, including those of the hippocampal
region, are clearly dependent of recipient age with
regard to exchange of fibers between the always
developing graft tissue and the host brain. In ac-
cordance with these general observations, MFs
were only seen to grow from dentate grafts into the
recipient brain and from the host brain fascia den-
tata into grafts under certain conditions. More
over, except for a minor projection of MFs grow-
ing from dentate grafts into the surrounding CA1
area (Sunde and Zimmer, 1981), ingrowth and
termination of MFs only occurred into the normal
CA3 target area.
Figure 8 illustrates one experimental setup,
where a small block of neonatal rat fascia dentata
had been grafted into the dentate area (Fig. 8B) or
next to CA3 (Fig. 8C) in other newborn rats with
the purpose of examining whether transplanted
dentata granule cells were able to restore the MF
projection of the recipient rats, which before the

Fig. 8. (A) Timm stained sections of hippocampus from adult rat, subjected to hippocampal x-irradiation as newborn, stop odentate
granule cell formation and lead to few MF terminals. (B) Timm stained section from the contralateral hemisphere of the rat shown in
A, depicting a well-integrated dentate transplant (tpl), derived from a small block of fascia dentata, grafted just after the x-irradiation.
From the graft, an apparently normal, laminar-specific MF projection (mf) developed, connecting dentate granule cells of the graft (g)
with the host CA3. The molecular layer of the graft (m) received a comparably laminar-specific projection of host entorhinal perforant
path projections, normalizing the Timm stained laminar appearance of the layer. (C) Slightly displaced dentate transplant (tpl), grafted
to x-irradiated newborn hippocampus just after irradiation as part of same experiment (Sunde et al., 1984). Encroaching on the
recipient CA3, MFs from the granule cells of this graft has entered adjacent parts of CA3 and project ‘‘downstream’’, but appear to
stop at the border with CA1. Surprisingly, no MFs from the graft projected in the ‘‘upstream’’ direction, i.e., toward the host fascia
dentata. Scale bar: 500 mm (See Color Plate 5.8 in color plate section.)
 101

transplantation had selectively x-irradiated over
the hippocampal region, resulting in a severely re-
duced number of granule cells and MFs for the
rest of the animal’s life (Fig. 8A). In the situation
with both graft and recipient being neonatal and
the graft positioned corresponding to the recipient
fascia dentata (Fig. 8B) there was extensive and
laminar specific outgrowth of host MFs into the
recipient CA3, but no outgrowth ‘‘backward’’ into
areas posterior to the graft. Note also in this ho-
motopically placed graft the extensive, laminar
specific ingrowth of host entorhinal perforant fib-
ers into the outer parts of the graft molecular
layer, illustrated by the presence of normal Timm
stained lamination of this layer (m, Fig. 8B.). With
placement of the dentate graft in the lateral ven-
tricle, encroaching on the distal parts of CA3
(Fig. 8C), graft MFs entered and innervated the
host CA3 in a specific and laminar fashion from
the site of entry and ‘‘downstream’’ to CA1. For
unknown reasons no graft MFs grew in the op-
posite of normal direction toward the residual host
fascia dentata.
Using the presence of the peptide cholecystoki-
nin (CCK) in mouse, but not rat, MF terminals,
Zimmer et al. (1988a, b) moreover demonstrated
that mouse MFs, originating from a newborn
mouse dentate xenograft placed next to the recip-
ient rat fascia dentata, were able to participate in
the normal development of the recipient MF sys-
tem, as illustrated by the presence of mouse xeno-
graft-derived, CCK immunoreactive MF terminals
in the adult host rat CA3 (Fig. 9). With proper
location of dentate grafts providing direct access

Fig. 9. Innervation of rat MF layer by mouse MF terminals from mouse dentate xenograft, topographically integrated with the
recipient rat fascia dentata. (A) Adult rat MF layer stained immunocytochemically for the peptide cholecystokinin (CCK), disclosing
some small-size, terminal-like structures with a preferential distribution among the CA3 pyramidal cells (p), and likely to arise from the
CCK-immunoreactive neurons (arrow heads). (B) Corresponding level of MF layer in contralateral hippocampus, which early post-
natally had received a xenograft of neonatal mouse fascia dentata, encroaching on the rat recipient fascia dentata for some distance
along the septotemporal axis. Corresponding to the levels of xenograft integration with the rat host fascia dentata and for some
additional distance temporally, mouse MF terminals, distinguished by their normal immunoreactivity for CCK, were present in the
host rat MF layer, as documented by their large MF terminal-like size (arrows), and their preferential suprapyramidal position, unlike
the smaller elements in the pyramidal cell layer (p) from the CCK-reaction intrinsic neurons (arrowheads). Scale bar: 50 mm.
102

Fig. 10. Adjacent Timm (A) and toluidine blue (B) stained sections from the dorsal hippocampus of adult rat, showing a focal lesion of
CA3 containing a small graft of fetal rat CA3 neurons, receiving a Timm stained, MF projection (asterisk, A) from the host fascia
dentata. In a two-step procedure, the rat received a localized injection of ibotenic acid, killing all CA3 neurons at the injection site. One
week later a cell suspension of late fetal rat CA3 neurons were grafted into the lesion site, where the afferent fibers, here dentate MFs,
still remained. Several weeks postgrafting, the CA3 graft (tpl) was found to be well integrated, yet not filling the lesion completely. Part
of lesioned CA3 pyramidal cell layer is thus still visible, but densely gliotic (p3). As demonstrated by Timm staining (asterisk, A), host
MFs have innervated the fetal CA3 graft. Scale bar: 100 mm.

for graft MFs to enter the recipient MF layer, Acknowledgment

graft-derived MFs may accordingly innervate the
CA3 area of the host hippocampus.
Ingrowth of adult recipient brain, glutamatergic
fibers into developing graft tissue are normally
very restricted, apparently because the gluta-
matergic fibers, unlike monoaminergic and choli-
nergic host fibers, are competing with the intrinsic
graft fibers, which are in their normal phase of
development. One exception to this is when grafts
are placed in so called ‘‘axon-sparing’’ lesions,
where recipient brain neurons have been killed by
ischemia or excitotoxins like ibotenic acid one or
more days prior to grafting. Using this paradigm,
Tønder et al. (1990) demonstrated that adult re-
cipient brain MFs, primed by preceding ibotenic
acid lesion of their normal target CA3 pyramidal
cells, indeed were able to innervate grafts of
late fetal CA3 pyramidal neurons, placed in
the lesioned CA3 area (Fig. 10). Adult MFs
with already established connection area can
accordingly after preceding axon-sparing lesions
of CA3 enter and terminate in fetal grafts of CA3
neurons.
The authors are grateful for the support received
by former and current colleagues during the writ-
ing of this review, primarily in terms of help with
provision of relevant references for reading and
discussions. Prof. Michael Frotscher, Freiburg,
Germany is particularly acknowledged for provid-
ing the EM pictures used as Fig. 1. Prof. Pico
Caroni, Basel, Switzerland is thanked for personal
communications on the newly emerging MF ter-
minal plasticity.
References
Acsády, L., Kamondi, A., Sik, A., Freund, T. and Buzsaki, G.
(1998) GABAergic cells are the major postsynaptic targets of
mossy fibers in the rat hippocampus. J. Neurosci., 18(9):
3386–3403.
Altman, J. and Das, G.D. (1965) Post-natal origin of micro-
neurones in the rat brain. Nature, 207(5000): 953–956.
Amaral, D.G. (1978) A Golgi study of cell types in the hilar
region of the hippocampus in the rat. J. Comp. Neurol.,
182(4 Pt 2): 851–914.

Amaral, D.G. (1979) Synaptic extensions from the mossy fibers
of the fascia dentata. Anat. Embryol. (Berl.), 155(3):
241–251.
Amaral, D.G. and Dent, J.A. (1981) Development of the mossy
fibers of the dentate gyrus. I. A light and electron microscopic
study of the mossy fibers and their expansions. J. Comp.
Neurol., 195(1): 51–86.
Amaral, D.G. and Witter, M.P. (1989) The three-dimensional
organization of the hippocampal formation: a review of an-
atomical data. Neuroscience, 31(3): 571–591.
Andersen, P., Bliss, T.V., Lomo, T., Olsen, L.I. and Skrede,
K.K. (1969) Lamellar organization of hippocampal excita-
tory pathways. Acta Physiol. Scand., 76(1): 4A–5A.
Angevine, J.B. Jr., (1965) Time of neuron origin in the hippo-
campal region. An autoradiographic study in the mouse.
Exp. Neurol., Suppl 2: 1–70.
Barber, R.P., Vaughn, J.E., Wimer, R.E. and Wimer, C.C.
(1974) Genetically-associated variations in the distribution of
dentate granule cell synapses upon the pyramidal cell dend-
rites in mouse hippocampus. J. Comp. Neurol., 156(4):
417–434.
Barthesaghi, R. and Ravasi, L. (1999) Pyramidal types
in the field CA2 of the guinea pig. Brain Res. Bull., 50:
263–273.
Bayer, S.A. (1980) Development of the hippocampal region in
the rat. I. Neurogenesis examined with 3H-thymidine auto-
radiography. J. Comp. Neurol., 190(1): 87–114.
Bayer, S.A. (1982) Changes in the total number of dentate
granule cells in juvenile and adult rats: a correlated volumet-
ric and 3H-thymidine autoradiographic study. Exp. Brain
Res., 46(3): 315–323.
Bernasconi-Guastalla, S., Wolfer, D.P. and Lipp, H.P. (1994)
Hippocampal mossy fibers and swimming navigation in mice:
correlations with size and left-right asymmetries. Hippocam-
pus, 4(1): 53–63.
Bischofberger, J., Engel, D., Frotscher, M. and Jonas, P. (2006)
Timing and efficacy of transmitter release at mossy fiber
synapses in the hippocampal network. Pflugers Arch., 453(3):
361–372.
Blackstad, T.W. (1956) Commissural connections of the hip-
pocampal region in the rat, with special reference to their
mode of termination. J. Comp. Neurol., 105(3): 417–537.
Blackstad, T.W. (1963) Ultrastructural studies on the hippo-
campal region. Prog. Brain Res., 3: 122–148.
Blackstad, T.W., Brink, K., Hem, J. and Jeune, B. (1970) Dis-
tribution of hippocampal mossy fibers in the rat. An exper-
imental study with silver impregnation methods. J. Comp.
Neurol., 138(4): 433–449.
Blackstad, T.W. and Kjaerheim, A. (1961) Special axo-
dendritic synapses in the hippocampal cortex: electron and
light microscopic studies on the layer of mossy fibers.
J. Comp. Neurol., 117: 133–159.
Blümcke, I., Schewe, J.C., Normann, S., Brustle, O., Schramm,
J., Elger, C.E. and Wiestler, O.D. (2001) Increase of nestin-
immunoreactive neural precursor cells in the dentate gyrus of
pediatric patients with early-onset temporal lobe epilepsy.
Hippocampus, 11(3): 311–321.
103
Blümcke, I., Suter, B., Behle, K., Kuhn, R., Schramm, J.,
Elger, C.E. and Wiestler, O.D. (2000) Loss of hilar mossy
cells in Ammon’s horn sclerosis. Epilepsia, 41(Suppl 6):
S174–S180.
Bruel-Jungerman, E., Laroche, S. and Rampon, C. (2005) New
neurons in the dentate gyrus are involved in the expression of
enhanced long-term memory following environmental en-
richment. Eur. J. Neurosci., 21(2): 513–521.
Cameron, H.A., Woolley, C.S., McEwen, B.S. and Gould, E.
(1993) Differentiation of newly born neurons and glia in the
dentate gyrus of the adult rat. Neuroscience, 56(2): 337–344.
Cassell, M.D. and Brown, M.W. (1984) The distribution of
Timm’s stain in the nonsulphide-perfused human hippocam-
pal formation. J. Comp. Neurol., 222(3): 461–471.
Chandy, J., Pierce, J.P. and Milner, T.A. (1995) Rat hippo-
campal mossy fibers contain cholecystokinin-like immunore-
activity. Anat. Rec., 243(4): 519–523.
Chicurel, M.E. and Harris, K.M. (1992) Three-dimensional
analysis of the structure and composition of CA3 branched
dendritic spines and their synaptic relationships with mossy
fiber boutons in the rat hippocampus. J. Comp. Neurol.,
325(2): 169–182.
Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1986) A light
and electron microscopic analysis of the mossy fibers of the
rat dentate gyrus. J. Comp. Neurol., 246(4): 435–458.
Commons, K.G. and Milner, T.A. (1996) Ultrastructural rela-
tionships between leu-enkephalin- and GABA-containing
neurons differ within the hippocampal formation. Brain Res.,
724(1): 1–15.
Cook, T.M. and Crutcher, K.A. (1985) Extensive target cell loss
during development results in mossy fibers in the regio su-
perior (CA1) of the rat hippocampal formation. Brain Res.,
353(1): 19–30.
Danscher, G. (1981) Histochemical demonstration of heavy
metals. A revised version of the sulphide silver method suit-
able for both light and electronmicroscopy. Histochemistry,
71(1): 1–16.
Danscher, G., Howell, G., Perez-Clausell, J. and Hertel, N.
(1985) The dithizone, Timm’s sulphide silver and the sele-
nium methods demonstrate a chelatable pool of zinc in CNS.
A proton activation (PIXE) analysis of carbon tetrachloride
extracts from rat brains and spinal cords intravitally treated
with dithizone. Histochemistry, 83(5): 419–422.
Danscher, G. and Zimmer, J. (1978) An improved Timm sul-
phide silver method for light and electron microscopic local-
ization of heavy metals in biological tissues. Histochemistry,
55(1): 27–40.
Dawirs, R.R., Teuchert-Noodt, G., Hildebrandt, K. and Fei, F.
(2000) Granule cell proliferation and axon terminal degra-
dation in the dentate gyrus of gerbils (Meriones unguiculatus)
during maturation, adulthood and aging. J. Neural Transm.,
107(6): 639–647.
De Paola, V., Arber, S. and Caroni, P. (2006) AMPA receptors
regulat edynamic equilibrium of presynaptic terminals in
mature hippocampal networks. Nat. Neurosci., 6: 491–500.
Derrick, B.E., York, A.D. and Martinez Jr., J.L. (2000) In-
creased granule cell neurogenesis in the adult dentate gyrus
104
following mossy fiber stimulation sufficient to induce long-
term potentiation. Brain Res., 857(1–2): 300–307.
Dimitrieva, N.I., Gozzo, S., Dmitriev, I.u.S., Lopatina, N.G.
and Kassil’, V.G. (1993) Neuroanatomical characteristics of
rat strains differing in the ability to form active avoidance
conditioned reflexes. Dokl. Akad. Nauk SSSR, 272(5):
1235–1238.
Doinikow, B. (1908) Beitrag zur histologie des Ammonshorns.
J. Psychol. Neurol. (Leipzig), 13: 166–202.
Duffy, C.J. and Rakic, P. (1983) Differentiation of granule cell
dendrites in the dentate gyrus of the rhesus monkey: a quan-
titative Golgi study. J. Comp. Neurol., 214(2): 224–237.
Eckenhoff, M.F. and Rakic, P. (1988) Nature and fate of pro-
liferative cells in the hippocampal dentate gyrus during the
life span of the rhesus monkey. J. Neurosci., 8(8): 2729–2747.
Eriksson, P.S., Perfilieva, E., Bjork-Eriksson, T., Alborn, A.M.,
Nordborg, C., Peterson, D.A. and Gage, F.H. (1998) Ne-
urogenesis in the adult human hippocampus. Nat. Med.,
4(11): 1313–1317.
Fitzpatrick, D. and Johnson, R.P. (1981) Enkephalin-like
immunoreactivity in the mossy fiber pathway of the hippo-
campal formation of the tree shrew (Tupaia glis). Neuro-
science, 6(12): 2485–2494.
Förster, E., Jossin, Y., Zhao, X., Frotscher, M. and Goffinet,
A.M. (2006) Recent progress in understanding the role
of Reelin in the radial neuronal migration, with specific
emphasis on the dentate gyrus. Eur. J. Neurosci., 23:
901–909.
Fredens, K. (1981) Genetic variation in the histoarchitecture of
the hippocampal region of mice. Anat. Embryol. (Berl.),
161(3): 265–281.
Frotscher, M. (1985) Mossy fibres form synapses with identified
pyramidal basket cells in the CA3 region of the guinea-pig
hippocampus: a combined Golgi-electron microscope study.
J. Neurocytol., 14(2): 245–259.
Frotscher, M. (1989) Mossy fiber synapses on glutamate dec-
arboxylase-immunoreactive neurons: evidence for feed-for-
ward inhibition in the CA3 region of the hippocampus. Exp.
Brain Res., 75(2): 441–445.
Frotscher, M., Jonas, P. and Sloviter, R.S. (2006) Synapses
formed by normal and abnormal hippocampal mossy fibers.
Cell Tissue Res., 326(2): 361–367.
Frotscher, M. and Seress, L. (1991) Basket cells in the monkey
fascia dentata: a Golgi/electron microscopic study. J. Ne-
urocytol., 20(11): 915–928.
Frotscher, M., Soriano, E. and Misgeld, U. (1994) Divergence
of hippocampal mossy fibers. Synapse, 16(2): 148–160.
Frotscher, M. and Zimmer, J. (1983) Lesion-induced mossy
fibers to the molecular layer of the rat fascia dentata: iden-
tification of postsynaptic granule cells by the Golgi-EM
technique. J. Comp. Neurol., 215(3): 299–311.
Frotscher, M. and Zimmer, J. (1986) Intracerebral transplants
of the rat fascia dentata: a Golgi/electron microscope
study of dentate granule cells. J. Comp. Neurol., 246(2):
181–190.
Gaarskjaer, F.B. (1978a) Organization of the mossy fiber sys-
tem of the rat studied in extended hippocampi. II.
Experimental analysis of fiber distribution with silver im-
pregnation methods. J. Comp. Neurol., 178(1): 73–88.
Gaarskjaer, F.B. (1978b) Organization of the mossy fiber sys-
tem of the rat studied in extended hippocampi. I. Terminal
area related to number of granule and pyramidal cells. J.
Comp. Neurol., 178(1): 49–72.
Gaarskjaer, F.B. (1981) The hippocampal mossy fiber system of
the rat studied with retrograde tracing techniques. Correla-
tion between topographic organization and neurogenetic
gradients. J. Comp. Neurol., 203(4): 717–735.
Gaarskjaer, F.B. (1985) The development of the dentate area
and the hippocampal mossy fiber projection of the rat. J.
Comp. Neurol., 241(2): 154–170.
Gaarskjaer, F.B. (1986) The organization and development of
the hippocampal mossy fiber system. Brain Res., 396(4):
335–357.
Gaarskjaer, F.B., Danscher, G. and West, M.J. (1982) Hippo-
campal mossy fibers in the regio superior of the European
hedgehog. Brain Res., 237(1): 79–90.
Gage, F.H. (2000) Mammalian neural stem cells. Science,
287(5457): 1433–1438.
Gage, F.H. (2002) Neurogenesis in the adult brain. J. Neurosci.,
22(3): 612–613.
Galimberti, I., Gogolia, N., Alberi, S., Santos, A.F., Muller, D.
and Caroni, P. (2006) Long-term rearrangements of hippo-
campal mossy fiber terminal connectivity in the adult regu-
lated by experience. Neuron, 50: 749–763.
Gall, C. (1984) The distribution of cholecystokinin-like
immunoreactivity in the hippocampal formation of the
guinea pig: localization in the mossy fibers. Brain Res.,
306(1–2): 73–83.
Gall, C. (1988) Seizures induce dramatic and distinctly different
changes in enkephalin, dynorphin, and CCK immunoreac-
tivities in mouse hippocampal mossy fibers. J. Neurosci., 6:
1852–1862.
Gall, C., Lauterborn, J., Isackson, P. and White, J. (1990) Sei-
zures, neuropeptide regulation, and mRNA expression in the
hippocampus. Prog. Brain Res., 83: 371–390.
Ganser, S. (1882) Vergleichend-anatomische studien über das
gehirn des maulwurfs. Morphol. Jahrb., 7: 591–725.
Geneser-Jensen, F.A., Haug, F.M. and Danscher, G. (1974)
Distribution of heavy metals in the hippocampal region of
the guinea pig. A light microscope study with Timm’s sulfide
silver method. Z. Zellforsch. Mikrosk. Anat., 147(4):
441–478.
Golgi, C. (1886) Sulla Fina Anatomia Degli Organi Centrali
Del Sistema Nervoso. U. Hoepli, Milano, p. 215.
Gould, E., Tanapat, P., McEwen, B.S., Flugge, G. and Fuchs,
E. (1998) Proliferation of granule cell precursors in the dent-
ate gyrus of adult monkeys is diminished by stress. Proc.
Natl. Acad. Sci. U.S.A., 95(6): 3168–3171.
Guidi, S., Ciani, E., Severi, S., Contestabile, A. and Bartesaghi,
R. (2005) Postnatal neurogenesis in the dentate gyrus of the
guinea pig. Hippocampus, 15(3): 285–301.
Hastings, N.B. and Gould, E. (1999) Rapid extension of axons
into the CA3 region by adult-generated granule cells. J.
Comp. Neurol., 413(1): 146–154.

Haug, F.M. (1967) Electron microscopical localization of the
zinc in hippocampal mossy fibre synapses by a modified
sulfide silver procedure. Histochemie, 8(4): 355–368.
Haug, F.M. (1974) Light microscopical mapping of the hippo-
campal region, the pyriform cortex and the corticomedial
amygdaloid nuclei of the rat with Timm’s sulphide silver
method. I. Area dentata, hippocampus and subiculum.
Z. Anat. Entwicklungsgesch., 145(1): 1–27.
Henze, D.A., Urban, N.N. and Barrionuevo, G. (2000) The
multifarious hippocampal mossy fiber pathway: a review.
Neuroscience, 98(3): 407–427.
Hjorth-Simonsen, A. (1977) Distribution of commissural affer-
ents to the hippocampus of the rabbit. J. Comp. Neurol.,
176(4): 495–513.
Holm, I.E. and Geneser, F.A. (1991) Histochemical
demonstration of zinc in the hippocampal region of the do-
mestic pig. III. The dentate area. J. Comp. Neurol., 308(3):
409–417.
Holm, I.E., Geneser, F.A. and Zimmer, J. (1992) Somatostatin-
and neuropeptide Y-like immunoreactivity in the dentate
area, hippocampus, and subiculum of the domestic pig. J.
Comp. Neurol., 322(3): 390–408.
Holm, I.E., Geneser, F.A. and Zimmer, J. (1993) Cholecysto-
kinin-, enkephalin-, and substance P-like immunoreactivity in
the dentate area, hippocampus, and subiculum of the do-
mestic pig. J. Comp. Neurol., 331(3): 310–325.
Holm, I.E. and West, M.J. (1994) Hippocampus of the domes-
tic pig: a stereological study of subdivisional volumes and
neuron numbers. Hippocampus, 4(1): 115–125.
Jensen, S., Sorensen, T., Møller, A.G. and Zimmer, J. (1984)
Intraocular grafts of fresh and freeze-stored rat hippocampal
tissue: a comparison of survivability and histological and
connective organization. J. Comp. Neurol., 227(4): 559–568.
Jessberger, S. and Kempermann, G. (2003) Adult-born hippo-
campal neurons mature into activity-dependent responsive-
ness. Eur. J. Neurosci., 18(10): 2707–2712.
Kempermann, G., Chesler, E.J., Lu, L., Williams, R.W. and
Gage, F.H. (2006) Natural variation and genetic covariance
in adult hippocampal neurogenesis. Proc. Natl. Acad. Sci.
U.S.A., 103(3): 780–785.
Kempermann, G., Kuhn, H.G. and Gage, F.H. (1997) Genetic
influence on neurogenesis in the dentate gyrus of adult mice.
Proc. Natl. Acad. Sci. U.S.A., 94(19): 10409–10414.
Kempermann, G., Kuhn, H.G. and Gage, F.H. (1998) Expe-
rience-induced neurogenesis in the senescent dentate gyrus. J.
Neurosci., 18(9): 3206–3212.
Keuker, J.I., Rochford, C.D., Witter, M.P. and Fuchs, E.
(2003) A cytoarchitectonic study of the hippocampal forma-
tion of the tree shrew (Tupaia belangeri). J. Chem. Ne-
uroanat., 26(1): 1–15.
Koelliker, A. (1896) Handbuch der Gewebelehre des Menchen
2. Nervensystem des Menchen und der Thiere. Wilhelm En-
gelmann, Leipzig, p. 874.
Kuhn, H.G., Dickinson-Anson, H. and Gage, F.H. (1996) Ne-
urogenesis in the dentate gyrus of the adult rat: age-related
decrease of neuronal progenitor proliferation. J. Neurosci.,
16(6): 2027–2033.
105
Lauder, J.M. and Mugnaini, E. (1980) Infrapyramidal mossy
fibers in the hippocampus of the hyperthyroid rat. A light
and electron microscopic study. Dev. Neurosci., 3(4–6):
248–265.
Laurberg, S. and Zimmer, J. (1980) Aberrant hippocampal
mossy fibers in cats. Brain Res., 188(2): 555–559.
Laurberg, S. and Zimmer, J. (1981) Lesion-induced sprouting
of hippocampal mossy fiber collaterals to the fascia dentata
in developing and adult rats. J. Comp. Neurol., 200(3):
433–459.
Lehmann, K., Butz, M. and Teuchert-Noodt, G. (2005) Offer
and demand: proliferation and survival of neurons in the
dentate gyrus. Eur. J. Neurosci., 21(12): 3205–3216.
Lein, E.S., Callaway, E.M., Albright, T.D. and Gage, F.H.
(2005) Redefining the boundaries of the hippocampal
CA2 subfield in the mouse using gene expression and
3-dimensional reconstruction. J. Comp. Neurol., 485(1):
1–10.
Lim, C., Blume, H.W., Madsen, J.R. and Saper, C.B. (1997)
Connections of the hippocampal formation in humans: I. The
mossy fiber pathway. J. Comp. Neurol., 385(3): 325–351.
Lipp, H.P., Schwegler, H. and Driscoll, P. (1984) Postnatal
modification of hippocampal circuitry alters avoidance learn-
ing in adult rats. Science, 225(4657): 80–82.
Lipp, H.P. and Wolfer, D.P. (1998) Genetically modified mice
and cognition. Curr. Opin. Neurobiol., 8(2): 272–280.
Lorente de Nó, R. (1934) Studies on the structure of the cer-
ebral cortex II. Continuation of the study of the ammonic
system. J. Psychol. Neurol. (Leipzig), 46: 113–177.
Lynch, G., Smith, R.L., Mensah, P. and Cotman, C. (1973)
Tracing the dentate gyrus mossy fiber system with horserad-
ish peroxidase histochemistry. Exp. Neurol., 40(2): 516–524.
Madeira, M.D., Paula-Barbosa, M., Cadete-Leite, A. and Tav-
ares, M.A. (1988) Unbiased estimate of hippocampal granule
cell numbers in hypothyroid and in sex-age-matched control
rats. J. Hirnforsch., 29(6): 643–650.
McCloskey, D.P., Hintz, T.M., Pierce, J.P. and Scharfman,
H.E. (2006) Stereological methods reveal the robust size and
stability of ectopic hilar granule cells after pilocarpine-in-
duced status epilepticus in the adult rat. Eur. J. Neurosci.,
24(8): 2203–2210.
McEwen, B.S. (1999) Stress and hippocampal plasticity. Ann.
Rev. Neurosci., 22: 105–122.
McGinty, J.F., Henriksen, S.J., Goldstein, A., Terenius, L. and
Bloom, F.E. (1983) Dynorphin is contained within hippo-
campal mossy fibers: immunochemical alterations after
kainic acid administration and colchicine-induced neurotox-
icity. Proc. Natl. Acad. Sci. U.S.A., 80(2): 589–593.
Mello, L.E., Cavalheiro, E.A., Tan, A.M., Kupfer, W.R., Pre-
torius, J.K., Babb, T.L. and Finch, D.M. (1993) Circuit
mechanisms of seizures in the pilocarpine model of chronic
epilepsy: cell loss and mossy fiber sprouting. Epilepsia, 34(6):
985–995.
Nek, N., Schwegler, H., Crusio, WE. and Frotscher, M. (1993)
Are the fine-structural characteristics of mouse hippocampal
mossy fiber synapses determined by the density of mossy fiber
axons? Neurosci. Lett., 158(1): 75–78.
106
van Praag, H., Schinder, A.F., Christie, B.R., Toni, N., Palmer,
T.D. and Gage, F.H. (2002) Functional neurogenesis in the
adult hippocampus. Nature, 415(6875): 1030–1034.
Raineteau, O., Hugel, S., Ozen, I., Rietschin, L., Sigrist, M.,
Arber, S. and Gahwiler, B.H. (2006) Conditional labeling of
newborn granule cells to visualize their integration into es-
tablished circuits in hippocampal slice cultures. Mol. Cell.
Neurosci., 32(4): 344–355.
Ramón y Cajal, S.R. (1893) Estructura del asta de Ammon.
Anal. Soc. Esp. Hist. Nat. (Madrid), 22: 53–114.
Ramón y Cajal, S.R. (1911) Histologie du Système Nerveux de
l’Homme et des Vertébrés, Vol. II. Maloine, Paris.
Ribak, C.E. and Peterson, G.M. (1991) Intragranular mossy
fibers in rats and gerbils form synapses with the somata and
proximal dendrites of basket cells in the dentate gyrus. Hip-
pocampus, 1(4): 355–364.
Sala, L. (1891) Zur feineren anatomie des grossen seepherde-
fusses. Z. Wiss. Zoll., 52: 18–45.
Sanchez-Andres, J.V., Palop, J.J., Ramirez, C., Nacher, J.,
Molowny, A. and Lopez-Gracia, C. (1997) Zinc-positive pre-
synaptic boutons of the rabbit hippocampus during early
postnatal development. Dev. Brain Res., 103(2): 171–183.
Sandi, C., Davies, H.A., Cordero, M.I., Rodriquez, J.J., Popov,
V.I. and Stewart, M.G. (2003) Rapid reversal of stress in-
duced loss of synapses in CA3 of rat hippocampal following
water maze training. Eur. J. Neurosci., 17: 2447–2456.
Sandler, R. and Smith, A. (1991) Coexistence of GABA and
glutamate in mossy fiber terminals of the primate hippocam-
pus: an ultrastructural study. J. Comp. Neurol., 303(2):
177–192.
Santarelli, L., Saxe, M., Gross, C., Surget, A., Battaglia, F.,
Dulawa, S., Weisstaub, N., Lee, J., Duman, R., Arancio, O.,
Belzung, C. and Hen, R. (2003) Requirement of hippocampal
neurogenesis for the behavioral effects of antidepressants.
Science, 301(5634): 805–809.
Schaffer, K. (1882) Beitrag zur histologie der Ammonshorn-
formation. Arch. Mikr. Anat., 39: 611–632.
Schwarzer, C., Williamson, J.M., Lothman, E.W., Vezzani, A.
and Sperk, G. (1995) Somatostatin, neuropeptide Y, ne-
urokinin B and cholecystokinin immunoreactivity in two
chronic models of temporal lobe epilepsy. Neuroscience,
69(3): 831–845.
Schwegler, H., Crusio, W.E., Lipp, H.P., Brust, I. and Mueller,
G.G. (1991) Early postnatal hyperthyroidism alters hippo-
campal circuitry and improves radial-maze learning in adult
mice. J. Neurosci., 11(7): 2102–2106.
Slomianka, L. and Geneser, F.A. (1997) Postnatal development
of zinc-containing cells and neuropil in the hippocampal re-
gion of the mouse. Hippocampus, 7(3): 321–340.
Sloviter, R.S., Dichter, M.A., Rachinsky, T.L., Dean, E.,
Goodman, J.H., Sollas, A.L. and Martin, D.L. (1996) Ba-
salin expression and duction of glutamate decarboxylase and
GABA in excitatory granule cells of the rat and monkey
hippocampal dentate gyrus. J. Comp. Neurol., 373(4):
593–618.
Sloviter, R.S., Zappone, C.A., Harvey, B.D. and Frotscher, M.
(2006) Kainic acid-induced recurrent mossy fiber innervation
of dentate gyrus inhibitory interneurons: possible anatomical
substrate of granule cell hyper-inhibition in chronically
epileptic rats. J. Comp. Neurol., 494(6): 944–960.
Song, H., Stevens, C.F. and Gage, F.H. (2002) Neural
stem cells from adult hippocampus develop essential prop-
erties of functional CNS neurons. Nat. Neurosci., 5(5):
438–445.
Soriano, E., Cobas, A., Lopez, C. and Berbel, P. (1983) Mor-
phology of mossy fiber collaterals in the hilus of the rat.
Neurosci. Lett. Suppl., 14: S352.
Stanfield, B.B. and Trice, J.E. (1988) Evidence that granule cells
generated in the dentate gyrus of adult rats extend axonal
projections. Exp. Brain Res., 72(2): 399–406.
Stengaard-Pedersen, K., Fredens, K. and Larsson, L.I. (1983)
Comparative localization of enkephalin and cholecystokinin
immunoreactivities and heavy metals in the hippocampus.
Brain Res., 273(1): 81–96.
Stirling, R.V. and Bliss, T.V. (1978) Hippocampal mossy fiber
development at the ultrastructural level. Prog. Brain Res., 48:
191–198.
Storm-Mathisen, J. (1981) Glutamate in hippocampal path-
ways. Adv. Biochem. Psychopharmacol., 27: 43–55.
Sunde, N., Laurberg, S. and Zimmer, J. (1984) Brain grafts can
restore irradiation-damaged neuronal connections in new-
born rats. Nature, 310(5972): 51–53.
Sunde, N. and Zimmer, J. (1981) Transplantation of central
nervous tissue. An introduction with results and implications.
Acta Neurol. Scand., 63(5): 323–335.
Sunde, N.A. and Zimmer, J. (1983) Cellular, histochemical and
connective organization of the hippocampus and fascia den-
tata transplanted to different regions of immature and adult
rat brains. Brain Res., 284(2–3): 165–191.
Sutula, T., Cascino, G., Cavazos, J., Parada, I. and Ramirez, L.
(1989) Mossy fiber synaptic reorganization in the epileptic
human temporal lobe. Ann. Neurol., 26(3): 321–330.
Sutula, T., He, X.X., Cavazos, J. and Scott, G. (1988) Synaptic
reorganization in the hippocampus induced by abnormal
functional activity. Science, 239(4844): 1147–1150.
Swanson, L.W., Wyss, J.M. and Cowan, W.M. (1978) An au-
toradiographic study of the organization of intrahippocam-
pal association pathways in the rat. J. Comp. Neurol., 181(4):
681–715.
Tamamaki, N., Abe, K. and Nojyo, Y. (1988) Three-dimen-
sional analysis of the whole axonal arbors originating from
single CA2 pyramidal neurons in the rat hippocampus with
the aid of a computer graphic technique. Brain Res.,
452(1–2): 255–272.
Tamamaki, N. and Nojyo, Y. (1991) Crossing fiber arrays in
the rat hippocampus as demonstrated by three-dimensional
reconstruction. J. Comp. Neurol., 303(3): 435–442.
Terrian, D.M., Hernandez, P.G., Rea, M.A. and Peters, R.I.
(1989) ATP release, adenosine formation, and modulation of
dynorphin and glutamic acid release by adenosine analogues
in rat hippocampal mossy fiber synaptosomes. J. Neuroc-
hem., 53(5): 1390–1399.
Thom, M., Martinian, L., Williams, G., Stoeber, K. and
Sisodiya, S.M. (2005) Cell proliferation and granule cell

dispersion in human hippocampal sclerosis. J. Neuropathol.
Exp. Neurol., 64(3): 194–201.
Timm, F. (1958) Histochemistry of the region of Ammon’s
horn. Z. Zellforsch. Mikrosk. Anat., 48(5): 548–555.
Tønder, N., Sørensen, T. and Zimmer, J. (1990) Grafting of
fetal CA3 neurons to excitotoxic, axon-sparing lesions of the
hippocampal CA3 area in adult rats. Prog. Brain Res., 83:
391–409.
Vaughn, J.E., Matthews, D.A., Barber, R.P., Wimer, C.C. and
Wimer, R.E. (1977) Genetically-associated variations in the
development of hippocampal pyramidal neurons may pro-
duce differences in mossy fiber connectivity. J. Comp. Ne-
urol., 173(1): 41–52.
Vida, I. and Frotscher, M. (2000) A hippocampal interneuron
associated with the mossy fiber system. Proc. Natl. Acad. Sci.
U.S.A., 97(3): 1275–1280.
West, J.R., Van Hoesen, G.W. and Kosel, K.C. (1982) A dem-
onstration of hippocampal mossy fiber axon morphology
using the anterograde transport of horseradish peroxidase.
Exp. Brain Res., 48(2): 209–216.
West, M.J., Gaarskjaer, F.B. and Danscher, G. (1984) The
Timm-stained hippocampus of the European hedgehog: a
basal mammalian form. J. Comp. Neurol., 226(4): 477–488.
West, M.J. and Gundersen, H.J. (1990) Unbiased stereological
estimation of the number of neurons in the human hippo-
campus. J. Comp. Neurol., 296(1): 1–22.
West, M.J., Slomianka, L. and Gundersen, H.J. (1991)
Unbiased stereological estimation of the total number of
107
neurons in the subdivisions of the rat hippocampus using the
optical fractionator. Anat. Rec., 231(4): 482–497.
Zimmer, J. (1973) Changes in the Timm sulfide silver
staining pattern of the rat hippocampus and fascia dentata
following early postnatal deafferentation. Brain Res., 64:
313–326.
Zimmer, J. (1974) Long term synaptic reorganization in rat
fascia dentata deafferented at adolescent and adult stages:
observations with the Timm method. Brain Res., 76(2):
336–342.
Zimmer, J., Finsen, B., Sorensen, T. and Poulsen, P.H. (1988a)
Xenografts of mouse hippocampal tissue. Exchange of lami-
nar and neuropeptide specific nerve connections with the host
rat brain. Brain Res. Bull., 3: 369–379.
Zimmer, J., Finsen, B., Sorensen, T. and Poulsen, P.H. (1988b)
Xenografts of mouse hippocampal tissue. Formation of nerve
connections between the graft fascia dentata and the host rat
brain. Prog. Brain Res., 78: 271–280.
Zimmer, J. and Gahwiler, B.H. (1984) Cellular and connective
organization of slice cultures of the rat hippocampus and
fascia dentata. J. Comp. Neurol., 228(3): 432–446.
Zimmer, J. and Gahwiler, B.H. (1987) Growth of hippocampal
mossy fibers: a lesion and coculture study of organotypic slice
cultures. J. Comp. Neurol., 264(1): 1–13.
Zimmer, J. and Haug, F.M. (1978) Laminar differentiation of
the hippocampus, fascia dentata and subiculum in developing
rats, observed with the Timm sulphide silver method. J.
Comp. Neurol., 179(3): 581–617.
Plate 5.2. Timm stained hippocampal sections from rat (A), guinea pig (B), cat (C), European hedgehog (D), dog (E) and man (F),
illustrating the distribution of intensely stained, black MF terminals in the different species, as well as other terminal fields and general
organization. Note species specific characteristics, like the ‘‘end bulb’’ at the terminal part of the MF layer in guinea pig (black asterisk,
B) and the layering of the dentate hilus with MF terminal-free, intermediate or plexiforme sublayer in the same species (white asterisk,
B), and MF projections into CA1 in some cats (arrows, C) and European hedgehog (D). Except for the black MF terminal staining, the
neuropil of the human hippocampus is virtually unstained, due to the use of special Timm staining procedures for non-perfused tissue
(Danscher and Zimmer, 1978). Abbreviations: CA1, hippocampal regio superior or subfield CA1; CA3, hippocampal regio inferior or
subfield CA3; FD, fascia dentata (or denate gyrus); g, dentate granule cell layer; h, dentate hilus or CA4; m, dentate molecular layer;
mf, MF layer; Sub, subiculum. Scale bars: 500 mm (For B/W version, see page 88 in the volume.)
Plate 5.3. Timm stained mouse hippocampal sections, illustrating strain differences in MFs terminal projections between BALB7C (A)
and C57B mice (B) and Reeler mutants (C). BALB/c mice have no infrapyramidal MFs, but a short intrapyramidal bundle (black
asterisk, A), compared to long infrapyramidal bundles in C57B mice (black asterisk, B). In Reeler mutant mice, granule and pyramidal
cell layers are disorganized, resulting in a corresponding distribution of MF terminals in the confluent granule cell layer/hilar area and
CA3. However, MFs still avoid CA1, similar to normal mice. Abbreviations: g, dentate granule cell layer; h, dentate hilus (CA4); m,
dentate molecular layer. Scale bar: 500 mm (For B/W version, see page 89 in the volume.)

Plate 5.5. Timm stained section from dorsoposterior level of young adult, rat hippocampus. When newborn, the rat received a
mechanical lesion of the CA3–CA1 transition (les) as well a transection of the entorhinal perforant path to the fascia dentata. The
CA3–CA1 lesion induced an aberrant ingrowth of MFs into CA1, illustrated by presence of small, Timm stained MF terminals
(asterisk) along the CA1 pyramidal cell layer. The removal of the perforant path projection to the outer molecular layer (m) induced
spread of associational-commissural projections from inner part of the layer, and induction of supragranular MF terminals (arrow)
(see Zimmer, 1973, 1974). Abbreviations: g, granule cell layer. Scale bar: 500 mm (For B/W version, see page 92 in the volume.)
Plate 5.6. A: Timm stained MF (collateral) terminals, found in normal adult rats to accompany dendrites (arrows), arising from
neurons in the limiting subzone just below the dentate granule cell layer or in the dentate hilus, and extending into and sometimes
through the dentate granule cell layer. (g) into the, commissural-associational terminal zone (c/a). B: Timm stained, dense supra-
granular MF terminal projection (sgr) found normally at temporal levels of the cat fascia dentata as well as in other species. h, dentate
hilus. Scale bar: 50 mm (A); 100 mm (B). (For B/W version, see page 94 in the volume.)

Plate 5.7. (A) Timm stained section from midposterior level of 5-day-old rat, demonstrating a distinct Timm stained MF layer in CA3
and the hilus at this age (arrows). (B, C) Parallel sections of hippocampal slice culture, derived from 7-day-old rat and grown for 3
weeks. Timm stain reveals the distribution of aberrant supragranular and normal CA3 MFs (arrow and mf, respectively, in B) and
thionine cell staining to visualize general cellular organization (C). Abbreviations: g, granule cell layer. Scale bars: 500 mm (For B/W
version, see page 95 in the volume.)
Plate 5.8. (A) Timm stained sections of hippocampus from adult rat, subjected to hippocampal x-irradiation as newborn, stop
odentate granule cell formation and lead to few MF terminals. (B) Timm stained section from the contralateral hemisphere of the rat
shown in A, depicting a well-integrated dentate transplant (tpl), derived from a small block of fascia dentata, grafted just after the x-
irradiation. From the graft, an apparently normal, laminar-specific MF projection (mf) developed, connecting dentate granule cells of
the graft (g) with the host CA3. The molecular layer of the graft (m) received a comparably laminar-specific projection of host
entorhinal perforant path projections, normalizing the Timm stained laminar appearance of the layer. (C) Slightly displaced dentate
transplant (tpl), grafted to x-irradiated newborn hippocampus just after irradiation as part of same experiment (Sunde et al., 1984).
Encroaching on the recipient CA3, MFs from the granule cells of this graft has entered adjacent parts of CA3 and project ‘‘down-
stream’’, but appear to stop at the border with CA1. Surprisingly, no MFs from the graft projected in the ‘‘upstream’’ direction, i.e.,
toward the host fascia dentata. Scale bar: 500 mm (For B/W version, see page 100 in the volume.)

Plate 6.1. Interneurons are the primary target of DG granule cells. Diagram representing the multiple types of excitatory contacts
made by granule cells. Within the dentate/hilar region collaterals of the mossy fibers (MFs) innervate both mossy cells and inhibitory
basket cells. In area CA3, the MFs contact pyramidal neurons via large expansions, but also excite interneurons through either
filopodial extensions from the large boutons or smaller en passant expansions along MF axons. (For B/W version, see page 110 in the
volume.)
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 6

Mossy fiber synaptic transmission: communication

$
from the dentate gyrus to area CA3

David B. Jaffe1, and Rafael Gutiérrez2

1
Department of Biology, University of Texas at San Antonio, One UTSA Circle, San Antonio, TX 78249, USA
2
Department of Physiology, Biophysics and Neurosciences, Center for Research and Advanced Studies, Mexico City,
Apartado Postal 14-740, Mexico D.F. 07000, Mexico

Abstract: Communication between the dentate gyrus (DG) and area CA3 of the hippocampus proper is
transmitted via axons of granule cells — the mossy fiber (MF) pathway. In this review we discuss and
compare the properties of transmitter release from the MFs onto pyramidal neurons and interneurons. An
examination of the anatomical connectivity from DG to CA3 reveals a surprising interplay between ex-
citation and inhibition for this circuit. In this respect it is particularly relevant that the major targets of the
MFs are interneurons and that the consequence of MF input into CA3 may be inhibitory or excitatory,
conditionally dependent on the frequency of input and modulatory regulation. This is further complicated
by the properties of transmitter release from the MFs where a large number of co-localized transmitters,
including GABAergic inhibitory transmitter release, and the effects of presynaptic modulation finely tune
transmitter release. A picture emerges that extends beyond the hypothesis that the MFs are simply ‘‘det-
onators’’ of CA3 pyramidal neurons; the properties of synaptic information flow from the DG have more
subtle and complex influences on the CA3 network.

Keywords: mossy fibers; synaptic transmission; CA3; co-transmitters; plasticity

Neural output from the dentate gyrus (DG) is
transmitted via axons of granule cells called the
mossy fiber (MF) pathway. The MFs not only
project into area CA3 of the hippocampus, but also
synapse proximally onto DG basket cells (provid-
ing local recurrent inhibition in the dentate) and
pyramidal-like neurons in the hilus (Johnston and
Amaral, 2004). They are so named because of the
large (up to 5 mm diameter) varicosities along the
axon. Ramon y Cajal (who named the pathway
$
Both authors contributed equally to this work.
Corresponding author. Tel.: +1 210 458 5843;
Fax: +1 210 458 5658; E-mail: david.jaffe@utsa.edu
originally) and Lorente de No both suggested that
the large varicosities made synaptic contacts onto
CA3 pyramidal neurons (Ramon y Cajal, 1911;
Lorente de No, 1934), later to be confirmed with
electron microscopy (Blackstad and Kjaerheim,
1961), thereby making this pathway an integral
component within the so-called, and overly sim-
plistic, ‘‘tri-synaptic’’ circuit.
In this review we will discuss and compare the
properties of transmitter release at the most well-
characterized MF synapses, those in area CA3
onto pyramidal neurons and interneurons. Our
discussion will focus on the potential roles of
co-localized transmitters and modulators as well
as how the properties of synaptic transmission

DOI: 10.1016/S0079-6123(07)63006-4 109

110

from the DG influence the primary target of
DG output — the CA3 region of the hippo-
campus.
MF anatomy: does form follow function?
The MFs form three types of synaptic contacts
onto its target neurons. First, and most notably,
are the large expansions that synapse onto CA3
pyramidal neurons. The large boutons appear at
approximately 150 mm intervals (Blackstad et al.,
1970) and a single granule cell contacts approxi-
mately 15 pyramidal neurons (each terminal
synapses onto a single pyramidal neuron). One
CA3 pyramidal neuron may receive up to a total
of approximately 50 MF inputs only (Claiborne
et al., 1986; Amaral et al., 1990). Second, from the
large expansions there may extend 2–3 filopodia
that make synaptic contacts onto interneurons
(Amaral, 1979; Acsady et al., 1998). These
so-called filopodia are motile and regulated by
glutamatergic neuromodulation (Tashiro et al.,
2003). Third, small terminals, resembling boutons
of other hippocampal neurons, also contact CA3
interneurons. As a result, it appears that by sheer
numbers the major target of the MFs in CA3 is
onto inhibitory interneurons, rather than pyram-
idal neurons (Fig. 1).
The large terminals of the MFs are unique
structures in a variety of ways. In addition to the
filopodial-like synaptic contacts onto interneu-
rons, mentioned above, the large boutons are
filled with a high density of vesicles (Blackstad and
Kjaerheim, 1961), express more than one active
zone, and encase a complex branched dendritic
spine emanating from CA3 pyramidal neurons —
generally referred to as a thorny excrescence
(Amaral and Dent, 1981; Chicurel and Harris,
1992). Putative synaptic sites on these boutons
were originally identified as both asymmetric and
symmetric, suggesting differences in their respec-
tive properties of transmission (Hamlyn, 1961).
Another important anatomical aspect of the MF
pathway is that they are generally restricted to
a narrow band running within stratum lucidum
where the large terminals make contacts onto the
most proximal portions of pyramidal neuron
apical dendrites (Brown and Johnston, 1983). Ad-
ditionally, there is a less extensive infrapyramidal
MF projection onto the proximal basal dendrites
of pyramidal neurons in CA3b and CA3c.
The first recordings of MF synaptic transmis-
sion were extracellular field potential recordings

Fig. 1. Interneurons are the primary target of DG granule cells. Diagram representing the multiple types of excitatory contacts made
by granule cells. Within the dentate/hilar region collaterals of the mossy fibers (MFs) innervate both mossy cells and inhibitory basket
cells. In area CA3, the MFs contact pyramidal neurons via large expansions, but also excite interneurons through either filopodial
extensions from the large boutons or smaller en passant expansions along MF axons. (See Color Plate 6.1 in color plate section.)

made in vivo. Stimulation of the DG triggered a
current sink restricted to s. lucidum (Gloor et al.,
1963), confirming the anatomical evidence that
MF input was restricted to a discrete layer,
and supporting the view that the synapse was
excitatory.
Two conflicting ideas emerge out of the ana-
tomical connectivity from the DG to area CA3,
discussed above. On the one hand, the proximal
location of the MF large boutons, and the fact that
their large terminals are copiously supplied with
vesicles, suggests that a single action potential
from a granule cell might have a strong excitatory
influence on CA3 pyramidal neurons. That is, a
single spike invading a single terminal might be
capable of triggering an action potential in a CA3
pyramidal neuron. This is the so-called ‘‘detonator
synapse’’ hypothesis. The circuitry of area CA3
resembles that of an auto-associative network
(Marr, 1971) reflected in the large number of
recurrent excitatory connections between CA3
pyramidal neurons (Johnston and Amaral, 2004).
The concept of the MFs as detonators led to the
hypothesis that these sparse inputs might serve as a
‘‘teacher’’ signal underlying forms of associative
learning (McNaughton and Morris, 1987; Rolls,
1989). The presence of a large supply of vesicles
suggests a large readily releasable pool (Hallermann
et al., 2003), and that the degree of transmitter
release could be maintained over time by the large
reserve pool of vesicles. The proximal location
of these synapses relative to the spike generating
zone also would limit any loss of depolarization
due to cable filtering (Johnston and Brown, 1983).
That individual MF synapses would have large
unitary synaptic strength would also be a require-
ment if the pathway as a whole were to be effective
at driving the postsynaptic cell, because — as
discussed above — each CA3 pyramidal neuron
receives only about 50 mossy inputs (Amaral et al.,
1990).
On the other hand, the MFs also make direct
synaptic contact onto inhibitory interneurons in
area CA3 (Fig. 1). Moreover, the ratio of synaptic
contacts onto interneurons is approximately four-
fold higher than onto pyramidal neurons. Assum-
ing that (i) a single granule cell contacts 15
pyramidal neurons, (ii) each large expansion
111
contacts directly or via its filopodia up to three
interneurons, and (iii) 15 interneurons are inner-
vated by the smaller boutons one-granule cell will
contact approximately 60 interneurons compared
with only 15 pyramidal neurons (a 4:1 ratio).
Therefore, it is conceivable that under certain con-
ditions the firing of granule cells would have a net
inhibitory effect on pyramidal neurons in CA3
(Bragin et al., 1995a, b; Penttonen et al., 1997;
Acsady et al., 1998).
Excitatory–inhibitory conductance sequence
Yamamoto (1972) was the first to utilize brain slice
methods and intracellular recording to study MF
synaptic transmission onto CA3 pyramidal neu-
rons. Stimulation of the granule cell layer triggered
a biphasic response composed of a small com-
pound excitatory postsynaptic potential (EPSP)
followed by a larger, overlapping compound in-
hibitory postsynaptic potential (IPSP), presumably
mediated either by feed-forward/feed-back inhibi-
tion or the direct stimulation of inhibitory inter-
neurons (Yamamoto, 1972). The IPSP itself was
biphasic, comprised of an initial GABAA receptor
and later GABAB receptor-mediated phase (Ogata
and Ueno, 1976; Knowles et al., 1984).
One of Yamamoto’s most important early
observations was that paired stimulation of the
fibers at short intervals (20–100 ms) markedly
potentiated the second EPSP. While low-frequency
stimulation (LFS) rarely triggered spikes,
frequency facilitation of the compound EPSP
was generally capable of reaching spike threshold,
in spite of the overlapping IPSPs. Frequency
facilitation occurred over a very wide range of
frequencies of stimulation (0.05–1 Hz), and could
potentiate responses as much as threefold (Salin
et al., 1996).
Johnston and Brown (1983) applied voltage-
clamp methods to the study of MF synaptic
transmission, recognizing that the advantageous
proximal location of MFs relative to the soma
permitted reasonable voltage-control and, in turn,
space clamp issues were minimized (Johnston and
Brown, 1983; Spruston and Johnston, 1992; Sprus-
ton et al., 1993). Measuring synaptic currents
112
under voltage-clamp also allowed better resolution
of the mixed excitatory–inhibitory conductance
sequence, even though it was recognized that there
was still considerable overlap between the excita-
tion and inhibition (Brown and Johnston, 1983;
Barrionuevo et al., 1986). Two important obser-
vations were made in these early studies. First, for
compound postsynaptic currents, the GABAA-
mediated inhibitory conductance was almost
fivefold larger than the excitatory conductance.
Second, the onset of the inhibitory conductance
appeared to be very close to that of the excitatory
response. A disynaptic GABAergic response, from
a feed-forward or feed-back circuit, should retard
the onset of the inhibitory response relative to the
excitatory conductance. It is therefore possible
that the paradigms used to stimulate the MF
pathway might directly trigger the release of
GABA, most likely from inhibitory neurons, onto
CA3 pyramidal neurons.
Properties of transmitter release from MF synapses
Granule cells discharge action potentials down
the MFs at basal rates less than 0.5 Hz (Jung and
McNaughton, 1993), though firing rates may reach
up to 50 Hz during certain types of behaviors
(Skaggs et al., 1996; Wiebe and Staubli, 1999;
Henze et al., 2002b) and conduction velocity is
approximately 7 m/s, consistent with the MFs
being an unmyelinated pathway (Langdon et al.,
1993). Upon reaching synaptic terminals, presy-
naptic action potentials trigger synaptic transmis-
sion by eliciting Ca2+ influx through multiple
types of voltage-gated Ca2+ channels. Blockade
of N-, P-, and R-type Ca2+ channels decreases fast
transmitter release, while dihydropyridine-sensi-
tive L-type Ca2+ channels do not (Kamiya et al.,
1988; Castillo et al., 1994; Nicoll et al., 1994;
Yamamoto et al., 1994; Wu et al., 1998; Miyazaki
et al., 2005). That said, L-type channels are present
in MF presynaptic terminals and allow Ca2+ entry
when MFs are stimulated (Tokunaga et al., 2004).
Although L-channels do not appear to participate
in fast transmitter release, they may play a role in
transmission during repetitive stimulation (Reuter,
1995).
To directly study presynaptic action potentials
and their interaction with voltage-gated ion chan-
nels, patch-clamp methods have recently been
applied to the large boutons of the MFs (Bischof-
berger et al., 2006). One of the first notable
observations was that the action potential, while
of short duration during LFS, widens during
high-frequency stimulation (HFS) due to the rapid
inactivation of a voltage-gated K+ conductance
(Geiger and Jonas, 2000). Frequency-dependent
spike broadening, and the associated increase in
Ca2+ influx, provides for, in part, a mechanism
underlying frequency-facilitation (Salin et al.,
1996). The density of Na+ channels in the boutons
appears relatively large, and may be important
for the efficient propagation of action potentials
(Engel and Jonas, 2005). The load resulting from
the greater surface area of the large boutons could
potentially extinguish or slow down spike propa-
gation in these unmyelinated fibers. Computer
simulations suggested that the high density and
properties of these Na+ channels ensures the
reliability of synaptic transmission.
Ca2+ in the presynaptic terminal is also affected
by other mechanisms. Intracellular Ca2+ stores
regulate presynaptic Ca2+ concentration (Liang
et al., 2002). Ca2+-induced Ca2+ release may
work in concert with spike-mediated Ca2+ influx
to raise intraterminal Ca2+ concentration. Inter-
estingly, although Ca2+-induced Ca2+ release ac-
counts for a large proportion of the presynaptic
signal (Scott and Rusakov, 2006), depletion of
Ca2+ stores has no effect on fast trans-
mitter release. Release of Ca2+ may, however,
play a role in frequency facilitation (Lauri et al.,
2003).
In a recent study, it was reported that EPSPs in
dentate granule cells, surprisingly, propagated
down hundreds of microns along the MFs and
were capable of modulating synaptic transmission
(Alle and Geiger, 2006). Furthermore, spike-
evoked EPSCs in CA3 pyramidal neurons were
enhanced when paired with a presynaptic wave-
form that mimicked an EPSP in the bouton. The
mechanism underling this type of modulation may
be related to a similar phenomenon observed
at the Calyx of Held (Awatramani et al., 2005),
but other studies suggest it may also be due to

voltage and not solely due to resting Ca2+ levels
(Ruiz et al., 2003).
Quantal nature of transmission at the MF-CA3
pyramidal neuron synapse
The strong frequency-dependent facilitation of
MF synaptic transmission onto CA3 pyramidal
neurons implies that the initial probability of
transmitter release is small. If the probability of
release were high, either a ceiling effect would limit
an increase in release (assuming that there is a
large reserve pool of primed vesicles) or, alterna-
tively, one could observe synaptic depression due
to the refractory period produced by a lack of
primed vesicles. Indeed, quantal analysis of uni-
tary synaptic responses onto CA3 pyramidal neu-
rons is consistent with the hypothesis that the
initial release probability is in fact low (von
Kitzing et al., 1994; Lawrence et al., 2004). This
is in spite of the fact that the number of release
sites at a single MF synapse is very high (Chicurel
and Harris, 1992), and that release at these
synapses is most likely multiquantal (Henze
et al., 1997). Indeed, in the study by Jonas et al.
(1993) quantal content (the number of quanta
released per action potential) ranged from 2–16,
with a mean of 8, values similar to that first
reported by Yamamoto (Yamamoto et al., 1991).
Others have found that the probability of release
was much lower, o0.3 (Lawrence et al., 2004).
There is also evidence, however, that large synap-
tic currents arise from the release of single quanta
(Henze et al., 2002a). This is in contrast to com-
missural/associational (C/A) synapses, where the
probability of release is higher and, concomitantly,
frequency facilitation is weaker (Salin et al., 1996).
Mean quantal size at MF-CA3 pyramidal
neuron synapses is approximately 150 pS, corre-
sponding to about 17 AMPA receptors. These
numbers are consistent with unitary responses
(unitary conductance ¼ quantal size
quantal
content), which generally have magnitudes of
approximately 1 nS. Corresponding unitary EPSP
amplitudes can be up to 10 mV (Jonas et al., 1993),
in contrast to C/A unitary responses that are
typically o1 mV (Debanne et al., 1998; Pavlidis
113
and Madison, 1999). Similarly large unitary EPSPs
have been demonstrated for MF-mossy cell
synapses (Scharfman et al., 1990).
Evidence for multiple neurotransmitters/modulators
in the granule cells
Glutamate is believed to be the primary excitatory
neurotransmitter released from the MFs (Crawford
and Connor, 1973; Terrian et al., 1988), and MF
EPSPs are blocked by glutamate receptor antago-
nists (Sawada et al., 1983). The neuronal gluta-
mate transporter (EAAC1) is the most abundant
uptake mechanism and is selectively enriched in
hippocampal principal neurons, including DG
granule cells (Rothstein et al., 1994). The presy-
naptic and postsynaptic actions of glutamate re-
leased from the MFs are discussed below.
Granule cells also contain and release several
neuromodulators of different chemical composi-
tion (Fig. 2). The most prominent are those of
peptidic nature, which are contained and released
from large dense-core vesicles. Dynorphin and
enkephalin, as well as their receptors, are present
in the MFs in rodents and humans, but there is
variation between mammalian species (Gall et al.,
1981; McGinty et al., 1983; Chavkin et al., 1985;
McLean et al., 1987; Terrian et al., 1988; Houser
et al., 1990; Chavkin, 2000). The actions of opioid
peptides in the hippocampus are inhibitory, medi-
ated by inhibition of Ca++ currents, in the case
of dynorphin, or activation of K+ currents, in
the case of enkephalins, and are a consequence
of activation of G-protein-coupled receptors
(Zieglgänsberger et al., 1979). In the rat MF sys-
tem, activation of m receptors facilitates MF LTP
in an indirect fashion. Activation of m receptors
depresses GABA release from interneurons, which
in turn leads to a failure of GABA to inhibit
glutamate release from MFs (Jin and Chavkin,
1999). On the other hand, the m opioid receptor
agonist DAMGO inhibits low-frequency stimu-
lated MF responses in Sprague-Dawley rats, as it
does in the guinea pig (Salin et al., 1995). Direct
actions of dynorphin on MF transmission have
been described in the guinea pig, but not in the rat.
Indeed the MFs possess k receptors, which, upon
114

Fig. 2. Summary of the pre- and post-synaptic constituents of the different MF terminals and of their plasticity. (A) Schematic
representation of the giant MF boutons, which contact CA3 pyramidal cells. These terminals are characterized by low probability of
basal release and multiple release sites. (B) Schematic representation of filopodial extensions and en passant contacts, which synapse on
to interneurons. These have a high probability of basal release and single release sites. Both types of MF terminals contain several
neuromodulators and the neurotransmitters glutamate and GABA. Note that both the presynaptic and postsynaptic sites contain
several ionotropic GluRs and metabotropic GluRs, which confer to the MF a high degree of plasticity and the capacity for synaptic
integration. (C) Relative expression of the different releasable contents and of the receptors and transporters during development, in
the adult (D), and after epileptic activity (E). Arrows and font size indicate relative differences between the three states. (See Color
Plate 6.2 in color plate section.)

activation, depress neurotransmitter release. Thus,
for example, in the guinea pig high frequency
stimulation of the MF causes a transient heterosy-
naptic inhibition of neighboring MF or perforant
path synapses in the dentate, which is mediated by
the synaptic release of dynorphin that activates
presynaptic k receptors (Weisskopf et al., 1993).
However, in the rat, neither exogenous nor en-
dogenous dynorphin affect MF neurotransmis-
sion, which is consistent with the finding that
k-receptor binding in this projection is dense in the
guinea pig and sparse in the Sprague-Dawley rat
(see Salin et al., 1995). Dynorphin significantly
inhibits MF responses in the hamster, mouse, and

in Long-Evans rats (Salin et al., 1995). Interest-
ingly, hippocampal opioid peptides are upregulat-
ed after seizure activity (McGinty et al., 1983;
Gall, 1988; Gall et al., 1990), whereas a decrease in
k receptors has been observed in epileptic human
cells (Jeub et al., 1999). Thus, opioid peptides are
localized in the MF, presumably for effective con-
trol of neurotransmitter release (Weisskopf et al.,
1993; Castillo et al., 1996; Drake et al., 1996;
Simmons and Chavkin, 1996).
Other peptides present in the MF are neuropep-
tide Y (NPY), neurokinin B and cholecystokinin
(Gall, 1984; Gall et al., 1990; Holm et al., 1993;
Tonder et al., 1994; Chandy et al., 1995; Schwarzer
and Sperk, 1995; Makiura et al., 1999). Like opioid
peptides, these peptides have been shown to exert
inhibitory actions. Neuropeptide Y is synthesized,
stored in, and released from the MFs, and inhibits
MF transmission by a presynaptic mechanism
(McQuiston and Colmers, 1996; McCarthy et al.,
1998) and their receptors are normally present in the
MFs themselves (Widdowson, 1993; Jacques
et al., 1997). Seizures increase the release of NPY
which in turn, tonically inhibits MF synaptic trans-
mission (Tu et al., 2005). Besides, seizures upregu-
late NPY Y2 receptors, which appear to mediate the
effects on MF (Vezzani and Sperk, 2004),
although Y5 receptors could play a role (Marsh
et al., 1999; Ho et al., 2000). This upregulation re-
flects a neuroprotective function of NPY that has
been shown in animal models of epilepsy and in the
epileptic human (Schwarzer et al., 1998; Patrylo
et al., 1999; Furtinger et al., 2001; Vezzani and
Sperk, 2004). By contrast, substance P, which is
contained in and released from the MFs activates
its receptors in the same MF, increases glutamate
release (Borhegyi and Leranth, 1997; Liu et al.,
1999).
The MFs also contain high levels of Zn++,
which can be released together with glutamate
in an activity-dependent manner (Stengaard-
Pedersen et al., 1981; Wenzel et al., 1997; Molnar
and Nadler, 2001; Qian and Noebels, 2005).
Exogenously applied Zn++ blocks glutamate
and GABA receptors (Westbrook and Mayer,
1987; Draguhn et al., 1990; Smart et al., 1994).
Indeed, Zn++ occupies a high-affinity binding site
on NMDA receptors (Vogt et al., 2000) and can
115
block GABAA receptors composed by a and b but
not g subunits (Draguhn et al., 1990). It has been
shown that synaptically released Zn++ from the
MF pathway strongly modulates NMDA (Vogt
et al., 2000) and GABAA receptors in CA3 (Ruiz
et al., 2004). Because MF release, besides gluta-
mate, GABA and Zn++ (Gutierrez, 2005), this
synapse is most suitable to study the effects of the
synaptically released Zn++ on GABA responses
elicited by the same terminals. Indeed, it was
found that Zn++ tonically depresses the inhibitory
actions of GABA, as it reaches a high concentra-
tion in the vicinity of the GABAA receptor on
CA3 pyramidal cells. Consequently, chelation of
Zn++ relieves this inhibition (Ruiz et al., 2004).
This is particularly important for developmental
and pathological processes, in particular, epilepsy.
Indeed, a higher input of Zn++ from sprouted
MFs after epileptic activity contributes to inhibit
reverberating excitatory activity in the DG
(Nadler, 2003; Tu et al., 2005).
Other interesting signaling molecules present in
the MF are brain derived neurotrophic factor
(BDNF) and nerve growth factor (NGF) (Gall and
Isackson, 1989; Gall and Lauterborn, 1992;
Lowenstein and Arsenault, 1996; Conner et al.,
1997; Smith et al., 1997). Granule cells contain
relatively high concentrations of BDNF and NGF,
which can be released in vivo by the MFs and
dendrites to affect other MFs or other cells, re-
spectively (Blöchl and Thoenen, 1995; Lowenstein
and Arsenault, 1996; Conner et al., 1997). Den-
dritic targeting of BDNF mRNA and its transla-
tion to protein can account for the release of this
neurotrophin to neighboring granule cell dendrites
(Tongiorgi et al., 2004). This suggests that the lig-
and–receptor interaction occurs by means of an
autocrine/paracrine mechanism. On the other
hand, anterograde transport of BDNF (Conner
et al., 1997) would effectively provide a mechanism
to release BDNF from MF terminals, where it
can increase neurotransmitter release (Altar and
DiStefano, 1998; Elmer et al., 1998). As with most
signaling molecules in the granule cells, BDNF
expression is increased after limbic seizures (Isack-
son et al., 1991). In the MF, a consequence of this
would be the activation of TrkB receptors located
in the same fibers (He et al., 2002; Scharfman,
116
2005). TrkB activation is thought to promote
Ca++ influx into terminals (Berninger et al., 1993)
and this is the mechanism that leads to enhanced
neurotransmitter release. Eventually, this produces
an enhancement of excitability that can lead to the
generation of epileptiform activity (Thoenen,
1995; Huang and Reichardt, 2001; Scharfman
et al., 2002).
Importantly, it has been shown that BDNF can
exert a rapid depolarization in hippocampal neu-
rons by promoting Na+ influx through TTX-
insensitive channels (Kafitz et al., 1999; Rose et al.,
2003). Another important issue is that BDNF may
be packaged preferentially in the large MF
boutons, although it does reside in the small ter-
minals also, so it can potentially modulate both
excitation and feed-forward inhibition in CA3
(Danzer and McNamara, 2004).
Besides a possible involvement of hippocampal
BDNF in excitability and synaptic plasticity
(Thoenen, 1995, 2000), neurotrophins have a
differentiating effect on interneurons in the hip-
pocampus (Marty et al., 1996a, b) and cortex
(Marty et al., 1997; Patz et al., 2003), by regulating
the expression of GABAergic markers. Neuronal
activity is the main activator of GAD expression
by neurotrophins, differentially modulating tran-
scription and translation in a context-dependent
manner (Patz et al., 2003). Interestingly, this is the
mechanism shown to underlie the maturation of
the GABAergic phenotype in hippocampal inter-
neurons (Marty et al., 1996a, b) and, as recently
shown, the expression of all GABAergic markers
in granule cells (Gomez-Lira et al., 2005).
Granule cells also contain high levels of adeno-
sine (Braas et al., 1986) and its A1 receptor
(Rivkees et al., 1995; Swanson et al., 1995). It
has been suggested that endogenous tonic activa-
tion of A1 receptors underlies the low basal prob-
ability of neurotransmitter release from the MFs
(Moore et al., 2003; also see Kukley et al., 2005).
Their activation produces a Gi/o protein-depend-
ent sustained inhibition of neurotransmitter
release (Moore et al., 2003). Other set of recep-
tors present in the MF terminals, and which
activation controls neurotransmitter release, are
the ATP receptors P2X (Armstrong et al., 2002)
and a adrenergic receptors (Scanziani et al., 1993).
Finally, the inhibitory amino acid GABA is
probably the most important and controversial
addition to the list of modulators and transmitters
that are co-localized in the granule cells. Sandler
and Smith (1991) provided the first data that sug-
gested the presence GABA in the ‘‘glutamatergic’’
MF in monkey and human hippocampi (Sandler
and Smith, 1991). They found GABA immunore-
activity in MF terminals that made asymmetric
synaptic contacts with spines arising from large
dendrites of CA3 pyramidal cells. They also
showed the colocalization of GABA and gluta-
mate within the same terminals with electron
microscopy. From this anatomical evidence, and
the idea that GAD was not present in the MF,
they concluded that GABA had to be either
incorporated from the extracellular space or orig-
inated from an alternative route of synthesis such
as g-hydroxybutyrate. Second, they suggested that
at least a component of the inhibitory synaptic
potentials evoked in pyramidal cells by DG stim-
ulation had to be of MF origin. In this way,
GABA released by the MF could modulate
the normal MF glutamatergic responses. On the
other hand, studies carried out on MF synapto-
somes provided the first neurochemical evidence
showing that MF terminals contained and re-
leased GABA (Terrian et al., 1988; Taupin et al.,
1994a, b). These authors, in agreement with
Sandler and Smith (1991), suggested that the
amino acid could be synthesized in the granule
cells from a route different from the GAD-
dependent pathway.
Evidence clarifying the presence and origin of
GABA in the MF was provided by Sloviter et al.
(1996). These authors conclusively demonstrated
that immunoreactivity for the amino acid and its
synthesizing enzyme, GAD, was normally present
in the MFs of rats, monkeys, and humans (Sloviter
et al., 1996). Therefore, if the granule cells had the
necessary enzyme for the synthesis of GABA and
GABA itself, the granule cells indeed synthesized
GABA that could probably be used as a neuro-
transmitter. Complementing and extending the
initial findings of Sandler and Smith (1991) and
Sloviter et al. (1996), Bergersen et al. (2003)
recently confirmed with immunogold the coexist-
ence of glutamate and GABA in MF synapses,

which also contained the respective receptors in
apposition to the presynaptic terminal (Bergersen
et al., 2003). This study demonstrated that both
amino acids coexist in all MF terminals examined
and that they have a close spatial relation to
synaptic vesicles. Indeed, both GABA and gluta-
mate were shown to be located at a distance that
suggests their presence inside vesicles and in the
release zones. The estimated concentration of
GABA, however, was much lower than that
of glutamate within the MF terminals and even
lower than that of inhibitory types of terminals
(Bergersen et al., 2003).
Interestingly, it was demonstrated that provok-
ing seizures by stimulation of the perforant path-
way for 24 upregulated the content of both
isoforms of GAD and GABA in the MFs, whereas
no changes were observed in area CA1 (Sloviter
et al., 1996). Other authors (Schwarzer and Sperk,
1995; Lehmann et al., 1996) showed that GAD67
and its mRNA (but not GAD65) were transiently
upregulated after seizures provoked by kainic acid
(KA) or by the kindling method in the rat. On the
other hand, the upregulation of GAD67 and its
mRNA in granule cells was further confirmed with
other seizure- and epilepsy-induction methods that
use chemical convulsants or electrical stimulation
(Ding et al., 1998; Makiura et al., 1999; Szabo
et al., 2000; Ramirez and Gutierrez, 2001). It was
also established that GAD67 could be upregulated
in an activity-dependent manner, in the absence of
epileptiform activity (Ramirez and Gutierrez,
2001; Gutierrez, 2002). Recently, a direct link be-
tween the presence of seizures and the concomitant
upregulation of GAD and endogenous GABA was
found in MF terminals of epileptic rats (Gomez-
Lira et al., 2002). An analysis of the expression of
GAD67 at different ages has revealed that GAD67
(but not GAD65 or GAD65 RNA) is also devel-
opmentally regulated in the MFs (Maqueda et al.,
2003). Indeed, it was shown that GAD67 is
expressed in the MFs early in life and then down-
regulated by days 23–24, after completion of
development (Gutierrez, 2003; Maqueda et al.,
2003). Likewise, GABA-immunoreactive cells with
characteristics of granule cells are found in the
stratum granulare of the DG of developing rats
but not of adults.
117
Contrary to the data showing the upregulation
of both isoforms of GAD by seizures (Sloviter
et al., 1996), several reports have shown that
GAD67 (and its mRNA) and not GAD65 is reg-
ulated in granule cells by increased activity
(Schwarzer and Sperk, 1995; Szabo et al., 2000;
Ramirez and Gutierrez, 2001; Maqueda et al.,
2003), Ca++ entry and BDNF activation (Gomez-
Lira et al., 2005). It has been proposed that the
cellular distribution of both enzymes differed and
this could be reflected in distinctive functions
within neurons (Erlander and Tobin, 1991), i.e.,
GAD65 may be in synaptic terminals, while
GAD67 is present in terminals, somata and dend-
rites (Kaufman et al., 1991). This suggests that one
isoform synthesizes a metabolic pool of GABA
(in the soma) and the other the releasable pool
(in the terminals). Despite the possible differences
in GAD65- and GAD67-originated GABA, there is
a strong correspondence of the expression of
GAD67 to that of GABA present within the gran-
ule cells and their MFs (Gomez-Lira et al., 2002).
All these studies suggest that GABA synthesis in
the MFs is a means to counteract the enhanced
excitability caused by epileptic activity.
Presynaptic modulation
As mentioned above, a number of neuromodula-
tors control transmitter release from the MFs.
Glutamate and GABA are also important presy-
naptic modulators of MF synaptic transmission
transmission. In particular, one of the unique
properties of MF synaptic transmission, in con-
trast to transmitter release from recurrent synapses
onto CA3 pyramidal neurons, is its sensitivity to
metabotropic glutamate receptor (mGluR) ago-
nists; mGluR agonists depress MF synaptic trans-
mission (Manzoni et al., 1995; Yoshino et al.,
1996). Because of this response, sensitivity to
mGluR agonists is a widely used assay to deter-
mine if a synaptic response is of MF origin. There
are species-specific differences with respect to pre-
synaptic mGluRs. MF glutamatergic neurotrans-
mission in the rat is sensitive to the group II
mGluR agonist, DCG-IV, but not to the group III
mGluR agonist, L-AP4 (Lanthorn et al., 1984;
118
Kamiya et al., 1996; Maccaferri et al., 1998). The
opposite is true for the guinea pig where L-AP4
strongly depresses MF transmission (Lanthorn
et al., 1984; Manzoni et al., 1995; Tong et al., 1996;
Min et al., 1998). Interestingly, the granule cells
and MF of the rat DG express groups II/III
mGluR mRNA (Ohishi et al., 1993; Ohishi et al.,
1995) and mGluR2,4,7 (Bradley et al., 1996;
Shigemoto et al., 1997; Lie et al., 2000). Electron
microscopy has revealed immunolabeling for the
group III mGluR predominantly in presynaptic
active zones of asymmetrical and symmetrical
synapses, whereas mGluR-II immunolabeling
was found in preterminal rather than terminal
portions of axons (Shigemoto et al., 1996, 1997).
Although it has been well established that acti-
vation of group II mGluR almost completely de-
presses MF glutamatergic transmission in the rat,
several reports show that MF GABAergic trans-
mission is strongly inhibited through activation of
group III mGluR (with L-AP4) both in the guinea
pig and the rat (Gutierrez, 2000; Walker et al.,
2001; Gutierrez, 2002, 2003; Romo-Parra et al.,
2003; Kasyanov et al., 2004; Safiulina et al., 2006).
This pharmacological profile is not only consistent
with neurotransmission of MF origin but it also
demonstrates that the modulation of MF
GABAergic transmission differs from MF glut-
amatergic transmission. This physiological evi-
dence gives a functional significance to the
anatomical findings showing that both types of
receptors are present in rat granule cells with a
distinct localization and function (Ohishi et al.,
1995; Shigemoto et al., 1996, 1997). These data,
together with the finding that activation of mGluR
produces a downregulation of the exocytotic
machinery (Kamiya and Ozawa, 1999), suggest
that group III mGluR are associated with mech-
anisms of GABA release, and also that there
is presynaptic segregation of mGluR receptors
according to the class of neurotransmitter to be
released.
Kainic acid has long been used as an epilepto-
genic agent and was recognized that it binds to the
MF with high affinity (Monaghan and Cotman,
1982; Represa et al., 1987). It is now clear that the
several KA receptors that are present in the MF
(Wisden and Seeburg, 1993; Darstein et al., 2003)
modulate plasticity at this synapse in a complex
manner (Schmitz et al., 2000).
Presynaptic GABAB receptors effectively inhibit
glutamate release from the MFs (Thompson and
Gahwiler, 1992), and it has recently shown that
ionotropic GABAA receptors also inhibit gluta-
mate (Ruiz et al., 2003) and GABA release
(Treviño and Gutierrez, 2005) and control the
excitability of this pathway (Kullmann et al., 2005;
Treviño and Gutierrez, 2005). Other subcortical
modulators also control the release of transmitter
from the MFs. Muscarinic acetylcholine receptors
inhibit transmitter release (Williams and Johnston,
1988, 1990) while nicotinic receptor activation
raises presynaptic Ca2+ levels and thereby may
enhance transmitter release (Gray et al., 1996; also
see Vogt and Regehr, 2001).
Co-localization of plasma membrane transporters
of glutamate and GABA
Glutamate transport is the major mechanism con-
trolling extracellular glutamate levels, preventing
excitotoxicity, and averting neural damage associ-
ated with hyperexcitability. As mentioned above,
the neuronal glutamate transporter (EAAC1) is
expressed in granule cells and, surprisingly, in a
number of GABAergic neurons (Rothstein et al.,
1994; He et al., 2002; Sepkuty et al., 2002). There-
fore, it has been suggested that besides controlling
extracellular glutamate levels, its function is linked
to GABA metabolism, because capture of gluta-
mate is essential for GABA synthesis (Sepkuty
et al., 2002). It is not surprising that seizures and
epilepsy have direct consequences on EAAC1 ex-
pression and function, as has been demonstrated
(Gorter et al., 2002; Zhang et al., 2004). On the
other hand, it has been proposed that the presence
of the membrane transporter of GABA, GAT-1, is
restricted to neurons that synthesize and release
GABA, and glial cells (Iversen and Kelly, 1975;
Radian et al., 1990; Ribak et al., 1996). Some
GABAergic cells, immunocytochemically charac-
terized by the presence of GAD67 or GABA, do
not contain or contain traces of GAT-1, but not
vice versa (Rattray and Priestley, 1993). However,
it has been found that GAT-1 is also localized in

the glutamatergic granule cells (Frahm et al., 2000;
Sperk et al., 2003), and it controls GABA uptake
to the MF terminals (Gomez-Lira et al., 2002) but
not GABA release at this synapse (Vivar and
Gutierrez, 2005; Safiulina et al., 2006).
Co-localization of the vesicular transporters for
glutamate (VGlut-1) and GABA (VGAT)
Because glutamate is a general metabolic substrate
and serves as the precursor of inhibitory transmit-
ter GABA, glutamate immunoreactivity is not
specific to glutamatergic neurons. Therefore, the
detection of glutamate vesicular transporter(s) has
been used to establish the glutamatergic phenotype
of neurons. As expected for glutamatergic neu-
rons, the MF terminals of the granule cells contain
the glutamate vesicular transporter VGlut-1
(Bellocchio et al., 1998; Kaneko et al., 2002). In
accordance with immunohistological data showing
that the granule cells express GAD and GABA, it
was found that they express VGAT mRNA in an
activity-dependent manner (Lamas et al., 2001;
Gutierrez, 2003; Gomez-Lira et al., 2005). How-
ever, these cells do not contain the transporter
protein (Chaudhry et al., 1998; Sperk et al., 2003;
also see Safiulina et al., 2006). Thus, the lack of
detection of VGAT indicates that its expression is
too low to be revealed with immunohistochemis-
try, or the existence of a yet unidentified trans-
porter. Recent experiments show that glutamate
and GABA are in close relation to vesicles in MFs
terminals (Bergersen et al., 2003) strongly suggest-
ing that VGAT or another related transporter
protein must be present in these terminals
(Chaudhry et al., 1998; Gomez-Lira et al., 2005).
Postsynaptic responses of CA3 pyramidal neurons
to MF glutamatergic input
As mentioned above, glutamate is believed to be the
primary excitatory neurotransmitter released from
the MFs (Crawford and Connor, 1973; Terrian
et al., 1988). The primary ionotropic glutamate re-
ceptors mediating the fast synaptic response at the
MF synapse are a-amino-3-hydroxy-5-methyl-4-
isoxazolepropionic acid (AMPA)-type receptors
119
(Lanthorn et al., 1984; Neuman et al., 1988; Ito
and Sugiyama, 1991; Jonas et al., 1993). Voltage-
clamp analysis of unitary and compound excitatory
postsynaptic currents (EPSC) onto CA3 pyramidal
neurons finds the equilibrium potential close to
0 mV, consistent with ionotropic glutamate recep-
tors (Brown and Johnston, 1983). Based on the fast
rise times and decay kinetics of the current, the re-
sponses are consistent with the properties of AMPA
receptors and therefore reflects an electrotonically
close synapse (Williams and Johnston, 1991; Jonas
et al., 1993).
Early studies using radio-labeled NMDA found
that the MF terminal field contains a low density
of NMDA receptors relative to other regions of
the hippocampal formation (Monaghan and
Cotman, 1985). Although lower than at other
synaptic sites, glutamate release from the MFs
activates a small, but measurable, NMDA com-
ponent that, as expected, exhibits voltage-depend-
ence and slower kinetics, which are characteristics
of NMDA receptor-mediated responses (Jonas
et al., 1993; Weisskopf and Nicoll, 1995).
In contrast to NMDA receptors, the MF ter-
minal field contains a sizable density of high-
affinity sites for KA binding (Foster et al., 1981;
Monaghan and Cotman, 1982) and KA channels
are expressed in CA3 pyramidal neurons (Egebjerg
et al., 1991; Werner et al., 1991). Focal application
of kainate into the MF terminal field induces a
strong depolarization in CA3 pyramidal neurons
(Sawada et al., 1988). Postsynaptic activation of
kainate receptors requires repetitive stimulation
due to their slow kinetics (Castillo et al., 1997;
Bortolotto et al., 2003). This, combined with the
strong frequency-dependent facilitation of release,
suggests that KA receptors contribute to the
activity-dependent enhancement of transmission
from granule cells to CA3 pyramidal neurons
discussed above.
Because of the arguments above that MFs re-
lease GABA and it acts as a classical neurotrans-
mitter, one would expect that GABA receptors
would be located in apposition to MF terminals.
In support of this prediction, there is evidence that
GABAA receptors cluster with glutamatergic
receptors in pyramidal cells in apposition to
both glutamatergic and GABAergic terminals
120
(Rao et al., 2000). It was later found, with im-
munogold detection, that AMPA and GABAA
receptors also colocalize at MF synapses in the
hilar region (Bergersen et al., 2003).
Finally, glutamate released from MFs triggers
the release of Ca++ from intracellular stores of
CA3 pyramidal cells via mGluR receptor activa-
tion (Miller et al., 1996; Jaffe and Brown, 1997).
Tetanic stimulation of the MFs triggers waves of
spike-independent increases in Ca++ concentra-
tion in the proximal apical dendrites of CA3
pyramidal neurons mediated by the activation of
type I mGluR receptors (Kapur et al., 2001).
Single presynaptic spikes appear to be insuffi-
cient to elicit Ca++ release (Reid et al., 2001).
Frequency-dependent mGluR release of Ca++ from
intracellular stores may be important for long-
lasting changes in MF synaptic efficacy, discussed
below (Yeckel et al., 1999; Wang et al., 2004).
Are MF synapses onto CA3 pyramidal neurons
detonators?
It is not surprising that LFS of the MFs fails to
routinely trigger spikes, given many of the points
made above. For example, the major target of the
MFs in area CA3 involves the activation of feed-
forward inhibitory circuits. Glutamate is co-
released with GABA, and a wide-array of other
neuromodulators, which potentially inhibit MF
transmission. CA3 pyramidal cells also have a
threshold that is 10–15 mV from resting potential
(Podlogar and Dietrich, 2006). There is a large
amplitude unitary EPSP (Jonas et al., 1993), so the
MFs may not relay synapses with a high safety
factor. An elegant demonstration of the ‘‘condi-
tional’’ nature of MF excitation of CA3 pyramidal
neurons was shown by Henze et al. (2002b). With
low frequency stimulation (o40 Hz), the likeli-
hood that a CA3 pyramidal neuron would dis-
charge following a granule cell was low. Only at
higher frequencies did CA3 pyramidal neurons
follow granule cell firing. Interestingly, this may
not be the case for all MF targets. Scharfman and
colleagues showed faithful following of granule
cell firing when the targets of MFs were hilar
mossy cells or hilar interneurons (Scharfman et al.,
1990), a difference that may contribute to hilar cell
vulnerability to excitotoxicity.
The MF-to-CA3 synapse can therefore be con-
sidered as ‘‘conditional detonators’’ in two do-
mains. The first, as described above, is the
temporal domain where presynaptic facilitation,
the slow membrane time constant of CA3 pyram-
idal neurons (Spruston and Johnston, 1992), and
possibly postsynaptic amplification by KA recep-
tors combine to boost EPSP amplitude and fire the
cell. The second domain reflects the spatial sum-
mation of inputs concomitantly with the MFs.
For example, if a MF EPSP occurred coincidently
with a perforant path EPSP, the summation of
these two would then be suprathreshold (Urban
and Barrionuevo, 1998). Although all synapses
could be considered as conditional detonators
given this definition, the mean strength of unitary
MF EPSPs endows them with a greater ability to
move the membrane potential closer to threshold.
The granule cells simultaneously release glutamate
and GABA: electrophysiological evidence
Indirect but compelling evidence has accumulated
over the last years of the co-release of glutamate
and GABA from the MFs. The first electrophys-
iological evidence of GABAergic transmission
from the MFs to CA3 agreed with the
immunohistochemical observations showing that
seizures transiently upregulated the expression of
GAD65 and GAD67 (Schwarzer and Sperk, 1995;
Sloviter et al., 1996). Indeed, it was shown that
stimulation of the MFs produced monosynaptic
GABA-mediated transmission in pyramidal cells
of CA3 in kindled epileptic but not in control
healthy rats (Gutierrez, 2000; Gutierrez and
Heinemann, 2001).
As mentioned above, feed-forward inhibition
results from activation of local inhibitory inter-
neurons by MF glutamatergic, excitatory synapses
(Crawford and Connor, 1973; Acsady et al.),
which in turn inhibit pyramidal cells (Yamamoto,
1972; Brown and Johnston, 1983; Buzsaki, 1984).
Thus, activation of the DG leads to mono-
synaptic excitation and disynaptic inhibition on
CA3 pyramidal neurons that, with intracellular

recordings, are observed as an EPSP/IPSP
sequence. Both these conductances are typically
blocked in the presence of blockers of glut-
amatergic transmission. In kindled epileptic ani-
mals, however, a bicuculline-sensitive IPSP is still
elicited in the presence of the glutamate receptor
blockers. This IPSP had the same latency as the
control EPSP and could be inhibited by activation
of metabotropic glutamate receptors (mGluR).
Expression of the monosynaptic IPSP was tran-
sient because it could be observed 24–48 h after the
last kindled seizure but was not present if the
experiment was carried out a month after the last
seizure. The kindled epileptic state is not necessary
for this monosynaptic IPSP, because a single sei-
zure (Gutierrez, 2000) or repeated LTP-like stim-
ulation could elicit the response (Gutierrez, 2002).
Moreover, the monosynaptic IPSP produced by
the activation of the DG persisted when perfusing
a medium with a low [Ca++] or the GABAB ago-
nist, baclofen. These manipulations depressed the
DG-evoked IPSP amplitude without altering its
onset latency or slope. This unequivocally estab-
lished that the IPSP was a monosynaptic response
because, under these conditions, the ability to
recruit local interneurons quickly enough to
trigger such a short-latency IPSP is unlikely. The
possibility of recruiting inhibitory interneurons
by activating electrical synapses was also dis-
carded (Gutierrez, 2000). Independently, Walker
et al. (2001) demonstrated that monosynaptic
GABAergic responses could be normally evoked
in CA3 pyramidal cells by MF activation in slices
of young guinea pigs (Walker et al., 2001). They
showed that these MF GABAergic responses had
the same pharmacological and plastic properties as
those reported for glutamatergic MF transmission,
i.e., there was strong frequency-dependent
potentiation, and they were sensitive to an
antagonist of the metabotropic glutamate recep-
tor that is expressed preferentially by MFs, L-AP4.
They were able to show that minimal stimulation
of the MF evoked glutamate only, GABA only,
and compound glutamate-GABA currents in py-
ramidal cells. This indicated that both responses,
glutamatergic and GABAergic, had a common
origin. In addition, both neurotransmitters could
be released synchronously or asynchronously,
121
discarding the possibility that they are packaged
in single vesicles. Moreover, the evidence of MF
GABA transmission onto both pyramidal cells
(Gutierrez, 2000; Gutierrez and Heinemann, 2001;
Walker et al., 2001), and interneurons of CA3
(Romo-Parra et al., 2003; Safiulina et al., 2006)
and onto the mossy cells in the hilar region
(Bergersen et al., 2003) excludes the possible
segregation of neurotransmitters according to the
target cell.
It has been shown that the putative release of
GABA from the MF also occurs during develop-
ment, when the granule cells express all the mark-
ers of the GABAergic phenotype, and provoke
GABA receptor-mediated responses in pyramidal
cells and interneurons of CA3. This occurs during
the first weeks of age, after which the GABAergic
phenotype shuts off in a clear-cut manner. Sup-
porting this notion, recent work by Gutiérrez et al.
(2003), Kasyanov et al. (2004) and Safiulina et al.
(2006) have shown that GABA is the main neuro-
transmitter released from MF terminals during the
first postnatal days (Kasyanov et al., 2004;
Safiulina et al., 2006). Thus, MFs contain two
different sets of low- and high-threshold fibers that
release GABA and GABA plus glutamate, respec-
tively. The first would disappear with maturation,
whereas the second would persist longer or would
reappear in pathological conditions, such as in ep-
ilepsy (Gutierrez, 2003; Safiulina et al., 2006). It is
possible that signaling by both, synaptic and non-
synaptic release of GABA, may play a crucial role
in tuning hippocampal network during postnatal
development (Gutierrez, 2003, 2005; Safiulina
et al., 2006). While the molecular mechanisms
that shut off the GABAergic phenotype at matu-
rity have not been disclosed, upregulation of
the GABAergic phenotype in the adult depends
on protein synthesis, Ca++ entry, and acti-
vation of TrkB receptors by BDNF (Gutierrez,
2002; Romo-Parra et al., 2003; Gomez-Lira et al.,
2005).
Long-term plasticity at the MF CA3 synapse
Like other excitatory synapses in the hippo-
campal formation, the MF synapse expresses
122
LTP in response to a brief episode of HFS
(Yamamoto et al., 1980). As described above, the
MF terminal field contains a lower density of
NMDA receptors compared with other areas of the
hippocampus. This observation motivated Harris
and Cotman (1986) to test whether LTP at the MF
synapse was dependent on NMDA receptors. They
found that in the presence of NMDA receptor
antagonists, HFS of the MFs was still capable of
triggering LTP (Harris and Cotman, 1986).
Although there is agreement that Ca++ is nec-
essary for the induction process, the mechanism
underlying the induction of NMDA-independent
LTP has been a source of controversy (Nicoll and
Schmitz, 2005). Briefly, many experiments point to
an induction mechanism that is solely presynaptic.
Here the working hypothesis is that HFS triggers
an increase in presynaptic Ca++ that, via calmod-
ulin, activates adenylyl cyclase (Zalutsky and
Nicoll, 1992; Weisskopf et al., 1994; Villacres
et al., 1998). An alternative presynaptic mecha-
nism is that presynaptic KA receptor activation
leads to Ca++ influx that triggers release of Ca++
from intracellular stores (Contractor et al., 2001;
Lauri et al., 2001; Bortolotto et al., 2003; Lauri
et al., 2003), although this remains controversial
(Breustedt and Schmitz, 2004).
Alternatively, there is evidence for a postsynaptic
component to the induction mechanism (Williams
and Johnston, 1989; Jaffe and Johnston, 1990).
Here, an interaction between Ca++ entry through
voltage-gated Ca++ channels and the activation
of mGluRs takes place (Kapur et al., 1998; Yeckel
et al., 1999; Kapur et al., 2001; Wang et al., 2004). It
may be that depending on stimulus conditions, in-
duction may have either a presynaptic or post-
synaptic locus (Urban and Barrionuevo, 1996).
In contrast to the induction process, there is lit-
tle controversy regarding the locus for the expres-
sion of MF LTP. Numerous studies indicate that
synaptic potentiation is mediated by a presynaptic
change in transmitter release (Zalutsky and Nicoll,
1990; Yamamoto et al., 1992; Zalutsky and Nicoll,
1992; Xiang et al., 1994; Reid et al., 2004). If
induction of MF LTP has a postsynaptic locus,
then expression must require a retrograde com-
munication. One possible mechanism involves
postsynaptic ephrinB receptors and presynaptic
B-ephrins reverse signaling (Contractor et al.,
2002; Armstrong et al., 2006).
‘‘What goes up must go down’’ is an apt phrase
to describe MF synaptic plasticity. LFS triggers
long-term depression (LTD) of MF synaptic in-
puts onto CA3 pyramidal neurons (Kobayashi
et al., 1996; Yokoi et al., 1996) where presynaptic
activation of mGluR2 receptors depresses cAMP
levels, countering the expression of LTP.
CA3-interneuron synapses
As discussed above, the majority of MF synaptic
contacts are onto GABAergic interneurons
(Acsady et al., 1998) of which there are a wide
variety of subtypes, typically characterized based
on a combination of parameters including cell
body location, axonal projection, morphology, co-
localized peptides, and calcium binding protein
content (Parra et al., 1998). Of particular interest
are a class of bipolar interneurons whose dendrites
primarily reside along and within s. lucidum
(Spruston et al., 1997; Vida and Frotscher, 2000),
in contrast to other interneurons where the den-
dritic tree projects orthogonally against the MFs
that receive most of their innervation from other
pathways. The studies discussed below were made
primarily from this specific subtype of interneuron.
Like the pyramidal cell synapse, MF transmis-
sion onto interneurons is reduced by the activation
of presynaptic group II mGluR receptors
(Maccaferri et al., 1998; Toth et al., 2000); Pelkey
et al., 2005), although other types of mGluRs may
play a role in synaptic plasticity at this synapse
(discussed below). In contrast to the pyramidal
neuron synapse, at MF-interneuron synapses
AMPA receptor expression is heterogeneous.
Glutamate released from the MFs may activate
receptors with or without GluR2 subunits, thereby
forming receptors with either Ca++ permeable or
impermeable subunits (Toth and McBain, 1998;
Toth et al., 2000).
Also different from the pyramidal neuron MF
synapse is the frequency response of the interneu-
ron, which is highly variable. The response can
range from a weak frequency-dependent facilita-
tion (much less pronounced than at the pyramidal

neuron synapse) to synaptic depression (Toth
et al., 2000).
Quantal transmission at interneuron synapses is
also different than at the pyramidal neuron
synapse. As expected from anatomy, these
synapses express a low number of release sites
(Acsady et al., 1998). The probability of release at
these sites, however, appears higher than at the
pyramidal neuron synapse (Lawrence et al., 2004).
This difference in probability of release is also
consistent with differences frequency facilitation
observed between the two synapses (i.e., a low in-
itial probability of release predicts large facilita-
tion). Interestingly, quantal size at interneuron
synapses is within the same ranges, if not higher,
than for pyramidal neuron synapse. As a result,
and combined with a more depolarized resting
potenial, unitary responses of interneuron
synapses are as capable of triggering action po-
tentials as MF inputs onto pyramidal neurons
(Lawrence et al., 2004), and this also appears to be
the case for MF transmission to hilar interneurons
(Scharfman et al., 1990).
Long-lasting plasticity at MF-interneuron
synapses also differ from their pyramidal neuron
counterparts. Most notably, HFS does not elicit
LTP at the MF-interneuron synapse (Maccaferri
et al., 1998). Rather, HFS triggers LTD
at synapses expressing Ca++-permeable AMPA
receptors, while NMDA receptors trigger the
induction of LTD at synapses containing Ca++-
impermeant AMPA receptors (Lei and McBain,
2002). Activation of mGluR7 receptors by L-AP4
produces a chemical version of LTD at this
synapse that is distinct from pyramidal neuron
contacts (Pelkey et al., 2005) and differentially
modulates voltage-gated Ca++ channels in the
presynaptic filopodia (Pelkey et al., 2006). If HFS
triggers LTD at the MF-interneuron synapse,
feed-forward inhibition should be depressed onto
CA3 pyramidal neurons following HFS. However,
fast synaptic inhibition onto CA3 pyramidal neu-
rons is not affected by LTP of the MFs (Griffith
et al., 1986) suggesting that a decrease in feed-for-
ward inhibition through interneurons of s. lucidum
is compensated by enhanced recurrent inhibition
or other forms of synaptic plasticity within the
CA3 network.
123
Communication from DG to CA3: exciting
yet inhibiting
A picture is emerging that relates the contribution
of MF transmission to excitation and inhibition
in area CA3, both in terms of the repertoire of
neurotransmitters/modulators released from the
fibers but also with respect to circuitry that is
required to take into account inhibitory neurons.
At low frequencies (o0.5 Hz), frequency facilitation
of the MF-CA3 pyramidal neuron synapse is weak
and the probability of eliciting a spike is low (Henze
et al., 2002b). Furthermore, in this frequency do-
main granule cells are likely to be firing inhibitory
interneurons preferentially relative to pyramidal
neurons CA3 (Bragin et al., 1995a, b; Penttonen
et al., 1997). As such, concomitant extrinsic input
from entorhinal cortex (via the perforant path) or
via recurrent excitation via the C/A pathway should
fail to excite sets of CA3 pyramidal neurons. But as
granule cell firing rates increase, such as when an
animal traverses a place field, frequency facilitation
can then depolarize CA3 pyramidal neurons
to spike threshold. Interestingly, after seizures,
MFs tonically inhibit b/g field oscillations and
spontaneous subthreshold membrane oscillations
of CA3 interneurons in the CA3 area through
GABA-medated signaling. Coincident stimulation
of the MFs at y and b/g frequencies produces a
frequency-dependent excitation of interneurons and
the inhibition of pyramidal cells. Indeed, these
effects are maximal at the b frequency, suggesting
a resonance phenomenon (Treviño et al., 2007).
Because of the properties of synaptic plasticity
within the auto-associative circuit of CA3, pattern
separation, pattern completion, and the encoding
of new information will then be dependent on the
precise timing of granule cell and pyramidal neuron
firing (Kohonen, 1977; Chattarji et al., 1989;
Zalutsky and Nicoll, 1990; August and Levy,
1999; Nakazawa et al., 2002; Rolls and Kesner,
2006).
References
Acsady, L., Kamondi, A., Sik, A., Freund, T. and Buzsaki, G.
(1998) GABAergic cells are the major postsynaptic targets
of mossy fibers in the rat hippocampus. J. Neurosci., 18:
3386–3403.
124
Alle, H. and Geiger, J.R. (2006) Combined analog and action
potential coding in hippocampal mossy fibers. Science, 311:
1290–1293.
Altar, C.A. and DiStefano, P.S. (1998) Neurotrophin traffick-
ing by anterograde transport. Trends Neurosci., 21: 433–437.
Amaral, D.G. (1979) Synaptic extensions from the mossy fibers
of the fascia dentata. Anat. Embryol. (Berl.), 155: 241–251.
Amaral, D.G. and Dent, J.A. (1981) Development of the mossy
fibers of the DG: I. A light and electron microscopic study of
the mossy fibers and their expansions. J. Comp. Neurol., 195:
51–86.
Amaral, D.G., Ishizuka, N. and Claiborne, B. (1990) Neurons,
numbers and the hippocampal network. Prog. Brain Res., 83:
1–11.
Armstrong, J.N., Brust, T.B., Lewis, R.G. and MacVicar, B.A.
(2002) Activation of presynaptic P2
7-like receptors de-
presses mossy fiber-CA3 synaptic transmission through p38
mitogen-activated protein kinase. J. Neurosci., 22:
5938–5945.
Armstrong, J.N., Saganich, M.J., Xu, N.J., Henkemeyer, M.,
Heinemann, S.F. and Contractor, A. (2006) B-ephrin reverse
signaling is required for NMDA-independent long-term po-
tentiation of mossy fibers in the hippocampus. J. Neurosci.,
26: 3474–3481.
August, D.A. and Levy, W.B. (1999) Temporal sequence com-
pression by an integrate-and-fire model of hippocampal area
CA3. J. Comput. Neurosci., 6: 71–90.
Awatramani, G.B., Price, G.D. and Trussell, L.O. (2005) Mod-
ulation of transmitter release by presynaptic resting potential
and background calcium levels. Neuron, 48: 109–121.
Barrionuevo, G., Kelso, S.R., Johnston, D. and Brown, T.H.
(1986) Conductance mechanism responsible for long-term
potentiation in monosynaptic and isolated excitatory synap-
tic inputs to hippocampus. J. Neurophysiol., 55: 540–550.
Bellocchio, E.E., Hu, H., Pohorille, A., Chan, J., Pickel, V.M.
and Edwards, R.H. (1998) The localization of the brain-
specific inorganic phosphate transporter suggests a specific
presynaptic role in glutamatergic transmission. J. Neurosci.,
18: 8648–8659.
Bergersen, L., Ruiz, A., Bjaalie, J.G., Kullmann, D.M. and
Gundersen, V. (2003) GABA and GABAA receptors at
hippocampal mossy fibre synapses. Eur. J. Neurosci., 18:
931–941.
Berninger, B., Garcia, D.E., Inagaki, N., Hahnel, C. and Lind-
holm, D. (1993) BDNF and NT-3 induce intracellular Ca++
elevation in hippocampal neurones. Neuroreport, 4:
1303–1306.
Bischofberger, J., Engel, D., Frotscher, M. and Jonas, P. (2006)
Timing and efficacy of transmitter release at mossy fiber
synapses in the hippocampal network. Pflugers Arch., 453:
361–372.
Blackstad, T.W., Brink, K., Hem, J. and Jeune, B. (1970) Dis-
tribution of hippocampal mossy fibers in the rat. An exper-
imental study with silver impregnation methods. J. Comp.
Neurol., 138: 433–449.
Blackstad, T.W. and Kjaerheim, A. (1961) Special axo-
dendritic synapses in the hippocampal cortex: electron and
light microscopic studies on the layer of mossy fibers.
J. Comp. Neurol., 117: 133–159.
Blöchl, A. and Thoenen, H. (1995) Characterization of nerve
growth factor (NGF) release from hippocampal neurons:
evidence for a constitutive and an unconventional sodium-
dependent regulated pathway. Eur. J. Neurosci., 7:
1220–1228.
Borhegyi, Z. and Leranth, C. (1997) Substance P innervation of
the rat hippocampal formation. J. Comp. Neurol., 384:
41–58.
Bortolotto, Z.A., Lauri, S., Isaac, J.T. and Collingridge, G.L.
(2003) Kainate receptors and the induction of mossy fibre
long-term potentiation. Philos. Trans. R. Soc. Lond. B Biol.
Sci., 358: 657–666.
Braas, K.M., Newby, A.C., Wilson, V.S. and Snyder, S.H.
(1986) Adenosine-containing neurons in the brain localized
by immunocytochemistry. J. Neurosci., 6: 1952–1961.
Bradley, S.R., Levey, A.I., Hersch, S.M. and Conn, P.J. (1996)
Immunocytochemical localization of group III metabotropic
glutamate receptors in the hippocampus with subtype-specific
antibodies. J. Neurosci., 16: 2044–2056.
Bragin, A., Jando, G., Nadasdy, Z., Hetke, J., Wise, K. and
Buzsaki, G. (1995a) Gamma (40–100 Hz) oscillation in the
hippocampus of the behaving rat. J. Neurosci., 15: 47–60.
Bragin, A., Jando, G., Nadasdy, Z., van Landeghem, M.
and Buzsaki, G. (1995b) Dentate EEG spikes and
associated interneuronal population bursts in the hippo-
campal hilar region of the rat. J. Neurophysiol., 73:
1691–1705.
Breustedt, J. and Schmitz, D. (2004) Assessing the role of
GLUK5 and GLUK6 at hippocampal mossy fiber synapses.
J. Neurosci., 24: 10093–10098.
Brown, T.H. and Johnston, D. (1983) Voltage-clamp analysis
of mossy fiber synaptic input to hippocampal neurons.
J. Neurophysiol., 50: 487–507.
Buzsaki, G. (1984) Feed-forward inhibition in the hippocampal
formation. Prog. Neurobiol., 22: 131–153.
Castillo, P.E., Malenka, R.C. and Nicoll, R.A. (1997) Kainate
receptors mediate a slow postsynaptic current in hippocam-
pal CA3 neurons. Nature, 388: 182–186.
Castillo, P.E., Salin, P.A., Weisskopf, M.G. and Nicoll, R.A.
(1996) Characterizing the site and mode of action of dynor-
phin at hippocampal mossy fiber synapses in the guinea pig.
J. Neurosci., 16: 5942–5950.
Castillo, P.E., Weisskopf, M.G. and Nicoll, R.A. (1994) The
role of Ca2+ channels in hippocampal mossy fiber synaptic
transmission and long-term potentiation. Neuron, 12:
261–269.
Chandy, J., Pierce, J.P. and Milner, T.A. (1995) Rat hippo-
campal mossy fibers contain cholecystokinin-like immunore-
activity. Anat. Rec., 243: 519–523.
Chattarji, S., Stanton, P.K. and Sejnowski, T.J. (1989) Com-
missural synapses, but not mossy fiber synapses, in hippo-
campal field CA3 exhibit associative long-term potentiation
and depression. Brain Res., 495: 145–150.
Chaudhry, F.A., Reimer, R.J., Bellocchio, E.E., Danbolt, N.C.,
Osen, K.K., Edwards, R.H. and Storm-Mathisen, J. (1998)

The vesicular GABA transporter, VGAT, localizes to synap-
tic vesicles in sets of glycinergic as well as GABAergic neu-
rons. J. Neurosci., 18: 9733–9750.
Chavkin, C. (2000) Dynorphins are endogenous opioid peptides
released from granule cells to act neurohumorly and inhibit
excitatory neurotransmission in the hippocampus. Prog.
Brain Res., 125: 363–367.
Chavkin, C., Shoemaker, W.J., McGinty, J.F., Bayon, A. and
Bloom, F.E. (1985) Characterization of the prodynorphin
and proenkephalin neuropeptide systems in rat hippo-
campus. J. Neurosci., 5: 808–816.
Chicurel, M.E. and Harris, K.M. (1992) Three-dimensional
analysis of the structure and composition of CA3 branched
dendritic spines and their synaptic relationships with mossy
fiber boutons in the rat hippocampus. J. Comp. Neurol., 325:
169–182.
Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1986) A light
and electron microscopic analysis of the mossy fibers of the
rat DG. J. Comp. Neurol., 246: 435–458.
Conner, J.M., Lauterborn, J.C., Yan, Q., Gall, C.M. and
Varon, S. (1997) Distribution of brain-derived neurotrophic
factor (BDNF) protein and mRNA in the normal adult rat
CNS: evidence for anterograde axonal transport. J. Neuro-
sci., 17: 2295–2313.
Contractor, A., Rogers, C., Maron, C., Henkemeyer, M.,
Swanson, G.T. and Heinemann, S.F. (2002) Trans-synaptic
Eph receptor-ephrin signaling in hippocampal mossy fiber
LTP. Science, 296: 1864–1869.
Contractor, A., Swanson, G. and Heinemann, S.F. (2001)
Kainate receptors are involved in short- and long-term plas-
ticity at mossy fiber synapses in the hippocampus. Neuron,
29: 209–216.
Crawford, I.L. and Connor, J.D. (1973) Localization and re-
lease of glutamic acid in relation to the hippocampal mossy
fibre pathway. Nature, 244: 442–443.
Danzer, S.C. and McNamara, J.O. (2004) Localization of
brain-derived neurotrophic factor to distinct terminals of
mossy fiber axons implies regulation of both excitation and
feedforward inhibition of CA3 pyramidal cells. J. Neurosci.,
24: 11346–11355.
Darstein, M., Petralia, R.S., Swanson, G.T., Wenthold, R.J.
and Heinemann, S.F. (2003) Distribution of kainate receptor
subunits at hippocampal mossy fiber synapses. J. Neurosci.,
23: 8013–8019.
Debanne, D., Gahwiler, B.H. and Thompson, S.M. (1998)
Long-term synaptic plasticity between pairs of individual
CA3 pyramidal cells in rat hippocampal slice cultures. J.
Physiol., 507(Pt 1): 237–247.
Ding, R., Asada, H. and Obata, K. (1998) Changes in extra-
cellular glutamate and GABA levels in the hippocampal CA3
and CA1 areas and the induction of glutamic acid dec-
arboxylase-67 in dentate granule cells of rats treated with
kainic acid. Brain Res., 800: 105–113.
Draguhn, A., Verdorn, T.A., Ewert, M., Seeburg, P.H. and
Sakmann, B. (1990) Functional and molecular distinction
between recombinant rat GABAA receptor subtypes by
Zn2+. Neuron, 5: 781–788.
125
Drake, C.T., Patterson, T.A., Simmons, M.L., Chavkin, C. and
Milner, T.A. (1996) Kappa opioid receptor-like immunore-
activity in guinea pig brain: ultrastructural localization in
presynaptic terminals in hippocampal formation. J. Comp.
Neurol., 370: 377–395.
Egebjerg, J., Bettler, B., Hermans-Borgmeyer, I. and Heine-
mann, S. (1991) Cloning of a cDNA for a glutamate receptor
subunit activated by kainate but not AMPA. Nature, 351:
745–748.
Elmer, E., Kokaia, Z., Kokaia, M., Carnahan, J., Nawa, H.
and Lindvall, O. (1998) Dynamic changes of brain-derived
neurotrophic factor protein levels in the rat forebrain after
single and recurring kindling-induced seizures. Neuroscience,
83: 351–362.
Engel, D. and Jonas, P. (2005) Presynaptic action potential
amplification by voltage-gated Na+ channels in hippocampal
mossy fiber boutons. Neuron, 45: 405–417.
Erlander, M.G. and Tobin, A.J. (1991) The structural and
functional heterogeneity of glutamic acid decarboxylase: a
review. Neurochem. Res., 16: 215–226.
Foster, A.C., Mena, E.E., Monaghan, D.T. and Cotman, C.W.
(1981) Synaptic localization of kainic acid binding sites.
Nature, 289: 73–75.
Frahm, C., Engel, D., Piechotta, A., Heinemann, U. and Drag-
uhn, A. (2000) Presence of gamma-aminobutyric acid trans-
porter mRNA in interneurons and principal cells of rat
hippocampus. Neurosci. Lett., 288: 175–178.
Furtinger, S., Pirker, S., Czech, T., Baumgartner, C., Ransm-
ayr, G. and Sperk, G. (2001) Plasticity of Y1 and Y2 recep-
tors and neuropeptide Y fibers in patients with temporal lobe
epilepsy. J. Neurosci., 21: 5804–5812.
Gall, C. (1984) The distribution of cholecystokinin-like
immunoreactivity in the hippocampal formation of the
guinea pig: localization in the mossy fibers. Brain Res., 306:
73–83.
Gall, C. (1988) Seizures induce dramatic and distinctly different
changes in enkephalin, dynorphin, and CCK immunoreac-
tivities in mouse hippocampal mossy fibers. J. Neurosci., 8:
1852–1862.
Gall, C., Brecha, N., Karten, H.J. and Chang, K.J. (1981) Lo-
calization of enkephalin-like immunoreactivity to identified
axonal and neuronal populations of the rat hippocampus. J.
Comp. Neurol., 198: 335–350.
Gall, C. and Lauterborn, J. (1992) The DG: a model system for
studies of neurotrophin regulation. Epilepsy Res. Suppl., 7:
171–185.
Gall, C., Lauterborn, J., Isackson, P. and White, J. (1990) Sei-
zures, neuropeptide regulation, and mRNA expression in the
hippocampus. Prog. Brain Res., 83: 371–390.
Gall, C.M. and Isackson, P.J. (1989) Limbic seizures increase
neuronal production of messenger RNA for nerve growth
factor. Science, 245: 758–761.
Geiger, J.R. and Jonas, P. (2000) Dynamic control of presy-
naptic Ca(2+) inflow by fast-inactivating K(+) channels in
hippocampal mossy fiber boutons. Neuron, 28: 927–939.
Gloor, P., Vera, C.L. and Sperti, L. (1963) Electrophysiological
Studies of Hippocampal Neurons. I. Configuration and
126
laminar analysis of the ‘‘resting’’ potential gradient, of the
main-transient response to perforant path, fimbrial and
mossy fiber volleys and of ‘‘spontaneous’’ activity. Elect-
roencephalogr. Clin. Neurophysiol., 15: 353–378.
Gomez-Lira, G., Lamas, M., Romo-Parra, H. and Gutierrez,
R. (2005) Programmed and induced phenotype of the hip-
pocampal granule cells. J. Neurosci., 25: 6939–6946.
Gomez-Lira, G., Trillo, E., Ramirez, M., Asai, M., Sitges, M.
and Gutierrez, R. (2002) The expression of GABA in mossy
fiber synaptosomes coincides with the seizure-induced ex-
pression of GABAergic transmission in the mossy fiber
synapse. Exp. Neurol., 177: 276–283.
Gorter, J.A., Van Vliet, E.A., Proper, E.A., De Graan, P.N.,
Ghijsen, W.E., Lopes Da Silva, F.H. and Aronica, E. (2002)
Glutamate transporters alterations in the reorganizing DG
are associated with progressive seizure activity in chronic
epileptic rats. J. Comp. Neurol., 442: 365–377.
Gray, R., Rajan, A.S., Radcliffe, K.A., Yakehiro, M. and Dani,
J.A. (1996) Hippocampal synaptic transmission enhanced by
low concentrations of nicotine. Nature, 383: 713–716.
Griffith, W.H., Brown, T.H. and Johnston, D. (1986) Voltage-
clamp analysis of synaptic inhibition during long-term po-
tentiation in hippocampus. J. Neurophysiol., 55: 767–775.
Gutierrez, R. (2000) Seizures induce simultaneous GABAergic
and glutamatergic transmission in the DG-CA3 system. J.
Neurophysiol., 84: 3088–3090.
Gutierrez, R. (2002) Activity-dependent expression of simulta-
neous glutamatergic and GABAergic neurotransmission
from the mossy fibers in vitro. J. Neurophysiol., 87:
2562–2570.
Gutierrez, R. (2003) The GABAergic phenotype of the ‘‘glut-
amatergic’’ granule cells of the DG. Prog. Neurobiol., 71:
337–358.
Gutierrez, R. (2005) The dual glutamatergic-GABAergic phe-
notype of hippocampal granule cells. Trends Neurosci., 28:
297–303.
Gutierrez, R. and Heinemann, U. (2001) Kindling induces
transient fast inhibition in the DG-CA3 projection. Eur. J.
Neurosci., 13: 1371–1379.
Gutierrez, R., Romo-Parra, H., Maqueda, J., Vivar, C., Ra-
mirez, M., Morales, M.A. and Lamas, M. (2003) Plasticity of
the GABAergic phenotype of the ‘‘glutamatergic’’ granule
cells of the rat DG. J. Neurosci., 23: 5594–5598.
Hallermann, S., Pawlu, C., Jonas, P. and Heckmann, M. (2003)
A large pool of releasable vesicles in a cortical glutamatergic
synapse. Proc. Natl. Acad. Sci. U.S.A., 100: 8975–8980.
Hamlyn, L.H. (1961) Electron microscopy of mossy fibre end-
ings in Ammon’s horn. Nature, 190: 645–646.
Harris, E.W. and Cotman, C.W. (1986) Long-term potentiation
of guinea pig mossy fiber responses is not blocked by N-
methyl D-aspartate antagonists. Neurosci. Lett., 70: 132–137.
He, X.P., Minichiello, L., Klein, R. and McNamara, J.O.
(2002) Immunohistochemical evidence of seizure-induced ac-
tivation of trkB receptors in the mossy fiber pathway of adult
mouse hippocampus. J. Neurosci., 22: 7502–7508.
Henze, D.A., Card, J.P., Barrionuevo, G. and Ben-Ari, Y.
(1997) Large amplitude miniature excitatory postsynaptic
currents in hippocampal CA3 pyramidal neurons are of
mossy fiber origin. J. Neurophysiol., 77: 1075–1086.
Henze, D.A., McMahon, D.B., Harris, K.M. and Barrionuevo,
G. (2002a) Giant miniature EPSCs at the hippocampal mossy
fiber to CA3 pyramidal cell synapse are monoquantal. J.
Neurophysiol., 87: 15–29.
Henze, D.A., Wittner, L. and Buzsaki, G. (2002b) Single gran-
ule cells reliably discharge targets in the hippocampal CA3
network in vivo. Nat. Neurosci., 5: 790–795.
Ho, M.W., Beck-Sickinger, A.G. and Colmers, W.F. (2000)
Neuropeptide Y(5) receptors reduce synaptic excitation in
proximal subiculum, but not epileptiform activity in rat hip-
pocampal slices. J. Neurophysiol., 83: 723–734.
Holm, I.E., Geneser, F.A. and Zimmer, J. (1993) Cholecysto-
kinin-, enkephalin-, and substance P-like immunoreactivity in
the dentate area, hippocampus, and subiculum of the do-
mestic pig. J. Comp. Neurol., 331: 310–325.
Houser, C.R., Miyashiro, J.E., Swartz, B.E., Walsh, G.O.,
Rich, J.R. and Delgado-Escueta, A.V. (1990) Altered pat-
terns of dynorphin immunoreactivity suggest mossy fiber re-
organization in human hippocampal epilepsy. J. Neurosci.,
10: 267–282.
Huang, E.J. and Reichardt, L.F. (2001) Neurotrophins: roles in
neuronal development and function. Annu. Rev. Neurosci.,
24: 677–736.
Isackson, P.J., Huntsman, M.M., Murray, K.D. and
Gall, C.M. (1991) BDNF mRNA expression is increased
in adult rat forebrain after limbic seizures: temporal
patterns of induction distinct from NGF. Neuron, 6:
937–948.
Ito, I. and Sugiyama, H. (1991) Roles of glutamate receptors in
long-term potentiation at hippocampal mossy fiber synapses.
Neuroreport, 2: 333–336.
Iversen, L.L. and Kelly, J.S. (1975) Uptake and metabolism of
gamma-aminobutyric acid by neurones and glial cells. Bioc-
hem. Pharmacol., 24: 933–938.
Jacques, D., Dumont, Y., Fournier, A. and Quirion, R. (1997)
Characterization of neuropeptide Y receptor subtypes in the
normal human brain, including the hypothalamus. Neuro-
science, 79: 129–148.
Jaffe, D. and Johnston, D. (1990) Induction of long-term po-
tentiation at hippocampal mossy-fiber synapses follows a
Hebbian rule. J. Neurophysiol., 64: 948–960.
Jaffe, D.B. and Brown, T.H. (1997) Calcium dynamics in
thorny excrescences of CA3 pyramidal neurons. J. Ne-
urophysiol., 78: 10–18.
Jeub, M., Lie, A., Blumcke, I., Elger, C.E. and Beck, H. (1999)
Loss of dynorphin-mediated inhibition of voltage-dependent
Ca2+ currents in hippocampal granule cells isolated from
epilepsy patients is associated with mossy fiber sprouting.
Neuroscience, 94: 465–471.
Jin, W. and Chavkin, C. (1999) Mu opioids enhance mossy fiber
synaptic transmission indirectly by reducing GABAB recep-
tor activation. Brain Res., 821: 286–293.
Johnston, D. and Amaral, D.G. (2004) Hippocampus. In:
Shepherd G.M. (Ed.), The Synaptic Organization of the
Brain. Oxford University Press, New York.

Johnston, D. and Brown, T.H. (1983) Interpretation of voltage-
clamp measurements in hippocampal neurons. J. Ne-
urophysiol., 50: 464–486.
Jonas, P., Major, G. and Sakmann, B. (1993) Quantal compo-
nents of unitary EPSCs at the mossy fibre synapse on CA3
pyramidal cells of rat hippocampus. J. Physiol., 472: 615–663.
Jung, M.W. and McNaughton, B.L. (1993) Spatial selectivity
of unit activity in the hippocampal granular layer.
Hippocampus, 3: 165–182.
Kafitz, K.W., Rose, C.R., Thoenen, H. and Konnerth, A.
(1999) Neurotrophin-evoked rapid excitation through TrkB
receptors. Nature, 401: 918–921.
Kamiya, H. and Ozawa, S. (1999) Dual mechanism for presy-
naptic modulation by axonal metabotropic glutamate recep-
tor at the mouse mossy fibre-CA3 synapse. J. Physiol.,
518(Pt 2): 497–506.
Kamiya, H., Sawada, S. and Yamamoto, C. (1988) Synthetic
omega-conotoxin blocks synaptic transmission in the hippo-
campus in vitro. Neurosci. Lett., 91: 84–88.
Kamiya, H., Shinozaki, H. and Yamamoto, C. (1996) Activa-
tion of metabotropic glutamate receptor type 2/3 suppresses
transmission at rat hippocampal mossy fibre synapses. J.
Physiol., 493(Pt 2): 447–455.
Kaneko, T., Fujiyama, F. and Hioki, H. (2002) Immunohisto-
chemical localization of candidates for vesicular glutamate
transporters in the rat brain. J. Comp. Neurol., 444: 39–62.
Kapur, A., Yeckel, M. and Johnston, D. (2001) Hippocampal
mossy fiber activity evokes Ca2+ release in CA3 pyramidal
neurons via a metabotropic glutamate receptor pathway.
Neuroscience, 107: 59–69.
Kapur, A., Yeckel, M.F., Gray, R. and Johnston, D. (1998)
L-type calcium channels are required for one form of
hippocampal mossy fiber LTP. J. Neurophysiol., 79:
2181–2190.
Kasyanov, A.M., Safiulina, V.F., Voronin, L.L. and Cherubini,
E. (2004) GABA-mediated giant depolarizing potentials as
coincidence detectors for enhancing synaptic efficacy in the
developing hippocampus. Proc. Natl. Acad. Sci. U.S.A., 101:
3967–3972.
Kaufman, D.L., Houser, C.R. and Tobin, A.J. (1991) Two
forms of the gamma-aminobutyric acid synthetic enzyme
glutamate decarboxylase have distinct intraneuronal distri-
butions and cofactor interactions. J. Neurochem., 56:
720–723.
von Kitzing, E., Jonas, P. and Sakmann, B. (1994) Quantal
analysis of excitatory postsynaptic currents at the hippocam-
pal mossy fiber-CA3 pyramidal cell synapse. Adv. Second
Messenger Phosphoprotein Res., 29: 235–260.
Knowles, W.D., Schneiderman, J.H., Wheal, H.V., Stafstrom,
C.E. and Schwartzkroin, P.A. (1984) Hyperpolarizing poten-
tials in guinea pig hippocampal CA3 neurons. Cell Mol.
Neurobiol., 4: 207–230.
Kobayashi, K., Manabe, T. and Takahashi, T. (1996) Presy-
naptic long-term depression at the hippocampal mossy fiber-
CA3 synapse. Science, 273: 648–650.
Kohonen, T. (1977) Associative Memory: A System Theoretic
Approach. Springer-Verlag, New York.
127
Kukley, M., Schwan, M., Fredholm, B.B. and Dietrich, D.
(2005) The role of extracellular adenosine in regulating mossy
fiber synaptic plasticity. J. Neurosci., 25: 2832–2837.
Kullmann, D.M., Ruiz, A., Rusakov, D.M., Scott, R., Semya-
nov, A. and Walker, M.C. (2005) Presynaptic, extrasynaptic
and axonal GABAA receptors in the CNS: where and why?
Prog. Biophys. Mol. Biol., 87: 33–46.
Lamas, M., Gomez-Lira, G. and Gutierrez, R. (2001) Vesicular
GABA transporter mRNA expression in the DG and in
mossy fiber synaptosomes. Brain Res. Mol. Brain Res., 93:
209–214.
Langdon, R.B., Johnson, J.W. and Barrionuevo, G. (1993) As-
ynchrony of mossy fibre inputs and excitatory postsynaptic
currents in rat hippocampus. J. Physiol., 472: 157–176.
Lanthorn, T.H., Ganong, A.H. and Cotman, C.W. (1984)
2-Amino-4-phosphonobutyrate selectively blocks mossy
fiber-CA3 responses in guinea pig but not rat hippocampus.
Brain Res., 290: 174–178.
Lauri, S.E., Bortolotto, Z.A., Bleakman, D., Ornstein, P.L.,
Lodge, D., Isaac, J.T. and Collingridge, G.L. (2001) A crit-
ical role of a facilitatory presynaptic kainate receptor in
mossy fiber LTP. Neuron, 32: 697–709.
Lauri, S.E., Bortolotto, Z.A., Nistico, R., Bleakman, D.,
Ornstein, P.L., Lodge, D., Isaac, J.T. and Collingridge,
G.L. (2003) A role for Ca2+ stores in kainate receptor-
dependent synaptic facilitation and LTP at mossy fiber
synapses in the hippocampus. Neuron, 39: 327–341.
Lawrence, J.J., Grinspan, Z.M. and McBain, C.J. (2004)
Quantal transmission at mossy fibre targets in the CA3
region of the rat hippocampus. J. Physiol., 554: 175–193.
Lehmann, H., Ebert, U. and Loscher, W. (1996) Immunocyto-
chemical localization of GABA immunoreactivity in dentate
granule cells of normal and kindled rats. Neurosci. Lett., 212:
41–44.
Lei, S. and McBain, C.J. (2002) Distinct NMDA receptors
provide differential modes of transmission at mossy fiber-
interneuron synapses. Neuron, 33: 921–933.
Liang, Y., Yuan, L.L., Johnston, D. and Gray, R. (2002) Cal-
cium signaling at single mossy fiber presynaptic terminals in
the rat hippocampus. J. Neurophysiol., 87: 1132–1137.
Lie, A.A., Becker, A., Behle, K., Beck, H., Malitschek, B.,
Conn, P.J., Kuhn, R., Nitsch, R., Plaschke, M., Schramm, J.,
Elger, C.E., Wiestler, O.D. and Blumcke, I. (2000) Up-reg-
ulation of the metabotropic glutamate receptor mGluR4 in
hippocampal neurons with reduced seizure vulnerability.
Ann. Neurol., 47: 26–35.
Liu, H., Mazarati, A.M., Katsumori, H., Sankar, R. and
Wasterlain, C.G. (1999) Substance P is expressed in hippo-
campal principal neurons during status epilepticus and plays
a critical role in the maintenance of status epilepticus. Proc.
Natl. Acad. Sci. U.S.A., 96: 5286–5291.
Lorente de No, E. (1934) Studies on the structure of the cer-
ebral cortex II. Continuation of the study of the ammonic
system. J. Psychol. Neurol., 46: 113–177.
Lowenstein, D.H. and Arsenault, L. (1996) The effects of
growth factors on the survival and differentiation of cultured
DG neurons. J. Neurosci., 16: 1759–1769.
128
Maccaferri, G., Toth, K. and McBain, C.J. (1998) Target-
specific expression of presynaptic mossy fiber plasticity.
Science, 279: 1368–1370.
Makiura, Y., Suzuki, F., Chevalier, E. and Onteniente, B.
(1999) Excitatory granule cells of the DG exhibit a double
inhibitory neurochemical content after intrahippocampal ad-
ministration of kainate in adult mice. Exp. Neurol., 159:
73–83.
Manzoni, O.J., Castillo, P.E. and Nicoll, R.A. (1995) Pharma-
cology of metabotropic glutamate receptors at the mossy
fiber synapses of the guinea pig hippocampus. Neurophar-
macology, 34: 965–971.
Maqueda, J., Ramirez, M., Lamas, M. and Gutierrez, R. (2003)
Glutamic acid decarboxylase (GAD)67, but not GAD65, is
constitutively expressed during development and transiently
overexpressed by activity in the granule cells of the rat. Ne-
urosci. Lett., 353: 69–71.
Marr, D. (1971) Simple memory: a theory of archicortex. Phi-
los. Trans. R. Soc. Lond., 262: 23–81.
Marsh, D.J., Baraban, S.C., Hollopeter, G. and Palmiter, R.D.
(1999) Role of the Y5 neuropeptide Y receptor in limbic sei-
zures. Proc. Natl. Acad. Sci. U.S.A., 96: 13518–13523.
Marty, S., Berninger, B., Carroll, P. and Thoenen, H. (1996a)
GABAergic stimulation regulates the phenotype of hippo-
campal interneurons through the regulation of brain-derived
neurotrophic factor. Neuron, 16: 565–570.
Marty, S., Berzaghi Mda, P. and Berninger, B. (1997) Ne-
urotrophins and activity-dependent plasticity of cortical in-
terneurons. Trends Neurosci., 20: 198–202.
Marty, S., Carroll, P., Cellerino, A., Castren, E., Staiger, V.,
Thoenen, H. and Lindholm, D. (1996b) Brain-derived ne-
urotrophic factor promotes the differentiation of various
hippocampal nonpyramidal neurons, including Cajal-Retzius
cells, in organotypic slice cultures. J. Neurosci., 16: 675–687.
McCarthy, J.B., Walker, M., Pierce, J., Camp, P. and White,
J.D. (1998) Biosynthesis and metabolism of native and ox-
idized neuropeptide Y in the hippocampal mossy fiber sys-
tem. J. Neurochem., 70: 1950–1963.
McGinty, J.F., Henriksen, S.J., Goldstein, A., Terenius, L. and
Bloom, F.E. (1983) Dynorphin is contained within hippo-
campal mossy fibers: immunochemical alterations after
kainic acid administration and colchicine-induced neurotox-
icity. Proc. Natl. Acad. Sci. U.S.A., 80: 589–593.
McLean, S., Rothman, R.B., Jacobson, A.E., Rice, K.C. and
Herkenham, M. (1987) Distribution of opiate receptor sub-
types and enkephalin and dynorphin immunoreactivity in the
hippocampus of squirrel, guinea pig, rat, and hamster. J.
Comp. Neurol., 255: 497–510.
McNaughton, B.L. and Morris, R.G.M. (1987) Hippocampal
synaptic enhancement and information storage within a dis-
tributed memory system. Trends Neurosci., 10: 408–415.
McQuiston, A.R. and Colmers, W.F. (1996) Neuropeptide Y2
receptors inhibit the frequency of spontaneous but not min-
iature EPSCs in CA3 pyramidal cells of rat hippocampus. J.
Neurophysiol., 76: 3159–3168.
Miller, L.D., Petrozzino, J.J., Golarai, G. and Connor, J.A.
(1996) Ca2+ release from intracellular stores induced by
afferent stimulation of CA3 pyramidal neurons in hippo-
campal slices. J. Neurophysiol., 76: 554–562.
Min, M.Y., Rusakov, D.A. and Kullmann, D.M. (1998)
Activation of AMPA, kainate, and metabotropic receptors
at hippocampal mossy fiber synapses: role of glutamate
diffusion. Neuron, 21: 561–570.
Miyazaki, K., Ishizuka, T. and Yawo, H. (2005) Synapse-to-
synapse variation of calcium channel subtype contributions
in large mossy fiber terminals of mouse hippocampus.
Neuroscience, 136: 1003–1014.
Molnar, P. and Nadler, J.V. (2001) Synaptically-released zinc
inhibits N-methyl-D-aspartate receptor activation at recurrent
mossy fiber synapses. Brain Res., 910: 205–207.
Monaghan, D.T. and Cotman, C.W. (1982) The distribution of
[3H]kainic acid binding sites in rat CNS as determined by
autoradiography. Brain Res., 252: 91–100.
Monaghan, D.T. and Cotman, C.W. (1985) Distribution of
N-methyl-D-aspartate-sensitive L-[3H]glutamate-binding
sites in rat brain. J. Neurosci., 5: 2909–2919.
Moore, K.A., Nicoll, R.A. and Schmitz, D. (2003) Adenosine
gates synaptic plasticity at hippocampal mossy fiber
synapses. Proc. Natl. Acad. Sci. U.S.A., 100: 14397–14402.
Nadler, J.V. (2003) The recurrent mossy fiber pathway of the
epileptic brain. Neurochem. Res., 28: 1649–1658.
Nakazawa, K., Quirk, M.C., Chitwood, R.A., Watanabe, M.,
Yeckel, M.F., Sun, L.D., Kato, A., Carr, C.A., Johnston, D.,
Wilson, M.A. and Tonegawa, S. (2002) Requirement for
hippocampal CA3 NMDA receptors in associative memory
recall. Science, 297: 211–218.
Neuman, R.S., Ben-Ari, Y., Gho, M. and Cherubini, E. (1988)
Blockade of excitatory synaptic transmission by 6-cyano-7-
nitroquinoxaline-2,3-dione (CNQX) in the hippocampus in
vitro. Neurosci. Lett., 92: 64–68.
Nicoll, R.A., Castillo, P.E. and Weisskopf, M.G. (1994) The
role of Ca2+ in transmitter release and long-term potentia-
tion at hippocampal mossy fiber synapses. Adv. Second
Messenger Phosphoprotein Res., 29: 497–505.
Nicoll, R.A. and Schmitz, D. (2005) Synaptic plasticity at hip-
pocampal mossy fibre synapses. Nat. Rev. Neurosci., 6:
863–876.
Ogata, N. and Ueno, S. (1976) Mode of activation of pyramidal
neurons by mossy fiber stimulation in thin hippocampal slices
in vitro. Exp. Neurol., 53: 567–584.
Ohishi, H., Akazawa, C., Shigemoto, R., Nakanishi, S. and
Mizuno, N. (1995) Distributions of the mRNAs for L-2-
amino-4-phosphonobutyrate-sensitive metabotropic gluta-
mate receptors, mGluR4 and mGluR7, in the rat brain. J.
Comp. Neurol., 360: 555–570.
Ohishi, H., Shigemoto, R., Nakanishi, S. and Mizuno, N.
(1993) Distribution of the mRNA for a metabotropic gluta-
mate receptor (mGluR3) in the rat brain: an in situ hybrid-
ization study. J. Comp. Neurol., 335: 252–266.
Parra, P., Gulyas, A.I. and Miles, R. (1998) How many subtypes
of inhibitory cells in the hippocampus? Neuron, 20: 983–993.
Patrylo, P.R., van den Pol, A.N., Spencer, D.D. and
Williamson, A. (1999) NPY inhibits glutamatergic excitation
in the epileptic human DG. J. Neurophysiol., 82: 478–483.

Patz, S., Wirth, M.J., Gorba, T., Klostermann, O. and
Wahle, P. (2003) Neuronal activity and neurotrophic factors
regulate GAD-65/67 mRNA and protein expression in
organotypic cultures of rat visual cortex. Eur. J. Neurosci.,
18: 1–12.
Pavlidis, P. and Madison, D.V. (1999) Synaptic transmission in
pair recordings from CA3 pyramidal cells in organotypic
culture. J. Neurophysiol., 81: 2787–2797.
Pelkey, K.A., Lavezzari, G., Racca, C., Roche, K.W. and
McBain, C.J. (2005) mGluR7 is a metaplastic switch con-
trolling bidirectional plasticity of feedforward inhibition.
Neuron, 46: 89–102.
Pelkey, K.A., Topolnik, L., Lacaille, J.C. and McBain, C.J.
(2006) Compartmentalized Ca2+ channel regulation at
divergent mossy-fiber release sites underlies target cell-
dependent plasticity. Neuron, 52: 497–510.
Penttonen, M., Kamondi, A., Sik, A., Acsady, L. and Buzsaki,
G. (1997) Feed-forward and feed-back activation of the DG
in vivo during dentate spikes and sharp wave bursts. Hip-
pocampus, 7: 437–450.
Podlogar, M. and Dietrich, D. (2006) Firing pattern of rat
hippocampal neurons: a perforated patch clamp study. Brain
Res., 1085: 95–101.
Qian, J. and Noebels, J.L. (2005) Visualization of transmitter
release with zinc fluorescence detection at the mouse hippo-
campal mossy fibre synapse. J. Physiol., 566: 747–758.
Radian, R., Ottersen, O.P., Storm-Mathisen, J., Castel, M. and
Kanner, B.I. (1990) Immunocytochemical localization of the
GABA transporter in rat brain. J. Neurosci., 10: 1319–1330.
Ramirez, M. and Gutierrez, R. (2001) Activity-dependent ex-
pression of GAD67 in the granule cells of the rat hippocam-
pus. Brain Res., 917: 139–146.
Ramon y Cajal, S. (1911) Histology of the nervous system of
man and vertebrates. Translated from Spanish to French by
L. Azoulay, from the French to English by N. Swanson and
L.W. Swanson (1995) Oxford University Press, New York.
Rao, A., Cha, E.M. and Craig, A.M. (2000) Mismatched
appositions of presynaptic and postsynaptic components in
isolated hippocampal neurons. J. Neurosci., 20: 8344–8353.
Rattray, M. and Priestley, J.V. (1993) Differential expression of
GABA transporter-1 messenger RNA in subpopulations of
GABA neurones. Neurosci. Lett., 156: 163–166.
Reid, C.A., Dixon, D.B., Takahashi, M., Bliss, T.V. and Fine,
A. (2004) Optical quantal analysis indicates that long-term
potentiation at single hippocampal mossy fiber synapses is
expressed through increased release probability, recruitment
of new release sites, and activation of silent synapses. J.
Neurosci., 24: 3618–3626.
Reid, C.A., Fabian-Fine, R. and Fine, A. (2001) Postsynaptic
calcium transients evoked by activation of individual hippo-
campal mossy fiber synapses. J. Neurosci., 21: 2206–2214.
Represa, A., Tremblay, E. and Ben-Ari, Y. (1987) Kainate
binding sites in the hippocampal mossy fibers: localization
and plasticity. Neuroscience, 20: 739–748.
Reuter, H. (1995) Measurements of exocytosis from single pre-
synaptic nerve terminals reveal heterogeneous inhibition by
Ca(2+)-channel blockers. Neuron, 14: 773–779.
129
Ribak, C.E., Tong, W.M. and Brecha, N.C. (1996) GABA
plasma membrane transporters, GAT-1 and GAT-3, display
different distributions in the rat hippocampus. J. Comp. Ne-
urol., 367: 595–606.
Rivkees, S.A., Price, S.L. and Zhou, F.C. (1995) Immunohisto-
chemical detection of A1 adenosine receptors in rat brain
with emphasis on localization in the hippocampal formation,
cerebral cortex, cerebellum, and basal ganglia. Brain Res.,
677: 193–203.
Rolls, E.T. (1989) Functions of neuronal networks in the hip-
pocampus and neocortex in memory. In: Byrne J.H. and
Berry W.O. (Eds.), Neural Models of Plasticity. Academic
Press, San Diego, CA.
Rolls, E.T. and Kesner, R.P. (2006) A computational theory of
hippocampal function, and empirical tests of the theory.
Prog. Neurobiol., 79: 1–48.
Romo-Parra, H., Vivar, C., Maqueda, J., Morales, M.A. and
Gutierrez, R. (2003) Activity-dependent induction of multi-
transmitter signaling onto pyramidal cells and interneurons
of hippocampal area CA3. J. Neurophysiol., 89: 3155–3167.
Rose, C.R., Blum, R., Pichler, B., Lepier, A., Kafitz, K.W. and
Konnerth, A. (2003) Truncated TrkB-T1 mediates neurotro-
phin-evoked calcium signalling in glia cells. Nature, 426:
74–78.
Rothstein, J.D., Martin, L., Levey, A.I., Dykes-Hoberg, M.,
Jin, L., Wu, D., Nash, N. and Kuncl, R.W. (1994) Local-
ization of neuronal and glial glutamate transporters. Neuron,
13: 713–725.
Ruiz, A., Fabian-Fine, R., Scott, R., Walker, M.C., Rusakov,
D.A. and Kullmann, D.M. (2003) GABAA receptors at hip-
pocampal mossy fibers. Neuron, 39: 961–973.
Ruiz, A., Walker, M.C., Fabian-Fine, R. and Kullmann, D.M.
(2004) Endogenous zinc inhibits GABA(A) receptors in a
hippocampal pathway. J. Neurophysiol., 91: 1091–1096.
Safiulina, V.F., Fattorini, G., Conti, F. and Cherubini, E.
(2006) GABAergic signaling at mossy fiber synapses in neo-
natal rat hippocampus. J. Neurosci., 26: 597–608.
Salin, P.A., Scanziani, M., Malenka, R.C. and Nicoll, R.A.
(1996) Distinct short-term plasticity at two excitatory
synapses in the hippocampus. Proc. Natl. Acad. Sci.
U.S.A., 93: 13304–13309.
Salin, P.A., Weisskopf, M.G. and Nicoll, R.A. (1995) A com-
parison of the role of dynorphin in the hippocampal mossy
fiber pathway in guinea pig and rat. J. Neurosci., 15:
6939–6945.
Sandler, R. and Smith, A.D. (1991) Coexistence of GABA and
glutamate in mossy fiber terminals of the primate hippocam-
pus: an ultrastructural study. J. Comp. Neurol., 303:
177–192.
Sawada, S., Higashima, M. and Yamamoto, C. (1988) Kainic
acid induces long-lasting depolarizations in hippocampal
neurons only when applied to stratum lucidum. Exp. Brain
Res., 72: 135–140.
Sawada, S., Takada, S. and Yamamoto, C. (1983) Selective
activation of synapses near the tip of drug-ejecting micro-
electrode, and effects of antagonists of excitatory amino acids
in the hippocampus. Brain Res., 267: 156–160.
130
Scanziani, M., Gahwiler, B.H. and Thompson, S.M. (1993)
Presynaptic inhibition of excitatory synaptic transmission
mediated by alpha adrenergic receptors in area CA3 of the
rat hippocampus in vitro. J. Neurosci., 13: 5393–5401.
Scharfman, H.E., Goodman, J.H., Sollas, A.L. and Croll, S.D.
(2002) Spontaneous limbic seizures after intrahippocampal
infusion of brain-derived neurotrophic factor. Exp. Neurol.,
174: 201–214.
Scharfman, H.E. (2005) Bain-derived neurotrophic factor and
epilepsy-A missing link? Epilepsy Curr., 5:83–88.
Scharfman, H.E., Kunkel, D.D. and Schwartzkroin, P.A.
(1990) Synaptic connections of dentate granule cells and
hilar neurons: results of paired intracellular recordings and
intracellular horseradish peroxidase injections. Neuroscience,
37: 693–707.
Schmitz, D., Frerking, M. and Nicoll, R.A. (2000) Synaptic
activation of presynaptic kainate receptors on hippocampal
mossy fiber synapses. Neuron, 27: 327–338.
Schwarzer, C., Kofler, N. and Sperk, G. (1998) Up-regulation
of neuropeptide Y-Y2 receptors in an animal model of tem-
poral lobe epilepsy. Mol. Pharmacol., 53: 6–13.
Schwarzer, C. and Sperk, G. (1995) Hippocampal granule cells
express glutamic acid decarboxylase-67 after limbic seizures
in the rat. Neuroscience, 69: 705–709.
Scott, R. and Rusakov, D.A. (2006) Main determinants of
presynaptic Ca2+ dynamics at individual mossy fiber-CA3
pyramidal cell synapses. J. Neurosci., 26: 7071–7081.
Sepkuty, J.P., Cohen, A.S., Eccles, C., Rafiq, A., Behar, K.,
Ganel, R., Coulter, D.A. and Rothstein, J.D. (2002) A
neuronal glutamate transporter contributes to neurotrans-
mitter GABA synthesis and epilepsy. J. Neurosci., 22:
6372–6379.
Shigemoto, R., Kinoshita, A., Wada, E., Nomura, S., Ohishi,
H., Takada, M., Flor, P.J., Neki, A., Abe, T., Nakanishi, S.
and Mizuno, N. (1997) Differential presynaptic localization
of metabotropic glutamate receptor subtypes in the rat hip-
pocampus. J. Neurosci., 17: 7503–7522.
Shigemoto, R., Kulik, A., Roberts, J.D., Ohishi, H., Nusser, Z.,
Kaneko, T. and Somogyi, P. (1996) Target-cell-specific con-
centration of a metabotropic glutamate receptor in the pre-
synaptic active zone. Nature, 381: 523–525.
Simmons, M.L. and Chavkin, C. (1996) k-Opioid receptor ac-
tivation of a dendrotoxin-sensitive potassium channel medi-
ates presynaptic inhibition of mossy fiber neurotransmitter
release. Mol. Pharmacol., 50: 80–85.
Skaggs, W.E., McNaughton, B.L., Wilson, M.A. and Barnes,
C.A. (1996) Theta phase precession in hippocampal neuronal
populations and the compression of temporal sequences.
Hippocampus, 6: 149–172.
Sloviter, R.S., Dichter, M.A., Rachinsky, T.L., Dean, E.,
Goodman, J.H., Sollas, A.L. and Martin, D.L. (1996) Basal
expression and induction of glutamate decarboxylase and
GABA in excitatory granule cells of the rat and monkey
hippocampal DG. J. Comp. Neurol., 373: 593–618.
Smart, T.G., Xie, X. and Krishek, B.J. (1994) Modulation of
inhibitory and excitatory amino acid receptor ion channels by
zinc. Prog. Neurobiol., 42: 393–441.
Smith, M.A., Zhang, L.X., Lyons, W.E. and Mamounas, L.A.
(1997) Anterograde transport of endogenous brain-derived
neurotrophic factor in hippocampal mossy fibers. Neurore-
port, 8: 1829–1834.
Sperk, G., Schwarzer, C., Heilman, J., Furtinger, S., Reimer,
R.J., Edwards, R.H. and Nelson, N. (2003) Expression of
plasma membrane GABA transporters but not of the vesic-
ular GABA transporter in dentate granule cells after kainic
acid seizures. Hippocampus, 13: 806–815.
Spruston, N., Jaffe, D.B., Williams, S.H. and Johnston, D.
(1993) Voltage- and space-clamp errors associated with the
measurement of electrotonically remote synaptic events. J.
Neurophysiol., 70: 781–802.
Spruston, N. and Johnston, D. (1992) Perforated patch-clamp
analysis of the passive membrane properties of three classes
of hippocampal neurons. J. Neurophysiol., 67: 508–529.
Spruston, N., Lubke, J. and Frotscher, M. (1997) Interneurons
in the stratum lucidum of the rat hippocampus: an anatom-
ical and electrophysiological characterization. J. Comp.
Neurol., 385: 427–440.
Stengaard-Pedersen, K., Fredens, K. and Larsson, L.I. (1981)
Enkephalin and zinc in the hippocampal mossy fiber system.
Brain Res., 212: 230–233.
Swanson, T.H., Drazba, J.A. and Rivkees, S.A. (1995) Adeno-
sine A1 receptors are located predominantly on axons in the
rat hippocampal formation. J. Comp. Neurol., 363: 517–531.
Szabo, G., Kartarova, Z., Hoertnagl, B., Somogyi, R. and
Sperk, G. (2000) Differential regulation of adult and embry-
onic glutamate decarboxylases in rat dentate granule cells
after kainate-induced limbic seizures. Neuroscience, 100:
287–295.
Tashiro, A., Dunaevsky, A., Blazeski, R., Mason, C.A. and
Yuste, R. (2003) Bidirectional regulation of hippocampal
mossy fiber filopodial motility by kainate receptors: a two-
step model of synaptogenesis. Neuron, 38: 773–784.
Taupin, P., Ben-Ari, Y. and Roisin, M.P. (1994a) Subcellular
fractionation on Percoll gradient of mossy fiber synapto-
somes: evoked release of glutamate, GABA, aspartate and
glutamate decarboxylase activity in control and degranulated
rat hippocampus. Brain Res., 644: 313–321.
Taupin, P., Zini, S., Cesselin, F., Ben-Ari, Y. and Roisin, M.P.
(1994b) Subcellular fractionation on Percoll gradient of
mossy fiber synaptosomes: morphological and biochemical
characterization in control and degranulated rat hippocam-
pus. J. Neurochem., 62: 1586–1595.
Terrian, D.M., Johnston, D., Claiborne, B.J., Ansah-Yiadom,
R., Strittmatter, W.J. and Rea, M.A. (1988) Glutamate and
dynorphin release from a subcellular fraction enriched in
hippocampal mossy fiber synaptosomes. Brain Res. Bull., 21:
343–351.
Thoenen, H. (1995) Neurotrophins and neuronal plasticity.
Science, 270: 593–598.
Thoenen, H. (2000) Neurotrophins and activity-dependent
plasticity. Prog. Brain Res., 128: 183–191.
Thompson, S.M. and Gahwiler, B.H. (1992) Comparison of the
actions of baclofen at pre- and postsynaptic receptors in the
rat hippocampus in vitro. J. Physiol., 451: 329–345.

Tokunaga, T., Miyazaki, K., Koseki, M., Mobarakeh, J.I.,
Ishizuka, T. and Yawo, H. (2004) Pharmacological dissection
of calcium channel subtype-related components of strontium
inflow in large mossy fiber boutons of mouse hippocampus.
Hippocampus, 14: 570–585.
Tonder, N., Kragh, J., Finsen, B.R., Bolwig, T.G. and Zimmer,
J. (1994) Kindling induces transient changes in neuronal
expression of somatostatin, neuropeptide Y, and calbindin in
adult rat hippocampus and fascia dentata. Epilepsia, 35:
1299–1308.
Tong, G., Malenka, R.C. and Nicoll, R.A. (1996) Long-term
potentiation in cultures of single hippocampal granule cells: a
presynaptic form of plasticity. Neuron, 16: 1147–1157.
Tongiorgi, E., Armellin, M., Giulianini, P.G., Bregola, G.,
Zucchini, S., Paradiso, B., Steward, O., Cattaneo, A. and
Simonato, M. (2004) Brain-derived neurotrophic factor
mRNA and protein are targeted to discrete dendritic lami-
nas by events that trigger epileptogenesis. J. Neurosci., 24:
6842–6852.
Toth, K. and McBain, C.J. (1998) Afferent-specific innervation
of two distinct AMPA receptor subtypes on single hippo-
campal interneurons. Nat. Neurosci., 1: 572–578.
Toth, K., Suares, G., Lawrence, J.J., Philips-Tansey, E. and
McBain, C.J. (2000) Differential mechanisms of transmission
at three types of mossy fiber synapse. J. Neurosci., 20:
8279–8289.
Treviño, M. and Gutierrez, R. (2005) The GABAergic projec-
tion of the DG to hippocampal area CA3 of the rat: pre-
and postsynaptic actions after seizures. J. Physiol., 567:
939–949.
Treviño, M., Vivar, C. and Gutierrez, R. (2007) b/g Oscillatory
activity in the CA3 hippocampal area is depressed by aber-
rant GABAergic transmission from the DG after seizures. J.
Neurosci., 27: 251–259.
Tu, B., Timofeeva, O., Jiao, Y. and Nadler, J.V. (2005) Spon-
taneous release of neuropeptide Y tonically inhibits recurrent
mossy fiber synaptic transmission in epileptic brain. J. Ne-
urosci., 25: 1718–1729.
Urban, N.N. and Barrionuevo, G. (1996) Induction of hebbian
and non-hebbian mossy fiber long-term potentiation by dis-
tinct patterns of high-frequency stimulation. J. Neurosci., 16:
4293–4299.
Urban, N.N. and Barrionuevo, G. (1998) Active summation of
excitatory postsynaptic potentials in hippocampal CA3 py-
ramidal neurons. Proc. Natl. Acad. Sci. U.S.A., 95:
11450–11455.
Vezzani, A. and Sperk, G. (2004) Overexpression of NPY and
Y2 receptors in epileptic brain tissue: an endogenous neuro-
protective mechanism in temporal lobe epilepsy? Neuropep-
tides, 38: 245–252.
Vida, I. and Frotscher, M. (2000) A hippocampal interneuron
associated with the mossy fiber system. Proc. Natl. Acad. Sci.
U.S.A., 97: 1275–1280.
Villacres, E.C., Wong, S.T., Chavkin, C. and Storm, D.R.
(1998) Type I adenylyl cyclase mutant mice have impaired
mossy fiber long-term potentiation. J. Neurosci., 18:
3186–3194.
131
Vivar, C. and Gutierrez, R. (2005) Blockade of the membranal
GABA transporter potentiates GABAergic responses evoked
in pyramidal cells by mossy fiber activation after seizures.
Hippocampus, 15: 281–284.
Vogt, K., Mellor, J., Tong, G. and Nicoll, R. (2000) The actions
of synaptically released zinc at hippocampal mossy fiber
synapses. Neuron, 26: 187–196.
Vogt, K.E. and Regehr, W.G. (2001) Cholinergic modulation of
excitatory synaptic transmission in the CA3 area of the hip-
pocampus. J. Neurosci., 21: 75–83.
Walker, M.C., Ruiz, A. and Kullmann, D.M. (2001) Monosy-
naptic GABAergic signaling from dentate to CA3 with a
pharmacological and physiological profile typical of mossy
fiber synapses. Neuron, 29: 703–715.
Wang, J., Yeckel, M.F., Johnston, D. and Zucker, R.S. (2004)
Photolysis of postsynaptic caged Ca2+ can potentiate and
depress mossy fiber synaptic responses in rat hippocampal
CA3 pyramidal neurons. J. Neurophysiol., 91: 1596–1607.
Weisskopf, M.G., Castillo, P.E., Zalutsky, R.A. and
Nicoll, R.A. (1994) Mediation of hippocampal mossy fiber
long-term potentiation by cyclic AMP. Science, 265:
1878–1882.
Weisskopf, M.G. and Nicoll, R.A. (1995) Presynaptic changes
during mossy fibre LTP revealed by NMDA receptor-medi-
ated synaptic responses. Nature, 376: 256–259.
Weisskopf, M.G., Zalutsky, R.A. and Nicoll, R.A. (1993) The
opioid peptide dynorphin mediates heterosynaptic depression
of hippocampal mossy fibre synapses and modulates long-
term potentiation. Nature, 365: 188.
Wenzel, H.J., Cole, T.B., Born, D.E., Schwartzkroin, P.A. and
Palmiter, R.D. (1997) Ultrastructural localization of zinc
transporter-3 (ZnT-3) to synaptic vesicle membranes within
mossy fiber boutons in the hippocampus of mouse and mon-
key. Proc. Natl. Acad. Sci. U.S.A., 94: 12676–12681.
Werner, P., Voigt, M., Keinanen, K., Wisden, W. and Seeburg,
P.H. (1991) Cloning of a putative high-affinity kainate re-
ceptor expressed predominantly in hippocampal CA3 cells.
Nature, 351: 742–744.
Westbrook, G.L. and Mayer, M.L. (1987) Micromolar con-
centrations of Zn2+ antagonize NMDA and GABA re-
sponses of hippocampal neurons. Nature, 328: 640–643.
Widdowson, P.S. (1993) Quantitative receptor autoradiography
demonstrates a differential distribution of neuropeptide-Y
Y1 and Y2 receptor subtypes in human and rat brain. Brain
Res., 631: 27–38.
Wiebe, S.P. and Staubli, U.V. (1999) Dynamic filtering of rec-
ognition memory codes in the hippocampus. J. Neurosci., 19:
10562–10574.
Williams, S. and Johnston, D. (1988) Muscarinic depression of
long-term potentiation in CA3 hippocampal neurons. Sci-
ence, 242: 84–87.
Williams, S. and Johnston, D. (1989) Long-term potentiation of
hippocampal mossy fiber synapses is blocked by postsynaptic
injection of calcium chelators. Neuron, 3: 583–588.
Williams, S. and Johnston, D. (1990) Muscarinic depression of
synaptic transmission at the hippocampal mossy fiber
synapse. J. Neurophysiol., 64: 1089–1097.
132
Williams, S.H. and Johnston, D. (1991) Kinetic properties
of two anatomically distinct excitatory synapses in hippo-
campal CA3 pyramidal neurons. J. Neurophysiol., 66:
1010–1020.
Wisden, W. and Seeburg, P.H. (1993) A complex mosaic of
high-affinity kainate receptors in rat brain. J. Neurosci., 13:
3582–3598.
Wu, L.G., Borst, J.G. and Sakmann, B. (1998) R-type Ca2+
currents evoke transmitter release at a rat central synapse.
Proc. Natl. Acad. Sci. U.S.A., 95: 4720–4725.
Xiang, Z., Greenwood, A.C., Kairiss, E.W. and Brown, T.H.
(1994) Quantal mechanism of long-term potentiation in hip-
pocampal mossy-fiber synapses. J. Neurophysiol., 71:
2552–2556.
Yamamoto, C. (1972) Activation of hippocampal neurons by
mossy fiber stimulation in thin brain sections in vitro. Exp.
Brain Res., 14: 423–435.
Yamamoto, C., Higashima, M., Sawada, S. and Kamiya, H.
(1991) Quantal components of the synaptic potential induced
in hippocampal neurons by activation of granule cells, and
the effect of 2-amino-4-phosphonobutyric acid. Hippocam-
pus, 1: 93–106.
Yamamoto, C., Matsumoto, K. and Takagi, M. (1980) Potent-
iation of excitatory postsynaptic potentials during and after
repetitive stimulation in thin hippocampal sections. Exp.
Brain Res., 38: 469–477.
Yamamoto, C., Sawada, S. and Kamiya, H. (1992) Enhance-
ment of postsynaptic responsiveness during long-term
potentiation of mossy fiber synapses in guinea pig hippo-
campus. Neurosci. Lett., 138: 111–114.
Yamamoto, C., Sawada, S. and Ohno-Shosaku, T. (1994) Sup-
pression of hippocampal synaptic transmission by the spider
toxin omega-agatoxin-IV-A. Brain Res., 634: 349–352.
Yeckel, M.F., Kapur, A. and Johnston, D. (1999) Multiple
forms of LTP in hippocampal CA3 neurons use a common
postsynaptic mechanism. Nat. Neurosci., 2: 625–633.
Yokoi, M., Kobayashi, K., Manabe, T., Takahashi, T., Sa-
kaguchi, I., Katsuura, G., Shigemoto, R., Ohishi, H., No-
mura, S., Nakamura, K., Nakao, K., Katsuki, M. and
Nakanishi, S. (1996) Impairment of hippocampal mossy fiber
LTD in mice lacking mGluR2. Science, 273: 645–647.
Yoshino, M., Sawada, S., Yamamoto, C. and Kamiya, H.
(1996) A metabotropic glutamate receptor agonist DCG-IV
suppresses synaptic transmission at mossy fiber pathway of
the guinea pig hippocampus. Neurosci. Lett., 207: 70–72.
Zalutsky, R.A. and Nicoll, R.A. (1990) Comparison of two
forms of long-term potentiation in single hippocampal neu-
rons. Science, 248: 1619–1624.
Zalutsky, R.A. and Nicoll, R.A. (1992) Mossy fiber long-term
potentiation shows specificity but no apparent cooperativity.
Neurosci. Lett., 138: 193–197.
Zhang, G., Raol, Y.S., Hsu, F.C. and Brooks-Kayal, A.R.
(2004) Long-term alterations in glutamate receptor and
transporter expression following early-life seizures are asso-
ciated with increased seizure susceptibility. J. Neurochem.,
88: 91–101.
Zieglgänsberger, W., French, E.D., Siggins, G.R. and Bloom,
F.E. (1979) Opioid peptides may excite hippocampal pyram-
idal neurons by inhibiting adjacent inhibitory interneurons.
Science, 205: 415–417.
Plate 5.8. (A) Timm stained sections of hippocampus from adult rat, subjected to hippocampal x-irradiation as newborn, stop
odentate granule cell formation and lead to few MF terminals. (B) Timm stained section from the contralateral hemisphere of the rat
shown in A, depicting a well-integrated dentate transplant (tpl), derived from a small block of fascia dentata, grafted just after the x-
irradiation. From the graft, an apparently normal, laminar-specific MF projection (mf) developed, connecting dentate granule cells of
the graft (g) with the host CA3. The molecular layer of the graft (m) received a comparably laminar-specific projection of host
entorhinal perforant path projections, normalizing the Timm stained laminar appearance of the layer. (C) Slightly displaced dentate
transplant (tpl), grafted to x-irradiated newborn hippocampus just after irradiation as part of same experiment (Sunde et al., 1984).
Encroaching on the recipient CA3, MFs from the granule cells of this graft has entered adjacent parts of CA3 and project ‘‘down-
stream’’, but appear to stop at the border with CA1. Surprisingly, no MFs from the graft projected in the ‘‘upstream’’ direction, i.e.,
toward the host fascia dentata. Scale bar: 500 mm (For B/W version, see page 100 in the volume.)

Plate 6.1. Interneurons are the primary target of DG granule cells. Diagram representing the multiple types of excitatory contacts
made by granule cells. Within the dentate/hilar region collaterals of the mossy fibers (MFs) innervate both mossy cells and inhibitory
basket cells. In area CA3, the MFs contact pyramidal neurons via large expansions, but also excite interneurons through either
filopodial extensions from the large boutons or smaller en passant expansions along MF axons. (For B/W version, see page 110 in the
volume.)
Plate 6.2. Summary of the pre- and post-synaptic constituents of the different MF terminals and of their plasticity. (A) Schematic
representation of the giant MF boutons, which contact CA3 pyramidal cells. These terminals are characterized by low probability of
basal release and multiple release sites. (B) Schematic representation of filopodial extensions and en passant contacts, which synapse on
to interneurons. These have a high probability of basal release and single release sites. Both types of MF terminals contain several
neuromodulators and the neurotransmitters glutamate and GABA. Note that both the presynaptic and postsynaptic sites contain
several ionotropic GluRs and metabotropic GluRs, which confer to the MF a high degree of plasticity and the capacity for synaptic
integration. (C) Relative expression of the different releasable contents and of the receptors and transporters during development, in
the adult (D), and after epileptic activity (E). Arrows and font size indicate relative differences between the three states. (For B/W
version, see page 114 in the volume.)
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 7

Development of cell and fiber layers in

the dentate gyrus

Michael Frotscher, Shanting Zhao and Eckart Förster

Institute of Anatomy and Cell Biology, University of Freiburg, Albertstr. 17, D-79104 Freiburg, Germany

Abstract: This chapter deals with the laminated organization of the dentate gyrus, particularly with the
molecular signals controlling its development. First, sites of granule cell generation, their modes and routes
of migration are described. This is followed by an analysis of the molecular determinants governing the
formation of a tightly packed granule cell layer that is normal in rodents and primates. Reelin, a protein
of the extracellular matrix, plays an important role for the proper migration and lamination of the granule
cells during development and for the maintenance of a laminated dentate gyrus in adulthood. Granule cell
positioning is crucial for the laminated termination of commissural/associational fibers to the dentate
gyrus, suggesting that the granule cells carry positional signals for these fibers. In contrast, not signals
of the target cells but molecules of the extracellular matrix, such as hyaluronan, underlie the layer-specific
termination of fibers from the entorhinal cortex. The molecular determinants controlling axonal pathfind-
ing and target recognition of the profusely terminating cholinergic and GABAergic subcortical afferents
still need to be elucidated.

Keywords: granule cell; reelin; entorhinal fibers; commissural/associational fibers; cholinergic fibers

Introduction et al., 2006; Christie and Cameron, 2006; Tashiro

When compared to other brain regions, which
largely develop prenatally, the dentate gyrus is char-
acterized by an ongoing postnatal development.
This enduring neurogenesis of dentate granule cells
was first described in the early studies by Altman
and coworkers (Altman, 1966; Altman and Das,
1966; Altman and Bayer, 1975; Bayer, 1980; Bayer
and Altman, 1987) and has recently been further
elaborated (e.g., Kempermann et al., 1997, 1998,
2003, 2004, review; Eriksson et al., 1998; Markakis
and Gage, 1999; van Praag et al., 2002, 2005;
Schmidt-Hieber et al., 2004; Lie et al., 2005; Aimone
Corresponding author. Tel.: +(49) 761 203-5056; Fax: +(49)
761 203-5054; E-mail: Michael.Frotscher@anat.uni-freiburg.de
et al., 2006; Zhao et al., 2006). It has been a focus of
these recent studies how postnatal granule cell for-
mation can be modified by external stimuli and how
the newly generated granule cells become integrated
into the mature network of the dentate gyrus.
In this chapter, we will not directly deal with the
very interesting phenomenon of postnatal neuro-
genesis in the dentate gyrus and the factors that
can modify this process, and the reader is referred
to the articles cited above. Here, we will address
fundamental issues of dentate gyrus development,
such as where granule cells are generated, how
they migrate to reach their destination, how the
formation of a compact granule cell layer is con-
trolled, and how major afferents to the dentate
gyrus are instructed to form distinct, laminated

DOI: 10.1016/S0079-6123(07)63007-6 133

134
termination zones for the contact with defined
segments of the granule cell’s dendritic arbor. It is
likely that many of the signals involved in these
processes apply to both early-generated granule
cells and granule cells generated in the adult
organism. As far as the development of afferent
projections to the granule cells is concerned,
we will deal with the projection from the ent-
orhinal cortex to the dentate gyrus, commissural/
associational (C/A) projections, and subcortical
projections, i.e., the cholinergic and GABAergic
projections from the medial septum to the dentate
gyrus.
Sites of granule cell generation
Like the pyramidal cells of the hippocampus
proper, the first granule cells are generated in the
ventricular zone. From there they migrate to form
the suprapyramidal blade of the dentate gyrus,
which develops before the infrapyramidal blade.
Later on in development, the secondary prolifer-
ation zone emerges within the future hilus. Obvi-
ously, this region contains stem cells from the
ventricular zone that have retained their prolifer-
ative capacity. Cells derived from this secondary
proliferation zone form the late developing infra-
pyramidal blade of the dentate gyrus and are
the source of all ongoing postnatal granule cell
neurogenesis.
How do early- and late-generated granule cells
reach their final destinations in the suprapyrami-
dal and infrapyramidal blades of the dentate
gyrus, respectively? No definitive answer to this
question is possible, since real-time microscopy
studies of newly generated, migrating granule cells
are not yet available. From studies in the neocor-
tex, we know that there are two major modes
of neuronal migration, nuclear translocation and
glia-guided neuronal migration (Nadarajah and
Parnavelas, 2002). Movement of the nucleus
within early-generated processes is generally as-
sumed to be an early form of neuronal migration,
allowing for a migration over short distances in the
yet small, immature embryonic brain. Later
on, when large distances need to be bridged, glia-
guided migration becomes the dominant form of
cell movement. It is not clear which of the two
forms predominates in the dentate gyrus and
whether the temporal sequence assumed for neo-
cortical neurons (Nadarajah and Parnavelas, 2002)
holds true for dentate granule cells. It is known,
however, that a radial glial scaffold forms in the
dentate gyrus that persists long into the postnatal
period, suggesting that radial glia-guided migra-
tion takes place in the dentate gyrus despite
the relatively short distance that granule cells
have to migrate from the secondary proliferation
zone to their destination in the granule cell layer.
Studies in mutants with defects in the reelin signa-
ling cascade provide further evidence for a role of
the radial glial scaffold in the proper migration of
dentate granule cells. In these mutants, the radial
glial scaffold is altered to a varying extent, and this
is accompanied by migration defects of the granule
cells (Förster et al., 2002, 2006a, b; Frotscher et al.,
2003; Weiss et al., 2003).
Formation of a compact granule cell layer
The lamination of neuronal cell bodies and fibers
is a characteristic feature of cortical organization.
In the dentate gyrus, where the granule cells form
a densely packed cell layer, this lamination is par-
ticularly striking. Moreover, the granule cells show
a clear, bipolar orientation with their dendrites
extending into the molecular layer and their axons,
the mossy fibers, invading the hilar region. This
characteristic organization of the dentate gyrus
and its predominant neurons, the granule cells, has
been recognized since the first Golgi studies of the
hippocampus (Fig. 1; Golgi, 1886; Koelliker, 1896;
Ramón y Cajal, 1911). What are the molecular
signals that control this well-ordered arrangement
of the granule cells?
Much of our knowledge about the formation of
granule cell lamination derives from the study
of various mouse mutants. In reeler mice lacking
the extracellular matrix protein reelin, the granule
cells do not form a compact cell layer but are
scattered all over the dentate gyrus. Moreover,
their dendrites have lost the uniform orientation
toward the hippocampal fissure and extend to var-
ious directions (Stanfield and Cowan, 1979;
 135

Fig. 1. Original drawing of the dentate gyrus and hippocampus proper by Camillo Golgi (Golgi, 1886). While Golgi was aware of the
laminated, bipolar arrangement of the granule cells, he did not correctly draw the course of the mossy fibers. As is known from
numerous more recent tracer studies, granule cell axons mainly run in stratum lucidum of CA3 and impinge on proximal dendritic
segments of pyramidal neurons. (See Color Plate 7.1 in color plate section.)

Drakew et al., 2002). Granule cell axons do
not form the normal compact projection to stra-
tum lucidum of CA3 and appear defasciculated
(Drakew et al., 2002). However, like in wildtype
mice, mossy fibers in the reeler mutant respect the
border with CA1. As mentioned above, the radial
glial scaffold is also altered in mutants with defects
in the reelin signaling pathway.
How can one explain these multiple effects
of reelin on the radial glial scaffold, granule cell
migration, and granule cell orientation? An expla-
nation is easier after a brief description of the
reelin signaling cascade. Reelin is synthesized and
secreted by early-generated Cajal-Retzius cells
populating the marginal zone of the cortex, which
in the dentate gyrus corresponds to the outer
molecular layer (Del Rio et al., 1997). Thus, reelin
is not ubiquitously present in the tissue but
at specific sites. It came as a surprise that lipo-
protein receptors, the very low-density lipopro-
tein receptor (VLDLR) and the apolipoprotein
E receptor 2 (ApoER2) are receptors for reelin.
Double-knockouts lacking these two lipopro-
tein receptors show a reeler-like phenotype
(Trommsdorff et al., 1999). Binding of reelin to
its lipoprotein receptors results in the phosphor-
ylation of disabled 1 (dab1), an adapter protein
interacting with the intracellular domains of
ApoER2 and VLDLR, respectively. The reelin
signaling cascade eventually involves cytoskeletal
proteins such as tau. For details and a survey of
recent literature, the reader is referred to relevant
review articles (Tissir and Goffinet, 2003; Förster
et al., 2006a, b).
136
Recent studies have provided evidence for a dual
role of reelin in the development of the dentate gyrus
(Zhao et al., 2004; Förster et al., 2006a, b). It has
been suggested that reelin is a positional signal for
the directed growth of radial glial fibers in the dent-
ate gyrus (Förster et al., 2002; Zhao et al., 2004). In
the absence of reelin, its lipoprotein receptors or
dab1, no regular radial glial scaffold is formed
(Weiss et al., 2003). Moreover, glial fibrillary acid
protein (GFAP)-positive glial cells of the hippo-
campus were found to express molecules of the
reelin signaling cascade and showed a preference for
reelin in a stripe choice assay (Förster et al., 2002;
Frotscher et al., 2003). It was concluded that the
migration defects in mutants of the reelin pathway
were due, at least in part, to a malformation of the
radial glial scaffold (Frotscher et al., 2003).
A second effect of reelin appears to be on neurons
directly. In the reeler mutant and in ApoER2 / /
VLDLR / mutants, the marginal zone is densely
populated by neurons, whereas it is almost cell-free
in wildtype animals. Thus, reelin was regarded as a
stop signal, preventing migrating neurons from in-
vading the marginal zone. Sanada et al. (2004) have
recently shown that phosphorylation of dab1 at
Fig. 2. Rescue of granule cell lamination in a slice culture of
reeler dentate gyrus is achieved by coculturing with a wildtype
culture providing a reelin-containing marginal zone. Two reeler
cultures (rl / 1 and rl / 2) are cocultured with a rat hippo-
campal slice. A compact cell layer (arrow) has only formed in
rl / 1, which was cocultured next to the outer molecular layer
of the rat dentate gyrus containing reelin-synthesizing Cajal-
Retzius cells. In rl / 2, which was cultured next to the stratum
oriens of CA1, the reeler-specific loose distribution of neurons
in the dentate gyrus is retained (arrowhead). Dashed lines rep-
resent borders between cultures. CA1, CA3, hippocampal re-
gions CA1 and CA3; DG, dentate gyrus; g, granule cell layer.
Scale bar: 200 mm (from Zhao et al., 2004, with permission).
(See Color Plate 7.2 in color plate section.)
tyrosine 220 and 232 leads to a detachment of the
migrating neuron from the radial fiber, thus termi-
nating the migration process. By coculturing slices
of reeler hippocampus with wildtype slices, such that
a reelin-containing marginal zone was proximal to
the reeler dentate gyrus, Zhao et al. (2004) were able
to show that a compact granule cell layer could
be rescued in the reeler dentate gyrus, leaving the
molecular layer free of neurons. No rescue was
achieved when a reeler slice was cocultured with
stratum oriens of CA1, which did not contain reelin-
synthesizing Cajal-Retzius cells (Fig. 2). These au-
thors also showed that a regularly oriented radial
glial scaffold could be rescued under these in vitro
conditions. It remains to be determined whether
reelin’s effect on radial glial cells or direct effect
on neurons, or both, is crucial for the formation of
a compact granule cell layer.
Formation of afferent fiber lamination in the dentate
gyrus
With the exception of some subcortical afferents,
major projections to the dentate gyrus show
a preference for distinct layers. Fibers originating
from neurons in layer 2 of the entorhinal cortex
(entorhinal fibers) are known to terminate in the
outer two-thirds of the molecular layer (Blackstad,
1958; Steward and Scoville, 1976; Amaral and
Witter, 1995). C/A fibers, derived from hilar mossy
cells projecting ipsilaterally and contralaterally,
terminate in the inner molecular layer (Blackstad
1956; Deller et al., 1995), and there is no overlap of
these two major projections to the dentate gyrus.
In the following paragraphs, we will summarize
what is known about the development of these
projections and the factors determining their
laminar specificities.
Entorhinal afferents
Fibers from the entorhinal cortex are among the
first afferents to reach the dentate gyrus in the
rodent (at E18/E19; Super and Soriano, 1994;
Ceranik et al., 1999). This early arrival implies that
their regional specificity and layer-specific termi-
nation in the outer two-thirds of molecular layer

cannot be controlled by their definitive target cells,
the granule cells, which are mainly generated after
birth (see above). How do the entorhinal axons
find their long way from the entorhinal cortex
across the hippocampal fissure and the subiculum?
Recent studies have indicated that early-generated
Cajal-Retzius cells located in the marginal zone
give rise to an early projection to the entorhinal
cortex, thereby providing a template for the later
outgrowth of entorhinal fibers. In fact, intracellu-
lar labeling studies as well as retrograde and
anterograde tracing substantiated this early dent-
ate-entorhinal projection of Cajal-Retzius cells
(Ceranik et al., 1999). Moreover, Cajal-Retzius
cells seem to serve as transient targets of early
arriving entorhinal fibers in the absence of the
granule cells. Electron microscopic studies showed
that entorhinal axons established synapses with
reelin-immunoreactive Cajal-Retzius cells in the
outer molecular layer, and selective removal of
Cajal-Retzius cells prevented the ingrowth of ent-
orhinal afferents (Del Rio et al., 1997; Frotscher
et al., 2001). Collectively, these findings provide
evidence for a role of Cajal-Retzius cells as a tem-
plate for outgrowing entorhinal axons. However,
reelin, secreted by Cajal-Retzius cells and forming
a component of the extracellular matrix in the
outer molecular layer, does not seem to be
required for the layer-specific ingrowth of ent-
orhinal axons. In the reeler mutant lacking reelin,
entorhinal axons find their correct laminar posi-
tion (Del Rio et al., 1997; Zhao et al., 2003).
Recent studies have provided evidence that other
molecules of the extracellular matrix such as
hyaluronan and molecules bound to it play
an important role in the laminar specificity of ent-
orhinal fibers. Treating slice cultures of hippo-
campus with hyaluronidase, which degrades
hyaluronan, abolishes the layer-specific termina-
tion of entorhinal fibers but not of C/A axons
(Zhao et al., 2003).
Commissural/associational fibers
C/A fibers represent ipsilateral and contralateral
projections of hilar mossy cells (Ribak et al., 1985;
Frotscher et al., 1991; Frotscher, 1992). C/A fibers
137
arrive in their termination zone, the inner molec-
ular layer, after the fibers from the entorhinal cor-
tex, i.e., at P2. This implies that many granule cells
have already developed proximal dendrites for the
contact with C/A fibers, and no transient target
neurons, as in the case of the entorhinal projec-
tion, are required. In fact, the laminated projection
of C/A fibers mirrors the laminated arrangement
of the granule cells in wildtype animals. Accord-
ingly, in mutants with defects in the reelin signa-
ling pathway showing migration defects of the
granule cells and no compact granular layer, the
projection of C/A fibers is diffuse and not lami-
nated (Deller et al., 1999; Gebhardt et al., 2002;
Zhao et al., 2003). A diffuse termination of C/A
fibers is most pronounced in the reeler mutant and
in double-knockouts lacking ApoER2 and
VLDLR, but is less prominent in single receptor
mutants which only show minor migration defects
of the granule cells. Zhao et al. (2003) used slice
cocultures to study the development of the com-
missural projection to the dentate gyrus. By
coculturing slices of reeler hippocampus and wild-
type hippocampus and tracing the ‘‘commissural’’
projection developing under these in vitro condi-
tions, they showed that the loose distribution of C/
A fibers is not a cell-autonomous effect of the
reeler mutation on the C/A projection. C/A fibers
originating from a reeler hippocampal culture
terminated with correct laminar specificity in the
inner molecular layer in a cocultured wildtype
slice, whereas C/A fibers from a wildtype culture
projecting to a reeler culture terminated profusely,
thus reflecting the scattered distribution of the
granule cells in the reeler culture (Fig. 3). Finally,
rescue of granule cell lamination in a reeler culture
by coculturing to a wildtype slice (see above) also
rescued the laminated termination of C/A fibers
(Zhao et al., 2004; Förster et al., 2006a, b). To-
gether these data indicate that the granule cells
carry positional signals for the laminated termina-
tion of C/A fibers. This contrasts to the guidance
cues relevant for the proper termination of ent-
orhinal fibers. As described, they are guided
to their termination zone by transient pioneer
neurons (Cajal-Retzius cells) and molecules of the
extracellular matrix. Thus, as one would expect, in
the reeler mutant with a scattered distribution of
138
the granule cells, the C/A fibers terminate diffu-
sely, but the entorhino-dentate fibers terminate
with laminar specificity in the outer molecular
layer forming a sharp border (Zhao et al., 2003).
Collectively, these findings show that different
types of molecular signal control laminar specifi-
city of commissural and entorhinal fibers to the
dentate gyrus, molecules associated with the cell
membranes of the granule cells and extracellular
matrix molecules, respectively. In light of these
results, the temporal sequence of fiber ingrowth
during development does not seem to determine
the layered termination of entorhinal and com-
missural afferents as was hypothesized by Bayer
and Altman (1987). This attractive temporal
hypothesis was challenged by Frotscher and
Heimrich (1993) when they used sequential cocul-
tures of hippocampus with its afferent regions.
Reversing the sequence of arrival of entorhinal
and commissural fibers did not change their layer-
specific termination in the dentate gyrus.
Cholinergic and GABAergic afferents from the
medial septum
While both the entorhinal fibers and C/A fibers are
characterized by their segregated, layer-specific
termination in the dentate gyrus, no such strictly
laminar termination is observed with subcortical
Fig. 3. Cocultures of two hippocampal slices to allow for the
formation of ‘‘commissural’’ projections, traced by anterograd-
ely transported biocytin injected into the hilar region of one of
the two hippocampal slices (sites of biocytin injections labeled
by asterisks). (A) Coculturing of two wildtype slices results in
the formation of a compact ‘‘commissural’’ projection in the
inner molecular layer (black arrow). Open arrow labels ipsilat-
eral mossy fiber projection in stratum lucidum of CA3. CA1,
CA3, hippocampal regions CA1 and CA3; g, granule cell layer.
(B) Coculture of wildtype (wt) hippocampus and reeler hippo-
campus. Biocytin-labeled ‘‘commissural’’ fibers from the wild-
type culture terminate profusely in the reeler culture, thus
reflecting the scattered distribution of their target neurons, the
granule cells. Open arrow labels the mossy fiber projection in
the wildtype culture. DG, dentate gyrus of the reeler culture
with scattered granule cells; P, pyramidal layer in the wildtype
culture; P1, P2, double pyramidal layer in the reeler culture.
(C) ‘‘Commissural’’ fibers from a reeler culture form a compact
projection in the inner molecular layer of the wildtype culture
(black arrow), indicating that reeler commissural fibers
project to the inner molecular layer as is normal, provided that
their target cells, the granule cells, form a tightly packed
layer. g, granule cell layer. Scale bar: 100 mm (applies to A–C)
(modified from Zhao et al., 2003, with permission; copyright by
the Society for Neuroscience). (See Color Plate 7.3 in color
plate section.)

afferents, such as the cholinergic and GABAergic
fibers from the medial septum and cat-
echolaminergic brain stem fibers (the latter not
dealt with in this chapter). The large cholinergic
neurons of the medial septum/diagonal band (MS/
DB) complex project via the fimbria-fornix to the
hippocampus where they terminate profusely in all
hippocampal layers (Frotscher and Leranth,
1985). A concentration of cholinergic (choline
acetyltransferase-immunoreactive) fibers is regu-
larly observed in the cell body layers, in the dent-
ate gyrus it is mainly in the subgranular zone. The
fine, varicose cholinergic fibers establish synaptic
contacts with a variety of postsynaptic structures
including cell bodies, dendritic shafts, and spines
(Frotscher and Leranth, 1986).
The GABAergic septohippocampal fibers, not
showing a laminated termination pattern like the
cholinergic fibers, specifically form synapses with
hippocampal GABAergic interneurons, thus serv-
ing a disinhibitory function (Freund and Antal,
1988). The cholinergic afferents that terminate
on both the hippocampal principal cells and
GABAergic interneurons do not have such a cell-
specific termination.
Little is known about the molecular signals con-
trolling the development of the septohippocampal
projection and the different target cell specificities
of cholinergic and GABAergic septohippocampal
fibers. Sema3C and its receptor neuropilin 2 have
been suggested to play a role in the guidance of
septal fibers to the hippocampus (Chedotal et al.,
1998; Steup et al., 2000; Skutella and Nitsch,
2001). Cholinergic fibers express the p75 neurotro-
phin receptor (p75NTR), and nerve growth factor
(NGF), one of the ligands for this receptor, is ex-
pressed by GABAergic interneurons in the hippo-
campus. Since GABAergic hippocampal neurons
project to the septum, they may thus provide both
a guiding scaffold and a neurotrophic source for
the ingrowing cholinergic fibers. The diffuse ter-
mination of septohippocampal fibers would then
result from the scattered distribution of GAB-
Aergic interneurons in the hippocampus and dent-
ate gyrus. That hippocampal neurons projecting
to the septum are involved in the guidance of
septohippocampal fibers is supported by tracer
studies on the development of these two
139
projections. The hippocampo-septal projection de-
velops early, clearly before the septohippocampal
projection (Linke et al., 1995) with hippocampal
fibers reaching the septum around E16. Septal fib-
ers arrive in the hippocampus 2–3 days later.
Moreover, it has been shown that the septal fibers
grow along hippocampo-septal axons (Linke and
Frotscher, 1993). The determinants of the different
target cell specificities of cholinergic and GAB-
Aergic septohippocampal fibers remain to be
elucidated.
Functional considerations
The functional significance of the laminated organ-
ization of the dentate gyrus is not known. In tem-
poral lobe epilepsy associated with Ammon’s horn
sclerosis, the lamination of the dentate gyrus is dis-
rupted, and the granule cells do not form a tightly
packed cell layer (granule cell dispersion, see
Houser, 1990). These structural changes may indi-
cate that a laminated dentate gyrus with a clear
segregation of the input site (granule cell dendrites
in the molecular layer) from the output site (granule
cell axons in the hilus) is required for the proper
function of this brain region. This hypothesis im-
plies that structural alterations, for instance migra-
tion defects during development, underlie the
development of the epileptic disorder.
Recent studies of animal models of epilepsy
have suggested a different scenario. Unilateral
injections of the glutamate agonist kainate into the
hippocampus of mice induced status epilepticus
and, with some delay, a characteristic granule cell
dispersion (Bouilleret et al., 1999; Heinrich et al.,
2006), thus suggesting that the dispersion of gran-
ule cells is a result of seizure activity rather than
being its source. How to explain, then, the loss of
granule cell lamination following seizure activity?
Haas et al. (2002) showed that the number
of reelin-synthesizing neurons in tissue samples of
hippocampus removed from epileptic patients for
therapeutical reasons was dramatically decreased
when compared to control samples. A similar loss
of reelin-synthesizing cells was observed following
unilateral kainate injection into the hippocampus
of adult mice (Heinrich et al., 2006). Granule cell
140
dispersion thus seems to be accompanied by a loss
of reelin-expressing cells both in human temporal
lobe epilepsy and in an animal model of the dis-
ease. During development, reelin is required for
the formation of a tightly packed granule cell
layer. The findings from epileptic tissue indicate
that reelin is also required for the maintenance of
a compact granule cell layer during adulthood.
Cajal-Retzius cells synthesizing reelin are very sen-
sitive to glutamate agonists (Del Rio et al., 1997).
Hence, epileptic seizures associated with excessive
glutamate release may affect Cajal-Retzius cells,
resulting in decreased reelin expression. In turn,
loss of reelin, a stop signal for neurons, may allow
the granule cells to leave their layer, resulting in
granule cell dispersion. A role for reelin in the
maintenance of a compact granule cell layer in
adult animals was confirmed by chronic unilateral
infusion of CR-50, a reelin-blocking antibody, into
the hippocampus of adult, naı̈ve animals. On the
injection side, but not on the control side, granule
cell dispersion developed. There was no granule
cell dispersion when the reelin-blocking CR-50
antibody was replaced by nonspecific IgG
(Heinrich et al., 2006). These data, as well as other
recent studies showing that sprouted mossy fibers
in epileptic animals may exert an inhibitory rather
than the commonly assumed excitatory effect on
hippocampal principal cells (Sloviter et al., 2006),
indicate that structural changes in the hippocam-
pus in epilepsy are likely to be the consequence
of the disease rather than its cause. However, the
functional significance of a strictly laminated dent-
ate gyrus, first noticed by Golgi and his contem-
poraries, still requires explanation.
Acknowledgments
Supported by the German Research Foundation
(SFB 505, TR-3) and Jung Foundation for Science
and Research.
References
Aimone, J.B., Wiles, J. and Gage, F.H. (2006) Potential role for
adult neurogenesis in the encoding of time in new memories.
Nat. Neurosci., 9: 723–727.
Altman, J. (1966) Autoradiographic and histological studies of
postnatal neurogenesis. II. A longitudinal investigation of the
kinetics, migration and transformation of cells incorporating
tritiated thymidine in infant rats, with special reference
to postnatal neurogenesis in some brain regions. J. Comp.
Neurol., 128: 431–474.
Altman, J. and Bayer, S.A. (1975) Postnatal development of
the hippocampal dentate gyrus under normal and experi-
mental conditions. In: Isaacson R.L. and Pribram K.H.
(Eds.), The Hippocampus, Vol. 2. Plenum Press, New
York, pp. 95–122.
Altman, J. and Das, G.D. (1966) Autoradiographic and histo-
logical studies of postnatal neurogenesis. I. A longitudinal
investigation of the kinetics, migration and transformation of
cells incorporating tritiated thymidine in neonate rats, with
special reference to postnatal neurogenesis in some brain
regions. J. Comp. Neurol., 126: 337–389.
Amaral, D.G. and Witter, M.P. (1995) The hippocampal
formation. In: Paxinos G. (Ed.), The Rat Nervous System
(2nd Ed). Academic Press, New York, pp. 443–494.
Bayer, S.A. (1980) Development of the hippocampal region in
the rat. I. Neurogenesis examined with 3H-thymidine auto-
radiography. J. Comp. Neurol., 190: 87–114.
Bayer, S.A. and Altman, J. (1987) Directions in neurogenetic
gradients and patterns of anatomical connections in the
telencephalon. Prog. Neurobiol., 29: 57–106.
Blackstad, T.W. (1956) Commissural connections of the hip-
pocampal region in the rat, with special reference to their
mode of termination. J. Comp. Neurol., 105: 417–537.
Blackstad, T.W. (1958) On the termination of some afferents to
the hippocampus and fascia dentata: an experimental study
in the rat. Acta Anat. (Basel), 35: 202–214.
Bouilleret, V., Ridoux, V., Depaulis, A., Marescaux, C., Nehlig,
A. and Le Gal La Salle, G. (1999) Recurrent seizures and
hippocampal sclerosis following intrahippocampal kainate
injection in adult mice: electroencephalography, histopathol-
ogy and synaptic reorganization similar to mesial temporal
lobe epilepsy. Neuroscience, 89: 717–729.
Ceranik, K., Deng, J., Heimrich, B., Lübke, J., Zhao, S.,
Förster, E. and Frotscher, M. (1999) Hippocampal Cajal-
Retzius cells project to the entorhinal cortex: retrograde
tracing and intracellular labelling studies. Eur. J. Neurosci.,
11: 4278–4290.
Chedotal, A., del Rio, J.A., Ruiz, M., He, Z., Borrell, V., de
Castro, F., Ezan, F., Goodman, C.S., Tessier-Lavigne, M.,
Sotelo, C. and Soriano, E. (1998) Semaphorins III and IV
repel hippocampal axons via two distinct receptors. Devel-
opment, 125: 4313–4323.
Christie, B.R. and Cameron, H.A. (2006) Neurogenesis in the
adult hippocampus. Hippocampus, 16: 199–207.
Del Rio, J.A., Heimrich, B., Borrell, V., Förster, E., Drakew,
A., Alcantara, S., Nakajima, K., Miyata, T., Ogawa, M.,
Mikoshiba, K., Derer, P., Frotscher, M. and Soriano, E.
(1997) A role for Cajal-Retzius cells and reelin in the devel-
opment of hippocampal connections. Nature, 385: 70–74.
Deller, T., Drakew, A. and Frotscher, M. (1999) Different
primary target cells are important for fiber lamination

in the fascia dentata: a lesson from reeler mutant mice. Exp.
Neurol., 156: 239–253.
Deller, T., Nitsch, R. and Frotscher, M. (1995) Phaseolus vul-
garis-Leucoagglutinin tracing of commissural fibers to the rat
dentate gyrus: evidence for a previously unknown commis-
sural projection to the outer molecular layer. J. Comp.
Neurol., 352: 55–68.
Drakew, A., Deller, T., Heimrich, B., Gebhardt, C., Del Turco,
D., Tielsch, A., Förster, E., Herz, J. and Frotscher, M. (2002)
Dentate granule cells in reeler mutants and VLDLR and
ApoER2 knockout mice. Exp. Neurol., 176: 12–24.
Eriksson, P.S., Perfilieva, E., Bjork-Eriksson, T., Alborn, A.M.,
Nordborg, C., Peterson, D.A. and Gage, F.H. (1998) Ne-
urogenesis in the adult human hippocampus. Nat. Med., 4:
1313–1317.
Förster, E., Jossin, Y., Zhao, S., Chai, X., Frotscher, M. and
Goffinet, A.M. (2006a) Recent progress in understand-
ing the role of Reelin in radial neuronal migration, with
specific emphasis on the dentate gyrus. Eur. J. Neurosci.,
23: 901–909.
Förster, E., Tielsch, A., Saum, B., Weiss, K.H., Johanssen, C.,
Graus-Porta, D., Müller, U. and Frotscher, M. (2002) Reel-
in, Disabled 1, and beta1 integrins are required for the for-
mation of the radial glial scaffold in the hippocampus. Proc.
Natl. Acad. Sci. U.S.A., 99: 13178–13183.
Förster, E., Zhao, S. and Frotscher, M. (2006b) Laminating the
hippocampus. Nat. Rev. Neurosci., 7: 259–267.
Freund, T.F. and Antal, M. (1988) GABA-containing neurons
in the septum control inhibitory interneurons in the hippo-
campus. Nature, 336: 170–173.
Frotscher, M. (1992) Application of the Golgi/electron
microscopy technique for cell identification in immunocyto-
chemical, retrograde labeling, and developmental studies
of hippocampal neurons. Microsc. Res. Technol., 23:
306–323.
Frotscher, M., Haas, C.A. and Förster, E. (2003) Reelin con-
trols granule cell migration in the dentate gyrus by acting on
the radial glial scaffold. Cereb. Cortex, 13: 634–640.
Frotscher, M. and Heimrich, B. (1993) Formation of layer-
specific fiber projections to the hippocampus in vitro. Proc.
Natl. Acad. Sci. U.S.A., 90: 10400–10403.
Frotscher, M. and Leranth, C. (1985) Cholinergic innervation
of the rat hippocampus as revealed by choline acetyltransf-
erase immunocytochemistry: a combined light and electron
microscopic study. J. Comp. Neurol., 239: 237–246.
Frotscher, M. and Leranth, C. (1986) The cholinergic innerva-
tion of the rat fascia dentata: identification of target struc-
tures on granule cells by combining choline acetyltransferase
immunocytochemistry and Golgi impregnation. J. Comp.
Neurol., 243: 58–70.
Frotscher, M., Seress, L., Abraham, H. and Heimrich, B. (2001)
Early generated Cajal-Retzius cells have different functions
in cortical development. In: Miyan J.A., Thorndyke M.,
Beesley P.W. and Bannister C.M. (Eds.), Brain Stem Cells.
BIOS Scientific Publishers Ltd., Oxford, pp. 43–49.
Frotscher, M., Seress, L., Schwerdtfeger, W.K. and Buhl, E.
(1991) The mossy cells of the fascia dentata: a comparative
141
study of their fine structure and synaptic connections in ro-
dents and primates. J. Comp. Neurol., 312: 145–163.
Gebhardt, C., Del Turco, D., Drakew, A., Tielsch, A., Herz, J.,
Frotscher, M. and Deller, T. (2002) Abnormal positioning of
granule cells alters afferent fiber distribution in the mouse
fascia dentata: morphologic evidence from reeler, apolipo-
protein E receptor 2-, and very low density lipoprotein re-
ceptor knockout mice. J. Comp. Neurol., 445: 278–292.
Golgi, C. (1886) Sulla fina anatomica degli organi centrali del
sistema nervoso. Hoepli, Milan.
Haas, C.A., Dudeck, O., Kirsch, M., Huszka, C., Kann, G.,
Pollak, S., Zentner, J. and Frotscher, M. (2002) Role for
reelin in the development of granule cell dispersion in tem-
poral lobe epilepsy. J. Neurosci., 22: 5797–5802.
Heinrich, C., Nitta, N., Flubacher, A., Müller, M., Fahrner, A.,
Kirsch, M., Freiman, T., Suzuki, F., Depaulis, A., Frotscher,
M. and Haas, C.A. (2006) Reelin deficiency and displacement
of mature neurons, but not neurogenesis, underlie the for-
mation of granule cell dispersion in the epileptic hippocam-
pus. J. Neurosci., 26: 4701–4713.
Houser, C.R. (1990) Granule cell dispersion in the dentate
gyrus of humans with temporal lobe epilepsy. Brain Res.,
535: 195–204.
Kempermann, G., Gast, D., Kronenberg, G., Yamaguchi, M.
and Gage, F.H. (2003) Early determination and long-term
persistence of adult-generated new neurons in the hippocam-
pus of mice. Development, 130: 391–399.
Kempermann, G., Jessberger, S., Steiner, B. and Kronenberg,
G. (2004) Milestones of neuronal development in the adult
hippocampus. Trends Neurosci., 27: 447–452.
Kempermann, G., Kuhn, H.G. and Gage, F.H. (1997) More
hippocampal neurons in adult mice living in an enriched en-
vironment. Nature, 386: 493–495.
Kempermann, G., Kuhn, H.G. and Gage, F.H. (1998) Expe-
rience-induced neurogenesis in the senescent dentate gyrus. J.
Neurosci., 18: 3206–3212.
Koelliker, A.v. (1896) Handbuch der Gewebelehre des Mensc-
hen. Nervensystem des Menschen und der Thiere (6th Ed).
Engelmann, Leipzig, Germany.
Lie, D.C., Colamarino, S.A., Song, H.J., Desire, L., Mira, H.,
Consiglio, A., Lein, E.S., Jessberger, S., Lansford, H., Dear-
ie, A.R. and Gage, F.H. (2005) Wnt signalling regulates adult
hippocampal neurogenesis. Nature, 437: 1370–1375.
Linke, R. and Frotscher, M. (1993) Development of the rat
septohippocampal projection: tracing with DiI and electron
microscopy of identified growth cones. J. Comp. Neurol.,
332: 69–88.
Linke, R., Pabst, T. and Frotscher, M. (1995) Development
of the hippocamposeptal projection in the rat. J. Comp.
Neurol., 351: 602–616.
Markakis, E.A. and Gage, F.H. (1999) Adult-generated neu-
rons in the dentate gyrus send axonal projections to field CA3
and are surrounded by synaptic vesicles. J. Comp. Neurol.,
406: 449–460.
Nadarajah, B. and Parnavelas, J.G. (2002) Modes of neuronal
migration in the developing cerebral cortex. Nat. Rev.
Neurosci., 3: 423–432.
142
van Praag, H., Schinder, A.F., Christie, B.R., Toni, N., Palmer,
T.D. and Gage, F.H. (2002) Functional neurogenesis in the
adult hippocampus. Nature, 415: 1030–1034.
van Praag, H., Shubert, T., Zhao, C. and Gage, F.H. (2005)
Exercise enhances learning and hippocampal neurogenesis in
aged mice. J. Neurosci., 25: 8680–8685.
Ramón y Cajal, S. (1911) Histologie du Système Nerveux de
l’Homme et des Vertébrés, Vol. 2. Maloine, Paris.
Ribak, C.E., Seress, L. and Amaral, D.G. (1985) The develop-
ment, ultrastructure and synaptic connections of the mossy
cells of the dentate gyrus. J. Neurocytol., 14: 835–857.
Sanada, K., Gupta, A. and Tsai, L.H. (2004) Disabled-1-
regulated adhesion of migrating neurons to radial glial fiber
contributes to neuronal positioning during early corticogen-
esis. Neuron, 42: 197–211.
Schmidt-Hieber, C., Jonas, P. and Bischofberger, J. (2004)
Enhanced synaptic plasticity in newly generated granule cells
of the adult hippocampus. Nature, 429: 184–187.
Skutella, T. and Nitsch, R. (2001) New molecules for hippo-
campal development. Trends Neurosci., 24: 107–113.
Sloviter, R.S., Zappone, C.A., Harvey, B.D. and Frotscher,
M. (2006) Kainic acid-induced recurrent mossy fiber inner-
vation of dentate gyrus inhibitory interneurons: possi-
ble anatomical substrate of granule cell hyperinhibition
in chronically epileptic rats. J. Comp. Neurol., 494:
944–960.
Stanfield, B.B. and Cowan, W.M. (1979) The morphology of
the hippocampus and dentate gyrus in normal and reeler
mice. J. Comp. Neurol., 185: 393–422.
Steup, A., Lohrum, M., Hamscho, N., Savaskan, N.E., Ninne-
mann, O., Nitsch, R., Fujisawa, H., Püschel, A.W. and
Skutella, T. (2000) Sema3C and netrin-1 differentially affect
axon growth in the hippocampal formation. Mol. Cell
Neurosci., 15: 141–155.
Steward, O. and Scoville, S.A. (1976) Cells of origin of
entorhinal cortical afferents to the hippocampus and fascia
dentata of the rat. J. Comp. Neurol., 169: 347–370.
Super, H. and Soriano, E. (1994) The organization of the
embryonic and early postnatal murine hippocampus. II.
Development of entorhinal, commissural and septal connec-
tions studied with the lipophilic tracer DiI. J. Comp. Neurol.,
344: 101–120.
Tashiro, A., Sandler, V.M., Toni, N., Zhao, C. and Gage, F.H.
(2006) NMDA-receptor-mediated, cell-specific integration of
new neurons in adult dentate gyrus. Nature, 442: 929–933.
Tissir, F. and Goffinet, A.M. (2003) Reelin and brain develop-
ment. Nat. Rev. Neurosci., 4: 496–505.
Trommsdorff, M., Gotthardt, M., Hiesberger, T., Shelton, J.,
Stockinger, W., Nimpf, J., Hammer, R.E., Richardson, J.E.
and Herz, J. (1999) Reeler/disabled-like disruption of neu-
ronal migration in knockout mice lacking the VLDL receptor
and ApoE receptor 2. Cell, 97: 689–701.
Weiss, K.H., Johanssen, C., Tielsch, A., Herz, J., Deller, T.,
Frotscher, M. and Förster, E. (2003) Malformation of the
radial glial scaffold in the dentate gyrus of reeler mice,
scrambler mice, and ApoER2/VLDLR-deficient mice. J.
Comp. Neurol., 460: 56–65.
Zhao, C., Teng, E.M., Summers Jr., R.G.,, Ming, G.L. and
Gage, F.H. (2006) Distinct morphological stages of dentate
granule neuron maturation in the adult mouse hippocampus.
J. Neurosci., 26: 3–11.
Zhao, S., Chai, X., Förster, E. and Frotscher, M. (2004) Reelin
is a positional signal for the lamination of dentate granule
cells. Development, 131: 5117–5125.
Zhao, S., Förster, E., Chai, X. and Frotscher, M. (2003)
Different signals control laminar specificity of commissural
and entorhinal fibers to the dentate gyrus. J. Neurosci.,
23: 7351–7357.
Plate 7.1. Original drawing of the dentate gyrus and hippocampus proper by Camillo Golgi (Golgi, 1886). While Golgi was aware of
the laminated, bipolar arrangement of the granule cells, he did not correctly draw the course of the mossy fibers. As is known from
numerous more recent tracer studies, granule cell axons mainly run in stratum lucidum of CA3 and impinge on proximal dendritic
segments of pyramidal neurons. (For B/W version, see page 135 in the volume.)

Plate 7.2. Rescue of granule cell lamination in a slice culture of reeler dentate gyrus is achieved by coculturing with a wildtype culture
providing a reelin-containing marginal zone. Two reeler cultures (rl/1 and rl/2) are cocultured with a rat hippocampal slice. A
compact cell layer (arrow) has only formed in rl/1, which was cocultured next to the outer molecular layer of the rat dentate gyrus
containing reelin-synthesizing Cajal-Retzius cells. In rl/2, which was cultured next to the stratum oriens of CA1, the reeler-specific
loose distribution of neurons in the dentate gyrus is retained (arrowhead). Dashed lines represent borders between cultures. CA1, CA3,
hippocampal regions CA1 and CA3; DG, dentate gyrus; g, granule cell layer. Scale bar: 200 mm (from Zhao et al., 2004, with
permission). (For B/W version, see page 136 in the volume.)
Plate 7.3. Cocultures of two hippocampal slices to allow for the formation of ‘‘commissural’’ projections, traced by anterogradely
transported biocytin injected into the hilar region of one of the two hippocampal slices (sites of biocytin injections labeled by asterisks).
(A) Coculturing of two wildtype slices results in the formation of a compact ‘‘commissural’’ projection in the inner molecular layer
(black arrow). Open arrow labels ipsilateral mossy fiber projection in stratum lucidum of CA3. CA1, CA3, hippocampal regions CA1
and CA3; g, granule cell layer. (B) Coculture of wildtype (wt) hippocampus and reeler hippocampus. Biocytin-labeled ‘‘commissural’’
fibers from the wildtype culture terminate profusely in the reeler culture, thus reflecting the scattered distribution of their target
neurons, the granule cells. Open arrow labels the mossy fiber projection in the wildtype culture. DG, dentate gyrus of the reeler culture
with scattered granule cells; P, pyramidal layer in the wildtype culture; P1, P2, double pyramidal layer in the reeler culture.
(C) ‘‘Commissural’’ fibers from a reeler culture form a compact projection in the inner molecular layer of the wildtype culture (black
arrow), indicating that reeler commissural fibers project to the inner molecular layer as is normal, provided that their target cells, the
granule cells, form a tightly packed layer. g, granule cell layer. Scale bar: 100 mm (applies to A–C) (modified from Zhao et al., 2003,
with permission; copyright by the Society for Neuro-science). (For B/W version, see page 138 in the volume.)
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 8

Genetic regulation of dentate gyrus morphogenesis

Guangnan Li and Samuel J. Pleasure

Department of Neurology, Programs in Neuroscience, Developmental Biology, University of California at San Francisco,
1550 4th Street, Rock Hall Room 448c, San Francisco, CA 94158, USA

Abstract: The dentate gyrus is one of the small number of forebrain areas that have continued adult
neurogenesis. During development the dentate gyrus acquires the capacity for neurogenesis by generating a
new neurogenic stem cell niche at the border between the hilus and dentate granule cell layer. This is in
distinction to the other prominent zone of continued neurogenesis in the subventricular zone where neurons
are born in a structure directly descended from the mid-gestation subventricular zone. The ability to
generate this newly formed dentate neurogenic niche is controlled by the action of a number of genes during
prenatal and early postnatal development that regulate the fate, survival, migration, expansion, and
differentiation of the cellular components of the dentate neurogenic niche. In this review, we provide an
updated framework discussing the molecular steps and genes involved in these early stages of dentate gyrus
formation. We previously described a molecular framework for dentate gyrus morphogenesis that can be
associated with specific gene defects (Li, G., Pleasure, S.J. (2005). Dev. Neurosci., 27, 93–99), and here we
add additional recently described molecular players and discuss this framework.

Neurogenesis in the adult dentate gyrus requires a
specialized stem cell niche in the subgranular zone
at the boundary between the hilus and dentate
granule cell layer. It is now clear that this niche
includes quiescent multipotent stem cells, transit-
amplifying precursor cells, and immature neurons,
and that the transitions between these stages of
neuronal maturation are regulated by signaling
molecules from families of proteins also power-
fully involved in dentate development (Pozniak
and Pleasure, 2006). We believe that the develop-
ment of this neurogenic stem cell niche is first es-
tablished in prenatal and early postnatal life. In
this review we will first provide an overview of the
neuroanatomy of dentate development and then
turn to consideration of the phenotypes of some of
Corresponding author. Tel.: þ 1 (415)502-5683;
Fax: þ 1 (415)502-4335; E-mail: sam.pleasure@ucsf.edu
the most significant mouse mutants with defects in
dentate development.
The neuroanatomy of the developing dentate gyrus
Studies using classic neuroanatomic labeling meth-
ods revealed specialized aspects of dentate gyrus
morphogenesis related to the development of the
postnatal neurogenic niche (Nowakowski and
Rakic, 1979, 1981; Eckenhoff and Rakic, 1984;
Altman and Bayer, 1990a, b). The first granule
neurons are born and the dentate precursor pool
expands in a specialized region of subventricular
zone that first becomes apparent adjacent to the
dentate notch at mid-gestation. Quickly after that,
the earliest dentate granule cells are born, migrate
to the nascent dentate gyrus, and provide some of
the scaffolding for the developing dentate granule
cell layers. Contemporaneously, dividing precursor

DOI: 10.1016/S0079-6123(07)63008-8 143

144
cells destined to settle in the hilus and subgranular
zone migrate from the dentate notch in a subpial
route toward the dentate anlage (in developmental
terms the anlage is the site of an incipient structure).
En route and in the dentate anlage these cells con-
tinue to divide and produce additional granule
neurons locally. The proliferative matrix in the hilus
beneath the condensing dentate granule cell layers
was termed the tertiary matrix by (Altman and
Bayer, 1990a, b) and is the site of a massive pro-
duction of granule neurons in the first week of
postnatal life. During the first postnatal week, the
hilus develops a radial glial matrix with radially
oriented fibers traversing the dentate granule cell
layer with their cell bodies in the subgranular zone
and their endfeet attached to the pial surface. By
P10, the tertiary matrix has reorganized so that
most precursor cells are dividing in the subgranular
zone at the border between the condensing granule
cell layer and the now identifiable hilus. Granule
cells continue to be born in this specialized niche,
although in decreasing numbers, throughout adult-
hood. Our current growing understanding of the
molecular steps of dentate development (Li and
Pleasure, 2005) is founded on this strong neuroan-
atomic underpinning (see Fig. 1 for an overview of
this process and some of the molecules involved).
The cortical hem in initial development of the
dentate gyrus and hippocampus
The initial formation of the hippocampus and
dentate gyrus shares features with the development
of the cortex as a whole since the hippocampal
formation occupies the caudomedial — most area
of the cortex and there is no evidence for a strict
compartmental boundary between the hippocam-
pus and neocortex. Recent studies have shown that
gradients of transcription factor expression are
critically important in specifying the cortex and
driving development of four pallial compartments
(Wilson and Rubenstein, 2000; Sur and Ruben-
stein, 2005). The hippocampus is derived from the
most medial pallial domain, adjacent to the region
forming the neocortex. Numerous recent studies
demonstrate that the gradients of these transcrip-
tion factors are controlled by diffusible cues
released from localized signaling centers (Grove
et al., 1998; Fukuchi-Shimogori and Grove, 2001;
Garel et al., 2003; Shimogori et al., 2004; Storm
et al., 2006). The cortical hem is the most caudo-
medial of these signaling centers and occupies a
small zone of neuroepithelium adjacent to the re-
gion of neuroepithelium where the dentate gyrus
arises (Fig. 1). The cortical hem releases Wnt and
BMP family members and a variety of axon guid-
ance molecules that are candidates for regulating
the development of the dentate gyrus and hippo-
campus at early stages (Furuta et al., 1997; Grove
et al., 1998; Bagri et al., 2002). In addition, mice
with early deletion of the cortical hem had dra-
matic loss of cortical volume and no identifiable
hippocampus, implying that factors from the hem
are necessary for cortical expansion (Monuki
et al., 2001; Yoshida et al., 2006).
Interestingly, the cortical hem was also recently
shown to be the source of most Cajal-Retzius neu-
rons that are generated in a wave of neurogenesis
very early in cortical development and then migrate
tangentially to cover the cerebral cortex in the most
superficial cortical layer, called the marginal zone
(Takiguchi-Hayashi et al., 2004; Bielle et al., 2005;
Yoshida et al., 2006; Zhao et al., 2006). Thus, many
Cajal-Retzius cells, critical regulators of cortical
migration via the release of reelin, are actually the
earliest born neurons from the medial pallium and
are produced from the neuroepithelium immedi-
ately adjacent to the region of neuroepithelium that
will generate the dentate granule neurons. Recent
studies have now shown that the tangential migra-
tion of these neurons is controlled by SDF1 pro-
duced by the meninges. and this allows these cells to
maintain their superficial cortical position (Borrell
and Marin, 2006; Paredes et al., 2006). Interestingly
the same secreted chemokine (SDF-1) later controls
the migration of cells to the dentate gyrus (Bagri
et al., 2002; Lu et al., 2002).
Transcription factors pattern the cortex and
hippocampus
Mice with mutations in transcription factors nor-
mally expressed at their highest levels in the most
caudomedial cortex have been found to have
 A B C D

Lhx2/FoxG1/Emx2/Lef1
Hip

DG
Wnts

p73
Hem
Bmps

Lhx5
CP

granule cells Cajal-Retzius cells

Secreted Transcription precursors
hippocampus fissure
Factors Factors
meninges

dentate notch fimbria

E11.5 E13.5 E17.5 P4

Fig. 1. Schematic diagram showing the anatomic events and some of the genes involved in dentate morphogenesis. (A) At E11.5 the cortical hem (Hem) is adjacent to the
dentate neuroepithelium (DG) and the development of both are regulated by Wnts, while the hem and choroid plexus (CP) are also regulated by BMPs. The expression
domains for Wnts and BMPs are shown with arrows to the left of the neuroepithelium. The dentate and hippocampal neuroepithelia (Hip) are regulated by the function of
a number of transcription factors (Foxg1, Lhx2, Emx2, and Lef1) discussed in the text, while hem development is controlled by p73 and Lhx5. The expression domains for
transcription factors are shown to the right of the neuroepithelium and their extent indicated with arrows. (B) At E13.5 the initial migration of precursors is underway
from the dentate notch to the forming dentate gyrus, this migration may be targeted in part by SDF1 expressed in the meninges. By this time, Cajal-Retzius cells, derived
from the hem, are already present in the marginal zone and their position is maintained by SDF1 signaling. (C) By E17.5 precursor cells and granule cells have begun to
mix, forming the tertiary matrix in the nascent hilus and there is continued dentate precursor migration along the subpial migratory course. (D) By P4 the radial
reorganization of the dentate is largely accomplished with condensing of the granule cell layers and positioning of precursors in the subgranular zone.

145
146
major defects in hippocampal development. Emx2
is a homeobox gene expressed in a high caudome-
dial to low anterolateral gradient, and mice with
Emx2 mutations have reduced size of caudal cor-
tical structures like the hippocampus and visual
cortex with compensatory increases in anterior
cortical structures (Bishop et al., 2003; Muzio and
Mallamaci, 2003), while mice with overexpression
of Emx2 have the reciprocal phenotype, with ex-
pansion of caudal cortical structures at the expense
of anterior structures (Hamasaki et al., 2004).
Consistent with their loss of posterior structures,
Emx2 null mice also lack a clearly distinguishable
dentate gyrus (the dentate gyrus forms from al-
most the most embryologically caudal portion of
the cortical neuroepithelium) and have reductions
in the production of Cajal-Retzius cells from the
cortical hem (the most caudal region of cortical
neuroepithelium) (Pellegrini et al., 1996; Yoshida
et al., 1997; Mallamaci et al., 2000; Shinozaki
et al., 2002). Another interesting patterning mol-
ecule is the homeobox gene Lhx5, which in the
cortex is expressed solely in the cortical hem and
Cajal-Retzius cells derived from the hem. Mouse
mutants for Lhx5 lack hem development entirely
and have dramatic secondary hippocampal defects
due to the loss of hem-derived patterning signals
(Zhao et al., 1999), as well as defects in hippo-
campal dependent learning (Paylor et al., 2001). In
addition, the transcription factor p73, which is se-
lectively expressed in the cortical hem and Cajal-
Retzius cells derived from it also appear to have
major hippocampal and cortical proliferation de-
fects due at least in large part due to disruption of
normal cortical hem development, although this
phenotype is less severe than the Lhx5 phenotype
with the most severe defects being dramatic
hypoplasia of the lower blade of the dentate gyrus
(Yang et al., 2000; Meyer et al., 2004).
Loss of function mutations of transcription fac-
tors expressed in the hem lead to hem defects (and
secondary hippocampal defects, e.g. the Lhx5 mu-
tant mice), but in contrast, mutation of transcription
factors normally excluded from the cortical hem
may result in phenotypes that seem to be almost the
opposite. Foxg1 is a forkhead transcription factor
and Lhx2 a homeobox gene, both of which are nor-
mally excluded from the hem, and mutation of
either of these leads to cortical hem expansion
(Porter et al., 1997; Bulchand et al., 2001; Monuki et
al., 2001; Hanashima et al., 2004; Muzio and Malla-
maci, 2005; Zhao et al., 2006). Amazingly, in the
case of Foxg1, mutant mice have a massive expan-
sion of all medial cortical structures so that essen-
tially all residual cortical structures have molecular
markers consistent with a hippocampal origin and
there is no discernible neocortex at all (Muzio and
Mallamaci, 2005).
The role of hem-derived signals in hippocampal and
dentate development
The studies discussed above indicate the central
importance of the cortical hem in controlling hip-
pocampal and dentate development at the earliest
stages, but what are the factors that the hem
makes? In addition to producing Cajal-Retzius
cells, the cortical hem is a source of multiple se-
creted Wnt proteins. The Wnts are a large family
(19 ligands in mice) of secreted glycoprotein mol-
ecules that act potently during neural development
(Ciani and Salinas, 2005). The Wnt signaling path-
way is very complex and significant discussion of it
is beyond the scope of this review, but most of the
hippocampal phenotypes of Wnt signaling mu-
tants or Wnt ligands thus far have been most
clearly associated with defects in the so-called ‘‘ca-
nonical’’ Wnt signaling pathway. Both of the most
studied Wnt signaling pathways involve Wnt lig-
and binding to one of the cognate receptor types,
the Frizzled genes, but after this the two pathways
diverge substantially. The canonical pathway re-
quires an additional receptor from the LRP family
(LRP6) to stabilize beta-catenin, which is then
transported to the nucleus where it acts with the
family of Tcf/Lef transcription factors to drive al-
terations in gene expression. The non-canonical
pathway has some transcriptional output but is
also involved in local cytoskeletal rearrangements
and axon guidance decisions and also utilizes
newly recognized receptors from the Ryk family in
conjunction with Frizzled receptors.
During early cortical development the earliest
Wnt ligand expressed in the cortical hem is Wnt3a
(Grove et al., 1998) and Wnt3a mutants entirely

lack discernible hippocampal morphogenesis, most
prominently due to a failure to expand neural pre-
cursors in the entire medial pallium (Lee et al.,
2000). Mutants in one of the downstream tran-
scription factors, Lef1, which is expressed exclu-
sively in the medial pallium in the cortex, almost
completely lack dentate granule neurons (Galceran
et al., 2000; Zhou et al., 2004). Interestingly, mice
with mutations in the canonical signaling pathway
co-receptor LRP6, which are hypomorphic for
multiple Wnt signaling pathways during develop-
ment, have a very similar phenotype to that of the
Lef1 mutants (Zhou et al., 2004). The detailed
analysis of both the Lef1 and LRP6 mutant phe-
notypes is most consistent with a relatively specific
deficit in expansion of early dentate precursors
leading to a failure to populate the dentate with
stem cells, and also a related defect in the prenatal
radial glial scaffolding (Zhou et al., 2004).
In addition to the production of Wnts, the cor-
tical hem also produces members of the extended
BMP family, although they are produced in higher
quantities by the choroid plexus epithelium adjacent
to the hem (Furuta et al., 1997; Grove et al., 1998;
Shimogori et al., 2004). Previous studies using BMP
ligand mutants demonstrated that there are defects
in the development of the medial cortical wall, but
early studies did not fully evaluate cortical develop-
ment in these mutants (Furuta et al., 1997). These
studies were quite difficult though, because of the
large number of BMPs expressed at the cortical
midline. However, a recent analysis of conditional
mutants, lacking all BMP receptor signaling during
cortical development after E10, showed that BMP
signaling was particularly important in development
of the choroid plexus (Hebert et al., 2002). However,
previous studies showing that BMP signaling is im-
portant earlier in regulating Foxg1 expression
(Furuta et al., 1997; Monuki et al., 2001) do dem-
onstrate that earlier in cortical development as the
cortex is initially patterned, there is likely to be an
important role for BMPs. This may not have been
apparent in the analysis of the BMP receptor con-
ditional mice because the Cre-driver line used was
Foxg1 and this does not start expressing until the
telencephalon is already partially specified (Hebert
et al., 2002). The important early role of BMPs in
cortical patterning was confirmed in a recent study
147
demonstrating that Twisted Gastrulation mutants
(combined with BMP4) have no dorsomedial telen-
cephalic structures (Zakin and De Robertis, 2004).
Since Twisted Gastrulation is an important extra-
cellular mediator that enhances BMP function (thus
loss of Twisted Gastrulation would mimic loss of
function for BMP signaling), this establishes a crit-
ical early role for BMPs in the developing medial
pallium and hippocampus, whether other later roles
for the BMPs can be distinguished remains to be
seen.
Reforming the neurogenic niche in the dentate gyrus
Once the dentate neuroepithelium is specified by the
actions of the signals discussed above, the next step
is the process of moving a cohort of neurogenic
stem cells with continued capacity for expansion to
the developing dentate gyrus. Previous studies have
shown that dentate precursors, along with early-
born granule neurons, migrate in a subpial stream
at the boundary between the fimbria and the me-
ninges during mid-gestation (Nowakowski and
Rakic, 1979; Altman and Bayer, 1990b; Bagri
et al., 2002). Genes that control either migration
of these cells directly or the organization of the ra-
dial glial network in this region disrupt morpho-
genesis of the dentate. One recent example of a
relatively specific medial cortical radial glial defect
comes from conditional FGFR1 mutants. These
mice have diminished precursor proliferation in the
dentate and defects in the radial glial scaffolding
projecting to the dentate (Ohkubo et al., 2004).
Even more interesting are the results of a recent in
depth analysis of these mice showing a defect in
radial glial precursor somal translocation in the
entire medial cortex (Smith et al., 2006). Since one
of the main distinctions between dentate develop-
ment and that of the rest of the cortex is that stem
cell/precursors must relocate to the developing
dentate from their original location in the neuro-
epithelium, this implies that these mice may actually
have a failure to properly seed the dentate gyrus
with appropriate numbers of dentate precursors.
This is somewhat similar to the canonical Wnt
signaling dentate defects seen from loss of Lef1 and
LRP6, which have proliferative failure of dentate
148
precursors and defects in the radial glial network as
well, but in these Wnt mutants it is not possible to
exclude earlier patterning defects (Galceran et al.,
2000; Zhou et al., 2004). We also recently described
a different Wnt signaling mechanism leading to loss
of some dentate precursors in the Frizzled9 mutant
mice, which have increased apoptotic cell death of
putative dentate precursors as they migrate prena-
tally, leading to an adult defect in granule cell
numbers (Zhao et al., 2005). This is a quite different
phenotype than in the Lef1 and LRP6 mutants, and
we are continuing to address whether this pheno-
type is due to a defect in non-canonical signaling or
compensation by other Frizzleds that uncovers a
later developmental role for canonical Wnt signa-
ling in other aspects of dentate development.
The subpial migratory route used by the dentate
precursors and immature granule cells to reach the
forming dentate gyrus is also a region with very
strong expression of the chemokine ligand SDF1,
which may regulate migration of these cells (Bagri
et al., 2002; Lu et al., 2002). Indeed, we and others
demonstrated that SDF1 serves as a direct chemo-
attractant for cells migrating in the subpial migra-
tory route and that the dentate gyrus is
depopulated of precursors without SDF-1, either
because of a migratory deficit or because of a pro-
liferative failure of the precursors (Bagri et al.,
2002; Lu et al., 2002). Interestingly, mice with de-
fects in integrin receptors or signaling systems
downstream of integrins in the migrating cells (e.g.
focal adhesion kinase — Fak) also have defects in
the organization of the subpial migratory pathway
frequently resulting in major defects of the lower
blade of the dentate (Forster et al., 2002; Beggs
et al., 2003; Niewmierzycka et al., 2005).
Mutants with defects in dentate granule cell
differentiation
During dentate gyrus development, from the ear-
liest embryonic stages to adulthood, the final step
is cell type-specific differentiation into neurons of
the appropriate phenotype with the appropriate
specialized markers [in the case of granule cells the
most specific marker is the divergent homeobox
gene Prox1 — (Pleasure et al., 2000)]. Recent work
has shown that this complex process depends on
intersecting networks of interacting transcription
factors regulated by combinations of extracellular
signaling cues. It has been difficult thus far to de-
termine the specific roles of these signaling path-
ways in granule cell differentiation because of their
pleiotrophic effects during earlier development
and at different stages of dentate morphogenesis.
Mice with mutations in NeuroD, a basic helix-
loop-helix transcription factor, have quite selective
defects in the differentiation of dentate granule
neurons without affecting other hippocampal neu-
rons (Miyata et al., 1999; Liu et al., 2000). This
selectivity is likely to be due to the expression of
Math2 and NeuroD2 (two closely related family
members) in essentially all other cortical regions
(Pleasure et al., 2000). In NeuroD mutants there are
still small numbers of neurons produced that even-
tually express Prox1 at low levels, and these cells
still project axons toward CA3 (the appropriate
target) and express Calbindin (another marker of
mature granule neurons) (Liu et al., 2000). This in-
dicates that there are likely to be other regulatory
genes involved in specifying granule neuron fate
and differentiation in addition to NeuroD.
Regulation of dentate granule cell layer density
The normal granule cell layer of mammals is a dis-
crete, compact layer with few granule cells located
outside its narrow boundaries. Reelin mutants
(and mice with mutations in Reelin receptors —
VLDLR and ApoER3 — or downstream signaling
molecules — Disabled) have a dispersed granule cell
layer, and a major defect in proliferative capacity of
dentate progenitors, but do not appear to have a
primary migratory defect because the earliest gran-
ule cells find the dentate gyrus just as they would
under normal conditions (Forster et al., 2002;
Frotscher et al., 2003; Zhao et al., 2004). The ab-
normalities in the Reelin mutant may be due to di-
rect effects of Reelin on granule neurons and their
organization, but it is also likely that the mecha-
nisms are related to a failure to properly organize
the radially organized radial glial network in the
early postnatal dentate gyrus. This is because there
are also direct effects of Reelin on the radial glial
Fig. 2. NeuroD mutants have radial glial scaffolding defects in the postnatal dentate. We examined the organization of the GFAP+ radially oriented dentate precursors
usually residing in the subgranular zone in NeuroD mutant mice. In control mice, Prox1 clearly marks the dentate granule cell layer with the adjacent subgranular zone
that serves as the focus of radially oriented GFAP+ fibers. In contrast, in NeuroD mutant mice there is only a small cap of Prox1+ granule cells in the dentate of P21
mutants and there are essentially no GFAP+ processes present in the mutant dentate. (See Color Plate 8.2 in color plate section.)

149
150
cells themselves (Frotscher et al., 2003). Interest-
ingly, recent studies of intractable epilepsy patients
(Haas et al., 2002) and animal models of epilepsy
(Heinrich et al., 2006) showed that dentate granule
cell layer dispersion is associated with loss of Reelin-
expressing cells in the dentate marginal zone, im-
plying a crucial role for Reelin in maintaining dent-
ate architecture in adulthood. However, since there
was no correlation with severity or duration of ep-
ilepsy comparing this group to the other patients
with pharmacologically tractable or other forms of
temporal lobe epilepsy one can not rule out the role
of preexisting developmental anomalies. Thus, in
some subsets of epilepsy patients, granule cell dis-
persion in temporal lobe epilepsy could be due to a
neurodevelopmental defect that led to a loss of
Reelin, or due to abnormalities in Cajal-Retzius cell
genesis or distribution (Haas et al., 2002).
We recently also noted that the residual granule
cells in NeuroD mutant mice are diffusely distrib-
uted as compared to the residual granule cells found
in Lef1 mutants. This implies that the few granule
cells made when NeuroD is missing either have a
cell autonomous defect in compaction or that their
might be a non-autonomous defect similar to the
radial glial phenotype in Reelin mice. In other
words, newborn granule cells lacking NeuroD may
fail to properly organize the forming dentate radial
glial scaffolding and secondarily cause a granule cell
dispersion phenotype. This would presumably be
non-autonomous because radial glial cells do not
express NeuroD (whereas they do express Reelin
receptors) and might point to a role for newborn
granule cells in organizing their own lamination. If
this is true, we would predict that NeuroD mice
would display radial glial defects even though these
cells never express NeuroD. While it is difficult to
resolve this issue without more specifically ex-
pressed markers and conditional alleles, we have
found that NeuroD mice do have radial glial
scaffolding defects (Fig. 2).
Continued use of developmental signaling systems in
the adult dentate gyrus
Recent studies have begun to establish that cues
regulating dentate development continue to function
in adult neurogenesis (Pozniak and Pleasure, 2006).
Sonic Hedgehog (Shh) is a crucial ligand in the es-
tablishment of the dentate neurogenic niche during
development (Machold et al., 2003) and clearly acts
directly on migratory precursors in the subpial mi-
gration pathway (Ahn and Joyner, 2005). It has now
become apparent that Shh regulates the prolifera-
tion of rapidly dividing adult dentate precursors and
quiescent stem cells as well (Lai et al., 2003; Ahn
and Joyner, 2005). In addition, Wnts have now been
shown to regulate the proliferation of committed
dentate precursors and Wnt signaling is required for
adult neurogenesis (Lie et al., 2005). It is likely that
over the next few years roles will become established
for most, if not all, developmental modulators of
dentate development in adult neurogenesis as well.
Acknowledgments
The authors would like to acknowledge the other
members of the Pleasure Lab for helpful discus-
sions. This work was funded by K02 MH074958,
R01 MH066084 and P50 NS35902 to S.J.P.
References
Ahn, S. and Joyner, A.L. (2005) In vivo analysis of quiescent
adult neural stem cells responding to Sonic hedgehog. Na-
ture, 437: 894–897.
Altman, J. and Bayer, S.A. (1990a) Migration and distribution
of two populations of hippocampal granule cell precursors
during the perinatal and postnatal periods. J. Comp. Neurol.,
301: 365–381.
Altman, J. and Bayer, S.A. (1990b) Mosaic organization of the
hippocampal neuroepithelium and the multiple germinal
sources of dentate granule cells. J. Comp. Neurol., 301:
325–342.
Bagri, A., Gurney, T., He, X., Zou, Y.R., Littman, D.R.,
Tessier-Lavigne, M. and Pleasure, S.J. (2002) The chemokine
SDF1 regulates migration of dentate granule cells. Develop-
ment, 129: 4249–4260.
Beggs, H.E., Schahin-Reed, D., Zang, K., Goebbels, S., Nave,
K.A., Gorski, J., Jones, K.R., Sretavan, D. and Reichardt,
L.F. (2003) FAK deficiency in cells contributing to the basal
lamina results in cortical abnormalities resembling congenital
muscular dystrophies. Neuron, 40: 501–514.
Bielle, F., Griveau, A., Narboux-Neme, N., Vigneau, S., Sigrist,
M., Arber, S., Wassef, M. and Pierani, A. (2005) Multiple
origins of Cajal-Retzius cells at the borders of the developing
pallium. Nat. Neurosci., 8: 1002–1012.

Bishop, K.M., Garel, S., Nakagawa, Y., Rubenstein, J.L. and
O’leary, D.D. (2003) Emx1 and Emx2 cooperate to regulate
cortical size, lamination, neuronal differentiation, develop-
ment of cortical efferents, and thalamocortical pathfinding.
J. Comp. Neurol., 457: 345–360.
Borrell, V. and Marin, O. (2006) Meninges control tangential
migration of hem-derived Cajal-Retzius cells via CXCL12/
CXCR4 signaling. Nat. Neurosci., 9: 1284–1293.
Bulchand, S., Grove, E.A., Porter, F.D. and Tole, S. (2001)
LIM-homeodomain gene Lhx2 regulates the formation of the
cortical hem. Mech. Dev., 100: 165–175.
Ciani, L. and Salinas, P.C. (2005) WNTs in the vertebrate
nervous system: from patterning to neuronal connectivity.
Nat. Rev. Neurosci., 6: 351–362.
Eckenhoff, M.F. and Rakic, P. (1984) Radial organization of
the hippocampal dentate gyrus: a Golgi, ultrastructural, and
immunocytochemical analysis in the developing rhesus mon-
key. J. Comp. Neurol., 223: 1–21.
Forster, E., Tielsch, A., Saum, B., Weiss, K.H., Johanssen, C.,
Graus-Porta, D., Muller, U. and Frotscher, M. (2002) Reelin,
Disabled 1, and beta 1 integrins are required for the formation
of the radial glial scaffold in the hippocampus. Proc. Natl.
Acad. Sci. U.S.A., 99: 13178–13183.
Frotscher, M., Haas, C.A. and Forster, E. (2003) Reelin con-
trols granule cell migration in the dentate gyrus by acting on
the radial glial scaffold. Cereb. Cortex, 13: 634–640.
Fukuchi-Shimogori, T. and Grove, E.A. (2001) Neocortex pat-
terning by the secreted signaling molecule FGF8. Science,
294: 1071–1074.
Furuta, Y., Piston, D.W. and Hogan, B.L. (1997) Bone
morphogenetic proteins (BMPs) as regulators of dorsal fore-
brain development. Development, 124: 2203–2212.
Galceran, J., Miyashita-Lin, E.M., Devaney, E., Rubenstein,
J.L. and Grosschedl, R. (2000) Hippocampus development
and generation of dentate gyrus granule cells is regulated by
LEF1. Development, 127: 469–482.
Garel, S., Huffman, K.J. and Rubenstein, J.L. (2003) Molecular
regionalization of the neocortex is disrupted in Fgf8 hypo-
morphic mutants. Development, 130: 1903–1914.
Grove, E.A., Tole, S., Limon, J., Yip, L. and Ragsdale, C.W.
(1998) The hem of the embryonic cerebral cortex is defined by
the expression of multiple Wnt genes and is compromised in
Gli3-deficient mice. Development, 125: 2315–2325.
Haas, C.A., Dudeck, O., Kirsch, M., Huszka, C., Kann, G.,
Pollak, S., Zentner, J. and Frotscher, M. (2002) Role for
reelin in the development of granule cell dispersion in tem-
poral lobe epilepsy. J. Neurosci., 22: 5797–5802.
Hamasaki, T., Leingartner, A., Ringstedt, T. and O’leary, D.D.
(2004) EMX2 regulates sizes and positioning of the primary
sensory and motor areas in neocortex by direct specification
of cortical progenitors. Neuron, 43: 359–372.
Hanashima, C., Li, S.C., Shen, L., Lai, E. and Fishell, G. (2004)
Foxg1 suppresses early cortical cell fate. Science, 303:
56–59.
Hebert, J.M., Mishina, Y. and Mcconnell, S.K. (2002) BMP
signaling is required locally to pattern the dorsal telence-
phalic midline. Neuron, 35: 1029–1041.
151
Heinrich, C., Nitta, N., Flubacher, A., Muller, M., Fahrner, A.,
Kirsch, M., Freiman, T., Suzuki, F., Depaulis, A., Frotscher,
M. and Haas, C.A. (2006) Reelin deficiency and displacement
of mature neurons, but not neurogenesis, underlie the for-
mation of granule cell dispersion in the epileptic hippocam-
pus. J. Neurosci., 26: 4701–4713.
Lai, K., Kaspar, B.K., Gage, F.H. and Schaffer, D.V. (2003)
Sonic hedgehog regulates adult neural progenitor prolifera-
tion in vitro and in vivo. Nat. Neurosci., 6: 21–27.
Lee, S.M., Tole, S., Grove, E. and Mcmahon, A.P. (2000) A
local Wnt-3a signal is required for development of the mam-
malian hippocampus. Development, 127: 457–467.
Li, G. and Pleasure, S.J. (2005) Morphogenesis of the dentate
gyrus: what we are learning from mouse mutants. Dev. Ne-
urosci., 27: 93–99.
Lie, D.C., Colamarino, S.A., Song, H.J., Desire, L., Mira, H.,
Consiglio, A., Lein, E.S., Jessberger, S., Lansford, H., Dear-
ie, A.R. and Gage, F.H. (2005) Wnt signalling regulates adult
hippocampal neurogenesis. Nature, 437: 1370–1375.
Liu, M., Pleasure, S.J., Collins, A.E., Noebels, J.L., Naya, F.J.,
Tsai, M.J. and Lowenstein, D.H. (2000) Loss of BETA2/
NeuroD leads to malformation of the dentate gyrus and
epilepsy. Proc. Natl. Acad. Sci. U.S.A., 97: 865–870.
Lu, M., Grove, E.A. and Miller, R.J. (2002) Abnormal devel-
opment of the hippocampal dentate gyrus in mice lacking the
CXCR4 chemokine receptor. Proc. Natl. Acad. Sci. U.S.A.,
99: 7090–7095.
Machold, R., Hayashi, S., Rutlin, M., Muzumdar, M.D., Nery,
S., Corbin, J.G., Gritli-Linde, A., Dellovade, T., Porter, J.A.,
Rubin, L.L., Dudek, H., Mcmahon, A.P. and Fishell, G.
(2003) Sonic hedgehog is required for progenitor cell mainte-
nance in telencephalic stem cell niches. Neuron, 39: 937–950.
Mallamaci, A., Mercurio, S., Muzio, L., Cecchi, C., Pardini,
C.L., Gruss, P. and Boncinelli, E. (2000) The lack of Emx2
causes impairment of Reelin signaling and defects of neu-
ronal migration in the developing cerebral cortex. J. Neuro-
sci., 20: 1109–1118.
Meyer, G., Cabrera Socorro, A., Perez Garcia, C.G., Martinez
Millan, L., Walker, N. and Caput, D. (2004) Developmental
roles of p73 in Cajal-Retzius cells and cortical patterning.
J. Neurosci., 24: 9878–9887.
Miyata, T., Maeda, T. and Lee, J.E. (1999) NeuroD is required
for differentiation of the granule cells in the cerebellum and
hippocampus. Genes Dev., 13: 1647–1652.
Monuki, E.S., Porter, F.D. and Walsh, C.A. (2001) Patterning
of the dorsal telencephalon and cerebral cortex by a roof
plate-Lhx2 pathway. Neuron, 32: 591–604.
Muzio, L. and Mallamaci, A. (2003) Emx1, emx2 and pax6 in
specification, regionalization and arealization of the cerebral
cortex. Cereb. Cortex, 13: 641–647.
Muzio, L. and Mallamaci, A. (2005) Foxg1 confines Cajal-Re-
tzius neuronogenesis and hippocampal morphogenesis to the
dorsomedial pallium. J. Neurosci., 25: 4435–4441.
Niewmierzycka, A., Mills, J., St-Arnaud, R., Dedhar, S. and
Reichardt, L.F. (2005) Integrin-linked kinase deletion from
mouse cortex results in cortical lamination defects resembling
cobblestone lissencephaly. J. Neurosci., 25: 7022–7031.
152
Nowakowski, R.S. and Rakic, P. (1979) The mode of migration
of neurons to the hippocampus: a Golgi and electron micro-
scopic analysis in foetal rhesus monkey. J. Neurocytol., 8:
697–718.
Nowakowski, R.S. and Rakic, P. (1981) The site of origin and
route and rate of migration of neurons to the hippocampal
region of the rhesus monkey. J. Comp. Neurol., 196:
129–154.
Ohkubo, Y., Uchida, A.O., Shin, D., Partanen, J. and Vac-
carino, F.M. (2004) Fibroblast growth factor receptor 1 is
required for the proliferation of hippocampal progenitor cells
and for hippocampal growth in mouse. J. Neurosci., 24:
6057–6069.
Paredes, M.F., Li, G., Berger, O., Baraban, S.C. and Pleasure,
S.J. (2006) Stromal-derived factor-1 (CXCL12) regulates la-
minar position of Cajal-Retzius cells in normal and dysplastic
brains. J. Neurosci., 26: 9404–9412.
Paylor, R., Zhao, Y., Libbey, M., Westphal, H. and Crawley,
J.N. (2001) Learning impairments and motor dysfunctions in
adult Lhx5-deficient mice displaying hippocampal disorgan-
ization. Physiol. Behav., 73: 781–792.
Pellegrini, M., Mansouri, A., Simeone, A., Boncinelli, E. and
Gruss, P. (1996) Dentate gyrus formation requires Emx2.
Development, 122: 3893–3898.
Pleasure, S.J., Collins, A.E. and Lowenstein, D.H. (2000)
Unique expression patterns of cell fate molecules delineate
sequential stages of dentate gyrus development. J. Neurosci.,
20: 6095–6105.
Porter, F.D., Drago, J., Xu, Y., Cheema, S.S., Wassif, C.,
Huang, S.P., Lee, E., Grinberg, A., Massalas, J.S., Bodine,
D., Alt, F. and Westphal, H. (1997) Lhx2, a LIM homeobox
gene, is required for eye, forebrain, and definitive erythrocyte
development. Development, 124: 2935–2944.
Pozniak, C.D. and Pleasure, S.J. (2006) A tale of two signals:
Wnt and Hedgehog in dentate neurogenesis. Sci. STKE, 319:
pe5.
Shimogori, T., Banuchi, V., Ng, H.Y., Strauss, J.B. and Grove,
E.A. (2004) Embryonic signaling centers expressing BMP,
WNT and FGF proteins interact to pattern the cerebral cor-
tex. Development, 131: 5639–5647.
Shinozaki, K., Miyagi, T., Yoshida, M., Miyata, T., Ogawa,
M., Aizawa, S. and Suda, Y. (2002) Absence of Cajal-Retzius
cells and subplate neurons associated with defects of tangen-
tial cell migration from ganglionic eminence in Emx1/2 dou-
ble mutant cerebral cortex. Development, 129: 3479–3492.
Smith, K.M., Ohkubo, Y., Maragnoli, M.E., Rasin, M.R.,
Schwartz, M.L., Sestan, N. and Vaccarino, F.M. (2006)
Midline radial glia translocation and corpus callosum for-
mation require FGF signaling. Nat. Neurosci., 9: 787–797.
Storm, E.E., Garel, S., Borello, U., Hebert, J.M., Martinez, S.,
Mcconnell, S.K., Martin, G.R. and Rubenstein, J.L. (2006)
Dose-dependent functions of Fgf8 in regulating telencephalic
patterning centers. Development, 133: 1831–1844.
Sur, M. and Rubenstein, J.L. (2005) Patterning and plasticity of
the cerebral cortex. Science, 310: 805–810.
Takiguchi-Hayashi, K., Sekiguchi, M., Ashigaki, S., Taka-
matsu, M., Hasegawa, H., Suzuki-Migishima, R., Yoko-
yama, M., Nakanishi, S. and Tanabe, Y. (2004) Generation
of reelin-positive marginal zone cells from the caudomedial
wall of telencephalic vesicles. J. Neurosci., 24: 2286–2295.
Wilson, S.W. and Rubenstein, J.L. (2000) Induction and
dorsoventral patterning of the telencephalon. Neuron, 28:
641–651.
Yang, A., Walker, N., Bronson, R., Kaghad, M., Oosterwegel,
M., Bonnin, J., Vagner, C., Bonnet, H., Dikkes, P., Sharpe,
A., Mckeon, F. and Caput, D. (2000) p73-deficient mice have
neurological, pheromonal and inflammatory defects but lack
spontaneous tumours. Nature, 404: 99–103.
Yoshida, M., Assimacopoulos, S., Jones, K.R. and Grove, E.A.
(2006) Massive loss of Cajal-Retzius cells does not disrupt
neocortical layer order. Development, 133: 537–545.
Yoshida, M., Suda, Y., Matsuo, I., Miyamoto, N., Takeda, N.,
Kuratani, S. and Aizawa, S. (1997) Emx1 and Emx2 func-
tions in development of dorsal telencephalon. Development,
124: 101–111.
Zakin, L. and De Robertis, E.M. (2004) Inactivation of mouse
Twisted gastrulation reveals its role in promoting Bmp4 ac-
tivity during forebrain development. Development, 131:
413–424.
Zhao, C., Aviles, C., Abel, R.A., Almli, C.R., Mcquillen, P. and
Pleasure, S.J. (2005) Hippocampal and visuospatial learning
defects in mice with a deletion of frizzled 9, a gene in the
Williams syndrome deletion interval. Development, 132:
2917–2927.
Zhao, C., Guan, W. and Pleasure, S.J. (2006) A transgenic
marker mouse line labels Cajal-Retzius cells from the cortical
hem and thalamocortical axons. Brain Res., 1077: 48–53.
Zhao, S., Chai, X., Forster, E. and Frotscher, M. (2004) Reelin
is a positional signal for the lamination of dentate granule
cells. Development, 131: 5117–5125.
Zhao, Y., Sheng, H.Z., Amini, R., Grinberg, A., Lee, E.,
Huang, S., Taira, M. and Westphal, H. (1999) Control of
hippocampal morphogenesis and neuronal differentiation by
the LIM homeobox gene Lhx5. Science, 284: 1155–1158.
Zhou, C.J., Zhao, C. and Pleasure, S.J. (2004) Wnt signaling
mutants have decreased dentate granule cell production and
radial glial scaffolding abnormalities. J. Neurosci., 24:
121–126.
Plate 8.2. NeuroD mutants have radial glial scaffolding defects in the postnatal dentate. We examined the organization of the
GFAP+ radially oriented dentate precursors usually residing in the subgranular zone in NeuroD mutant mice. In control mice, Prox1
clearly marks the dentate granule cell layer with the adjacent subgranular zone that serves as the focus of radially oriented GFAP+
fibers. In contrast, in NeuroD mutant mice there is only a small cap of Prox1+ granule cells in the dentate of P21 mutants and there
are essentially no GFAP+ processes present in the mutant dentate. (For B/W version, see page 149 in the volume.)
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 9

Ultrastructure and synaptic connectivity of cell types

in the adult rat dentate gyrus

Charles E. Ribak and Lee A. Shapiro

Department of Anatomy & Neurobiology, University of California at Irvine, Irvine, CA 92697-1275, USA

Abstract: The rat hippocampal dentate gyrus is an extensively studied structural component of the limbic
system. It is the first station in the classical tri-synaptic circuit of the hippocampus in that its major input
arises from the entorhinal cortex via the perforant pathway. The second part of this circuit arises from the
projection cells of the dentate gyrus, the granule cells, which send their axons to the pyramidal cells of CA3.
Within the dentate gyrus, there also is an extensive inhibitory network of cells that are involved in syn-
chronizing the rhythmic firing of the granule cells. This chapter provides a review of the ultrastructural
features and synaptic connectivity of both projection cells and local circuit neurons in the dentate gyrus.

Keywords: granule cells; basket cells; mossy cells; chandelier cells; commissural cells; fusiform cells;
GABAergic interneurons

Granule cells: the principal cells of the dentate gyrus
The principal cells of the dentate gyrus are the
granule cells, and they are mainly found in the
granule cell layer, which is up to
7–8 cells thick
(Ramon y Cajal, 1911; Lorente de Nó, 1934;
Laatsch and Cowan, 1966). Granule cells are
reported to exist infrequently in the two other parts
of the dentate gyrus, the molecular layer and hilus
(Amaral, 1978; Gaarskjaer and Laurberg, 1983;
Marti-Subirana et al., 1986; Dashtipour et al.,
2001). It should be noted that the vast majority of
the cell bodies in the granule cell layer are granule
cells, but not all of the cell bodies are of this type
(Fig. 1). In the adult rat, granule cells (Fig. 1) dis-
play a round nucleus
10–12 mm in diameter, a thin
shell of perikaryal cytoplasm, and both symmetric
and asymmetric axosomatic synapses (Blackstad,
Corresponding author. Tel.: +1 949-824-5494 & +1 949-824-
4558; Fax: +1 949-824-8549; E-mail: ribak@uci.edu
1963; Seress and Ribak, 1985; Lubbers and Frotsc-
her, 1987). The nucleus is usually round or slightly
oval and has heterochromatin clumping against its
nuclear envelope. It is uncommon to observe any
infoldings of the nucleus of granule cells.
Processes arise from the somata of granule cells
and can be easily distinguished by the poles from
which they arise. The axons of the granule cells,
the well-described mossy fibers, arise from the hi-
lar pole of the granule cell body (Ramon y Cajal,
1911), although one report indicated that the axon
can arise from the apical dendrite (Yan et al.,
2001). At their origin, they display the typical fea-
tures of an axon initial segment with few organ-
elles, bundles of microtubules, a subaxolemmal
density and infrequent axon initial segment
synapses (Steward and Ribak, 1986). It is known
that mossy fibers arborize in the hilus and stratum
lucidum of CA3. In addition, mossy fibers can be
found either intragranular or supragranular. The
intragranular fibers are often observed orthogonal

DOI: 10.1016/S0079-6123(07)63009-X 155

156

NN

GC
D

GC GL
A

Fig. 1. An electron micrograph obtained from the part of the granule cell layer (GL) adjacent to the hilus (H) illustrates the
ultrastructural features of granule cells (GC), radial glial cells (A) and a doublecortin-immunolabeled newborn neuron (NN). The
mature granule cells (GC) located within the GL display a round or oval nucleus and a thin shell of perikaryal cytoplasm. The cell body
of the doublecortin-immunolabeled cell (NN) is much smaller than that of the granule cells (GC). Note that a doublecortin-labeled
apical process (white arrow) is adjacent to the radial process of an astrocyte (A), identified as a radial glial cell based on other
characteristics. Another astrocyte (A) with similar ultrastructural features is shown at the lower part of this image within the GL where
a proximal apical dendrite (D) from another mature granule cell apposes it. Many myelinated axons are present in the neuropil of the
hilus (H) and the lower part of the GL. Scale bar ¼ 2 mm.

to the granule cell layer where they form asym-
metric synapses with the apical dendrites and so-
mata of basket cells (Ribak and Peterson, 1991).
The supragranular fibers are located in the inner
molecular layer and are found to sprout there after
seizures, but are infrequent in the normal adult rat
(Nadler et al., 1980; Buckmaster et al., 2002).
These supragranular mossy fibers target the den-
drites of granule cells to contribute to recurrent
excitatory circuits after seizures. In addition, the
hilar basal dendrites observed following seizures
are also postsynaptic to mossy fibers (Ribak et al.,
2000). Thus, seizures can alter the morphology and
connectivity of granule cells.
The apical dendrites of granule cells extend ra-
dially through the granule cell layer (Fig. 1) and
then arborize in the molecular layer and their tips
eventually grow to reach the hippocampal fissure.
These dendrites arise as a thick process with or-
ganelles that are typical of the perikaryal

cytoplasm (Fig. 1), including cisternae of the gran-
ular endoplasmic reticulum and Golgi complex
(Blackstad, 1963; Laatsch and Cowan, 1966). The
apical dendrites also display spines that are more
common in the molecular layer, where perforant
pathway axons target their outer two-thirds por-
tion. In addition, there are several types of inter-
neurons that synapse on the apical dendrites of
granule cells in the molecular layer and the granule
cell layer. These interneurons will be described in
detail later in this chapter.
In light of the relatively recent discovery of
functional neurogenesis of granule cells in the
adult rat dentate gyrus, it is necessary to describe
the ultrastructural features of this population of
newborn neurons. The newborn granule cells have
been described using doublecortin (DCX)-
immunolabeled preparations and electron micros-
copy (Shapiro et al., 2005). These newborn granule
cells are initially found in the subgranular zone
with no dendritic processes and are cradled by an
astrocyte, considered to be a radial glial cell
(Shapiro et al., 2005). The diameter of these
DCX-labeled cells is only
6 mm, which is smaller
than that of the mature granule cells in the layer
(Fig. 1). Also, these newborn cells have only a thin
shell of perikaryal cytoplasm with few organelles.
As the DCX-labeled newborn neuron migrates to-
ward the granule cell layer, it extends an apical
dendritic process along the radial process of the
radial glial cells (Fig. 1). At this stage, the DCX-
labeled cell body has a larger diameter and a
thicker shell of perikaryal cytoplasm, relative to
the DCX-labeled cells that lack processes in the
subgranular zone. It should be noted that the
dendrites in the molecular layer are much thinner
than the apical dendrites of the mature granule
cells (Shapiro et al., 2007). Although basal den-
drites are not normally present on mature granule
cells in the adult rat dentate gyrus, they are a
transient feature of newborn granule cells in the
young and adult rat (Seress and Pokorny, 1981;
Ribak and Seress, 1990; Ribak et al., 2004), and
persist following seizures (Ribak et al., 2000). It is
pertinent to note that in human and non-human
primates, basal dendrites are a persistent feature of
some mature granule cells in the adult dentate
gyrus (Seress and Mrzljak, 1987).
157
Radial glial cells in the adult dentate gyrus: the
mothers of adult born granule cells
Of the many types of progenitor cells in the adult
rat dentate gyrus, the radial glial cells appear to
have an intimate relationship with the newborn
granule cells. These radial glial cells are immuno-
reactive for GFAP, are found in the hilus, sub-
granular zone, in the granule cell layer, at the
border between the granule cell layer and the mo-
lecular layer and in the molecular layer (Kosaka
and Hama, 1986). In this relationship, the radial
glial cell and its non-radial watery cytoplasmic
processes have been shown to encompass most of
the newborn granule cell at the border between the
subgranular zone and the granule cell layer (Fig.
1). These glial processes contain only sparse or-
ganelles and bundles of intermediate glial filaments
(Shapiro et al., 2005). The region of the newborn
neuron that is not surrounded by the radial glial
cell is shown to give rise to an apical process that
grows toward the granule cell layer. In this rela-
tionship, the radial process of the radial glial cell
provides a scaffold for the apical dendrite from the
newborn neuron to grow through the granule cell
layer and into the molecular layer (Shapiro et al.,
2005).
Basket cells: a plexus of inhibitory axons in the
granule cell layer
There are five distinct types of basket cells that
have been described in the rat dentate gyrus
(Ribak and Seress, 1983). These basket cells (Figs.
2 and 3) are located in the granule cell layer or the
hilus, within 50 mm of the base of the granule cell
layer and may express glutamate decarboxylase,
GABA, or the calcium-binding protein, parvalbu-
min (Ribak et al., 1978, 1990; Seress and Ribak,
1983; Lubbers and Frotscher, 1987; Nitsch et al.,
1990; Halasy and Somogyi, 1993). The shapes of
these cells can be pyramidal (Fig. 2), horizontal,
fusiform or multipolar. The ultrastructural fea-
tures of all five types of basket cells are very similar
(Ribak and Anderson, 1980; Ribak and Seress,
1983). They have a somal size of 20–30 mm with
large nuclei containing euchromatin and
158

B
G
G

Fig. 2. (A) shows the cell body and two proximal basal dendrites (arrows) of a pyramidal basket cell that was processed by the Golgi/
electron microscopy (EM) method and was previously identified at the light microscopic level. Note the gold particles subjacent to its
plasma membrane. Only a small part of its nucleus (N) is shown but its perikaryal cytoplasm is packed with many organelles. Note the
relatively larger cell body of the basket cell as compared to that of the neighboring granule cells (G). Cell bodies of two oligodendroglia
(O) flank this basket cell at the border between the granule cell layer and the hilus (H). (B) is another basket cell processed with the
Golgi/EM method to show the characteristic features of the nuclei of these cells. These features include nuclear infoldings (black
arrows), a prominent nucleolus and intranuclear filaments (arrowhead). Like the basket cell in (A), this cell also has gold particles
(white arrow) beneath its cell membrane while the surrounding granule cells (G) are not labeled. Scale bars ¼ 2 mm in A and 3 mm in B.
(Reprinted with permission from Ribak and Seress, 1983.)

intranuclear rods and sheets, and displaying ex-
tensive nuclear infoldings (Figs. 2 and 3). Their
thick shell of perikaryal cytoplasm has abundant
cisternae of granular endoplasmic reticulum,
which aggregate to form Nissl bodies, a well-de-
veloped Golgi complex, numerous mitochondria,
free ribosomes and lysosomes. Golgi-electron mi-
croscopic preparations revealed details about the
basket cell axons, showing that they have an ex-
tensive arborization concentrated mostly in the
granule cell layer, and to a lesser extent at the
border with the molecular layer. The axon termi-
nals of basket cells contain numerous mi-
tochondria and a sparse number of pleomorphic
synaptic vesicles. These terminals were shown to
form symmetric synapses, most commonly with
the somata and proximal apical dendrites of gran-
ule cells (Ribak and Seress, 1983). In rare instances
the basket cell axon terminal was observed to form
a synapse with a spine (Ribak and Seress, 1983). It
is important to emphasize that through these con-
nections, these basket cells provide feedback inhi-
bition to the granule cells of the dentate gyrus. In
contrast, basket cells have not been observed to
synapse on axon initial segments in adult rats,
however, examples of this were documented in
young (postnatal 10–16 days) rats (Seress and
Ribak, 1990).
The dendrites of basket cells are aspinous or
sparsely spinous and are found in all layers of the
dentate gyrus (Ribak and Seress, 1983). The asp-
inous dendrites have a mixture of asymmetric and
symmetric synapses on their surfaces. In the hilus,
the basal dendrites of pyramidal and fusiform
basket cells have the greatest concentration of
axon terminals apposed to their surfaces, and most
of these axon terminals contain round synaptic
vesicles and form asymmetric synapses. Some of
the axon terminals that synapse on the basal
dendrites of pyramidal basket cells were identified
as mossy fiber terminals (Ribak and Seress, 1983).
The basket cells receive afferents from at least two
other fiber systems besides the mossy fibers from
granule cells. They include the commissural axons
from the contralateral dentate gyrus that terminate
in the inner molecular layer and beneath the gran-
ule cell layer, and perforant path fibers that ter-
minate in the outer molecular layer (Seress and
159
Ribak, 1984; Zipp et al., 1989). In addition to
forming synapses on granule cell apical dendrites,
these latter afferents also form synapses onto
basket cell apical dendrites. These connections
were suggested to function as a feed-forward in-
hibitory circuit for granule cells (Seress and Ribak,
1984; Zipp et al., 1989; Kneisler and Dingledine,
1995a). Other studies have shown that CA3 and
the septum provide excitatory and inhibitory in-
put, respectively, to basket cells (Gulyas et al.,
1990; Scharfman, 1994; Kneisler and Dingledine,
1995b). Therefore, the GABAergic inhibitory bas-
ket cells mediate both feedback and feed-forward
inhibition in the dentate gyrus (Kneisler and
Dingledine, 1995a).
Mossy cells: the cells with great convergence of
mossy fibers
The mossy cells represent a distinct population of
neurons that dominate the landscape of the dentate
gyrus hilar region (Ribak et al., 1985; Frotscher
et al., 1991). The mossy cells are characterized by
large triangular or multipolar somata (25–30 mm di-
ameter). These somata contain a round nucleus
without infoldings, a thick shell of perikaryal cyto-
plasm and abundant organelles (Fig. 4). Arising
from these somata are three or four thick primary
dendrites (Fig. 4), which bifurcate to form an elab-
orate dendritic plexus that is mostly restricted to the
hilus. These cells have complex spines known as
thorny excrescences on their somata and proximal
dendrites (Fig. 4), while their distal dendrites have
typical spines. The majority of axon terminals form-
ing synapses onto these dendrites arise from mossy
fibers of granule cells (Fig. 4). Therefore, this mas-
sive granule cell input provides a large convergence
to mossy cells, and this input was shown to be ex-
citatory to mossy cells (Scharfman et al., 1990).
The axons of these mossy cells bifurcate and the
major projection is to the distant portions of the
inner molecular layer (Soltesz et al., 1993). The
axon terminals of mossy cells were observed to
make asymmetric synapses onto postsynaptic tar-
gets in the hilus and molecular layer of the dentate
gyrus (Ribak et al., 1985; Scharfman et al., 1990)
and showed immunoreactivity primarily for
160

Fig. 3. Illustrations of a pyramidal basket cell that was immunolabeled for parvalbumin. (A) shows a semi-thin section of this cell in
the lower part of the granule cell layer (GL) next to the hilus (H). Its apical dendrite branches in the inner molecular layer (ML). The
small punctate labeling in the GL probably represents the basket cell axons that outline the cell bodies of unlabeled granule cells. The
electron micrograph in (B) shows the soma of this parvalbumin-immunolabeled pyramidal basket cell with the typical features of this
cell type. These include nuclear infoldings (arrowheads), intranuclear rod (open arrow), axosomatic synapses (closed arrow) and a
thick perikaryal cytoplasm. Note the size difference between this soma and the somata of the adjacent granule cells (G). (C) is an
enlargement of the axosomatic symmetric synapses (arrows) denoted in (B) with an arrow. The axon terminals (T) forming these
synapses with the parvalbumin-immunolabeled basket cell body have pleomorphic vesicles and several mitochondria. (D) is another
example of axosomatic synapses with the same basket cell body but these are located at the transition with its apical dendrite and these
terminals (T) form asymmetric synapses (arrows). Note that this part of the soma displays more immunolabeling. Scale bar in A ¼
10 mm in A, 2.5 mm in B, 0.1 mm in C and D. (Reprinted with permission from Ribak et al., 1990.)
 161

Fig. 4. (A) shows the cell body and two thick proximal dendrites (D) of a mossy cell that was processed by the Golgi/EM method and
was previously identified at the light microscopic level. Most of its cell body is occupied by a large round nucleus (N). A large somal
spine (arrow) is found to the right of the cell body. Two dendrites (D) arise from the cell body. (B) shows a portion of a dendrite (D)
from this mossy cell. Note the numerous gold-labeled spines arising from this dendrite. Many of these spines are postsynaptic to large
axon terminals (M) that resemble mossy fiber tufts. (C) is an enlargement of a mossy fiber (M and outlined by dashed line) that forms a
synapse (arrow) with a gold-labeled dendrite of this mossy cell and with one of its gold-labeled spines (S). Scale bars ¼ 4 mm in A,
1.5 mm in B and 0.5 mm in C. (Reprinted with permission from Ribak et al., 1985.)

glutamate, but never for GABA (Soriano and
Frotscher, 1994). Within the hilus, glutamate-pos-
itive mossy cell axon terminals targeted GABA-
positive dendritic shafts of hilar interneurons
and GABA-negative dendritic spines. Mossy cell
axons in the inner molecular layer form asymmet-
ric synapses with dendritic spines associated
with GABA-negative, granule cell dendrites. Thus,
excitatory (glutamatergic) mossy cell termi-
nals contact GABAergic interneurons and non-
GABAergic neurons in the hilar region and
GABA-negative dendrites of granule cells in the
molecular layer. This pattern of connectivity is
consistent with the hypothesis that mossy cells
provide excitatory feedback to granule cells in a
dentate gyrus associational network and also ac-
tivate local hilar inhibitory elements (Wenzel et al.,
1997).
Fusiform cells (spiny and aspiny) in the hilus
The fusiform cells of the rat dentate gyrus are lo-
cated in the hilus, within 100 mm of the base of the
granule cell layer (Fig. 5). These cells are one of the
most prevalent cell types in the subgranular zone
of the dentate gyrus (Ribak and Seress, 1988). The
fusiform cells are bipolar, with oval cell bodies,
and their dendrites typically run parallel to the
granule cell layer, and within the subgranular zone
(Fig. 5). The dendrites of these cells are either
spiny or sparsely spiny. The spiny fusiform cell
receives the majority of its input from the axon
collaterals of the mossy fiber axons from the gran-
ule cells. Interestingly, the spines are not only ob-
served on the dendrites, but are also observed on
the soma (Fig. 5). The spiny fusiform cell has a
soma that is similar in appearance to the mossy
162

Fig. 5. Illustrations of two types of fusiform cells that were processed by the Golgi/EM method and were previously identified at the
light microscopic level. (A) is an electron micrograph of a sparsely spiny fusiform cell body and one of its proximal dendrites (D). The
nucleus (N) displays a prominent nucleolus and numerous infoldings (large arrows). The perikaryal cytoplasm contains many cisternae
of granular endoplasmic reticulum (ER) that are organized into stacks or Nissl bodies. A few small somal spines (small arrows) are also
shown. (B) is a light micrograph of the same sparsely spiny fusiform cell shown in (A) to illustrate its flattened soma (large arrow) and
its dendrites that run parallel to the granule cell layer (small arrows). (C) shows a spiny fusiform cell body with many gold-labeled
somal spines (arrows). The nucleus (N) in this section occupies less area than would the nucleus obtained from a section through the
center of the soma. The cisternae of the granular endoplasmic reticulum (ER) do not form parallel stacks. Electron-dense lysosomes as
well as mitochondria are randomly distributed in the perikaryal cytoplasm. A small capillary (C) is also shown in the upper left corner
of the micrograph. (D) is an enlargement of the two somal spines (S) indicated by the top arrow in (C). Large axon terminals packed
with agranular round vesicles form asymmetric synapses (large arrows) with the gold-labeled spines. In contrast, the somal surface is
contacted by a small terminal that appears to form a symmetric synapse (small arrow). Scale bars ¼ 20 mm in (A), 2 mm in (B), 2 mm in
(C) and 0.5 mm in (D). (Reprinted with permission from Ribak and Seress, 1988.)

cell because it displays Nissl bodies and little or no
nuclear infolding (Fig. 5). In contrast, the fusiform
cells with sparsely spiny dendrites have somata
with nuclear infoldings (Fig. 5) like that of basket
cells and few Nissl bodies. It should also be noted
that this latter cell type displays a variety of
axodendritic synapses, suggesting a more diverse
synaptic input than that of the spiny fusiform cell.
Several studies have demonstrated the presence
of somatostatin immunolabeling in hilar cells with
the morphology of fusiform cells (Bakst et al.,
1986; Milner and Bacon, 1989; Leranth et al.,

1990). These studies provided a valuable insight
into the axonal projections of the fusiform cell. It
was shown that the somatostatin-labeled axons
formed a dense plexus in the outer molecular layer
where they made symmetric synapses onto the
granule cell dendrites (Bakst et al., 1986; Milner
and Bacon, 1989; Leranth et al., 1990). In addi-
tion, somatostatin-labeled axon terminals in the
hilus were also reported (Milner and Bacon, 1989).
It should be noted that more than 90% of the hilar
neurons expressing somatostatin are GABAergic
(Esclapez and Houser, 1995) and that
50% of
them express neuropeptide Y (Deller and Leranth,
1990).
Chandelier cells: a GABAergic cell-type targeting
axon initial segments of granule cells
Chandelier cells are observed throughout the
cerebral cortex, but were first described in the
hippocampus by Kosaka (1980) and Somogyi et al.
(1983a, b), and later in the dentate gyrus by
Soriano and Frotscher (1989). All chandelier cells
share two features that are unique among inhib-
itory interneurons; their axons form rows of
boutons and they make symmetric synapses ex-
clusively on axon initial segments.
The chandelier cells of the dentate gyrus have
their somata located within or immediately adja-
cent to the granule cell layer. The cell bodies and
dendrites of these neurons exhibit several of the
characteristic ultrastructural features of non-gran-
ule cells, including a large perikaryal cytoplasm,
nuclear infoldings, intranuclear inclusions and a
large number of synapses on the soma and aspiny
dendrites. Within the dentate gyrus, chandelier cell
axons densely innervate the granule cell layer with
radially oriented terminal rows, and also form an
extensive plexus in the hilus. The dendrites of the
chandelier cells extend radially through the mo-
lecular layer and can reach as far as the hippo-
campal fissure (Soriano and Frotscher, 1989;
Soriano et al., 1990; Han et al., 1993; Buhl et al.,
1994). Occasionally, basal dendrites from chande-
lier cells cross the granule cell layer toward the
hilus (Soriano et al., 1990). There are also three
classes of chandelier cells that have their somata in
163
the hilus. The three chandelier cell types observed
in the hilus projected to granule cells at the same
septo–temporal level where their cell bodies were
located. It should also be noted that the chandelier
cells exhibit immunolabeling for glutamate de-
carboxylase, GABA-transporter-1 and parvalbu-
min (Ribak et al., 1990, 1996; Soriano et al., 1990).
Thus, the dentate gyrus chandelier cell provides a
spatially selective innervation of granule cells and
complements that of the basket cell that provides
GABAergic inhibition to cell bodies and proximal
dendrites.
Commissural neurons: communication and
synchronization across hemispheres involving
excitatory and inhibitory neurons
Within the dentate gyrus, there are at least two
classes of cells that send projections to the con-
tralateral dentate gyrus. Using retrograde labeling
with horseradish peroxidase (HRP), Seroogy et al.
(1983) showed different electron microscopic fea-
tures for two classes of labeled commissural neu-
ron. The first type consisted of cells with somata
that exhibited round or oval nuclei with no intra-
nuclear inclusions and had exclusively symmetric
synapses on their somata. The main dendrites of
those neurons were thick and tapering. This type
had features that resembled the morphology of the
mossy cell. A subsequent study using combined
retrograde transport of HRP with Golgi/electron
microscopy (EM) confirmed this suggestion
(Frotscher, 1992). The second type of labeled neu-
ronal soma had infolded nuclei containing intra-
nuclear rods or sheets, displayed both symmetric
and asymmetric synapses on its soma and had
dendrites that were less thick and generally asp-
inous. This type of commissurally labeled cell had
features that were similar to the dentate gyrus
basket cell, a local circuit neuron associated with
GABAergic inhibition, as described above. An-
other line of evidence supported the possibility
that GABAergic neurons had commissural pro-
jections. In a quantitative study of GABAergic
neurons in the hilus of the dentate gyrus, Seress
and Ribak (1983) showed that 60% of the hilar
neurons are GAD-positive. Because previous
164
studies indicated that 80% of hilar neurons give
rise to associational and commissural pathways,
many GABAergic neurons in the hilus were sug-
gested to be projection neurons, and a subsequent
combined tracer and immunofluorescence study
showed several double-labeled GABAergic neu-
rons in the hilus contralateral to the injection site
(Ribak et al., 1986). Another confirmation of this
conclusion came with the elegant intracellular
labeling study of Han et al. (1993). In that study,
the axons of several classes of GABAergic inter-
neurons were mapped in the rat dentate gyrus, in-
cluding one that had its axon terminals distributed
in the commissural and associational pathway ter-
mination field. Because they named the cell using
both its cell body location and axon terminal field,
and the cell body of these neurons resided in the
hilus, this cell type was referred to as the hilus/
commissural-associational pathways (HICAP)
cell (Han et al., 1993). Therefore, the commissu-
ral projection in the rat includes both excitatory
(mossy cells) and inhibitory (HICAP cells)
neurons.
Other GABAergic interneuron types
Two other GABAergic neuron types were observed
in the dentate gyrus and were described simulta-
neously with the HICAP cell (Han et al., 1993).
One of these is another type of hilar cell that has an
axon, which ramifies in the perforant path terminal
field in the outer two-thirds of the molecular layer.
This hilar cell with axon terminals distributed in
the perforant path termination field was named the
hilus/perforant pathway (HIPP) cell. Electron mi-
croscopy of this cell type revealed sparsely spiny
dendrites that were covered with many synaptic
boutons on both their shafts and their spines but
no details were provided about the HIPP cell’s
somal features (Han et al., 1993). It should be
noted that this cell type has its axon terminal field
in the same location as the somatostatin-labeled
cell described above in the fusiform cell section.
That cell type was also GABAergic (Esclapez and
Houser, 1995).
The other GABAergic cell type described by
Han et al. (1993) had its cell body in the molecular
layer and its dendritic and axonal domains were
confined to the perforant path terminal zone. This
cell type was referred to as the molecular layer/
perforant pathway (MOPP) cell and its axon made
symmetric synapses exclusively onto dendritic
shafts, 60% of which were shown to emit spines
(Halasy and Somogyi, 1993). The smooth dend-
rites of the MOPP cell were also restricted to the
outer two-thirds of the molecular layer, where they
received both GABA-negative and GABA-positive
synaptic inputs. Similar to the HIPP cells, ultra-
structural details about the MOPP cell’s soma
were not included in the description.
Conclusion
There have been many studies on the cell types in
the dentate gyrus and their synaptic connections.
Because the dentate gyrus is the first station in
the classical tri-synaptic circuit of the hippocam-
pus, understanding its anatomical organization is
essential to understand the function of this struc-
ture. Here we have outlined the ultrastructural
features and synaptic connections of the principal
cell type (the granule cells) and the projection and
local circuit neurons that make up the dentate
gyrus.
Acknowledgments
The authors are grateful to the following collab-
orators who contributed significantly to one or
several studies in this review: Drs. David Amaral,
Nich Brecha, Khashayar Dashtipour, James Fal-
lon, Michael Frotscher, Mathew Korn, J. Victor
Nadler, Robert Nitsch, Andre Obenaus, Maxine
Okazaki, Gary Peterson, Kihachi Saito, Larry
Schmued, Laszlo Seress, Kim Seroogy, Igor Spi-
gelman, Oswald Steward, Winnie Tong, James
Vaughn and Xiao-Xin Yan. This review is dedi-
cated to the mentor of C.E.R., Alan Peters, who
perfected the Golgi/EM method for the analysis
of short axonal connections. We acknowledge
the support from NIH grant R01-NS38331 (to
C.E.R.) and NIH training grant T32-NS45540 (for
L.A.S.).

References
Amaral, D.G. (1978) A Golgi study of cell types in the hilar
region of the hippocampus in the rat. J. Comp. Neurol., 195:
851–914.
Bakst, I., Avendano, C., Morrison, J.H. and Amaral, D.G.
(1986) An experimental analysis of the origins of somatosta-
tin-like immunoreactivity in the dentate gyrus of the rat. J.
Neurosci., 6: 1452–1462.
Blackstad, T. (1963) Ultrastructural studies on the hippocampal
region. Prog. Brain Res., 3: 122–148.
Buckmaster, P.S., Zhang, G.F. and Yamawaki, R. (2002) Axon
sprouting in a model of temporal lobe epilepsy creates a pre-
dominantly excitatory feedback circuit. J. Neurosci., 22:
6650–6658.
Buhl, E.H., Han, Z.S., Lorinczi, Z., Stezhka, V.V., Karnup,
S.V. and Somogyi, P. (1994) Physiological properties of an-
atomically identified axo-axonic cells in the rat hippocampus.
J. Neurophysiol., 71: 1289–1307.
Dashtipour, K., Tran, P.H., Okazaki, M.M., Nadler, J.V. and
Ribak, C.E. (2001) Ultrastructural features and synaptic
connections of hilar ectopic granule cells in the rat dentate
gyrus are different from those of granule cells in the granule
cell layer. Brain Res., 890: 261–271.
Deller, T. and Leranth, C. (1990) Synaptic connections of
neuropeptide Y (NPY) immunoreactive neurons in the hilar
area of the rat hippocampus. J. Comp. Neurol., 300:
433–447.
Esclapez, M. and Houser, C.R. (1995) Somatostatin neurons
are a subpopulation of GABA neurons in the rat dentate
gyrus: evidence from colocalization of pre-prosomatostatin
and glutamate decarboxylase messenger RNAs. Neurosci-
ence, 64: 339–355.
Frotscher, M. (1992) Application of the Golgi/electron micros-
copy technique for cell identification in immunocytochemi-
cal, retrograde labeling, and developmental studies of
hippocampal neurons. Microsc. Res. Tech., 23: 306–323.
Frotscher, M., Seress, L., Schwerdtfeger, W.K. and Buhl, E.
(1991) The mossy cells of the fascia dentata: a comparative
study of their fine structure and synaptic connections in ro-
dents and primates. J. Comp. Neurol., 312: 145–163.
Gaarskjaer, F.B. and Laurberg, S. (1983) Ectopic granule cells
of hilus fasciae dentatae projecting to the ipsilateral regio
inferior of the rat hippocampus. Brain Res., 274: 11–16.
Gulyas, A.I., Gorcs, T.J. and Freund, T.F. (1990) Innervation
of different peptide-containing neurons in the hippocampus
by GABAergic septal afferents. Neuroscience, 37: 31–44.
Halasy, K. and Somogyi, P. (1993) Subdivisions in the multiple
GABAergic innervation of granule cells in the dentate gyrus
of the rat hippocampus. Eur. J. Neurosci., 5: 411–429.
Han, Z.S., Buhl, E.H., Lorinczi, Z. and Somogyi, P. (1993) A
high degree of spatial selectivity in the axonal and dendritic
domains of physiologically identified local-circuit neurons in
the dentate gyrus of the rat hippocampus. Eur. J. Neurosci.,
5: 395–410.
Kneisler, T.B. and Dingledine, R. (1995a) Spontaneous and
synaptic input from granule cells and the perforant path to
165
dentate basket cells in the rat hippocampus. Hippocampus, 5:
151–164.
Kneisler, T.B. and Dingledine, R. (1995b) Synaptic input from
CA3 pyramidal cells to dentate basket cells in rat hippocam-
pus. J. Physiol., 487: 125–146.
Kosaka, T. (1980) The axon initial segment as a synaptic site:
ultrastructure and synaptology of the initial segment of the
pyramidal cell in the rat hippocampus (CA3 region). J. Ne-
urocytol., 9: 861–882.
Kosaka, T. and Hama, K. (1986) Three-dimensional structure
of astrocytes in the rat dentate gyrus. J. Comp. Neurol., 249:
242–260.
Laatsch, R.H. and Cowan, W.M. (1966) Electron microscopic
studies of the dentate gyrus of the rat. I. Normal structure
with special reference to synaptic organization. J. Comp.
Neurol., 128: 359–396.
Leranth, C., Malcolm, A.J. and Frotscher, M. (1990) Afferent
and efferent synaptic connections of somatostatin-immuno-
reactive neurons in the rat fascia dentata. J. Comp. Neurol.,
295: 111–122.
Lorente de Nó, R. (1934) Studies on the structure of the
cerebral cortex. II. Continuation of the study of the ammonic
system. J. Psychol. Neurol. Lpz., 46: 113–177.
Lubbers, K. and Frotscher, M. (1987) Fine structure and
synaptic connections of identified neurons in the rat fascia
dentata. Anat. Embryol. (Berl.), 177: 1–14.
Marti-Subirana, A., Soriano, E. and Garcia-Verdugo, J.M.
(1986) Morphological aspects of the ectopic granule-like cel-
lular populations in the albino rat hippocampal formation: a
Golgi study. J. Anat., 144: 31–47.
Milner, T.A. and Bacon, C.E. (1989) Ultrastructural localiza-
tion of somatostatin-like immunoreactivity in the rat dentate
gyrus. J. Comp. Neurol., 290: 544–560.
Nadler, J.V., Perry, B.W. and Cotman, C.W. (1980) Selective
reinnervation of hippocampal area CA1 and the fascia den-
tata after destruction of CA3-CA4 afferents with kainic acid.
Brain Res., 182: 1–9.
Nitsch, R., Soriano, E. and Frotscher, M. (1990) The parval-
bumin-containing nonpyramidal neurons in the rat hippo-
campus. Anat. Embryol. (Berl.), 181: 413–425.
Ramon y Cajal, S. (1911) Histologie du Systeme Nerveux de
l’Homme et des Vertebres, Vol. 2. Maloine, Paris.
Ribak, C.E. and Anderson, L. (1980) Ultrastructure of the py-
ramidal basket cells in the dentate gyrus of the rat. J. Comp.
Neurol., 192: 903–916.
Ribak, C.E., Korn, M.J., Shan, Z. and Obenaus, A. (2004)
Dendritic growth cones and recurrent basal dendrites are
typical features of newly generated dentate granule cells in
the adult hippocampus. Brain Res., 1000: 195–199.
Ribak, C.E., Nitsch, R. and Seress, L. (1990) Proportion of
parvalbumin-positive basket cells in the GABAergic inner-
vation of pyramidal and granule cells of the rat hippocampal
formation. J. Comp. Neurol., 300: 449–461.
Ribak, C.E. and Peterson, G.M. (1991) Intragranular mossy
fibers in rats and gerbils form synapses with the somata and
proximal dendrites of basket cells in the dentate gyrus.
Hippocampus, 1: 355–364.
166
Ribak, C.E. and Seress, L. (1983) Five types of basket cell in the
hippocampal dentate gyrus: a combined Golgi and electron
microscopic study. J. Neurocytol., 12: 577–597.
Ribak, C.E. and Seress, L. (1988) A Golgi-electron microscopic
study of fusiform neurons in the hilar region of the dentate
gyrus. J. Comp. Neurol., 271: 67–78.
Ribak, C.E. and Seress, L. (1990) Postnatal development of the
light and electron microscopic features of basket cells in the
hippocampal dentate gyrus of the rat. Anat. Embryol. (Berl.),
181: 547–565.
Ribak, C.E., Seress, L. and Amaral, D.G. (1985) The develop-
ment, ultrastructure and synaptic connections of the mossy
cells of the dentate gyrus. J. Neurocytol., 14: 835–857.
Ribak, C.E., Seress, L., Peterson, G.M., Seroogy, K.B., Fallon,
J.H. and Schmued, L.C. (1986) A GABAergic inhibitory
component within the hippocampal commissural pathway. J.
Neurosci., 6: 3492–3498.
Ribak, C.E., Tong, W.M. and Brecha, N.C. (1996) GABA
plasma membrane transporters, GAT-1 and GAT-3, display
different distributions in the rat hippocampus. J. Comp.
Neurol., 367: 595–606.
Ribak, C.E., Tran, P.H., Spigelman, I., Okazaki, M.M. and
Nadler, J.V. (2000) Status epilepticus-induced hilar basal
dendrites on rodent granule cells contribute to recurrent ex-
citatory circuitry. J. Comp. Neurol., 428: 240–253.
Ribak, C.E., Vaughn, J.E. and Saito, K. (1978) Immunocyto-
chemical localization of glutamic acid decarboxylase in neu-
ronal somata following colchicine inhibition of axonal
transport. Brain Res., 140: 315–332.
Scharfman, H.E. (1994) Synchronization of area CA3 hippo-
campal pyramidal cells and non-granule cells of the dentate
gyrus in bicuculline-treated rat hippocampal slices. Neuro-
science, 59: 245–257.
Scharfman, H.E., Kunkel, D.D. and Schwartzkroin, P.A.
(1990) Synaptic connections of dentate granule cells and hi-
lar neurons: results of paired intracellular recordings and in-
tracellular horseradish peroxidase injections. Neuroscience,
37: 693–707.
Seress, L. and Mrzljak, L. (1987) Basal dendrites of granule
cells are normal features of the fetal and adult dentate gyrus
of both monkey and human hippocampal formations. Brain
Res., 405: 169–174.
Seress, L. and Pokorny, J. (1981) Structure of the granular layer
of the rat dentate gyrus. J. Anat., 133: 181–195.
Seress, L. and Ribak, C.E. (1983) GABAergic cells in the dent-
ate gyrus appear to be local circuit and projection neurons.
Exp. Brain Res., 50: 173–182.
Seress, L. and Ribak, C.E. (1984) Direct commissural connec-
tions to the basket cells of the hippocampal dentate gyrus:
anatomical evidence for feed-forward inhibition. J. Neuro-
cytol., 13: 215–225.
Seress, L. and Ribak, C.E. (1985) A substantial number of
asymmetric axosomatic synapses is a characteristic of the
granule cells of the hippocampal dentate gyrus. Neurosci.
Lett., 56: 21–26.
Seress, L. and Ribak, C.E. (1990) The synaptic connections of
basket cell axons in the developing rat hippocampal forma-
tion. Exp. Brain Res., 81: 500–508.
Seroogy, K.B., Seress, L. and Ribak, C.E. (1983) Ultrastructure
of commissural neurons of the hilar region in the hippocam-
pal dentate gyrus. Exp. Neurol., 82: 594–608.
Shapiro, L.A., Korn, M.J., Shan, Z. and Ribak, C.E. (2005)
GFAP-expressing radial glia-like cell bodies are involved in a
one-to-one relationship with doublecortin-immunolabeled
newborn neurons in the adult dentate gyrus. Brain Res.,
1040: 81–91.
Shapiro, L.A., Upadhyaya, P. and Ribak, C.E. (2007) Spatio-
temporal profile of dendritic outgrowth from newly born
granule cells in the adult rat dentate gyrus. Brain Res., 1149:
30–37.
Soltesz, I., Bourassa, J. and Deschenes, M. (1993) The behavior
of mossy cells of the rat dentate gyrus during theta oscilla-
tions in vivo. Neuroscience, 57: 555–564.
Somogyi, P., Nunzi, M.G., Gorio, A. and Smith, A.D. (1983a)
A new type of specific interneuron in the monkey hippocam-
pus forming synapses exclusively with the axon initial seg-
ments of pyramidal cells. Brain Res., 259: 137–142.
Somogyi, P., Smith, A.D., Nunzi, M.G., Gorio, A., Takagi, H.
and Wu, J.Y. (1983b) Glutamate decarboxylase immunore-
activity in the hippocampus of the cat: distribution of
immunoreactive synaptic terminals with special reference to
the axon initial segment of pyramidal neurons. J. Neurosci.,
3: 1450–1468.
Soriano, E. and Frotscher, M. (1989) A GABAergic axo-axonic
cell in the fascia dentata controls the main excitatory hippo-
campal pathway. Brain Res., 503: 170–174.
Soriano, E. and Frotscher, M. (1994) Mossy cells of the rat
fascia dentata are glutamate-immunoreactive. Hippocampus,
4: 65–69.
Soriano, E., Nitsch, R. and Frotscher, M. (1990) Axo-axonic
chandelier cells in the rat fascia dentata: Golgi-electron mi-
croscopy and immunocytochemical studies. J. Comp.
Neurol., 293: 1–25.
Steward, O. and Ribak, C.E. (1986) Polyribosomes associated
with synaptic specializations on axon initial segments: local-
ization of protein-synthetic machinery at inhibitory synapses.
J. Neurosci., 6: 3079–3085.
Wenzel, H.J., Buckmaster, P.S., Anderson, N.L., Wenzel, M.E.
and Schwartzkroin, P.A. (1997) Ultrastructural localization
of neurotransmitter immunoreactivity in mossy cell axons
and their synaptic targets in the rat dentate gyrus. Hippo-
campus, 7: 559–570.
Yan, X.X., Spigelman, I., Tran, P.H. and Ribak, C.E. (2001)
Atypical features of rat dentate granule cells: recurrent basal
dendrites and apical axons. Anat. Embryol. (Berl.), 203:
203–209.
Zipp, F., Nitsch, R., Soriano, E. and Frotscher, M. (1989)
Entorhinal fibers form synaptic contacts on parvalbumin-
immunoreactive neurons in the rat fascia dentata. Brain Res.,
495: 161–166.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 10

Morphological development and maturation of

granule neuron dendrites in the rat dentate gyrus

Omid Rahimi1 and Brenda J. Claiborne2,

1
Department of Cellular and Structural Biology, University of Texas Health Science Center at San Antonio, San Antonio,
TX 78229, USA
2
Department of Biology, University of Texas at San Antonio, One UTSA Circle, San Antonio, TX 78249, USA

Abstract: The first granule neurons in the dentate gyrus are born during late embryogenesis in the rodent,
and the primary period of granule cell neurogenesis continues into the second postnatal week. On the day of
birth in the rat, the oldest granule neurons are visible in the suprapyramidal blade and exhibit rudimentary
dendrites extending into the molecular layer. Here we describe the morphological development of the
dendritic trees between birth and day 14, and we then review the process of dendritic remodeling that occurs
after the end of the second week. Data indicate that the first adult-like granule neurons are present on day
7, and, furthermore, physiological recordings demonstrate that some granule neurons are functional at this
time. Taken together, these results suggest that the dentate gyrus may be incorporated into the hippo-
campal circuit as early as the end of the first week. The dendritic trees of the granule neurons, however,
continue to increase in size until day 14. After that time, the dendritic trees of the oldest granule neurons are
sculpted and refined. Some dendrites elongate while others are lost, resulting in a conservation of total
dendritic length. We end this chapter with a review of the quantitative aspects of granule cell dendrites in
the adult rat and a discussion of the relationship between the morphology of a granule neuron and the
location of its cell body within stratum granulosum and along the transverse axis of the dentate gyrus.

Keywords: dendritic trees; spines; filopodia; neonates; hippocampus

Introduction forms synapses on pyramidal neurons in the CA3

Granule neurons are the principal cell type in the
dentate gyrus, and their cell bodies are located in
stratum granulosum of the suprapyramidal and in-
frapyramidal blades. Dendrites extend from the
apical pole of the granule cell body into the over-
lying molecular layer, and the axon, or mossy fiber,
exits from the basal pole. The axon gives rise to
collateral branches in the hilar region and then
Corresponding author. Tel.: +1 210 458 5487;
Fax: 1 210 458 5669; E-mail: brenda.claiborne@utsa.edu
region of the hippocampus proper. The apical
dendrites of the granule neurons bifurcate as they
traverse the molecular layer, and the vast majority
of terminal branches reach the top of the layer in the
adult. The dendritic trees of most granule neurons
are elliptical, and all dendrites of granule neurons in
the adult dentate gyrus are covered with spines.
The primary period of granule cell neurogenesis
occurs over a two- to three-week period in the ro-
dent, beginning in late embryogenesis and contin-
uing through the second postnatal week. In the rat,
although a few granule neurons are born as early as

DOI: 10.1016/S0079-6123(07)63010-6 167

168
embryonic day 14, over 80% are born after the
birth of the animal (which occurs at about embry-
onic day 21) and neurogenesis peaks near the end of
the first week of life (Bayer and Altman, 1974;
Schlessinger et al., 1975). It is worth noting that the
granule neurons are the last cells to be generated in
the hippocampal formation, and it is well known
that granule cell neurogenesis continues into adult-
hood (Altman and Das, 1965; Kaplan and Hinds,
1977). Here we focus on granule neurons that are
generated in the neonatal rat. From recent evidence
in the mouse, it appears that adult-generated gran-
ule neurons progress through a similar set of stages
as they develop and mature (Zhao et al., 2006).
Any description of the development and matu-
ration of the granule neurons must take into ac-
count the temporal and spatial gradients of
granule cell neurogenesis (Schlessinger et al.,
1975; Cowan et al., 1980, 1981). The earliest born
granule neurons form stratum granulosum in the
septal portion of the dentate gyrus, and neurons
that are generated later form the more temporal
portions of the dentate gyrus. This gradient is re-
ferred to as the septotemporal gradient. A second
gradient exists along the transverse axis of the
dentate gyrus and is of considerable importance
for developmental and morphological studies. As
the first granule neurons are born, they form the
cell-body layer at the tip of the suprapyramidal
blade. As additional neurons are generated, they
form the cell-body layer in the middle of the sup-
rapyramidal blade and then the portions of stra-
tum granulosum closest to the crest region. This
gradient continues as more neurons are generated
such that the youngest neurons make up stratum
granulosum in the infrapyramidal blade. A third
gradient exists within stratum granulosum. The
neurons that are born first move into their final
position at the top of the cell-body layer near the
molecular layer, and the younger neurons move
into position beneath them such that they are lo-
cated in the bottom portion of stratum granulo-
sum near the hilar border. This developmental
pattern is in contrast to the ‘‘inside-out’’ pattern
found in other areas of the mammalian cerebral
cortex in which the later-generated neurons move
through the earlier-generated cells to occupy po-
sitions at the top of the cell-body layer.
Thus the oldest granule neurons are most likely
to be found at the top of stratum granulosum near
the distal tip of the suprapyramidal blade at the
septal pole, whereas the youngest neurons are lo-
cated predominantly in the infrapyramidal blade
near the temporal pole and in the deeper portions
of stratum granulosum along the entire extent of
the transverse axis of the dentate gyrus. In the rat,
the suprapyramidal blade begins to form in late
embryogenesis as the first granule neurons are
born — it is visible as a separate structure on the
day of birth and consists of a cell-body layer and a
relatively thin molecular layer (Cowan et al.,
1980). The molecular layer increases greatly in
width over the first several weeks. It is less than
100 mm at day 4 and increases to just over 200 mm
at day 14; it averages approximately 300 mm in
width in young adult rats (Loy et al., 1977; Clai-
borne et al., 1990; Rihn and Claiborne, 1990). The
infrapyramidal blade is barely visible on the day of
birth and grows more slowly than the suprapy-
ramidal blade during the first week, increasing
from approximately 45 mm in width on day 4 to
approximately 110 mm on day 10; it measures be-
tween 205 and 240 mm in young adult rats (Loy et
al., 1977; Claiborne et al., 1990).
Here we describe the development and matura-
tion of the dendritic trees of the granule neurons in
the rat, and we review the quantitative data on
dendritic morphology in the adult. We consider
the developmental period to encompass the time
from the birth of the animal through day 14. By
day 14, the oldest granule neurons have assumed
their adult form and size. From day 14 to 60,
however, the neurons go through a period of mat-
uration during which the dendritic tree is sculpted
and refined and the density of spines continues to
increase. The dendritic trees of granule neurons
appear to be mature by day 60: unpublished data
from our lab indicate that they do not undergo any
quantitative changes between 60 and 180 days.
Development of granule neuron dendrites
Because of the prolonged time-course of granule
cell neurogenesis in neonatal rats, a wide range of
dendritic morphologies are observed on any one

day during the first and second postnatal weeks
(Fricke, 1975; Seress and Pokorny, 1981; Wenzel
et al., 1981; Lübbers and Frotscher, 1988; Liu
et al., 1996, 2000; Ye et al., 2000; Jones et al.,
2003). It is possible, however, to identify distinct
stages in the development of granule cell dendritic
trees, either by comparing the various neuronal
structures present on a single day or by examining
the progression of dendritic morphologies over the
course of the neonatal period. As an example of
the first approach, Lübbers and Frotscher (1988)
identified several stages of developing granule neu-
rons in a 5-day-old rat. Neurons in the earliest
stages of development had only rudimentary den-
dritic trees, each consisting of two or more pri-
mary apical dendrites exiting from the cell body,
along with one or more basal branches. The pri-
mary apical dendrites gave rise to higher order
branches that were relatively short and thin and
that exhibited varicosities and the occasional
growth cone. Spines were not present on these
immature cells. In contrast, the most mature gran-
ule neurons on day 5 had a more elaborate apical
tree with longer branches that exhibited numerous
varicosities as well as occasional growth cones and
filopodia. A few spines were present on the apical
dendrites.
Based on data from Lübbers and Frotscher
(1988) and a number of other investigators, here
we describe a sequence of stages that characterize
the morphological development of granule neuron
dendrites in the young rat. A variety of techniques
have been used to examine developing granule neu-
rons during the first few weeks of life, including
Golgi impregnation (Fricke, 1975; Duffy and
Teyler, 1978a, b; Wenzel et al., 1981; Lübbers and
Frotscher, 1988; Zafirov et al., 1994), intracellular
injection (Liu et al., 1996, 2000; Ye et al., 2000), and
retrograde labeling (Jones et al., 2003). Fricke
(1975) characterized the dendritic structures of
Golgi-stained granule neurons in tissue from rats
at postnatal days 1, 2, 4, 8, 12, and 20, as well as in
tissue from adult rats over the age of 60 days. Other
labs using Golgi-impregnated material have exam-
ined a similar range of ages. Duffy and Teyler
(1978a, b) analyzed granule neurons in sections
from rats at 7, 14, 30, 60, and 210 days of age,
Wenzel et al. (1981) determined morphologies at a
169
variety of ages between days 0 and 180, and Zafirov
et al. (1994) described granule neurons at days 5,
10, 15, and 20. Trommer and colleagues (Liu et al.,
1996, 2000; Ye et al., 2000) injected granule neurons
in slices from rats between the ages of 5 and 32 days
with biocytin, whereas Jones et al. (2003) used ret-
rograde labeling with DiI to analyze the dendritic
structures of granule neurons in rats between the
ages of 2 and 9 days. The latter group focused on
the oldest granule neurons — those that were lo-
cated near the tip of the suprapyramidal blade and
in the top portion of stratum granulosum.
On the day of birth and on postnatal day 1 in the
rat, granule neurons are visible in the suprapyram-
idal blade of Golgi-stained tissue and exhibit rudi-
mentary trees. Some have only a few short, stubby
branches, whereas others have longer and more
numerous dendrites (Fricke, 1975; Wenzel et al.,
1981). The dendrites are smooth and growth cones
are not typically observed. Wenzel et al. (1981)
considered granule neurons on the day of birth to
be in the first stage of development and labeled
them as primitive or early neuroblasts. As noted by
Fricke (1975), it is a bit surprising that all of the
stained granule neurons at this age have such sparse
dendritic trees — although most granule neurons
are likely to be only a few days old at this time, a
few neurons are born as early as embryonic day 14
and are already 7 days old on the day of birth.
On postnatal days 2 and 3, the oldest granule
neurons, in the suprapyramidal blade, have one or
more primary apical dendrites and varying num-
bers of shorter, higher-order branches (Fig. 1A
and B; Jones et al., 2003). Granule cells located at
the very top of stratum granulosum have several
primary dendrites whereas those located a bit
deeper in the layer are more likely to have only one
thick primary dendrite emerging from the cell
body, as described previously for granule neurons
in adult rodents (Fricke, 1975; Desmond and
Levy, 1982; Green and Juraska, 1985; Claiborne et
al., 1990). In the developing rat, the primary dend-
rites of both the superficial and the deep granule
cells tend to branch at or near the top of stratum
granulosum, and the diameters of the dendrites
change abruptly at branch points. The majority of
branches are still relatively smooth at this age,
with only occasional varicosities or growth cones.
170
Fig. 1. Photomontage (A) and drawings (B and C) made from
serial micrographs of DiI-labeled granule neurons from a 3-
day-old (A, B) and a 6-day-old rat (C). Note the presence of
immature features on dendrites at this age. Solid-curved arrow,
growth cone; solid-straight arrow, varicosity; open-straight ar-
row, filopodia; long-filled arrow, abrupt diameter change; open
arrowheads, continuation of axon; a, axon; c, axon collateral.
The montage and the drawings are shown at the same magni-
fication. Scale bar ¼ 50 mm. (Adapted with permission from
Jones et al., 2003.)
A few dendrites, however, exhibit numerous
growth cones, varicosities, and filopodia. Most in-
vestigators report the absence of dendritic spines
at these early ages, although Wenzel et al. (1981)
noted ‘‘spine-resembling’’ structures at day 0 and a
few spines at day 2. Basal dendrites are common
on neurons at days 2 and 3. They tend to be
thinner than apical dendrites but exhibit the same
immature features, including growth cones, vari-
cosities, and filopodia. Basal dendrites are con-
sidered to be an immature feature of granule
neurons in the rodent — they are not commonly
found on adult granule neurons (Seress and Po-
korny, 1981; Lübbers and Frotscher, 1988).
As development proceeds, the most mature
granule neurons exhibit more extensive dendritic
trees with longer branches and a full complement
of immature features (Fricke, 1975; Duffy and
Teyler, 1978a, b; Wenzel et al., 1981; Lübbers and
Frotscher, 1988; Jones et al., 2003). On day 4,
dendrites have larger varicosities and an increased
number of filopodia as compared to neurons in
younger animals (Jones et al., 2003). Diameter
changes are abrupt at branch points, and most
dendrites terminate before reaching the top of the
molecular layer. Basal dendrites are present on the
majority of cells. Although Jones et al. (2003) did
not report any spines on granule neuron dendrites
at day 4, Fricke (1975) noted the occasional spine
on Golgi-stained dendrites at this age.
On day 5, the oldest granule neurons are char-
acterized by numerous apical branches with many
filopodia and varicosities and several basal
branches (Fig. 2A; Wenzel et al., 1981; Lübbers
and Frotscher, 1988; Zafirov et al., 1994; Jones et
al., 2003). Abrupt diameter changes and growth
cones are present but are less numerous than on
day 4. Spines are scattered throughout the dendri-
tic trees of most of the older neurons on day 5,
although they are found in clusters on a few dend-
rites on a small number of cells (Seress and Po-
korny, 1981; Wenzel et al., 1981; Lübbers and
Frotscher, 1988; Zafirov et al., 1994; Jones et al.,
2003). Wenzel et al. (1981) characterized granule
neurons between 5 and 8 days of age as interme-
diate neuroblasts and considered this period to be
the second stage of granule neuron development.
The most mature granule neurons observed on
day 6 differ somewhat from those in the younger
animals (Fig. 1C; Jones et al., 2003). Growth cones
and varicosities are smaller and less numerous,
whereas the number of filopodia is greater (Fig.
2B). Abrupt diameter changes and basal dendrites
are less frequent. Most dendrites on these neurons
reach the hippocampal fissure at the top of the

Fig. 2. Stacked series of individual images taken with a con-
focal laser-scanning microscope. Dendrites are from a 5-day-
old rat (A) and a 6-day-old rat (B). The dendrite shown in B is
from one of the most mature neurons seen on day 6. Although
many spines are present, immature features are still observed on
day 6. Open arrow, filopodia; filled arrow, spines. Scale
bar ¼ 10 mm. (Adapted with permission from Jones et al.,
2003.)
molecular layer, and spines are present in varying
densities and on varying numbers of dendritic
branches.
Jones et al. (2003) observed the first adult-like
granule neurons on day 7 (Fig. 3). These neurons
were considered to be adult-like because they ex-
hibited the three primary characteristics of adult
granule neurons: they were devoid of immature
features except for occasional varicosities or filopo-
dia, the vast majority of the dendrites reached the
top of the molecular layer, and all dendrites were
covered with spines. The cell bodies of the adult-like
granule neurons were located towards the top of the
granule cell layer in the suprapyramidal blade and
were most likely between 7 and 10 days old at this
time. Other investigators had suggested that adult-
like granule neurons were present at about this time.
171
Fricke (1975) illustrated a granule neuron with nu-
merous spines from an 8-day-old rat. Wenzel et al.
(1981) noted that granule neurons in 10-day-old rats
resembled adult neurons, although they reported
the presence of growth cones and varicosities and
suggested that such features were indicative of con-
tinued dendritic growth. They considered granule
cells at this age to be in the third stage of devel-
opment and labeled them maturing or young neu-
rons. Zafirov et al. (1994) reported that some
granule neurons display extensive dendritic trees
with spines and without immature features on day
10. Similarly, Liu et al. (2000) described granule
neurons with at least some adult-like features on
day 7 and illustrated an adult-like granule cell from
a 12-day-old rat.
Thus it is now clear that a small proportion of
the oldest granule neurons exhibit adult-like mor-
phological features by the end of the first postnatal
week in the rat. These data, in combination with
results from other studies, suggest that the dentate
gyrus may be functional at this early age. For ex-
ample, our in vivo studies demonstrated that long-
term potentiation (LTP) and long-term depression
(LTD) can be elicited at medial perforant path
synapses onto the granule neurons at day 7
(O’Boyle et al., 2004), confirming earlier in vitro
studies (Duffy and Teyler, 1978b; Trommer et al.,
1995). It is also worth noting that the granule cell
afferents are in their approximate adult locations
at this time, and that the dendritic trees of hilar
interneurons are well developed (Cowan et al.,
1980; Seay-Lowe and Claiborne, 1992). Further-
more, the axons (or mossy fibers) of the oldest
granule neurons reach region CA3 of the hippo-
campus proper on the day of birth or soon there-
after (Minkwitz, 1976; Stirling and Bliss, 1978;
Jones et al., 2003), and granule cell stimulation can
induce LTD in CA3 pyramidal neurons by day 7
(Battistin and Cherubini, 1994). Taken together,
these data suggest that the traditional view of the
dentate gyrus as a ‘‘late developing’’ structure may
be incorrect; at least some of the granule neurons
are capable of functioning within the hippocampal
circuit by the end of the first postnatal week, at
about the same time that adult-like properties are
first observed in the pyramidal neurons of the hip-
pocampus proper.
172
Fig. 3. Drawing made from serial micrographs of a DiI-labeled granule neuron from a 7-day-old rat. Note the absence of immature
features and the prevalence of spines. A varicosity is visible on the axon (short, filled arrow); open arrowhead, continuation of axon.
Scale bar ¼ 50 mm. (Adapted with permission from Jones et al., 2003.)
Although granule neurons with adult-like fea-
tures are present by the end of the first week, it is
important to note that granule neuron dendrites
continue to elongate during the second postnatal
week (Fricke, 1975; Duffy and Teyler, 1978a, b;
Seress and Pokorny, 1981; Wenzel et al., 1981;
Zafirov et al., 1994). Using one of the first com-
puter-microscope systems designed for quantita-
tive, three-dimensional morphological studies of
single neurons, Fricke (1975) analyzed the dendri-
tic trees of developing Golgi-impregnated granule
neurons over the first postnatal month in the rat.
He found that the total dendritic length (defined as
the sum of the lengths of all individual dendritic
segments) of a granule neuron increased between
day 8 and day 12, from an average of approxi-
mately 500 mm on day 8 to an average of approx-
imately 1000 mm on day 12, with a wide variability
at both ages. Duffy and Teyler (1978a, b) also re-
ported that granule neurons increased dramati-
cally in size between day 7 and day 14, and Zafirov
et al. (1994) showed that total dendritic lengths
increased from approximately 300 mm at day 5 to
approximately 465 mm on day 10, with no signifi-
cant increase between days 10 and 15.
As granule neuron dendrites continue to elon-
gate after day 7, spine densities and synaptic con-
tacts also increase. As noted above, spines are
observed on granule neurons on days 4 and 5 in
the rat, and in some cases, spines have been re-
ported as early as days 2 or 3. Jones et al. (2003)
considered spines to be protrusions that were less
than 2 mm in length (whereas longer protrusions
were considered to be filopodia) and noted that
dendritic spines on neonatal granule neurons
ranged in shape from short, stubby protrusions,
either with or without heads, to those with long,
thin necks ending in definitive spine heads (Des-
mond and Levy, 1985; Trommald and Hulleberg,
1997). Counts of spines on DiI-labeled dendrites
located in the middle of the molecular layer re-
vealed an average of 0.40 spines/mm in 5-day-old
rats and 0.57 spines/mm in 6-day-old animals
(Jones et al., 2003). Densities increased to 0.81
spines/mm by day 7, although they were still far
below the values reported for dendrites in the same

region in the adult rat (1.66 spines/mm; Desmond
and Levy, 1985). Zafirov et al. (1994) also dem-
onstrated that spine densities increased after day 5.
These densities, however, were slightly lower than
those reported by Jones et al. (2003), perhaps be-
cause they calculated spine densities across the en-
tire dendritic tree. Based on measurements of the
total dendritic length and the total number of
spines per cell, they found that spine densities in-
creased from day 5 through day 15, with 0.11
spines/mm on day 5, 0.47 spines/mm on day 10, and
0.67 spines/mm on day 15. Duffy and Teyler
(1978a, b) also reported that the number of spines
per granule neuron more than doubled between
day 7 and day 14, increasing from approximately
100 spines/cell on day 7 to approximately 255
spines/cell on day 14. Data from the above studies
are substantiated by the detailed analysis of spine
densities on dendrites of various orders over the
course of granule neuron development and matu-
ration done by Wenzel et al. (1981). For example,
they reported that densities on 4th order branches
increased from 0.05 spines/mm on day 5 to 0.12
spines/mm on day 8 and to 0.18 spines/mm on
day 10.
Electron microscope studies demonstrate that
only a few synapses are present in the molecular
layer of the suprapyramidal blade on postnatal
day 1 (0.4 synapses/100 mm2; Cowan et al., 1980).
There is, however, a dramatic increase in synapses
over the first 10 days (Cowan et al., 1980). On days
4 and 5, there are approximately 2–3 synapses/
100 mm2 and, by day 10, densities have increased to
11.0 synapses/100 mm2 (Crain et al., 1973; Cowan
et al., 1980). In the molecular layer of the infra-
pyramidal blade, synapses are present on day 5
(0.8 synapses/100 mm2) and synaptic density in-
creases about sixfold by day 10 (6.1 synapses/
100 mm2; Cowan et al., 1980). In addition to this
considerable increase in synapse densities, the mo-
lecular layer also increases in volume approxi-
mately fourfold between days 5 and 10, resulting in
approximately a 16-fold increase in absolute
synapse number. It is also worth noting that a
high percentage of synapses are found on dendritic
shafts during the neonatal period (Cowan et al.,
1981). On day 5, about 50% of synapses are onto
dendritic shafts whereas on day 10, approximately
173
40% are onto shafts. The percentage of shaft
synapses continues to decline into adulthood, and
by day 41, only approximately 10% of synapses
are found on dendritic shafts with the remainder
contacting dendritic spines. In addition to
synapses in the molecular layer, synaptic contacts
are found on granule cell bodies in the neonatal
rat. A fair proportion is symmetric and stain for
glutamate-decarboxylase (Lubbers and Frotscher,
1988; Seress et al., 1989).
In summary, during the first week of life, granule
neuron dendrites undergo a sequence of morpho-
logical changes. Immature features appear and then
regress, dendrites elongate, and spines and synapses
develop. By day 7, the oldest granule neurons ex-
hibit adult-like characteristics immature features
have regressed, the vast majority of dendrites reach
the top of the molecular layer, and all dendrites are
covered with spines. During the second postnatal
week, dendrites continue to elongate and spine and
synaptic densities increase dramatically. By day 14,
a considerable number of granule neurons have at-
tained adult-like characteristics. It also appears that
the large en passant boutons of the mossy fibers
have attained an adult-like shape and complexity by
day 14, even as the number of boutons continue to
increase into young adulthood (Stirling and Bliss,
1978; Amaral and Dent, 1981).
Maturation of granule neuron dendritic trees
While granule neurons in 14-day-old rats qualita-
tively resemble adult cells, quantitative studies sug-
gest that their dendrites continue to change as the
animal matures. Early work indicated that the den-
dritic tree might increase in size after the second
week. Fricke (1975) reported that the total dendritic
lengths of Golgi-stained granule neurons increased
after day 12, reaching adult values of a little over
1500 mm on day 20. Duffy and Teyler (1978a, b)
also reported an increase in granule cell dendritic
length between days 14 and 30. In addition, they
found a slight but not statistically significant de-
crease between days 30 and 60 and another slight
increase at day 210 — the average length at 210
days was approximately the same as that at day 30.
It is not clear from the methods, however, whether
174
the reported lengths reflect measurements of the
sum of all dendritic branch lengths or the length of
the longest dendrite. Duffy and Teyler (1978a, b)
state that they measured ‘‘from mid-soma to the
most distal dendritic process’’, and the reported
values are quite similar to the width of the molec-
ular layer (Rihn and Claiborne, 1990). Thus, these
lengths may reflect the distance from the soma to
the most distal dendritic tips. Liu et al. (2000), using
intracellular labeling techniques, also found that
granule cells enlarged after day 14 although the in-
creases in total dendritic length were not statisti-
cally significant.
Whereas the studies described above suggest
that dendritic trees increase in size during the third
and fourth weeks, Wenzel et al. (1981) reported
that granule cell dendritic length and the degree of
branching reached adult proportions at day 15. To
resolve this issue, Rihn and Claiborne (1990) used
intracellular labeling techniques and three-dimen-
sional analyses to quantify the dendritic trees of
only the oldest granule neurons in the suprapy-
ramidal blade between days 14 and 60. The somata
of these neurons were located in the top portion of
the granule cell layer, between the tip and the
middle of the suprapyramidal blade. Neurons
from rats of five age groups were examined: the
first group included rats between the ages of 14
and 19 days and the oldest group included animals
between 50 and 60 days. The dendritic trees from
rats in each age group exhibited similar structures,
having between 1 and 4 primary apical dendrites
that were covered with spines (Fig. 4). Several
qualitative differences, however, were apparent
between neurons in the youngest rats and those in
the oldest animals. For example, neurons in the
youngest rats had thicker dendrites in the proximal
and middle thirds of the molecular layer, and they
exhibited a higher number of short, terminal seg-
ments in the distal third of the layer.
Analyses of these maturing granule neurons
demonstrated that both dendritic growth and re-
gression occurred between days 14 and 60, leading
to a conservation of total dendritic length during
this period (Rihn and Claiborne, 1990). The mo-
lecular layer of the suprapyramidal blade increased
by approximately 50% (from an average of 205 to
305 mm) between days 14 and 60, and the vast
majority of dendrites reached the top of the mo-
lecular layer in animals of all ages, suggesting that
individual dendrites elongated as the animal ma-
tured. The number of dendritic segments, however,
decreased from an average of 36 segments in the
14- to 19-day-old rats to an average of 28 segments
in the 50- to 60-day-old animals, with the majority
of the decrease occurring by day 29. (Segments
were defined as a length of dendrite between an
origin and a branch point, between two branch
points, or between a branch point and a distal
termination point.) As a consequence of this con-
current branch elongation and loss, the average
total dendritic length did not change significantly
between days 14 and 60 (3086 and 3417 mm, re-
spectively). A similar process of dendritic loss with
no change in total dendritic length has been doc-
umented for nonpyramidal neurons in the rat vis-
ual cortex (Parnavelas and Uylings, 1980).
Thus, although granule cells have attained their
adult dendritic lengths by postnatal day 14, den-
dritic remodeling occurs between days 14 and 60
with the elongation of some branches and the loss
of others. In addition, the number of spines and
synapses increase (Duffy and Teyler, 1978a, b;
Cowan et al., 1981; Wenzel et al., 1981; Zafirov et
al., 1994). Duffy and Teyler (1978a, b) showed a
slight increase in the number of spines per cell be-
tween days 14 and 30. Values remained stable be-
tween days 30 and 60 and then increased again
between days 60 and 210. Seress and Pokorny
(1981) noted that spine densities increased between
days 10 and 25, and Zafirov et al. (1994) reported
that spine densities (calculated on the basis of total
spines per cell divided by the total dendritic length)
increased from 0.67 spines/mm at day 15 to 1.16
spines/mm at day 20. Wenzel et al. (1981) found
that spine densities increased dramatically on
dendrites in almost all branch orders between
days 15 and 30. For example, densities increased
from approximately 0.2 to 0.59 spines/mm on 4th
order branches and from 0.14 to 0.56 spines/mm on
6th order branches. Furthermore, they found that
densities also continued to increase slightly after
day 30; when all branch orders were considered
together, densities increased from approximately
0.5 to approximately 0.6 spines/mm between days
30 and 180. Synaptic density in the molecular layer

Fig. 4. Camera lucida drawing of a granule neuron from a 53-day-old rat. Note the significant increase in size as compared to the
adult-like neuron from a 7-day-old rat shown in Fig. 3. Scale bar ¼ 50 mm.
also increased with maturation (Cowan et al.,
1980). In the molecular layer of the suprapyram-
idal blade, densities increased from 11 synapses/
100 mm2 on day 10 to 36 synapses/100 mm2 on day
21. Only a slight increase was seen after day 21;
densities were 37 synapses/100 mm2 on day 41.
Similarly, the percentage of spine synapses in-
creased considerably from day 10 to 21, but did
not exhibit much of a change between days 21 and
41. Spine synapses comprise 47% of all synapses
on day 10, 84% on day 21, and 88% on day 41.
In summary, the total dendritic lengths of the
oldest granule neurons in the suprapyramidal
blade appear to be established by day 14. Quan-
titative changes in the dendritic tree do occur after
this time, however; some dendrites elongate while
others are lost, thus leading to a conservation of
total length. Interestingly, there is one other report
of regression in the hippocampal formation during
175
maturation. The axons or mossy fibers of the
granule neurons exhibit filopodial-like extensions
that elongate during the first two postnatal weeks,
reach a peak in length on day 14, and then de-
crease to adult lengths by day 28 (Amaral, 1979).
The time course of their growth and retraction is
remarkably similar to the time course of dendritic
branch growth and regression in the oldest granule
neurons — whether the two processes are gov-
erned by the same or similar mechanisms is not yet
known. In this regard, preliminary data from our
lab indicate that dendritic branch loss may be
affected by incoming neuronal activity. Specifi-
cally, we found that blockade of N-methyl-D-as-
partate (NMDA) glutamate receptors between
days 14 and 24 did not affect dendritic growth,
but did result in a decrease in dendritic branch loss
in the oldest granule neurons (Blake and Clai-
borne, 1995). Because branches continued to grow
176
and fewer branches were lost, the granule neurons
in the treated animals had greater total dendritic
lengths than did those in the controls. It is worth
noting that the glutamate released from synapses
on granule cell dendrites in the middle third of the
molecular layer binds to NMDA receptors, and
that the afferents to this region arise from the ent-
orhinal cortex. The entorhinal cortex, in turn, re-
ceives inputs from the visual, auditory, and
somatosensory cortices, and a number of signifi-
cant events occur in the maturation of these sen-
sory systems soon after day 14 in the rat. For
example, the eyes open approximately at day 15
and adult sensitivities to sound are reached be-
tween days 16 and 18 (Tilney, 1933; Crowley and
Hepp-Reymond, 1966). Therefore, it is not sur-
prising that granule cell remodeling begins shortly
after day 14 and may be governed by afferent in-
puts from the entorhinal cortex.
Granule neuron dendrites in the adult rat
Quantitative data on the dendritic trees of adult
granule neurons are based on two- and three-di-
mensional measurements of both Golgi-impreg-
nated and intracellularly labeled cells. Fricke
(1975) was the first to make use of a computer-
microscope system to analyze Golgi-stained neu-
rons in three dimensions, whereas Desmond and
Levy (1982) developed a novel probabilistic
method for quantifying dendritic trees of granule
neurons from Golgi-stained tissue in two dimen-
sions. They corrected for cut dendrites at the edge
of sections by estimating their branching and ter-
mination patterns based on the patterns observed
for intact dendrites. Corrected values were com-
pared with data from tracings of six neurons (two
from each region) that were followed through se-
rial sections. Only granule neurons located in the
middle 80% of the longitudinal extent of the hip-
pocampus and with a minimum of cut dendrites in
the proximal third of the molecular layer were in-
cluded. Claiborne et al. (1990) and Rihn and Clai-
borne (1990) applied the computational techniques
first used by Fricke (1975) to quantify the dendritic
trees of intracellularly labeled granule neurons in
relatively thick transverse slices (400 mm) of the
hippocampal formation. They analyzed only those
neurons that were completely stained, had no cut
dendrites in the proximal portion of the molecular
layer, and had fewer than two-severed dendrites in
the distal two-thirds of the layer.
Here we review the quantitative parameters of
adult granule neurons reported by the above in-
vestigators and by others employing similar tech-
niques. We first discuss the available data on the
entire population of granule neurons, and we then
review the relationship between the morphology of
a granule neuron and the location of its cell body
within stratum granulosum and within the two
blades of the dentate gyrus.
The cell body of an adult granule neuron in the
rat is approximately 10 mm wide and approxi-
mately 19 mm long (Claiborne et al., 1990). Be-
tween 1 and 5 primary apical dendrites exit from
the apical pole of the cell body and bifurcate rel-
atively close to the soma (Fig. 5). Fricke (1975)
noted that most branching occurs within 100 mm of
the soma and within the vicinity of the boundary
between stratum granulosum and stratum mole-
culare. Desmond and Levy (1982) found that the
majority of primary segments branch within the
granule cell layer and first ninth of the molecular
layer. Furthermore, they reported that nearly all
first-order branch points occur within 50 mm of the
cell body and within the first 30 mm of the molec-
ular layer. Primary dendrites bifurcate into higher
order segments, and up to eighth-order branches
have been reported, with an average maximum
branch order of 5.7 (Claiborne et al., 1990). Gran-
ule neurons have a total of approximately 30 den-
dritic segments, with reported numbers ranging
from 22 to 40 (Fricke, 1975; Seress and Pokorny,
1981; Desmond and Levy, 1982; Green and
Juraska, 1985; Claiborne et al., 1990; Rihn and
Claiborne, 1990).
Based on Golgi impregnations, Fricke (1975)
reported an average total dendritic length of
1602 mm (range from 773 to 2445 mm) for adult
granule neurons, whereas Seress and Pokorny
(1981) reported an average of 2405 mm, and Green
and Juraska (1985) found total dendritic lengths
ranging from 1161 to 1279 mm. After correcting for
cut dendrites, Desmond and Levy (1982) calcu-
lated an average total dendritic length of 3662 mm
 177

Fig. 5. Computer-generated plots of three-dimensional reconstructions of the dendritic trees of granule cells located at various
positions along the transverse axis of the dentate gyrus. Each neuron is from a different animal, and the total dendritic length of each
dendritic tree is indicated. Note that the total dendritic lengths of the trees in the suprapyramidal blade are greater than those of the
trees in the infrapyramidal blade. CA3, field CA3 of the hippocampus proper; GL, granule cell layer. Scale bar ¼ 100 mm. (Adapted
with permission from Claiborne et al., 1990.)

for Golgi-stained neurons. Claiborne and col-
leagues reported similar averages of 3221 and
3417 mm, respectively, for intracellularly labeled
granule neurons analyzed in two different studies
(Fig. 5; Claiborne et al., 1990; Rihn and Claiborne,
1990). Desmond and Levy (1982) reported that
approximately 12% of the total length was found
within stratum granulosum, approximately 25%
within the proximal third of the molecular layer,
and the remaining portion was restricted to the
distal two-thirds. Similarly, data from Claiborne et
al. (1990) indicated that 30% of the total length
was found in the granule cell layer and proximal
third of the molecular layer, 30% in the middle
third and 40% in the distal third. These results
suggest that more dendritic length is available for
synapses from the entorhinal afferents that
terminate in the distal two-thirds of the layer than
for commissural and associational afferents that
make contacts in the inner third of the layer.
The majority of synapses onto the granule neu-
ron dendrites occur on spines in the adult. As dis-
cussed above, a number of groups have shown that
spine and synaptic densities increase throughout
granule neuron development and maturation, and
their reported values for densities in maturing
and young adult rats are reviewed above (Duffy
and Teyler, 1978a, b; Cowan et al., 1981; Seress and
Pokorny, 1981; Wenzel et al., 1981; Zafirov et al.,
1994). Other investigators have reported spine den-
sities only for granule neurons in young adult rats.
Fricke (1975) noted that spine densities were low on
dendrites of adult granule neurons within the first
20 mm of the soma and within 50 mm of dendritic
178
terminations. He found a maximum of approxi-
mately 1.3 spines/mm on dendrites in the remaining
portion of the dendritic tree. Desmond and Levy
(1985) reported that spines were infrequent on
dendrites within stratum granulosum and that there
were three major types of spines on adult granule
cell dendrites in the molecular layer: stubby, mush-
room-shaped, and thin. They calculated spine den-
sities for each type in the proximal, middle, and
distal thirds of the molecular layer of the two
blades. In both blades and in all three portions of
the layer, densities were highest for the thin spines
and were lowest for mushroom-shaped protrusions.
When all spine types were included and corrections
made for obscured spines, densities averaged 1.6
spines/mm on dendrites of neurons in the suprapy-
ramidal blade and 1.3 spines/mm on those in the
infrapyramidal blade. These values are similar to
those reported by Fricke (1975) but are much
greater than those reported by Wenzel et al. (1981;
see above) perhaps because different criteria were
used for the inclusion of spines (Desmond and
Levy, 1985).
In summary, the dendritic trees of granule neu-
rons in young adult rats are composed of approx-
imately 30 spine-covered segments and have total
dendritic lengths of approximately 3400 mm. It is
not surprising that these total lengths are much
less than those of the relatively large pyramidal
neurons in the hippocampus proper. Reports of
the total dendritic lengths of pyramidal neurons in
region CA1 vary from an average of 10,800 to
17,400 mm (Ishizuka et al., 1995; Mainen et al.,
1996; Pyapali and Turner, 1996; Pyapali et al.,
1998; Megias et al., 2001), whereas the total
lengths for pyramidal neurons in region CA3
range from a low of 7000 mm for pyramidal neu-
rons in region CA3c to a high of 19,800 mm for
those in region CA3a (Ishizuka et al., 1995; Turner
et al., 1995; Gonzales et al., 2001). It is also not
surprising that the granule neurons are the most
electrotonically compact of the three major classes
of hippocampal neurons (Carnevale et al., 1997).
Given the gradients of granule cell neurogenesis,
it is of interest that several parameters of granule
neuron dendrites are correlated with the location
of the parent cell body in the granule cell layer —
both along the transverse axis of the dentate gyrus
and within the depth of the granule cell layer. Data
demonstrate that the dendritic trees of neurons in
the suprapyramidal blade are larger than those in
the infrapyramidal blade (Fig. 5). Fricke (1975)
reported that suprapyramidal neurons have
greater total dendritic lengths than do those in
the infrapyramidal blade (1674 mm vs. 1482 mm),
and Desmond and Levy (1982) confirmed this re-
sult for Golgi-stained granule neurons that were
traced through multiple sections (3107 mm for sup-
rapyramidal neurons vs. 2078 mm for infrapyram-
idal blade cells). Similarly, Claiborne et al. (1990)
showed that suprapyramidal granule neurons have
greater total dendritic lengths (3478 mm vs.
2793 mm), more dendritic segments (31 vs. 27),
and wider dendritic spreads in the transverse plane
of the hippocampal formation (347 mm vs. 288 mm)
than do the infrapyramidal blade neurons. These
results demonstrate that suprapyramidal granule
neurons have more dendritic length available for
afferent contacts than do neurons in the opposite
blade. In addition, Desmond and Levy (1985) re-
ported that spine density is greater on dendrites of
suprapyramidal neurons (1.6 spines/mm vs. 1.3
spines/mm), and this result, combined with the in-
creased length of the suprapyramidal neurons,
suggests that neurons in this blade receive many
more synaptic contacts than do neurons in the in-
frapyramidal blade.
Overall, these data indicate that the dendritic
trees of granule neurons in the suprapyramidal
blade are likely to be larger than those in the in-
frapyramidal blade and to receive more afferent
contacts. It is of considerable interest that the axon
trajectories of the granule neurons also vary with
granule cell location. Axons of granule neurons
located toward the tip of the suprapyramidal blade
traverse stratum radiatum of region CA3 before
entering stratum lucidum, avoiding both the hilar
region and the proximal part of the pyramidal cell
layer, whereas axons of neurons in the crest and
the infrapyramidal blade travel through the hilar
region and contact pyramidal neurons in the prox-
imal portion of field CA3 (Claiborne et al., 1986).
It is likely that such morphological distinctions in
the granule neurons, in combination with physio-
logical differences (Scharfman et al., 2002; Chawla
et al., 2005), may lead to functional differences

between the two blades. This idea is best exempli-
fied by recent results showing that induction of the
activity-regulated, immediate early gene Arc fol-
lowing behavioral stimulation occurs in the sup-
rapyramidal blade, but not in the infrapyramidal
blade (Chawla et al., 2005).
Not only do some morphological parameters
vary according to the location of the cell body
along the transverse axis of the dentate gyrus, sev-
eral dendritic parameters also differ according to
the depth of the parent cell body in the granule cell
layer. Neurons with somata located at the top of
the layer tend to have multiple primary branches
whereas those with cell bodies located deeper in
the layer tend to have only one primary dendrite
(Fricke, 1975; Seress and Pokorny, 1981; Des-
mond and Levy, 1982; Green and Juraska, 1985;
Claiborne et al., 1990). Granule neurons in the
superficial half of the granule cell layer also exhibit
nearly twice the density of axosomatic synapses as
do cells in the bottom half of the layer (Lee et al.,
1982). In addition, Green and Juraska (1985)
showed that Golgi-stained granule neurons with
somata in the superficial third of the granule cell
layer had greater maximum widths in the trans-
verse plane of the dentate gyrus, more branches in
orders 1 through 3, and fewer branches in orders 4
through 6. These neurons also had greater total
dendritic lengths (1279 vs. 1161) than did deeper
neurons; 80% of the neurons in their sample were
located in the suprapyramidal blade. In contrast,
Claiborne et al. (1990) did not find a statistically
significant difference between the average total
lengths of neurons with somata in the superficial
half of the granule cell layer and of those with
somata in the bottom half of the layer in either
blade. When only neurons in the suprapyramidal
blade were analyzed, however, they found that
superficial granule neurons had more primary
dendrites (2.4 vs. 1.5) and larger transverse spreads
(378 mm vs. 293 mm) than deeper neurons. In ad-
dition, 42% of the total length of superficial cells
was in the distal third of the layer as compared to
37% of the length of the deeper cells. For neurons
in the infrapyramidal blade, the only significant
difference was that superficial cells had larger
transverse spreads than did deeper neurons
(311 mm vs. 244 mm).
179
Thus, a number of structural differences exist
between the dendritic trees of neurons located in
the superficial portion of stratum granulosum and
those located in the deeper aspects of the layer. At
least one of the differences may have functional
consequences. Carnevale et al. (1997) demon-
strated that the single primary dendrites of deeper
neurons attenuated voltage signals spreading from
the soma out to the dendrites, thereby reducing the
effect of somatic events, including action poten-
tials, on molecular layer synapses. Given that
adult-generated granule neurons migrate into the
lower portion of the granule cell layer, it will be of
interest to determine whether other functional
differences correlate with the depth of granule cell
bodies in stratum granulosum in adult animals.
Summary
Granule neurons in the dentate gyrus of the rat
undergo periods of development and maturation
before reaching their final adult form and size. In
the rat, the primary period of neurogenesis begins
during late embryogenesis and continues over the
first two weeks of life. The oldest granule neurons
occupy the suprapyramidal blade whereas later
generated cells form the infrapyramidal blade. On
the day of birth, granule neurons in the suprapy-
ramidal blade exhibit a few sparse dendrites. Ru-
dimentary trees appear over the next few days and
by day 4, dendritic branching is quite extensive.
Immature features are abundant at this time and
include abrupt diameter changes, varicosities, filo-
podia, growth cones, and basilar dendrites. On days
5 and 6, elaborate dendritic trees are present, there
is a reduction in the frequency of immature fea-
tures, and some spines are visible on the most ma-
ture cells. On day 7, the first few adult-like trees are
present on granule neurons in the suprapyramidal
blade. Their dendrites reach the top of the molec-
ular layer and no longer exhibit immature features
except for an occasional filopodium or varicosity.
All of the branches are covered with spines. Phys-
iological recordings demonstrate that some granule
neurons are functional at this time, suggesting that
the dentate gyrus may be incorporated into the
hippocampal circuit by the end of the first week.
180
After day 7, the dendritic trees increase in size, and,
by day 14, numerous adult-like granule cells are
present. The oldest granule neurons then undergo a
process of refinement. Many dendrites continue to
elongate during this period, and other branches are
lost; the simultaneous growth and regression results
in a conservation of total dendritic length after day
14. Spine densities, however, continue to increase.
Adult granule neurons in the rat have approxi-
mately 30 dendritic segments and total dendritic
lengths of approximately 3400 mm. A variety of
morphological features are correlated with the lo-
cation of the granule cell body, both along the
transverse axis of the dentate gyrus and within the
depth of the granule cell layer. Recent evidence
suggests that there are functional distinctions be-
tween the two blades of the dentate gyrus; it will be
of considerable interest to determine whether differ-
ences in granule cell morphologies underlie these
distinctions.
Acknowledgment
This work was supported by NIH grant #GM08194
to BJC.
References
Altman, J. and Das, G.D. (1965) Autoradiographic and histo-
logical evidence of postnatal hippocampal neurogenesis in
rats. J. Comp. Neurol., 124(3): 319–335.
Amaral, D.G. (1979) Synaptic extensions from the mossy fibers
of the fascia dentata. Anat. Embryol. (Berl.), 155: 241–251.
Amaral, D.G. and Dent, J.A. (1981) Development of the mossy
fibers of the dentate gyrus: I. A light and electron microscopic
study of the mossy fibers and their expansions. J. Comp.
Neurol., 195: 51–86.
Battistin, T. and Cherubini, E. (1994) Developmental shift from
long-term depression to long-term potentiation at the mossy
fibre synapses in the rat hippocampus. Eur. J. Neurosci.,
6(11): 1750–1755.
Bayer, S.A. and Altman, J. (1974) Hippocampal development
in the rat: cytogenesis and morphogenesis examined with
autoradiography and low-level X-irradiation. J. Comp.
Neurol., 158(1): 55–79.
Blake, N.M.J. and Claiborne, B.J. (1995) Possible role of the N-
methyl-D-aspartate (NMDA) receptor in the remodeling of
dendritic arbors during development. Soc. Neurosci. Abstr.,
21: 1292.
Carnevale, N.T., Tsai, K.Y., Claiborne, B.J. and Brown, T.H.
(1997) Comparative electrotonic analysis of three classes
of rat hippocampal neurons. J. Neurophysiol., 78(2):
703–720.
Chawla, M.K., Guzowski, J.F., Ramirez-Amaya, V., Lipa, P.,
Hoffman, K.L., Marriott, L.K., Worley, P.F., McNaughton,
B.L. and Barnes, C.A. (2005) Sparse, environmentally selec-
tive expression of Arc RNA in the upper blade of the rodent
fascia dentata by brief spatial experience. Hippocampus,
15(5): 579–586.
Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1986) A light
and electron microscopic analysis of the mossy fibers of the
rat dentate gyrus. J. Comp. Neurol., 246: 435–458.
Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1990) Quan-
titative, three-dimensional analysis of granule cell dendrites
in the rat dentate gyrus. J. Comp. Neurol., 302: 206–219.
Cowan, W.M., Stanfield, B.B. and Amaral, D.G. (1981) Fur-
ther observations on the development of the dentate gyrus.
In: Cowan W.M. (Ed.), Studies in Developmental Neurobi-
ology: Essays in Honor of Viktor Hamburger. Oxford Uni-
versity Press, New York, pp. 395–435.
Cowan, W.M., Stanfield, B.B. and Kishi, K. (1980) The devel-
opment of the dentate gyrus. Curr. Top. Dev. Biol., 15:
103–157.
Crain, B., Cotman, C., Taylor, D. and Lynch, G. (1973) A
quantitative electron microscopic study of synaptogenesis in
the dentate gyrus of the rat. Brain Res., 63: 195–204.
Crowley, D.E. and Hepp-Reymond, M.C. (1966) Development
of cochlear function in the ear of the infant rat. J. Comp.
Physiol., 62: 427–432.
Desmond, N.L. and Levy, W.B. (1982) A quantitative anatom-
ical study of the granule cell dendritic fields of the rat dentate
gyrus using a novel probabilistic method. J. Comp. Neurol.,
212: 131–145.
Desmond, N.L. and Levy, W.B. (1985) Granule cell dendritic
spine density in the rat hippocampus varies with spine shape
and location. Neurosci. Lett., 54: 219–224.
Duffy, C.J. and Teyler, T.J. (1978a) Development of habitua-
tion in the dentate gyrus of rat: physiology and anatomy.
Brain Res. Bull., 3: 305–310.
Duffy, C.J. and Teyler, T.J. (1978b) Development of potent-
iation in the dentate gyrus of rat: physiology and anatomy.
Brain Res. Bull., 3: 425–430.
Fricke, R.A. (1975) Studies on the morphology and develop-
ment of the hippocampus and dentate gyrus. Ph.D. Disser-
tation, Washington University, St. Louis, MO.
Gonzales, R.B., DeLeon Galvan, C.J., Rangel, Y.M. and Clai-
borne, B.J. (2001) Distribution of thorny excrescences on
CA3 pyramidal neurons in the rat hippocampus. J. Comp.
Neurol., 430(3): 357–368.
Green, E.J. and Juraska, J.M. (1985) The dendritic morphology
of hippocampal dentate granule cells varies with their posi-
tion in the granule cell layer: a quantitative Golgi study. Exp.
Brain Res., 59: 582–586.
Ishizuka, N., Cowan, W.M. and Amaral, D.G. (1995) A quan-
titative analysis of the dendritic organization of pyramidal
cells in the rat hippocampus. J. Comp. Neurol., 362: 17–45.

Jones, S.P., Rahimi, O., O’Boyle, M.P., Diaz, D.L. and Clai-
borne, B.J. (2003) Maturation of granule cell dendrites after
mossy fiber arrival in hippocampal field CA3. Hippocampus,
13: 413–427.
Kaplan, M.S. and Hinds, J.W. (1977) Neurogenesis in the adult
rat: electron microscopic analysis of light radioautographs.
Science, 197: 1092–1094.
Lee, K.S., Gerbrandt, L. and Lynch, G. (1982) Axo-somatic
synapses in the normal and x-irradiated dentate gyrus: fac-
tors affecting the density of afferent innervation. Brain Res.,
249: 51–56.
Liu, X., Tilwalli, S., Ye, G., Lio, P.A., Pasternak, J.F. and
Trommer, B.L. (2000) Morphologic and electrophysiologic
maturation in developing dentate gyrus granule cells. Brain
Res., 856: 202–212.
Liu, Y.B., Lio, P.A., Pasternak, J.F. and Trommer, B.L. (1996)
Developmental changes in membrane properties and postsy-
naptic currents of granule cells in rat dentate gyrus. J.
Neurophysiol., 76(2): 1074–1088.
Loy, R., Lynch, G. and Cotman, C.W. (1977) Development of
afferent lamination in the fascia dentata of the rat. Brain
Res., 121: 229–243.
Lübbers, K. and Frotscher, M. (1988) Differentiation of gran-
ule cells in relation to GABAergic neurons in the rat fascia
dentata. Combined Golgi/EM and immunocytochemical
studies. Anat. Embryol. (Berl.), 178: 119–127.
Mainen, Z.F., Carnevale, N.T., Zador, A.M., Claiborne, B.J.
and Brown, T.H. (1996) Electrotonic architecture of hippo-
campal CA1 pyramidal neurons based on three-dimensional
reconstructions. J. Neurophysiol., 76(3): 1904–1923.
Megias, M., Emri, Z., Freund, T.F. and Gulyas, A.I. (2001)
Total number and distribution of inhibitory and excitatory
synapses on hippocampal CA1 pyramidal cells. Neurosci-
ence, 102: 527–540.
Minkwitz, H.G. (1976) Zur Entwicklung der Neuronenstruktur
des Hippocampus während der pär- und postnatalen Onto-
genese der Albinoratte. J. Hirnforsch., 17: 213–231.
O’Boyle, M.P., Do, V., Derrick, B.E. and Claiborne, B.J. (2004)
In vivo recordings of long-term potentiation and long-term
depression in the dentate gyrus of the neonatal rat.
J. Neurophysiol., 91: 613–622.
Parnavelas, J.G. and Uylings, H.B. (1980) The growth of non-
pyramidal neurons in the visual cortex of the rat: a morpho-
metric study. Brain Res., 193(2): 373–382.
Pyapali, G.K., Sik, A., Penttonen, M., Buzsaki, G. and Turner,
D.A. (1998) Dendritic properties of hippocampal CA1 py-
ramidal neurons in the rat: intracellular staining in vivo and
in vitro. J. Comp. Neurol., 391: 335–352.
Pyapali, G.K. and Turner, D.A. (1996) Increased dendritic
extent in hippocampal CA1 neurons from aged F344 rats.
Neurobiol. Aging, 17(4): 601–611.
Rihn, L.L. and Claiborne, B.J. (1990) Dendritic growth and
regression in rat dentate granule cells during late postnatal
development. Brain Res. Dev. Brain Res., 54: 115–124.
181
Scharfman, H.E., Sollas, A.L., Smith, K.L., Jackson, M.B. and
Goodman, J.H. (2002) Structural and functional asymmetry
in the normal and epileptic rat dentate gyrus. J. Comp. Ne-
urol., 454: 424–439.
Schlessinger, A.R., Cowan, W.M. and Gottlieb, D.I. (1975) An
autoradiographic study of the time of origin and the pattern
of granule cell migration in the dentate gyrus of the rat. J.
Comp. Neurol., 159(2): 149–175.
Seay-Lowe, S.L. and Claiborne, B.J. (1992) Morphology of in-
tracellularly labeled interneurons in the dentate gyrus of the
immature rat. J. Comp. Neurol., 324: 23–36.
Seress, L., Frotscher, M. and Ribak, C.E. (1989) Local circuit
neurons in both the dentate gyrus and Ammon’s horn es-
tablish synaptic connections with principal neurons in five
day old rats: a morphological basis for inhibition in early
development. Exp. Brain Res., 78: 1–9.
Seress, L. and Pokorny, J. (1981) Structure of the granular layer
of the rat dentate gyrus. A light microscopic and Golgi study.
J. Anat., 133(2): 181–195.
Stirling, R.V. and Bliss, T.V. (1978) Hippocampal mossy fiber
development at the ultrastructural level. Prog. Brain Res., 48:
191–198.
Tilney, F. (1933) Behavior in its relation to the development of
the brain. II. Correlation between the development of the
brain and behavior in the albino rat from embryonic states to
maturity. Bull. Neurol. Inst. N.Y., 3: 252–358.
Trommald, M. and Hulleberg, G. (1997) Dimensions and den-
sity of dendritic spines from rat dentate granule cells based
on reconstructions from serial electron micrographs.
J. Comp. Neurol., 377(1): 15–28.
Trommer, B.L., Kennelly, J.J., Colley, P.A., Overstreet, L.S.,
Slater, N.T. and Pasternak, J.F. (1995) AP5 blocks LTP in
developing rat dentate gyrus and unmasks LTD. Exp.
Neurol., 131: 83–92.
Turner, D.A., Li, X.G., Pyapali, G.K., Ylinen, A. and Buzsaki,
G. (1995) Morphometric and electrical properties of recon-
structed hippocampal CA3 neurons recorded in vivo.
J. Comp. Neurol., 356: 580–594.
Wenzel, J., Stender, G. and Duwe, G. (1981) Zur Entwicklung
der Neuronenstruktur der Fascia dentata bei der Ratte. Ne-
urohistologish-morphometrische, ultrastrukturelle and ex-
perimentelle Untersuchungen. J. Hirnforsch., 22: 629–683.
Ye, G.L., Liu, X.S., Pasternak, J.F. and Trommer, B.L. (2000)
Maturation of glutamatergic neurotransmission in dentate
gyrus granule cells. Brain Res. Dev. Brain Res., 124:
33–42.
Zafirov, S., Heimrich, B. and Frotscher, M. (1994) Dendritic
development of dentate granule cells in the absence of
their specific extrinsic afferents. J. Comp. Neurol., 345:
472–480.
Zhao, C., Teng, E.M., Summers Jr., R.G., Ming, G.L. and
Gage, F.H. (2006) Distinct morphological stages of dentate
granule neuron maturation in the adult mouse hippocampus.
J. Neurosci., 26(1): 3–11.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 11

Physiological studies of human dentate granule cells

Anne Williamson1, and Peter R. Patrylo2

1
Department of Neurosurgery, Yale University School of Medicine, New Haven, CT 06518, USA
2
Department of Physiology, Southern Illinois University, Carbondale, IL 62901, USA

Abstract: The availability of human hippocampi obtained through surgery (usually for treatment of tem-
poral lobe epilepsy) has allowed us to investigate the properties of the human dentate in a way that cannot
be done with other brain regions. The dentate has been the primary focus of these studies because of its
relative preservation in all patient specimens. Moreover, there is extensive synaptic reorganization of
numerous neurotransmitter systems in this the fascia dentate (dentate gyrus and the hilus) in humans with
specific forms of TLE. These changes are not evident in tissue from patients with seizure that begin outside
the hippocampus, and, as a result, this tissue provides an invaluable resource for comparisons. Physio-
logical data using both slices and acutely dissociated cells demonstrate that the granule cells have mem-
brane properties similar to those of rodents although there are specific changes that appear to be associated
with seizures. Similarly, in the non-sclerotic hippocampi, the synaptic properties are similar to those
reported in rodents. There are also a number of parallels between the findings in humans and in status
animal models of temporal lobe epilepsy. This review will cover analyses of membrane properties as well as
of glutamatergic, GABAergic, and neuromodulatory systems. Thus, while there are a number of issues that
invariably arise with studies of pathological human tissue, this tissue is ideally suited to verify and refine
animal models of temporal lobe epilepsy. In addition, one can argue that human tissue provides the only
resource to evaluate the ways that granule cells recorded from laboratory animals approximate human
granule cell physiology.

Keywords: temporal lobe epilepsy; hippocampal slice; mossy fiber sprouting; neuropeptides; inhibition

Introduction the work of Schwartzkroin and Prince (1976), and

We have a greater understanding of the physio-
logical properties of the dentate gyrus than any
other region of the human brain because of the
availability of tissue resected from patients with
pharmacoresistant temporal lobe epilepsy (TLE).
Numerous groups have studied resected human
hippocampi over the past 20 years, starting with
Corresponding author. Tel.: +1 203 785 5327;
Fax: +1 203 737 -2159; E-mail: anne.williamson@yale.edu
extending to the present day. The majority of these
studies have used physiological methods, using
both slices and acutely dissociated cells. More re-
cent studies have begun to use molecular biological
and imaging approaches to expand our under-
standing of the human dentate gyrus. This review
will focus on the physiology of the human dentate
gyrus with special attention to the changes ob-
served in patients with the most common pathol-
ogy associated with temporal lobe epilepsy, medial
temporal lobe sclerosis (MTS) (de Lanerolle et al.,
2003).

DOI: 10.1016/S0079-6123(07)63011-8 183

184
While there are a number of issues associated
with studies of human material, we argue that use
of human tissue is useful because it provides vir-
tually the only resource for comparison to rodent
tissue. In addition, tissue from patients with
intractable TLE provides a critical foundation
for the development of accurate animal models of
TLE.
The major issues to consider are variability of
the patient population and the lack of true con-
trols. Other, more minor concerns, include varia-
bility in the surgical resection of the tissue and
hence tissue viability, and the fact that the area of
the hippocampus that is resected may be limited.
The vast majority of hippocampi that are avail-
able for physiological study are resected for treat-
ment of medically intractable seizures and thus are
not from normal brain. However, within the pop-
ulation of TLE patients, there are distinct groups
that can provide useful comparisons. While rap-
idly obtained autopsy material can serve as a non-
pathological control for anatomical studies, true
control material for physiological studies is ex-
tremely rare. The bulk of available tissue has the
pathological features of MTS which include pro-
found neuronal loss with the trend being CA1>
hilusCA3>CA2DG. In severely sclerotic hippo-
campi, up to 80% of neurons are lost; in a rep-
resentative series of patients, the mean neuronal
loss (for all hippocampal regions) was 64% nor-
malized to autopsy controls (Kim et al., 1990;
de Lanerolle et al., 2003). There is a concomitant
reactive gliosis in MTS hippocampi; however, the
role of reactive astrocytes in epileptic tissue is only
now being investigated (Binder and Steinhauser,
2006).
The best available comparison material for typ-
ical MTS patients is those patients with no evi-
dence of a lesion and where there is no atrophy
detectable on MRI. There is no consensus on the
terminology for this patient group. For this dis-
cussion, these patients will be referred to as par-
adoxical TLE (PTLE). This group of patients
provides the most reasonable comparison material
for the MTS tissue for several reasons. First, the
confounds of previous seizure experience and ex-
posure to anticonvulsants can be nullified with this
group. Second, there is little region-specific cell
loss and no evidence for synaptic reorganization.
See Cohen-Gadol et al. (2005) and de Lanerolle
et al. (2003) for detailed descriptions of this pa-
thology. Figure 1A shows quantitative cell counts
for autopsy controls, MTS and PTLE material in
several hippocampal regions. Note that there is no
significant difference between the cell counts for
the PTLE group and the non-pathological autopsy
controls in all four regions sampled.
A second comparison population includes
hippocampi from patients with extrahippocampal
lesions. In these cases, the hippocampus is re-
moved to limit seizure recurrence. This tissue is
referred to as Mass-Associated TLE (MaTLE).
The epileptogenic lesions include gliomas (usually
oligodendrogliomas or low grade astrocytomas),
cavernous malformations, and cysts (Beaumont
and Whittle, 2000). We have noted no physiolog-
ical differences in tissue from these two non-
sclerotic populations and therefore will group
them. For clarity, we will refer to these as com-
parison hippocampi. Similarly, granule cells from
MTS or comparison hippocampi will be referred
to as MTS or comparison granule cells.
Membrane properties of human dentate granule
cells
There are several published reports, which describe
the membrane properties of human dentate gran-
ule cells relative to rodents, conducted in slices.
Based on routinely obtained physiological meas-
ures, human granule cells can be successfully
recorded in slices and are comparable in their rest-
ing membrane potentials and input resistances to
those of rodents (Isokawa et al., 1991; Williamson
et al., 1993, 1999; Dietrich et al., 1999a). Thus,
any differences between rodent and human tissue
cannot be attributed solely to slice ‘‘health.’’
However, there are differences in the firing
properties of these cells that may be a feature of
human tissue. Williamson et al. (1993) showed
that, using sharp electrode recordings, there was
reduced spike frequency accommodation between
guinea pig and human granule cells from non-
sclerotic hippocampi. No other changes were
noted between the rodent and comparison tissue.
 185

Fig. 1. Characteristics of epileptic human dentate. A compares the cell loss in different pathologies between the dentate granule cells,
CA fields and the hilus. Note that the dentate is less affected than the CA1 region in the MTS tissue, and that there is little difference
between the autopsy controls and the comparison material. Data courtesy of Dr. Jung H. Kim. B shows the presence of graded
bursting activity in response to molecular layer stimulation (stimulus strength in mA). The arrows show examples of spontaneous
excitatory activity, which is another common physiological finding in MTS tissue that is rare in comparison material. The resting
membrane potential in the comparison granule cell was 58 mV and 55 in the MTS tissue. Spontaneous events are likely to be
glutamatergic, not GABAergic, because the RMP is depolarized to ECl.
186
These data suggest that there may be some differ-
ences in intracellular calcium dynamics between
the two species, but could also reflect changes due
to anticonvulsant drug use.
A similar study was carried out by Dietrich et al.
(1999) using MTS tissue. In this study, two pop-
ulations of granule cells were described that could
be discriminated by the characteristics of the fast
afterhyperpolarization (AHP) and the loss of a
medium duration AHP. There were predictable
differences in spike frequency adaptation between
the two groups of cells that correlated with the
extent of the AHPs. The majority of the cells that
had shorter AHPs also demonstrated synaptic
hyperexcitability (see below); however, 21% had
physiological features comparable to rodents
(and primates (St John et al., 1997)) indicating
that some of the granule cells in sclerotic hippo-
campi retain normal membrane and synaptic
properties.
Other biophysical properties of human hippo-
campal neurons have been studied using
acutely dissociated cells. This preparation is ad-
vantageous because it allows high quality voltage
clamp recordings; however, the possible effect of
the enzyme treatment on extracellular proteins is
a potential limitation. Using this approach pro-
vides the best means to address the properties of
Na+, Ca2+, and K+ channels in human granule
cells.
Na+ currents
Several studies have examined the properties of
Na+ currents in acutely isolated granule cells from
epileptic patients. It is notable that there were very
large currents in these cells compared to cells from
rat tissue or from non-focal neocortical tissue. The
other noteable feature of the Na+ currents was
that there was a component of the current that had
a slow recovery from inactivation (Reckziegel
et al., 1998). In a subsequent study of granule
cells from epileptic patients, it was shown that
there was a limited effect of the anticonvulsant
carbemazepine on these channels (Reckziegel
et al., 1999). Comparable studies have not been
conducted in non-epileptic cells, so it is difficult to
determine if these differences in Na+ channel
properties reflect treatment with anticonvulsants,
many of which act on Na+ channels. The changes
in the Na+ channel transcripts that have been
reported in human granule cells may account for
some of these physiological changes (Lombardo
et al., 1996).
K+ currents
With few exceptions, there are few differences
between the K+ channels in human granule cells
and those in rats. These currents could be sepa-
rated into at least four types by their kinetic and
pharmacological properties. These include at least
one voltage-dependent current similar to those
observed in mammalian hippocampal neurons,
and two Ca2+-dependent K+ currents that
most probably correspond to SK- and BK-type
currents. The sole significant difference was that a
classical A-type current could be detected in some
patients with MTS but not in comparison material
(Beck et al., 1996).
Ca2+ currents
Beck et al. (1998, 1999) and Nagerl et al. (2000)
demonstrated that granule cells express N-, L-,
and T-type Ca2+ channels based on their phar-
macological sensitivity. Similar to studies in
epileptic rats, the only significant change in these
properties was an increase in the T-type Ca
channel density. This increased density of a low
threshold Ca2+ current may lead to greater Ca2+
entry during synaptic activity, and thus modulate a
variety of systems, including GABAA receptor-
mediated inhibition (Isokawa, 1998).
HCN currents
Finally, the ability to generate rhythmic activity in
neurons can be regulated by the hyperpolariza-
tion-activated cyclic nucleotide gated current
(HCN). While a detailed analysis of this current
has not been performed in human tissue, there is
evidence for enhanced HCN mRNA in human

dentate granule cells from sclerotic hippocampi.
This finding parallels that seen in the rat (Bender
et al., 2003). The consequences of excess HCN
could lead to granule cells with physiological
properties closer to CA1 pyramidal cells,
i.e., more depolarized membrane potentials, and
cells that are more likely to have rebound activity
following recovery from a hyperpolarization.
Moreover, in conjunction with the enhanced T
channel density described above, granule cells
from MTS hippocampi will be more prone to
oscillatory activity, analogous to thalamic relay
cells. While prominent oscillations have not been
reported in human granule cells, subtle changes
in ion channels may allow for enhanced neural
synchronization.
Synaptic properties
Excitatory synaptic activity
A consistent observation in granule cells from
patients with MTS is that bursts, i.e., a depolar-
izing envelope with three or more action potentials
riding on it, can be evoked by stimulation in the
outer molecular layer. This stimulus will primarily
activate the perforant path, but will also activate
any recurrent mossy fibers. It is difficult to selec-
tively activate the medial vs. lateral perforant path
in the human, so we refer below to the site of
stimulation, rather than a specific fiber branch of
the perforant pathway. The percentage of granule
cells that display the stimulus-induced bursts
varies widely between patients, as well as in pub-
lished reports (Franck et al., 1995; Williamson
et al., 1995a; Cohen-Gadol et al., 2005). In our
sharp electrode recordings from 190 cells (75
patients), we noted bursting in 70/102 (68.6%)
cells from 41 patients with MTS. In contrast, this
type of activity was occurred in only 15/72 (20.5%)
cells recorded from comparison hippocampi.
However, it is important to note that we never
saw greater than a three action potential burst
in the comparison material while four or more
were routinely observed in the MTS tissue. Thus,
in a large sample, bursting is a characteristic of
hippocampi from patients with MTS, but is not
exclusively found in this group of patients.
187
Examples of this graded bursting activity are
shown in Figs. 1B and 2A.
The evoked bursting activity in slices is depend-
ent on NMDA receptor activation (Isokawa and
Levesque, 1991; Franck et al., 1995; Cohen-Gadol
et al., 2005). In our studies, we noted a 53.6% and
47.5% decrease in the area of the evoked response
in MTS granule cells following application of the
NMDA receptor antagonists amino phosphono-
valeric acid (APV; 30 mM) or Zn2+ (200 mM), re-
spectively. In contrast, there was only a 6.4%
change after APV and 18.5% decrease following
Zn2+in the comparison tissue (Williamson and
Spencer, 1995). These data are consistent with
voltage clamp recordings in MTS slices (Isokawa
and Levesque, 1991) showing an APV-sensitive
component of the evoked EPSC. In addition,
studies using acutely dissociated granule cells
described prolonged NMDA channel openings in
cells from MTS patients. In this study it was
noteable that calcineurin did not modulate the
open channel conductance in these cells, similar
to kindled rat granule cells (Lieberman and
Mody, 1999). This may be related to the loss of
calbindin D28 K in these cells (Magloczky et al.,
1997).
Role of mossy fiber sprouting in granule cell
excitability
Robust recurrent mossy fiber sprouting into the
inner third of the molecular layer in MTS hippo-
campi was described by a number of authors in the
early 1980s using a variety of techniques, including
Timm stain and dynorphin immunoreactivity
(de Lanerolle et al., 1989; Sutula et al., 1989;
Houser et al., 1990). This is not seen in the com-
parison tissue or in autopsy material. However,
defining the role of the sprouted mossy fibers
in epileptogenesis continues to remain controver-
sial (e.g., Patrylo and Dudek, 1998; Wuarin
and Dudek, 2001; Harvey and Sloviter, 2005;
Frotscher et al., 2006; Sloviter et al., 2006).
We have used a variety of approaches to inves-
tigate the possible effect of mossy fiber sprouting
on human granule cell excitability. As in rat mod-
els of TLE with mossy fiber sprouting, robust
188

Fig. 2. Granule cell hyperexcitability is associated with mossy fiber sprouting. A shows representative traces comparing the response to
molecular layer (orthodromic) and hilar (antidromic) stimulation in a patient with MTS. Note that there is a greater response to
maximal hilar stimulation than to stimulation of outer molecular layer. Moreover, note that hilar stimulation produced a synaptic
response prior to an antidromic spike. These data support the hypothesis that mossy fibers are primarily excitatory in resected tissue. B
shows Timm staining from the region of the dentate where the cell shown in A was recorded. Note the robust supergranular labeling
indicating the presence of sprouted fibers. (C1) Stimulation protocols have the problem of activating multiple fiber systems. Therefore,
in the presence of 30 mM bicuculline, we applied a bolus of glutamate to the granule cell layer at a distance from the recorded cell and
were able to evoke a long lasting barrage of EPSPs. The glutamate was applied at a distance from the recorded cell and thus we were
not observing direct glutamate effects. C2 shows that a similar response was not observed when the glutamate was applied to the hilus.
(See Color Plate 11.2 in color plate section.)

synaptic events can be evoked from granule which then form functional excitatory contacts
cells with hilar (antidromic) stimulation in MTS onto granule cells (Masukawa et al., 1992). An
hippocampi, and this is likely to represent the an- example comparing responses to orthodromic and
tidromic activation of recurrent mossy fibers, antidromic stimulation in the same cell, recorded

from sclerotic hippocampus, is shown in Fig. 2A.
This hippocampus was characterized by robust
mossy fiber sprouting, as shown in Fig. 2B. Note
that progressively longer and more complex events
occurred following the antidromic spike, consist-
ent with the hypothesis that there is a complex
network of recurrent fibers that can easily activate
a given granule cell.
A second approach was glutamate application,
using microdrops, to the granule cells located at a
distance from the recorded neuron. Presumably
the granule cells would be interconnected by
recurrent circuits in the MTS tissue, and the
GABAA receptor antagonist bicuculline (30 mM)
was added to ensure that inhibition would
not mask the effects of recurrent excitatory
circuits. As shown in Fig. 2C, the microdrop
application was followed by a transient increase in
the frequency of spontaneous activity (presumably
EPSPs). This type of response was not observed
when a glutamate was applied to the hilus.
These data demonstrate that there are robust
polysynaptic connections made by the recurrent
mossy fibers onto granule cells in MTS hippo-
campi.
We were unable to evoke synaptic activity with
hilar stimulation or with glutamate microdrops
applied to the granule cell layer in the dentate of
comparison hippocampi, supporting the hypothe-
sis that this activity is mediated by synaptic reor-
ganization. Moreover, they also indicate that the
excitatory mossy cells of the hilus are not able to
induce this type of graded excitation, presumably
because of the longitudinal orientation of their
axonal arbors. However, little is known about
mossy cells in the human dentate.
An important caveat to the results described
above is that effects of hilar stimulation, electrical
or chemical, could have been due to activation of
basal dendrites of granule cells. Primates are
known to have basal dendrites (see chapter by L.
Seress in this volume). Approximately 40% of hu-
man granule cells from epileptic tissue have been
reported to have basal dendrites, and they extend
into the hilus, where they could be activated by
either electrical or chemical stimulation (von
Campe et al., 1997). Basal dendrites do not exist
normally in the rat, and are greatly increased in
189
animal model of MTS (Ribak et al., 2000; Dash-
tipour et al., 2002). However, the data shown in
Fig. 2C suggest that activation of recurrent mossy
fibers is more effective in synchronizing granule
cell activity than activating granule cell basal
dendrites.
An additional clue to the role of mossy fiber
sprouting in generating hyperexcitability is pro-
vided by a distinct group of patients that have cell
loss predominantly in CA1 with less damage in the
dentate gyrus, hilus and CA3, and there is little to
no mossy fiber sprouting (as determined by dynor-
phin immunohistochemistry; (de Lanerolle et al.,
1997, 2003). In this group, the ability to generate
bursts with perforant path stimulation was com-
parable to that seen in typical MTS patients; how-
ever, there was significantly less asynchronous
synaptic activity and fewer spontaneous baseline
EPSPs (de Lanerolle et al., 1997). Therefore, only
specific aspects of the observed hyperexcitability
can be attributed to robust sprouting. We cannot
rule out a possible contribution of minimal sprout-
ing that was not detected by the immunostaining,
however.
In addition to sprouting, a common feature of
the MTS dentate is granule cell dispersion (Hous-
er, 1990). The mechanisms for this dispersion have
included deficiencies in Reelin (Haas et al., 2002;
Heinrich et al., 2006) leading to a broad cell body
layer. However, it appears that the wider granule
cell layer is linked with MTS, because the width
of the granule cell layers are correlated with the
degree of sprouting. However, total cell loss is
the greatest predictor of the density of sprouting
(El Bahh et al., 1999).
The enhanced width of the granule cell layer in
MTS hippocampi may reflect an increase in dent-
ate granule cell neurogenesis, which occurs after
seizures in laboratory animals, especially animal
models of TLE (Scharfman, 2004). Neural pro-
genitors have been described in the human dentate
gyrus (Roy et al., 2000) and thus it is likely that
different populations of granule cells are found in
the human hippocampus that are of different ages.
How the physiological differences in cells, related
to their age and degree of incorporation into the
hippocampal circuitry, has not been addressed in
human material.
190
GABAergic inhibition
The possible role of altered inhibition in epileptic
tissue has been extensively discussed in the liter-
ature (Bernard et al., 1998, 2000). A significant
problem in studying GABAergic inhibition in the
dentate gyrus is the enormous complexity of the
interneuronal systems found within the hilus and
in the dentate gyrus, with different populations of
interneurons subserving different functions and
contacting distinct regions of the granule cells
(Freund and Buzsaki, 1996).
There are numerous changes in interneuronal
networks in the MTS hippocampus compared to
either comparison tissue or autopsy controls (de
Lanerolle et al., 1989, 1992; Magloczky et al.,
2000; Wittner et al., 2002). The primary changes
include a loss of hilar interneurons that co-localize
specific peptides and GABA. Those cells immuno-
reactive for neuropeptide Y (NPY) and somatosta-
tin (SST) appear to be especially vulnerable
(de Lanerolle et al., 1992; Mathern et al., 1995),
while cells immunoreactive for parvalbumin and
calbindin are preserved (Magloczky et al., 2000).
However, despite these findings, there is no sig-
nificant loss of immunoreactivity for glutamic acid
decarboxylase, the GABA synthetic enzyme (Babb
et al., 1989; Mathern et al., 1995). Adding to
the complexity of describing the changes seen in
inhibitory systems are the alterations in fiber path-
ways. This includes sprouting of NPY, SS, and SP
systems (de Lanerolle et al., 1992; Mathern et al.,
1995). Many of these changes were recently
reviewed by Magloczky and Freund (2005).
Physiologically, several groups have reported a
decrease in evoked GABAA receptor-mediated in-
hibition in MTS hippocampi relative to compar-
ison material (Isokawa, 1998; Williamson et al.,
1999; Fig. 3). The degree of reduction is similar to
that seen in the pilocarpine rat model (Kobayashi
and Buckmaster, 2003). As with the membrane
properties, no significant differences in polysynap-
tically IPSP/C conductances have been reported in
parallel studies between control rats and the
comparison material in humans (Isokawa, 1996;
Williamson et al., 1999). It was interesting that
both polysynaptically and monosynaptically
evoked GABAA receptor-mediated events were
reduced to the same degree, suggesting that a
decreased afferent input onto the remaining inter-
neurons does not explain the decreased inhibitory
efficacy. The decreased evoked responses may be
to due the loss the SS and NPY immunoreactive
interneurons that target the dendritic arbor rather
than the cell body, because there was no change
in the density of parvalbumin-immunoreactive
terminals at the cell body of granule cells in
tissue removed from patients with MTS compared
to controls (Seress et al., 1993).
Finally, the decrease in evoked inhibition is not
due to changes in GABA receptor density or re-
sponsiveness. The studies by Shumate et al. (1998)
and Gibbs et al. (1996) clearly show that acutely
dissociated cells have an enhanced response to ap-
plied GABA and that the pharmacology of the
GABA responses demonstrate an increased sensi-
tivity to Zn2+, suggesting changes in the g subunit
composition of these receptors. These pharmaco-
logical experiments are consistent with the mRNA
studies from single granule cells showing a signifi-
cant increase in the GABA a5 receptor subunit
(Brooks-Kayal et al., 1999). Later studies revealed
that there is significant diversity in the patterns of
GABAA receptors expressed by different granule
cells in pediatric patients with intractable TLE
(Porter et al., 2005). Thus, the specific character-
istics of GABA receptor-mediated responses may
be quite heterogeneous between patients.
Taken together, these data indicate that, like the
changes in excitatory systems, there is a complex
interplay between changes in pre- and postsynaptic
elements. The decrease in evoked (polysynaptic
and monosynaptic) IPSPs could reflect presynap-
tic modulation of GABA release, or changes in
the amount of GABA available for release despite
increases in postsynaptic receptor density.
Synaptic plasticity
Several forms of synaptic plasticity are expressed
in the dentate gyrus and have been intensively
studied in rodents. These include long-term po-
tentiation (LTP), induced by brief periods of high
frequency stimulation, and long-term depression
(LTD), induced by low frequency stimuli. While
 191

Fig. 3. Decreased poly- and monosynaptic GABAA receptor-mediated inhibition in MTS hippocampi. We noted significant decreases
in molecular layer-evoked inhibitory responses. A shows events delivered in normal medium; the stimulus intensity was set at twice that
needed to evoke a single action potential. The d in A and B indicate the points at which the fast and slow IPSPs were measured. B
shows monosynaptic responses evoked in presence of APV and CNQX (30 mM each). The stimulus intensity was set to the greatest that
did not evoke an antidromic action potential. C and D shows grouped IPSP conductance data for polysynaptically and monosy-
naptically-evoked IPSPs, respectively. Note the profound and significant decrease in the IPSP conductance in the MTS tissue relative
to comparison material. po0.05; po0.001.

the link between learning and memory and these
forms of synaptic plasticity remains controversial,
the availability of human material provides an ad-
ditional avenue to examine this issue, because
memory deficits are a common clinical feature of
MTS (Helmstaedter et al., 2003).
Both LTP and LTD can be reliably obtained in
slices from epileptic human temporal neocortex
with pharmacology comparable to that seen in
rodents (Chen et al., 1996). However, both forms
of synaptic plasticity are impaired in patients with
MTS relative to patients without this pathology
(Beck et al., 2000). Moreover, as shown in Fig. 4,
there is a direct correlation between the degree of
LTP impairment and a measure of neuropsycho-
logical function, the selective reminding task
(SRT) (Bell et al., 2005). For these studies,
200 Hz stimuli were delivered in a theta burst
pattern and the change in the response was meas-
ured 30 min following the tetani. This value was
then correlated with each patient’s SRT z score on
the selective reminding task, a measure of laterali-
zed hippocampal function. Taken together, these
data support the hypothesis that altered synaptic
plasticity is associated with the cognitive dysfunc-
tion in MTS patients.
Neurotransmitter transport
Another aspect of neurotransmission that has been
investigated in the human dentate is GABA
192

Fig. 4. Altered synaptic plasticity in the hippocampi of patients with MTS. A shows examples of field PSPs recorded in the molecular
layer of a MTS patient. Note that there was pronounced depression of the evoked response following five trains of 100 Hz stimuli (100
stimuli/train delivered at 30 s intervals). B1 shows the group data for seven patients and B2 shows the cumulative probability plot
demonstrating that depression was the predominant response. C shows that those patients that did not express LTD with this
stimulation protocol had normal SRT z scores (control is set to 0) while those that did show LTD had markedly lower SRT scores.

uptake, commonly assessed using focal GABA
application (Williamson et al., 1995b; Patrylo
et al., 2001). These studies showed prolonged
responses to GABA in the MTS dentate, which
was not evident in the CA2 region (Telfeian et al.,
1999). Moreover, there was a blunted effect
of NO-711, a selective GAT-1 inhibitor, in the
MTS tissue compared to comparison material.
Finally, the responses to NPA, a GAT inhibitor
that induces heterotransport (Honmou et al.,
1995), were also reduced in MTS hippocampi.
These data were interpreted as impaired GABA
uptake. Parallel findings were made in the kainic
acid rat model of TLE (Patrylo et al., 2001); how-
ever, see Frahm et al. (2003). GAT-1 is found both
on neurons and on astrocytes, so these data do
not address the role of gliosis in the regulation
of GABA uptake. These physiological data are
somewhat consistent with anatomical data show-
ing a loss of GAT-1 around granule cell somata
(where the GABA was applied), but an increase in
the dendrites (Mathern et al., 1999; Lee et al.,
2006).
There are a number of possible consequences of
altered GABA transport, which could affect the
overall excitability of the dentate. For example,
granule cells with the GABAA d receptor subunit
could be involved in tonic inhibition (Peng et al.,
2004). Moreover, a slowed GABA clearance will
also allow for binding to GABAB receptors, many
of which are presynaptic, and can limit neuro-
transmitter release (Sperk et al., 2004). Conversely,
there may be a reduced efficacy of GABA in
this tissue due to GABA receptor desensitization.
Finally, under conditions of high interneuronal
activity, a reduction in neuronal uptake could
limit the availability of GABA for repackaging
and release. Which of these possibilities dominates
under normal conditions has yet to be resolved.
Glutamate transport is somewhat more com-
plex. The bulk of glutamate uptake is mediated
by glial transporters (Danbolt, 2001) and there is
no real consensus about changes in MTS vs. non-
MTS hippocampi (see Binder and Steinhauser,
2006 for review). For example, there is little
evidence for changes in glial transporter density

when studied with Western blot (e.g., Eid et al.,
2004), however, immunohistochemical studies
show decreases in EAAT-2 (Mathern et al., 1999;
Proper et al., 2002). However, there are no pub-
lished studies measuring glial glutamate transport
in human astrocytes that can shed light on these
conflicting anatomical findings.
Neuromodulation
There is now anatomical and physiological evi-
dence for multiple neuromodulatory systems in the
human dentate gyrus. Three systems will be de-
scribed here, with the understanding that there are
numerous other modulators that may be impor-
tant in regulating the excitatory tone of the human
dentate.
Metabotropic glutamate receptors (mGluRs)
Several lines of evidence suggest that excitability in
epileptic human tissue is constrained by metabo-
tropic glutamate receptors. First, extracellular
glutamate levels are very high in MTS hippocampi
(Cavus et al., 2005) so that there is ample trans-
mitter available to bind mGluRs. Second, Dietrich
et al. (1999b) demonstrated that there is a reduced
efficacy of L-AP4 (a type 2 mGluR agonist) in
MTS tissue vs. comparison material. Similar find-
ings were reported by Patrylo et al. (1999). One
possibility is that there is little additional effect of
mGluR agonists because the receptors are satu-
rated. As shown in Fig. 5, there is relatively little
effect of L-AP4 compared to other presynaptic
modulators, such as NPY.
The anatomic distribution of mGluR5 has been
studied in the human dentate and a significant
upregulation of this postsynaptic receptor has
been observed by several investigators (Tang et
al., 2001; Notenboom et al., 2006). However, the
specific physiological effects of mGluR5 receptors
in human tissue have not been investigated.
Neuropeptides
As in rodents, there is tremendous diversity
within the interneuronal population in terms of
193
neuropeptide expression and numerous changes
in peptide distributions have been observed as
described above.
Fewer physiological studies have been per-
formed relative to the anatomical ones. The best
studied peptide is NPY which, as in rodent studies,
is a potent modulator of excitatory activity
(rodent, Patrylo et al., 1999). While there is a
loss of NPY-immunopositive cells in MTS hippo-
campi, there is evidence for axonal sprouting of
NPY-immunoreactive (IR) fibers that cover the
entire molecular layer, instead of being limited to
the inner third of the molecular layer (de Lanerolle
et al., 1989). When NPY was exogenously applied
to human granule cells, there were no changes
in the membrane properties, but a significant
decrease in the evoked response. It was notable
that NPY had a greater effect in cells with greater
excitability: there was a 60% decrease in the
evoked response when there were three or more
evoked action potentials, but only a 34% decrease
in cells where only one action potential could be
evoked (Patrylo et al., 1999). This is consistent
with the observations that NPY receptors are
expressed on mossy fibers, but not on the perfo-
rant path inputs (Klapstein and Colmers, 1993),
suggesting that there is a greater effect in those
hippocampi with robust synaptic reorganization.
Another peptide that is reorganized in MTS
hippocampi is somatostatin. The SS-IR hilar
interneurons appear to be highly susceptible to
seizures in both human MTS and animal models
(Robbins et al., 1991; Buckmaster and Schwa-
rtzkroin, 1995; Vezzani and Hoyer, 1999). As with
NPY, the primary effects of SS are on excitatory,
rather than inhibitory neurotransmission (Tallent
and Siggins, 1997). SS receptors are expressed at
specific synapses such as associational-comissural
fibers, but not on perforant path fibers. As shown
in Fig. 5A, there was only a modest effect of SS on
orthodromic and evoked hyperexcitability. The
ability of SS to reduce hyperexcitability was much
less than that of NPY (Fig. 5B). However, this
may reflect the downregulation of postsynaptic
SS receptors in the outer dendritic regions MTS
hippocampi (Csaba et al., 2005).
These are two examples of the physiological
effects numerous peptides that are known to exist
194

Fig. 5. Modulation of granule cell synaptic response by the neuropeptides NPY and somatostatin and glutamate receptor antagonists.
A shows that NPY (1 mM) dramatically reduced the orthodromic response in an MTS case. In contrast, somatostatin (SS) had a less
robust effect on the evoked response. Grouped data showing the percent change for NPY and SS are shown in C1. The effect of NPY
was significantly greater than that of SS (po0.05). C2 shows the percent change in area mediated by glutamate receptor antagonists
APV (30 mM) and the mGluR group 2 antagonist L-AP4 (100 mM) in MTS tissue.

in hilar interneurons and which exhibit some re-
organization. Much more work will need to be
done to determine how these changes in both fiber
and receptor distributions will affect the overall
synaptic output. It is also important to note that
there are significant differences between humans
and rodents in many of these systems, and thus
data from rodent models may not accurately pre-
dict the situation in the human (Magloczky et al.,
2000).
Metabolism
A consistent feature of epileptic tissue is hypome-
tabolism. This has been assessed in vivo using a
variety of methods including glucose or O2 use
with PET and SPECT, as well as phosphocreatine
and NAA (a neuronal mitochondrial marker) with
MRS (Casse et al., 2002; Van Paesschen, 2004a, b;
Pan et al., 2005). However, none of these in vivo
approaches allows for a regional assessment of
metabolism.
The cellular basis for this hypometabolism
remains somewhat unclear; however, several
recent studies have begun to examine the issue.
In the elegant studies of Kann et al. (2005),
electrical stimulation was used to metabolically
activate the tissue and the changes in the levels
of NAD(P)H autofluorescence, as well as in mi-
tochondrial potential, were studied. These exper-
iments demonstrated that there was a pronounced
drop in NADH fluorescence and a smaller over-
shoot during stimulation in the dentate as well as
in other hippocampal areas. Despite this effect,
fluorescence imaging of mitochondrial potentials

indicated that individual mitochondria exhibited
normal voltage changes to these stimuli. Overall,
these results were interpreted to indicate reduced
mitochondrial function, because small overshoots
are associated with decreased efficacy of mi-
tochondrial NAD(P)H+ reduction. An alternate
explanation is that these changes reflect the altered
neuronal-glial metabolic coupling know to exist in
human TLE (Petroff et al., 2002). However, it is
important to note that the dentate is less metabol-
ically impaired in MTS than other hippocampal
regions, due in part to the differences in neuronal
loss between the CA fields and the dentate (Kunz
et al., 2000).
Despite these imaging studies, there have been
few studies on the interaction between metabolism
and neuronal function in either animals or hu-
mans. Our initial studies in this field demonstrated
that there were strong correlations between the
levels of phosphocreatine and the extent of neu-
ronal bursting as well as with the IPSP conduct-
ance. For both variables, lower PCr was associated
with greater excitability (Williamson et al., 2005).
Finally, together with the mitochondrial imaging
data, these results suggest that the intimate rela-
tionship between excitability and energetics may
be altered in specific ways in MTS hippocampi.
Conclusions
While physiological/functional studies of human
tissue are restricted to surgical specimens, the va-
rieties in pathology allow for some conclusions to
be made about the changes that are specific to
MTS. These include robust synaptic reorganiza-
tion that promotes hyperexcitability, and a de-
crease in the efficacy of evoked GABA inhibition.
The alterations in GABAergic and neuromodula-
tory systems are quite complex and the specific
consequences of these changes will be difficult to
separate. However, because there are distinct
differences between these systems in humans and
in animal models (Magloczky and Freund, 2005),
these studies will require human material. The
availability of viable human specimens allows us
the unique opportunity to study the physiological
changes associated with TLE and provide insight
195
into other pathologies with hippocampal involve-
ment, such as Alzheimer’s disease.
References
Babb, T.L., Pretorius, J.K., Kupfer, W.R. and Crandall, P.H.
(1989) Glutamate decarboxylase-immunoreactive neurons
are preserved in human epileptic hippocampus. J. Neurosci.,
9: 2562–2574.
Beaumont, A. and Whittle, I.R. (2000) The pathogenesis of
tumour associated epilepsy. Acta Neurochir. (Wien.), 142:
1–15.
Beck, H., Blumcke, I., Kral, T., Clusmann, H., Schramm, J.,
Wiestler, O.D., Heinemann, U. and Elger, C.E. (1996) Prop-
erties of a delayed rectifier potassium current in dentate
granule cells isolated from the hippocampus of patients with
chronic temporal lobe epilepsy. Epilepsia, 37: 892–901.
Beck, H., Steffens, R., Elger, C.E. and Heinemann, U. (1998)
Voltage-dependent Ca2+ currents in epilepsy. Epilepsy Res,
32: 321–332.
Beck, H., Steffens, R., Heinemann, U. and Elger, C.E. (1999)
Ca(2+)-dependent inactivation of high-threshold Ca(2+)
currents in hippocampal granule cells of patients with chronic
temporal lobe epilepsy. J Neurophysiol, 82: 946–954.
Bell, B.D., Fine, J., Dow, C., Seidenberg, M. and Hermann,
B.P. (2005) Temporal lobe epilepsy and the selective remind-
ing test: the conventional 30-minute delay suffices. Psychol.
Assess., 17: 103–109.
Bender, R.A., Soleymani, S.V., Brewster, A.L., Nguyen, S.T.,
Beck, H., Mathern, G.W. and Baram, T.Z. (2003) Enhanced
expression of a specific hyperpolarization-activated cyclic
nucleotide-gated cation channel (HCN) in surviving dentate
gyrus granule cells of human and experimental epileptic hip-
pocampus. J. Neurosci., 23: 6826–6836.
Bernard, C., Cossart, R., Hirsch, J.C., Esclapez, M. and
Ben-Ari, Y. (2000) What is GABAergic inhibition? How is it
modified in epilepsy? Epilepsia, 41(Suppl 6): S90–S95.
Bernard, C., Esclapez, M., Hirsch, J.C. and Ben-Ari, Y. (1998)
Interneurones are not so dormant in temporal lobe epilepsy:
a critical reappraisal of the dormant basket cell hypothesis.
Epilepsy Res., 32: 93–103.
Binder, D.K. and Steinhauser, C. (2006) Functional changes in
astroglial cells in epilepsy. Glia, 54: 358–368.
Brooks-Kayal, A.R., Shumate, M.D., Jin, H., Lin, D.D.,
Rikhter, T.Y., Holloway, K.L. and Coulter, D.A. (1999)
Human neuronal gamma-aminobutyric acid(A) receptors:
coordinated subunit mRNA expression and functional
correlates in individual dentate granule cells. J. Neurosci.,
19: 8312–8318.
Buckmaster, P.S. and Schwartzkroin, P.A. (1995) Interneurons
and inhibition in the dentate gyrus of the rat in vivo.
J. Neurosci., 15: 774–789.
Casse, R., Rowe, C.C., Newton, M., Berlangieri, S.U.
and Scott, A.M. (2002) Positron emission tomography and
epilepsy. Mol. Imag. Biol., 4: 338–351.
196
Cavus, I., Kasoff, W.S., Cassaday, M.P., Jacob, R., Gueor-
guieva, R., Sherwin, R.S., Krystal, J.H., Spencer, D.D. and
Abi-Saab, W.M. (2005) Extracellular metabolites in the cortex
and hippocampus of epileptic patients. Ann. Neurol., 57:
226–235.
Chen, W.R., Lee, S., Kato, K., Spencer, D.D., Shepherd, G.M.
and Williamson, A. (1996) Long-term modifications of
synaptic efficacy in the human inferior and middle temporal
cortex. Proc. Natl. Acad. Sci. U.S.A., 93: 8011–8015.
Cohen-Gadol, A.A., Bradley, C.C., Williamson, A., Kim, J.H.,
Westerveld, M., Duckrow, R.B. and Spencer, D.D. (2005)
Normal magnetic resonance imaging and medial temporal
lobe epilepsy: the clinical syndrome of paradoxical temporal
lobe epilepsy. J. Neurosurg., 102: 902–909.
Csaba, Z., Pirker, S., Lelouvier, B., Simon, A., Videau, C.,
Epelbaum, J., Czech, T., Baumgartner, C., Sperk, G. and
Dournaud, P. (2005) Somatostatin receptor type 2 undergoes
plastic changes in the human epileptic dentate gyrus. J. Ne-
uropathol. Exp. Neurol., 64: 956–969.
Danbolt, N.C. (2001) Glutamate uptake. Prog. Neurobiol., 65:
1–105.
Dashtipour, K., Yan, X.X., Dinh, T.T., Okazaki, M.M.,
Nadler, J.V. and Ribak, C.E. (2002) Quantitative and mor-
phological analysis of dentate granule cells with recurrent
basal dendrites from normal and epileptic rats. Hippocam-
pus, 12: 235–244.
Dietrich, D., Clusmann, H., Kral, T., Steinhauser, C., Blumcke,
I., Heinemann, U. and Schramm, J. (1999a) Two electro-
physiologically distinct types of granule cells in epileptic hu-
man hippocampus. Neuroscience, 90: 1197–1206.
Dietrich, D., Kral, T., Clusmann, H., Friedl, M. and Schramm,
J. (1999b) Reduced function of L-AP4-sensitive metabotropic
glutamate receptors in human epileptic sclerotic hippocam-
pus. Eur. J. Neurosci., 11: 1109–1113.
Eid, T., Thomas, M.J., Spencer, D.D., Runden-Pran, E., Lai,
J.C., Malthankar, G.V., Kim, J.H., Danbolt, N.C., Ottersen,
O.P. and de Lanerolle, N.C. (2004) Loss of glutamine syn-
thetase in the human epileptogenic hippocampus: possible
mechanism for raised extracellular glutamate in mesial tem-
poral lobe epilepsy. Lancet, 363: 28–37.
El Bahh, B., Lespinet, V., Lurton, D., Coussemacq, M., Le Gal
La Salle, G. and Rougier, A. (1999) Correlations between
granule cell dispersion, mossy fiber sprouting, and hippo-
campal cell loss in temporal lobe epilepsy. Epilepsia, 40:
1393–1401.
Frahm, C., Stief, F., Zuschratter, W. and Draguhn, A. (2003)
Unaltered control of extracellular GABA-concentration
through GAT-1 in the hippocampus of rats after pilocar-
pine-induced status epilepticus. Epilepsy Res., 52: 243–252.
Franck, J.E., Pokorny, J., Kunkel, D.D. and Schwartzkroin,
P.A. (1995) Physiologic and morphologic characteristics of
granule cell circuitry in human epileptic hippocampus. Epile-
psia, 36: 543–558.
Freund, T.F. and Buzsaki, G. (1996) Interneurons of the hip-
pocampus. Hippocampus, 6: 347–470.
Frotscher, M., Jonas, P. and Sloviter, R.S. (2006) Synapses
formed by normal and abnormal hippocampal mossy fibers.
Cell Tissue Res, 326: 361–367.
Gibbs III, J.W., Zhang, Y.F., Kao, C.Q., Holloway, K.L., Oh,
K.S. and Coulter, D.A. (1996) Characterization of GABAA
receptor function in human temporal cortical neurons. J.
Neurophysiol., 75: 1458–1471.
Haas, C.A., Dudeck, O., Kirsch, M., Huszka, C., Kann, G.,
Pollak, S., Zentner, J. and Frotscher, M. (2002) Role for
reelin in the development of granule cell dispersion in tem-
poral lobe epilepsy. J. Neurosci., 22: 5797–5802.
Harvey, B.D. and Sloviter, R.S. (2005) Hippocampal granule
cell activity and c-Fos expression during spontaneous sei-
zures in awake, chronically epileptic, pilocarpine-treated rats:
implications for hippocampal epileptogenesis. J. Comp. Ne-
urol., 488: 442–463.
Heinrich, C., Nitta, N., Flubacher, A., Muller, M., Fahrner, A.,
Kirsch, M., Freiman, T., Suzuki, F., Depaulis, A., Frotscher,
M. and Haas, C.A. (2006) Reelin deficiency and displacement
of mature neurons, but not neurogenesis, underlie the for-
mation of granule cell dispersion in the epileptic hippocam-
pus. J. Neurosci., 26: 4701–4713.
Helmstaedter, C., Kurthen, M., Lux, S., Reuber, M. and Elger,
C.E. (2003) Chronic epilepsy and cognition: a longitudinal
study in temporal lobe epilepsy. Ann. Neurol., 54: 425–432.
Honmou, O., Kocsis, J.D. and Richerson, G.B. (1995) Gaba-
pentin potentiates the conductance increase induced by nip-
ecotic acid in CA1 pyramidal neurons in vitro. Epilepsy Res.,
20: 193–202.
Houser, C.R. (1990) Granule cell dispersion in the dentate
gyrus of humans with temporal lobe epilepsy. Brain Res.,
535: 195–204.
Houser, C.R., Miyashiro, J.E., Swartz, B.E., Walsh, G.O.,
Rich, J.R. and Delgado-Escueta, A.V. (1990) Altered pat-
terns of dynorphin immunoreactivity suggest mossy fiber re-
organization in human hippocampal epilepsy. J. Neurosci.,
10: 267–282.
Isokawa, M. (1996) Decrement of GABAA receptor-mediated
inhibitory postsynaptic currents in dentate granule cells in
epileptic hippocampus. J. Neurophysiol., 75: 1901–1908.
Isokawa, M. (1998) Modulation of GABAA receptor-mediated
inhibition by postsynaptic calcium in epileptic hippocampal
neurons. Brain Res., 810: 241–250.
Isokawa, M., Avanzini, G., Finch, D.M., Babb, T.L. and Le-
vesque, M.F. (1991) Physiologic properties of human dentate
granule cells in slices prepared from epileptic patients. Ep-
ilepsy Res., 9: 242–250.
Isokawa, M. and Levesque, M.F. (1991) Increased NMDA re-
sponses and dendritic degeneration in human epileptic hip-
pocampal neurons in slices. Neurosci. Lett., 132: 212–216.
Kann, O., Kovacs, R., Njunting, M., Behrens, C.J., Otahal, J.,
Lehmann, T.N., Gabriel, S. and Heinemann, U. (2005) Met-
abolic dysfunction during neuronal activation in the ex vivo
hippocampus from chronic epileptic rats and humans. Brain,
128: 2396–2407.

Kim, J.H., Guimaraes, P.O., Shen, M.Y., Masukawa, L.M.
and Spencer, D.D. (1990) Hippocampal neuronal density in
temporal lobe epilepsy with and without gliomas. Acta
Neuropathol. (Berl.), 80: 41–45.
Klapstein, G.J. and Colmers, W.F. (1993) On the sites of pre-
synaptic inhibition by neuropeptide Y in rat hippocampus in
vitro. Hippocampus, 3: 103–111.
Kobayashi, M. and Buckmaster, P.S. (2003) Reduced inhibition
of dentate granule cells in a model of temporal lobe epilepsy.
J. Neurosci., 23: 2440–2452.
Kunz, W.S., Kudin, A.P., Vielhaber, S., Blumcke, I., Zusch-
ratter, W., Schramm, J., Beck, H. and Elger, C.E. (2000)
Mitochondrial complex I deficiency in the epileptic focus of
patients with temporal lobe epilepsy. Ann. Neurol., 48:
766–773.
de Lanerolle, N.C., Brines, M., Williamson, A., Kim, J.H. and
Spencer, D.D. (1992) Neurotransmitters and their receptors
in human temporal lobe epilepsy. Epilepsy Res. Suppl., 7:
235–250.
de Lanerolle, N.C., Kim, J.H., Robbins, R.J. and Spencer,
D.D. (1989) Hippocampal interneuron loss and plasticity in
human temporal lobe epilepsy. Brain Res., 495: 387–395.
de Lanerolle, N.C., Kim, J.H., Williamson, A., Spencer, S.S.,
Zaveri, H.P., Eid, T. and Spencer, D.D. (2003) A retrospec-
tive analysis of hippocampal pathology in human temporal
lobe epilepsy: evidence for distinctive patient subcategories.
Epilepsia, 44: 677–687.
de Lanerolle, N.C., Williamson, A., Meredith, C., Kim, J.H.,
Tabuteau, H., Spencer, D.D. and Brines, M.L. (1997) Dynor-
phin and the kappa 1 ligand [3H]U69,593 binding in the hu-
man epileptogenic hippocampus. Epilepsy Res., 28: 189–205.
Lee, T.S., Bjornsen, L.P., Paz, C., Kim, J.H., Spencer, S.S.,
Spencer, D.D., Eid, T. and de Lanerolle, N.C. (2006) GAT1
and GAT3 expression are differently localized in the human
epileptogenic hippocampus. Acta Neuropathol. (Berl.), 111:
351–363.
Lieberman, D.N. and Mody, I. (1999) Properties of single
NMDA receptor channels in human dentate gyrus granule
cells. J. Physiol., 518(Pt 1): 55–70.
Lombardo, A.J., Kuzniecky, R., Powers, R.E. and Brown, G.B.
(1996) Altered brain sodium channel transcript levels in hu-
man epilepsy. Brain Res. Mol. Brain Res., 35: 84–90.
Magloczky, Z. and Freund, T.F. (2005) Impaired and repaired
inhibitory circuits in the epileptic human hippocampus.
Trends Neurosci., 28: 334–340.
Magloczky, Z., Halasz, P., Vajda, J., Czirjak, S. and Freund,
T.F. (1997) Loss of Calbindin-D28K immunoreactivity from
dentate granule cells in human temporal lobe epilepsy. Ne-
uroscience, 76: 377–385.
Magloczky, Z., Wittner, L., Borhegyi, Z., Halasz, P., Vajda, J.,
Czirjak, S. and Freund, T.F. (2000) Changes in the distribu-
tion and connectivity of interneurons in the epileptic human
dentate gyrus. Neuroscience, 96: 7–25.
Masukawa, L.M., Uruno, K., Sperling, M., O’connor, M.J. and
Burdette, L.J. (1992) The functional relationship between
197
antidromically evoked field responses of the dentate gyrus
and mossy fiber reorganization in temporal lobe epileptic
patients. Brain Res., 579: 119–127.
Mathern, G.W., Babb, T.L., Pretorius, J.K. and Leite, J.P.
(1995) Reactive synaptogenesis and neuron densities for ne-
uropeptide Y, somatostatin, and glutamate decarboxylase
immunoreactivity in the epileptogenic human fascia dentata.
J. Neurosci., 15: 3990–4004.
Mathern, G.W., Mendoza, D., Lozada, A., Pretorius, J.K.,
Dehnes, Y., Danbolt, N.C., Nelson, N., Leite, J.P., Chimelli,
L., Born, D.E., Sakamoto, A.C., Assirati, J.A., Fried, I.,
Peacock, W.J., Ojemann, G.A. and Adelson, P.D. (1999)
Hippocampal GABA and glutamate transporter immunore-
activity in patients with temporal lobe epilepsy. Neurology,
52: 453–472.
Nagerl, U.V., Mody, I., Jeub, M., Lie, A.A., Elger, C.E. and
Beck, H. (2000) Surviving granule cells of the sclerotic human
hippocampus have reduced Ca(2+) influx because of a loss
of calbindin-D(28k) in temporal lobe epilepsy. J Neurosci.,
20: 1831–1836.
Notenboom, R.G., Hampson, D.R., Jansen, G.H., Van Rijen,
P.C., Van Veelen, C.W., Van Nieuwenhuizen, O. and De
Graan, P.N. (2006) Up-regulation of hippocampal metabo-
tropic glutamate receptor 5 in temporal lobe epilepsy pa-
tients. Brain, 129: 96–107.
Pan, J.W., Kim, J.H., Cohen-Gadol, A., Pan, C., Spencer, D.D.
and Hetherington, H.P. (2005) Regional energetic dysfunc-
tion in hippocampal epilepsy. Acta Neurol. Scand., 111:
218–224.
Patrylo, P.R. and Dudek, F.E. (1998) Physiological unmasking
of new glutamatergic pathways in the dentate gyrus of hip-
pocampal slices from kainate-induced epileptic rats. J. Ne-
urophysiol., 79: 418–429.
Patrylo, P.R., Spencer, D.D. and Williamson, A. (2001) GABA
uptake and heterotransport are impaired in the dentate gyrus
of epileptic rats and humans with temporal lobe sclerosis. J.
Neurophysiol., 85: 1533–1542.
Patrylo, P.R., Van Den Pol, A.N., Spencer, D.D. and Will-
iamson, A. (1999) NPY inhibits glutamatergic excitation in
the epileptic human dentate gyrus. J. Neurophysiol., 82:
478–483.
Peng, Z., Huang, C.S., Stell, B.M., Mody, I. and Houser, C.R.
(2004) Altered expression of the delta subunit of the GABAA
receptor in a mouse model of temporal lobe epilepsy. J. Ne-
urosci., 24: 8629–8639.
Petroff, O.A., Errante, L.D., Rothman, D.L., Kim, J.H. and
Spencer, D.D. (2002) Glutamate-glutamine cycling in the
epileptic human hippocampus. Epilepsia, 43: 703–710.
Porter, B.E., Zhang, G., Celix, J., Hsu, F.C., Raol, Y.H., Tel-
feian, A., Gallagher, P.R., Coulter, D.A. and Brooks-Kayal,
A.R. (2005) Heterogeneous GABAA receptor subunit ex-
pression in pediatric epilepsy patients. Neurobiol. Dis., 18:
484–491.
Proper, E.A., Hoogland, G., Kappen, S.M., Jansen, G.H.,
Rensen, M.G., Schrama, L.H., Van Veelen, C.W., Van Rijen,
198
P.C., Van Nieuwenhuizen, O., Gispen, W.H. and De Graan,
P.N. (2002) Distribution of glutamate transporters in the
hippocampus of patients with pharmaco-resistant temporal
lobe epilepsy. Brain, 125: 32–43.
Reckziegel, G., Beck, H., Schramm, J., Elger, C.E. and Urban,
B.W. (1998) Electrophysiological characterization of Na+
currents in acutely isolated human hippocampal dentate
granule cells. J. Physiol., 509(Pt 1): 139–150.
Reckziegel, G., Beck, H., Schramm, J., Urban, B.W. and Elger,
C.E. (1999) Carbamazepine effects on Na+ currents in hu-
man dentate granule cells from epileptogenic tissue. Epile-
psia, 40: 401–407.
Ribak, C.E., Tran, P.H., Spigelman, I., Okazaki, M.M. and
Nadler, J.V. (2000) Status epilepticus-induced hilar basal
dendrites on rodent granule cells contribute to recurrent ex-
citatory circuitry. J. Comp. Neurol., 428: 240–253.
Robbins, R.J., Brines, M.L., Kim, J.H., Adrian, T., de Lane-
rolle, N., Welsh, S. and Spencer, D.D. (1991) A selective loss
of somatostatin in the hippocampus of patients with tempo-
ral lobe epilepsy. Ann. Neurol., 29: 325–332.
Roy, N.S., Wang, S., Jiang, L., Kang, J., Benraiss, A., Harri-
son-Restelli, C., Fraser, R.A., Couldwell, W.T., Kawaguchi,
A., Okano, H., Nedergaard, M. and Goldman, S.A. (2000) In
vitro neurogenesis by progenitor cells isolated from the adult
human hippocampus. Nat. Med., 6: 271–277.
Scharfman, H.E. (2004) Functional implications of seizure-in-
duced neurogenesis. Adv. Exp. Med. Biol., 548: 192–212.
Schwartzkroin, P.A. and Prince, D.A. (1976) Microphysiology
of human cerebral cortex studied in vitro. Brain Res., 115:
497–500.
Seress, L., Gulyas, A.I., Ferrer, I., Tunon, T., Soriano, E. and
Freund, T.F. (1993) Distribution, morphological features,
and synaptic connections of parvalbumin- and calbindin
D28k-immunoreactive neurons in the human hippocampal
formation. J. Comp. Neurol., 337: 208–230.
Shumate, M.D., Lin, D.D., Gibbs III, J.W., Holloway, K.L.
and Coulter, D.A. (1998) GABA(A) receptor function in ep-
ileptic human dentate granule cells: comparison to epileptic
and control rat. Epilepsy Res., 32: 114–128.
Sloviter, R.S., Zappone, C.A., Harvey, B.D. and Frotscher, M.
(2006) Kainic acid-induced recurrent mossy fiber innervation
of dentate gyrus inhibitory interneurons: possible anatomical
substrate of granule cell hyper-inhibition in chronically ep-
ileptic rats. J. Comp. Neurol., 494: 944–960.
Sperk, G., Furtinger, S., Schwarzer, C. and Pirker, S. (2004)
GABA and its receptors in epilepsy. Adv. Exp. Med. Biol.,
548: 92–103.
St John, J.L., Rosene, D.L. and Luebke, J.I. (1997) Morphol-
ogy and electrophysiology of dentate granule cells in the
rhesus monkey: comparison with the rat. J. Comp. Neurol.,
387: 136–147.
Sutula, T., Cascino, G., Cavazos, J., Parada, I. and Ramirez, L.
(1989) Mossy fiber synaptic reorganization in the epileptic
human temporal lobe. Ann. Neurol., 26: 321–330.
Tallent, M.K. and Siggins, G.R. (1997) Somatostatin depresses
excitatory but not inhibitory neurotransmission in rat CA1
hippocampus. J. Neurophysiol., 78: 3008–3018.
Tang, F.R., Lee, W.L. and Yeo, T.T. (2001) Expression of the
group I metabotropic glutamate receptor in the hippocampus
of patients with mesial temporal lobe epilepsy. J. Ne-
urocytol., 30: 403–411.
Telfeian, A.E., Spencer, D.D. and Williamson, A. (1999) Lack
of correlation between neuronal hyperexcitability and elect-
rocorticographic responsiveness in epileptogenic human
neocortex. J. Neurosurg., 90: 939–945.
Van Paesschen, W. (2004a) Ictal SPECT. Epilepsia, 45(Suppl
4): 35–40.
Van Paesschen, W. (2004b) Qualitative and quantitative imag-
ing of the hippocampus in mesial temporal lobe epilepsy with
hippocampal sclerosis. Neuroimag. Clin. North Am., 14:
373–400.
Vezzani, A. and Hoyer, D. (1999) Brain somatostatin: a can-
didate inhibitory role in seizures and epileptogenesis. Eur. J.
Neurosci., 11: 3767–3776.
Von Campe, G., Spencer, D.D. and de Lanerolle, N.C. (1997)
Morphology of dentate granule cells in the human epilepto-
genic hippocampus. Hippocampus, 7: 472–488.
Williamson, A., Patrylo, P.R., Pan, J., Spencer, D.D. and He-
therington, H. (2005) Correlations between granule cell phys-
iology and bioenergetics in human temporal lobe epilepsy.
Brain, 128: 1199–1208.
Williamson, A., Patrylo, P.R. and Spencer, D.D. (1999)
Decrease in inhibition in dentate granule cells from
patients with medial temporal lobe epilepsy. Ann. Neurol.,
45: 92–99.
Williamson, A. and Spencer, D. (1995) Zinc reduces dentate
granule cell hyperexcitability in epileptic humans. Neurore-
port, 6: 1562–1564.
Williamson, A., Spencer, D.D. and Shepherd, G.M. (1993)
Comparison between the membrane and synaptic properties
of human and rodent dentate granule cells. Brain Res., 622:
194–202.
Williamson, A., Spencer, S.S. and Spencer, D.D. (1995a) Depth
electrode studies and intracellular dentate granule cell re-
cordings in temporal lobe epilepsy. Ann. Neurol., 38:
778–787.
Williamson, A., Telfeian, A.E. and Spencer, D.D. (1995b) Pro-
longed GABA responses in dentate granule cells in slices
isolated from patients with temporal lobe sclerosis. J. Ne-
urophysiol., 74: 378–387.
Wittner, L., Eross, L., Szabo, Z., Toth, S., Czirjak, S., Halasz,
P., Freund, T.F. and Magloczky, Z.S. (2002) Synaptic reor-
ganization of calbindin-positive neurons in the human hip-
pocampal CA1 region in temporal lobe epilepsy.
Neuroscience, 115: 961–978.
Wuarin, J.P. and Dudek, F.E. (2001) Excitatory synaptic input
to granule cells increases with time after kainate treatment. J.
Neurophysiol., 85: 1067–1077.
Plate 11.2. Granule cell hyperexcitability is associated with mossy fiber sprouting. A shows representative traces comparing the
response to molecular layer (orthodromic) and hilar (antidromic) stimulation in a patient with MTS. Note that there is a greater
response to maximal hilar stimulation than to stimulation of outer molecular layer. Moreover, note that hilar stimulation produced a
synaptic response prior to an antidromic spike. These data support the hypothesis that mossy fibers are primarily excitatory in resected
tissue. B shows Timm staining from the region of the dentate where the cell shown in A was recorded. Note the robust supergranular
labeling indicating the presence of sprouted fibers. (C1) Stimulation protocols have the problem of activating multiple fiber systems.
Therefore, in the presence of 30 mM bicuculline, we applied a bolus of glutamate to the granule cell layer at a distance from the
recorded cell and were able to evoke a long lasting barrage of EPSPs. The glutamate was applied at a distance from the recorded cell
and thus we were not observing direct glutamate effects. C2 shows that a similar response was not observed when the glutamate was
applied to the hilus. (For B/W version, see page 188 in the volume.)
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 12

Hilar mossy cells: functional identification and

activity in vivo

Darrell A. Henze1 and György Buzsáki2,

1
Merck Research Laboratories, West Point, PA 19486, USA
2
Center for Molecular and Behavioral Neuroscience, Rutgers, The State University of New Jersey, Newark, NJ, USA

Abstract: Network oscillations are proposed to provide the framework for the ongoing neural compu-
tations of the brain. Thus, an important aspect of understanding the functional roles of various cell classes
in the brain is to understand the relationship of cellular activity to the ongoing oscillations. While many
studies have characterized the firing properties of cells in the hippocampal network including granule cells,
pyramidal cells and interneurons, information about the activity of dentate mossy cells in the intact brain
is scant. Here we review the currently available information and describe biophysical properties and
network-related firing patterns of mossy cells in vivo. These new observations will assist in the extracellular
identification of this unique cell type and help elucidate their functional role in behaving animals.

Keywords: mossy cell; hilus; network patterns; sharp wave; gamma; slow oscillation

Introduction Mossy cells are perhaps the most underinvesti-

The granule cell is the most numerous cell type
of the dentate gyrus (1 million in one hemisphere
in the rat, (Amaral et al., 1990), serves to integrate
entorhinal input and sends its messages to three
major target cell populations. These three cell
groups include mossy cells (MCs) of the hilus,
CA3 pyramidal cells and interneurons of both
dentate and CA3 regions. Although the MCs are
the least numerous cell type in the hilar region,
their unique properties suggest that they are
likely to be a very important cell type of the dent-
ate region. This chapter examines the properties
of the MCs and their possible role in overall
hippocampal function.
Corresponding author. Tel.: +1 (973) 353-1080 ext. 3131;
Fax: +1 (973) 353-1820; E-mail: buzsaki@andromeda.rutgers.edu
gated neurons in the hippocampal formation. Part
of the scarcity of information can be explained by
the low number of MCs (approximately 10,000 in
the rat) (Amaral et al., 1990). Unlike other prin-
cipal (excitatory) cell types of the hippocampus,
MCs do not form recognizable layers with densely
packed somata. Instead, they are scattered in
the hilar region under the granule cell layer,
making their in vivo accessibility for physiological
studies difficult. Most of what we know about the
functions of MCs comes from the pioneering
in vitro studies of Scharfman (Scharfman and
Schwartzkroin, 1988; Scharfman et al., 1990, 2001;
Scharfman, 1991, 1992a, b, 1994a–c, 1995), in vivo
studies of Schwartzkroin et al. (Buckmaster et al.,
1992, 1993, 1996; Buckmaster and Schwartzkroin
1995; Wenzel et al., 1997) and pathoanatomical
studies of epilepsy and ischemia by Sloviter et al.
(Sloviter, 1989, 1991a, 1994; Sloviter et al., 1991,

DOI: 10.1016/S0079-6123(07)63012-X 199

200

2003; Goodman et al., 1993; Zappone and
Sloviter, 2004). Behavioral correlates of MCs are
unknown due to the lack of criteria for the reliable
identification of MCs with extracellular methods.
A major goal of this chapter is to provide an over-
view of the available knowledge about the firing
patterns and biophysical properties of anatomi-
cally identified MCs and their network-related
behavior, and discuss how this information can
facilitate studies on MCs in the intact, behaving
animal.
Mossy cell anatomy
Mossy cells of the hilus were first recognized by
Cajal (1911) and Lorente de No (1934) for their
dendrites covered with large spines. These cells
were later given the name ‘‘mossy cells’’ by Amaral
(1978) due to the ‘‘mossy’’ appearance of their
large spines (see Fig. 1A). It has also been dem-
onstrated that all MCs selectively stain for GluR2/
3 receptors as opposed to other cells of the hilus
(Petralia and Wenthold, 1992; Leranth et al., 1996;
Fujise and Kosaka, 1999). Interestingly, there ap-
pears to be differences in the peptide content of the
mouse and hamster MC population, with ventral
MCs containing calretinin (Murakawa and
Kosaka, 2001) which is largely absent in MCs of
the dorsal hilus. This is in contrast to the rat where
no MCs contain calretinin. In addition, calcitonin
gene related peptide has been reported to selec-
tively label MCs in the rat (Freund et al., 1997).
Because these peptide markers are conspicuously
absent in the pyramidal cells of the hippocampus
proper, their differential staining can be taken as
clear justification of separating MCs of the dentate
gyrus and pyramidal neurons of the Ammon’s
horn.
The MCs are usually multipolar and have
tapering dendrites that largely remain restricted
to the hilus proper, although occasional dendrites
are observed in the molecular layer of the dentate
gyrus in both rats (Scharfman, 1991), and more of-
ten in primates (Frotscher et al., 1991; Buckmaster
and Amaral, 2001). These dendrites in the molec-
ular layer can receive direct input from the ent-
orhinal cortex bypassing the granule cells. This
variability in direct EC input is likely to be im-
portant for physiological function. All CA3 py-
ramidal cells, including those with mostly
horizontal dendrites residing in zone 3 of Amaral
(1978), send at least one dendritic branch to the
stratum lacunosum-moleculare and therefore

A B

Ca3 PC
Layer

C 20 mV
0.6 nA
50 ms

DG granule cell layer

Fig. 1. Hilar mossy cell and associated basic properties. (A) Biocytin labeled mossy cell. Note the large spines covering the soma and
proximal dendrites (arrows). Example current step evoked traces from the mossy cell shown in A. (B) Responses to a series of
hyperpolarizing steps from a resting potential of
57 mV. Note the robust ‘‘sag’’ observed for the larger hyperpolarizing steps and the
long charging curve for the smallest hyperpolarizing step. For this cell, the in vivo input resistance was 52 MO. (C) The response of this
cell to a strong depolarizing step (1.0 nA). Note that there is only very weak accommodation observed.

receive inputs from layer 2 neurons of the ent-
orhinal cortex. In contrast, MCs without dendrites
in the molecular layer generally receive only indi-
rect information from the entorhinal input relayed
by the granule cells (but see Kohler, 1985; Deller,
1998).
The proximal dendrites and somata of MCs are
covered by large ‘‘thorny excrescences’’ (Fig. 1).
While the thorny excrescences at first seem similar
to those of the proximal apical dendrite of CA3
pyramidal cells, they are qualitatively different in
that they resemble ‘‘clusters of spheres’’ (Amaral,
1978) where as CA3 excrescences appear more
irregular and thorny in their shape (Chicurel and
Harris, 1992). As in CA3 pyramidal cells, the
thorny excrescences of MCs receive synaptic input
from the mossy fibers of the granule cells (Amaral,
1978; Murakawa and Kosaka, 2001). Recent find-
ings suggest that the mossy fiber input to CA3 is a
mixed glutamatergic/GABAergic input for the first
three weeks of development in the rat or the young
guinea pig (Walker et al., 2001). Under normal
conditions in the adult, there is no detectable
GABAergic transmission. However, mossy fiber
GABAergic transmission is restored in the adult
after periods of strong repetitive stimulation or
hyperexcitability such as epilepsy (see Gutierrez,
2003, for review). There are also significant GAD-
positive, inhibitory inputs to the somatic region of
MCs suggesting a strong perisomatic inhibitory
input to these cells (Acsady et al., 2000; Murakawa
and Kosaka, 2001). In fact, the perisomatic inhi-
bition to MCs is very strong (15–40 times more
synapses per soma) compared to the inhibitory
cells of the hilus. The perisomatic inhibitory in-
nervation of MCs is primarily from terminals that
contain parvalbumin or cholecystokinin (CCK)
(Acsady et al., 2000), similar to the pyramidal
cells (Freund and Buzsáki, 1996). There is also a
perisomatic and proximal dendritic input from
cholinergic fibers (Deller et al., 1999) which may
serve to provide excitatory tone to the MCs during
oscillations such as theta. The more distal, or
so-called peripheral dendrites, are innervated by a
variety of inputs including more mossy fiber inputs
onto regular small spines as well as other anatom-
ically uncharacterized inputs making asymmetric
and symmetric synapses (Frotscher et al., 1991).
201
It is possible that at least some of the input to the
more peripheral dendrites is from CA3 pyramidal
cells which extend axons back into the hilus and
granule cell layer (Li et al., 1994). However, direct
anatomical evidence for a CA3-mossy cell com-
munication is still lacking (but see below for dis-
cussion of functional data). The dendrites of MCs
are extensive, extending several hundred microm-
eters in both medio-lateral and dorso-caudal di-
rections, suggesting a largely cylindrical dendritic
tree arborizing largely in the subgranular zone.
The large span of the dendritic arbor suggests
that MCs are innervated by spatially distributed
granule cells. The 100:1 ratio of granule cells to
MCs predicts that a typical MC receives inputs
from as many as a hundred granule cells. Given
the relative sparse distribution of mossy fiber
terminals in the hilus, it is also likely that each
granule cell innervates only one or two MCs with
large mossy fiber boutons. This low divergence
and convergence should be contrasted to the wide-
spread reciprocal innervation of granule cells by
the MCs (see below).
The MCs make glutamatergic (Soriano and
Frotscher, 1994; Wenzel et al., 1997) asymmetric
synapses with both excitatory and inhibitory
postsynaptic targets (Frotscher et al., 1991;
Buckmaster et al., 1996). In general, a single
MC’s synaptic targets can be divided into three
classes: local hilar targets; longitudinal targets
(located >1 mm either septally or temporally); and
contralateral targets. These three target classes have
been observed in both rodents and primates. Excit-
atory innervation of CA3 pyramidal cells by the
MCs has been repeatedly suggested (Buckmaster
et al., 1996; Buckmaster and Amaral, 2001) but
conclusive anatomical evidence for this alleged
connection is missing.
The main local hilar targets are mostly aspiny
dendrites of interneurons and potentially dendrites
of other MCs, although anatomical proof for
mutual MC communication is lacking. CA3 re-
current collaterals also innervate aspiny interneu-
rons in the hilus (Buckmaster et al., 1996) but
conspicuously avoid spiny interneurons of which
there are many in the hilus (Wittner et al., 2006).
The major bulk of the axon cloud of MCs (>90%
of ipsilateral synaptic contacts) target dentate
202
granule cell dendrites in the inner third of the
molecular layer of the dentate gyrus rostral and
caudal to the soma and dendrites of the parent
MC (Buckmaster et al., 1992, 1996; Wenzel et al.,
1997). This longitudinal arrangement is of major
physiological consequence because it eliminates
the possibility of a local granule cell–mossy
cell–granule cell recurrent excitatory loop. Gran-
ule cells, which drive discharge of the MCs, do
not receive excitatory information about the firing
status of their targets. Instead, the output of
the activated MC is distributed over the septal-
temporal extent of the dentate. In these target
areas, MCs innervate both granule cells and inter-
neurons, giving rise to potentially interesting phys-
iological scenarios. First, a granule cell discharging
a target MC can silence competing granule cells
over large territories via the feed-forward excita-
tion of interneurons by the MC. However, if this
were the sole function of such widespread axon
arborization, there would be no need for MC
excitatory innervation of granule cells. Through
the latter numerically dominant excitatory path-
way, the possibility exists that MCs can synchro-
nize spatially distinct granule cells in the
septotemporal axis, although the functional effi-
cacy of the mossy cell–granule cell synapse is not
well characterized (but see Scharfman, 1995).
From the latter perspective, the main function of
MCs would be to integrate functions of large
numbers of active granules cells in the septotem-
poral axis of the hippocampus.
The contralateral cellular targets of MCs are
largely undefined. However, in addition to inner-
vating granule cells, at least hilar neuropeptide Y
(NPY) containing cells may receive input from
contralateral MCs (Deller and Leranth, 1990).
It will be interesting to learn the extent and iden-
tity of the contralateral targets and how they
compare to the predominant granule cell targets
ipsilaterally.
Functional cellular connectivity of MCs
The anatomical connectivity described above pro-
vides a prospective framework for understanding
the functional role of the MCs in normal
hippocampal function. This can be summarized
most simply as primarily providing distributed
excitatory feedback to dentate granule cells and
secondarily providing excitatory drive to local
inhibitory interneurons of the hilus. The functional
importance of the contralateral projection is
harder to predict since the anatomical targets
are not yet characterized. A body of work
by Scharfman and colleagues (Scharfman and
Schwartzkroin, 1988; Scharfman et al., 1990,
2001; Scharfman, 1991, 1992a, b, 1994a–c, 1995)
has provided functional data to support the pre-
dictions of the anatomical connectivity to and
from MCs. For example, MCs that have dendrites
that extend into the DG molecular layer have a
lower threshold to fire in response to perforant
path stimulation (Scharfman, 1991). A challenging
series of studies in ventral slices used paired
recordings from anatomically confirmed hilar,
dentate and CA3 pyramidal cells to investigate
the functional connectivity in this region. Paired
recording of granule cells or CA3 pyramidal cells
with MCs showed that single action potentials in
either GCs or PCs can evoke EPSPs in MCs
(Scharfman, 1994b). Another paired recording
study demonstrated that MCs monosynaptically
excite both granule cells and inhibitory interneu-
rons. In addition, evidence was observed for
polysynaptic inhibition of GCs in response to
MC activity (Scharfman, 1995).
The hypothesized physiological function of the
low convergence and divergence of granule cells
onto CA3 pyramidal cells is to disperse (‘‘orthog-
onalize’’) the entorhinal information onto the large
recurrent system of CA3 neurons during the en-
coding of memories and provide multiple but
sparse representation (Treves and Rolls, 1992,
1994). In the retrieval process, the auto-associative
CA3 recurrent system, in turn, can recover the
whole memory representation from partial or
fragmental information (‘‘pattern completion’’)
(McNaughton and Morris, 1987; Kanerva, 1988).
Given this model of memory encoding and re-
trieval, we can ask what role might MCs play in
this system? Although the anatomical connectivity
between granule cells and MCs is similar to the
granule cell–CA3 connections, given the small
number of MCs and the lack of their reciprocal

excitation with one another makes it unlikely
that MCs are part of the pattern completion
mechanism. Instead, they may assist in increasing
the sparseness of the memory representation in the
recurrent CA3 system by the hypothesized feed
forward suppression of granule cells in the septo-
temporal axis or by allowing a sparse but coordi-
nated transfer of information from different
segments of the entorhinal cortex onto the CA3
system. This may be an important combinatorial
mechanism because the metric of spatial represen-
tation increases in the dorso-caudal axis of the
entorhinal cortex with corresponding projections
to different segments of the septotemporal
axis of the hippocampus (Hafting et al., 2005).
The synchronizing mechanism of granule cells by
the MCs in the longitudinal axis would secure the
simultaneous but distributed representation of the
different sampling metric of the entorhinal cortex
in the associative CA3 network. This hypothesis
differs from an alternative proposed by Buckmas-
ter and Schwartzkroin (1994) dubbed the ‘‘granule
cell association’’ hypothesis. In that hypothesis,
the MCs provide the necessary links to form
associative connections within the dentate network
analogous to the associational collaterals of
pyramidal cells in the CA3 region.
Cellular properties of MCs
The basic cellular properties of MCs are quite dis-
tinctive from other cell types in the hilar region.
The MCs tend to have higher input resistance and
strong inward rectification in response to hyper-
polarization. Figure 1B shows a typical MC from a
set of 10 cells we have recorded in vivo in response
to a series of hyperpolarizing steps. Each anatom-
ically verified MC recorded in urethane anestheti-
zed rats had action potentials that crossed 0 mV, a
mean resting membrane potential of 5871.8 mV,
and a mean input resistance of 6178.6 MO
(means7SEM). The resting potential is similar
to that of CA3 (63 mV; n ¼ 84) and CA1
(65 mV; n ¼ 280) pyramidal cells but different
from the more hyperpolarized granule cells
(74 mV; n ¼ 41). Of the principal cell types,
MCs had the largest input resistance
203
(CA3 ¼ 53.8 mO; n ¼ 8; CA1 ¼ 48.4 MO (Henze
and Buzsaki, 2001)). The smallest hyperpolarizing
step applied (0.2 nA) resulted in a hyperpolari-
zation of
22 mV and the time constant is best fit
with a double exponential with tau1 ¼ 7.5 ms and
tau2 ¼ 250 ms. A strong inward rectification can
be seen with a step injection of 0.4 nA that results
ultimately in a maximum hyperpolarization of
35 mV. The hyperpolarizations of the membrane
potential by small step currents were best fit by a
double exponential in seven of the nine cells where
it could be measured; the mean time constant
values were 15.9 and 188.2 ms. The remaining two
MCs were best fit with single exponentials with
time-constants of 32 and 26 ms. Figure 1C shows
that the MCs typically do not show burst firing in
response to depolarizing steps (1.0 nA) and only
show weak accommodation for the duration of
the step depolarization. This behavior can be
contrasted to the typical burst pattern and strong
spike accommodation in response to current steps
in CA3 pyramidal cells (personal observations,
Bilkey and Schwartzkroin, 1990; Buckmaster
et al., 1993; Scharfman, 1993b).
The background synaptic activity in MCs has
been reported to be quite high, both in vivo and in
vitro (Strowbridge et al., 1992; Scharfman,
1993a).We also have observed high background
activity that included some very large events
(>10 mV; Fig. 2). It is likely that these giant PSPs
with fast rise times arise from the mossy fiber
synaptic inputs to the MCs reflecting either
synchronous multivesicular release from a com-
plex MF bouton or perhaps the release of large
individual quanta as has been reported in CA3
pyramidal cells (e.g. see Henze et al., 2002).
Although the magnitude of the giant PSPs is quite
variable, it is unlikely that it reflects varying
convergence of activity from multiple granule
cells. First, the convergence of granule cells
onto MCs is low (
100). Second, the density of
mossy boutons in the hilus is quite low. Third,
intracellularly labeled neighboring granule cells
never showed spatially clustered boutons that
would otherwise suggest common targets (Acsady
et al., 2000).
One feature of MC activity that we have
observed that has not been previously described
204
A
10 mV
0.1 s
B
2 mV
0.01 s
Fig. 2. (A) Intracellular recording of a hilar mossy cell resting
at 55 mV (solid line). Notice the large spontaneous post-
synaptic potentials that occur from the baseline that sometimes
lead to action potentials. (B) Higher resolution depiction of two
large EPSPs from the box in (A).
is their relatively high background firing rate
(e.g. see Figs. 3, 5, and 9C). Our findings under
urethane anesthesia are partly consistent with
Scharfman (1993a), who showed that there is
often a high rate of firing but it waxes and wanes in
vitro. However, it is not consistent with other studies
in vivo (Soltesz et al., 1993) under ketamine/
xylazine where rates of less than 1 Hz were reported.
The source of this high firing rate is not well under-
stood because granule cells under urethane an-
esthesia are typically hyperpolarized and fire at a
low rate. One hypothesis is that spontaneous release
from MF boutons can lead to MC action potentials
(Henze et al., 2002). If the high firing pattern is
confirmed by investigations in the freely behaving
animal, it can be contrasted with the more clustered
firing patterns of CA3 pyramidal cells. Figure 3 il-
lustrates the autocorrelograms calculated from our
MC recordings. These autocorrelograms should fa-
cilitate the comparison of the spike dynamics of
MCs under anesthesia and in the drug-free animal in
future studies.
Another essential piece of information in the
process of physiological identification of MCs is
the extracellular features and shape of the extra-
cellular action potential. To this end, we have re-
corded from a single MC using simultaneous
intracellular and extracellular electrodes (Fig. 4).
The waveform we have observed is somewhat
unusual in that the extracellular unit waveform has
a rounded trough that has not been observed in
other hippocampal unit waveforms. It is possible
that the rounded (wide) pattern derives from the
summed extracellular currents from the soma
and the thick dendrites of MCs. In support of this
interpretation, the extracellular spikes were
recorded by three shanks of the silicon probe, a
total distance of 300 mm. We would suggest that
the horizontally wide current distribution of MC
spikes may be used as a distinguishing feature in
extracellular recordings from the granule cells and
other nearby smaller size interneurons.
MC cellular activity in a network
Although there is a relative lack of information
about the physiological role of MCs under nor-
mal conditions, their role has been a frequently
discussed topic of debate in the epilepsy literature.
This is because MCs are often observed to be
reduced in post-mortem tissue taken from people
who have had temporal lobe epilepsy (TLE)
(Sloviter et al., 1991). Sloviter and colleagues pro-
posed the so-called ‘‘dormant basket cell’’ hypoth-
esis of epilepsy (Sloviter, 1991b). The dormant
basket cell hypothesis holds that the importance of
the excitatory tone provided by the MCs is to
provide excitation of hilar inhibitory interneurons
which in turn then provide strong inhibition of
granule cells. When MCs are lost due to neurode-
generation associated with TLE, the inhibitory
basket cells lose their tonic excitatory drive result-
ing in a net disinhibition of dentate gyrus granule
cells (Sloviter, 1994; Sloviter et al., 2003). A
contrasting hypothesis has been called the ‘‘irrita-
ble mossy cell’’ hypothesis as proposed by Soltesz
and colleagues. In this view, it is not the loss of
MCs that leads to a net excitation of granule cells,
instead it is the remaining MCs that have higher
firing rates and provide uncontrolled excitatory
feedback to the dentate gyrus granule cells thus
exacerbating the epileptic process (Santhakumar
et al., 2000; Ratzliff et al., 2002).
Both of the ‘‘dormant basket cell’’ and ‘‘irritable
mossy cell’’ hypotheses have their attractions.
Recent studies by both camps have provided
 205

A F

-0.10 -0.05 0.00 0.05 0.10 -0.10 -0.05 0.00 0.05 0.10

B G

-0.10 -0.05 0.00 0.05 0.10 -0.10 -0.05 0.00 0.05 0.10

C H

-0.10 -0.05 0.00 0.05 0.10 -0.10 -0.05 0.00 0.05 0.10

D I

-0.10 -0.05 0.00 0.05 0.10

-0.10 -0.05 0.00 0.05 0.10

E J

-0.10 -0.05 0.00 0.05 0.10 -0.10 -0.05 0.00 0.05 0.10

Fig. 3. Hilar mossy cells show a variety of firing patterns as assessed by autocorrellograms (A–J). The autocorrelogram for each of the
10 MCs in our dataset was calculated from spontaneous spiking from the resting potential (no injected current). The majority of the
cells show a pattern that is more reminiscent of that seen for repetitively firing interneurons than the more bursty firing of CA1 or CA3
pyramidal cells (e.g. Csicsvari et al., 1998).
206

Spontaneous Evoked
20 mV
0.5 mS
IC

EC

50 microV 40 microV
0.5ms 0.5ms

Fig. 4. Extracellular waveforms of mossy cells during spontaneously and evoked spiking. A recording was obtained where the
intracellularly recorded mossy cell was also observed on the extracellular electrode. The extended shape of the mossy cell dendrites
allowed the extracellular signal to be observed on three shanks of the extracellular silicon probe (150 mm between shanks). The
extracellular electrode track is indicated by the white arrows. The soma of the mossy cell is indicated by the yellow arrow. There was a
difference in the shape of the action potentials, both intracellular and extracellular, suggesting differences in the site of spike initiation
for spontaneous and evoked spikes. (See Color Plate 12.4 in Color Plate Section.)

evidence to support the respective theories. Three
days after kainic acid induced status epilepticus,
there is reduction in inhibition and thus increase in
excitability of the dentate gyrus that correlates
with the degree of MC loss (Zappone and Sloviter,
2004). However, Ratzliff et al. (2004) showed that
in the hippocampal slice, if they acutely removed
MCs via manual destruction, there was a net
reduction in the excitability of the dentate gyrus
presumably due to the loss of direct excitatory
input from MCs. In all likelihood, both hypotheses
are at least partially correct and strict interpreta-
tion of these studies is confounded by the technical
challenges of studying the MCs in vivo under non-
pathological conditions. The intrinsic firing rates
of MCs in the intact unanesthetized animal is not
known, and both low (Soltesz et al., 1993) and
high firing rates (Figs. 3, 5, and 9C) have been
observed under anesthesia. Bulk stimulation of
fiber pathways such as the perforant path input to
the dentate gyrus is inherently not physiological in
that the precise synchrony of synaptic input that
results very rarely, if ever, happens in natural
processing. As such, the ratio of synchronously
activated excitatory to inhibitory inputs may be
very different than that observed during normal
ongoing hippocampal function.
Although the hilus also contains a large variety
of interneurons (Amaral, 1978; Mizumori et al.,
1990; Halasy and Somogyi, 1993; Buhl et al., 1994;
Sik et al., 1997), the contribution of these inter-
neurons to the rich variety of dentate area network

patterns is not known. Our overriding hypothesis
is that the main goal of both MCs and hilar
interneurons is to control network patterns during
various behaviors. This rationale has been suc-
cessfully applied to characterize CA1 interneurons
(Csicsvari et al., 1999; Klausberger et al., 2003,
2004, 2005), where the initial network pattern-
based classification (theta phase and sharp waves)
in freely behaving rats was followed by juxtacel-
lular labeling of similarly characterized interneu-
rons under anesthesia. We suggest that a similar
approach in the dentate gyrus can be equally fruit-
ful. Here we have preliminary yet more advanced
knowledge of the network contribution of MCs
under anesthesia than in the behaving animals.
In the dentate area, several distinct oscillatory
patterns are present.
Network oscillations are known to involve a
rhythmic pattern of periods of active inhibition
that are counterbalanced by periods of permissive
excitability (e.g. theta in CA1) (Buzsaki, 2002).
Perhaps the unique role of the MCs is to provide
active excitation to granule cells in the periods
between the rhythmic inhibitory inputs driven by
interneurons. Although all the various patterns are
mediated by a limited set of excitatory pathways, it
is expected that MCs and the various interneuron
groups may differentially participate in these net-
work patterns because the dynamics of activation
can differentially affect neurons with different
properties. The following patterns can be used to
characterize the network contribution of MCs and
contrast them to granule cells, CA3 pyramidal cells
and hilar area interneuron types.
1. In the exploring rat and during REM sleep,
large amplitude theta oscillations are present
in the dentate gyrus. Theta waves in the dent-
ate region are coherent with but phase-shifted
by approximately 2701 relative to the theta
oscillation in the CA1 pyramidal layer
(Buzsaki et al., 1983).
2. Concurrent with theta, another prominent
pattern in the dentate area is the gamma fre-
quency oscillation (40–100 Hz). The power of
gamma oscillations is strongly phase modu-
lated by the slower theta rhythm and both
theta and gamma oscillations in the dentate
207
gyrus depend mainly on inputs from the
entorhinal cortex (Bragin et al., 1995a;
Penttonen et al., 1998). In addition to the
classical gamma frequency, slower (12–40 Hz;
beta) oscillations are also observed, often
with a larger amplitude in the hilus. It is not
clear whether the slow oscillation is simply a
slower version of gamma or comprises a
physiologically distinct rhythm.
3. The largest amplitude dentate gyrus event is
the ‘‘dentate spike’’. This is a short duration
(o60 ms), large amplitude (>0.5–2.5 mV)
field potential characterized by synchronous
discharge of granule cells and interneurons
and suppression of CA3 pyramidal cells
(Bragin et al., 1995b; Penttonen et al.,
1997). Two types of dentate spikes have been
distinguished. The first type is a short burst of
gamma oscillation consisting of 2–5 waves,
one of which of excessively high amplitude
with large sinks in the outer third of the
dentate molecular layer. The second type,
observed in a subset of animals, has a some-
what different voltage vs. depth profile with a
large sink located in the middle third of the
dentate molecular layer. Our unpublished
observations suggest that type 2 dentate
spikes occur when thalamocortical high volt-
age spindles (Buzsaki et al., 1988) invade the
dentate area.
4. Neocortical slow oscillations (Steriade et al.,
1990) also exert an impact on the firing
patterns of the hippocampus, likely by way of
the entorhinal cortex (Isomura et al., 2006;
Wolansky et al., 2006). During the UP state
of slow oscillations, gamma power in the
dentate gyrus and spiking of neurons in-
creases dramatically. In contrast, dentate
gamma activity decreases during the DOWN
state (corresponding to delta waves of deep
sleep in the neocortex) but CA3 pyramidal
cells may increase their firing rates and
generate gamma oscillations (Isomura et al.,
2006)
5. Sharp waves-ripple complexes (SPW) are truly
self-organized endogenous hippocampal
events that occur during slow-wave sleep,
immobility and consummatory behaviors
208

(Buzsaki et al., 1983). They arise in the CA3 A

recurrent system and can spread to the CA1
region and the dentate region. The transient 1
excitation of CA1 neurons gives rise to a 0.2 mV
short-lived fast oscillation (‘‘ripple’’) 10 mV
0.5 s
(O’Keefe and Nadel, 1978; Buzsáki et al.,
1992; Ylinen et al., 1995). No ripples are as-
sociated with SPW in the dentate area but 2
putative interneurons and, occasionally,
granule cells are depolarized and discharge
in synchrony with CA1 ripples, although
the typical response is hyperpolarization
(Penttonen et al., 1997).
B
We submit that these five unique population
patterns can be used in future studies to function- 1
ally classify dentate region neurons in the freely
behaving animal. In turn, these same patterns can
0.1 mV
be used in anesthetized rats where they can be 10 mV
labeled by intracellular or juxtacellular methods. 0.5 s
The network-based classification should be com-
bined with the spike dynamics and wave-shape
2
features of extracellular spikes. Fortunately, the
repertoire of rat hippocampal network activity
patterns under urethane anesthesia largely reflects
what is observed in the drug-free rat. Using this
approach we have been able to observe how MCs
behave in relation to some of these naturally
occurring network patterns. Fig. 5. (A) Example of mossy cell inhibition by transition to
theta rhythm evoked by tail pinch in urethane anesthetized rat.
Extracellular recording (A1) from area CA1 recorded simulta-
Theta neously with a hilar mossy cell (A2). A tail pinch was applied at
the arrow and the mossy cell responded by slowing its firing
rate. (B) Example of mossy cell excitation by transition to theta
It has been previously reported that MC mem- rhythm evoked by tail pinch in urethane anesthetized rat. Ex-
brane potential shows rhythmic oscillations in the tracellular recording (B1) from area CA3 recorded simultane-
theta band that are phase-locked to the extracel- ously with (B2) a hilar mossy cell that was excited by tail pinch
lular theta oscillation in the contralateral hippo- (arrow) evoked theta.
campus (Soltesz et al., 1993). However, as noted
above, this study reported a very low (o1 Hz) peak of the locally derived theta oscillation
basal firing rate for the MCs. Nevertheless, our (Fig. 6) and coherent with the discharge of some
observations support the involvement of MCs in interneuron types. We predict that MCs in the
theta oscillations. Individual MCs can either be behaving rat will keep a similar relationship to
depolarized (8 of 10) or hyperpolarized (2 of 10) by local hilar/CA3c theta oscillations.
a transition from slow wave sleep to theta evoked
by a tail pinch (Fig. 5). This behavior is similar Beta/gamma oscillations
to pyramidal cells (Kamondi et al., 1998). In
addition, the membrane potential of MCs shows a The power of gamma frequency oscillation in
co-variation with the extracellular field, with peak the hilus is phase modulated by the slower theta
depolarization and discharge slightly after the (Bragin et al., 1995a). This gamma frequency
 209

Fig. 6. (A) Continuous theta/delta power ratio from an extracellular electrode located in the hilus. The increases in theta power at
45
and 200 s was induced via tail pinch. (B) rastergrams of isolated units recorded from three extracellular tetrodes in the hilus/area CA3c.
(C) Intracellular membrane potential recorded from a mossy cell during this same recording epoch while passing hyperpolarizing
current to maintain the resting potential near 80 mV. Notice that during the periods of theta power increase, the mossy cell
experiences a depolarization from the 80 mV holding potential. (D) Two examples of average mossy cell membrane potential time
aligned to the two ‘‘theta on’’ putative interneuronal units indicated by the start symbols in (C). The upper trace is the average
membrane potential of the MC. The lower trace is the average wide-band extracellular trace. The autocorrelogram and waveform for
the isolated IN units are also shown as insets.
210

A Gamma cycle

Theta oscillation

20µV
EC unit for MC
200 ms

1mV
B
20 ms

25µV
1ms

Fig. 7. (A) Average of extracellular hilar field potential aligned to the peaks of the intracellular MC spontaneous action potentials
recorded over a 200 sec recording period. This is the same case as in Fig. 4 where the extracellular electrode picked up the same MC as
was recorded intracellularly. The extracellular unit correlate of the intracellular AP is indicated. Notice that the MC fires on the end of
the gamma cycle waveform and there is an overall theta oscillation present that is time aligned to the intracellular spike. (B) Red traces:
MC membrane potential averaged and time aligned on the spike times of isolated interneurons recorded in the hilus/CA3c. Left
column: IN unit autocorrelogram, black traces: average IN unit waveforms. Note similar phase relationship of the interneuron-timed
intracellular gamma frequency oscillation.

modulation is also evident in the relationship be-
tween the local field and fast spiking interneurons
and their effect on the MC (Fig. 7). When the ac-
tion potentials of MCs are used as reference for
averaging local field potentials, phaselocking of
MC spikes to the gamma oscillation, superimposed
on the peak of theta waves becomes evident
(Fig. 7A). Furthermore, when nearby fast-spiking
putative interneurons are used as the reference,
membrane oscillation of MCs in the gamma fre-
quency band becomes evident (Fig. 7B). Similar
phase-locked behavior has been also observed in
the beta oscillation frequency range as well.
Slow oscillations
Slow oscillations arise in neocortical networks
(Steriade et al., 1993a–c) and spread to the
 211

Fig. 8. (A) CA3 unit activity and MC membrane potential fluctuations. Mossy cells show synaptic activity correlated with unit activity
in CA3 (2/2 cells recorded with extracellular electrode placed in CA3c. Upper traces show extracellular wideband recording middle
trace bandpass filtered between 800 Hz and 3 kHz. Lower traces show intracellular membrane potential. (B) Cross-correlation between
spontaneous MC spikes recorded intracellularly and all extracellular units combined. Lower panels are shuffled controls. (C) Hilar/
CA3c population bursts correlate with spontaneous mossy cell spikes. Hilar/CA3c bursts were defined as five unit spikes with less than
20 ms inter-spike interval. The spikes could originate from any unit. Shuffled correlograms are shown for controls. (D) Cross-
correlation between spontaneous mossy cell spikes and all hilar/CA3 units combined for a second extracellular/intracellular recording
pair.

entorhinal cortex and subiculum from where
they can invade the dentate gyrus as well (Isomura
et al., 2006; Wolansky et al., 2006). At the single
cell level, most cortical neurons show a bimodal
distribution of the membrane potential and these
UP and DOWN states alternate relatively rhyth-
mically at 0.5–2 Hz. Slow oscillations are most
prominent under anesthesia but are also present in
deep stages of slow wave sleep (Achermann and
Borbely, 1997), where the transient delta waves
correspond to the DOWN (or silent) state. At the
transition of the entorhinal DOWN–UP state, the
surge of excitation induces a strong discharge of
granule cells and also gamma frequency oscilla-
tions (Isomura et al., 2006). Often one of these
gamma waves becomes excessively large and has
been referred to as the dentate spike (Bragin et al.,
1995b). In this case, the MC firing is likely timed
by feed forward excitation from dentate gyrus
granule cells that are driven by the entorhinal
input. The surge of activity in the input from
entorhinal cortex is also reflected by the sudden
212

Fig. 9. CA1 sharpwave/ripples are associated with inhibition of mossy cells. (A) Example of a CA1 sharpwave/ripple complex
recorded extracellularly in the pyramidal layer of CA1. (B) Intracellular spiking activity of mossy cell is suppressed during SPW. (C, D)
Temporal expansion of region in box in A and B. Mossy cell membrane response during SPW/ripples has a reversal near 60 mV.
(E, F) Average extracellular SPW and mossy cell membrane potential from a hyperpolarized baseline (E; n ¼ 12 SPWs) and a more
depolarized baseline (F; n ¼ 11 SPWs). (G) Plot of the peak change in mossy cell membrane potential during CA1 SPWs from 15
recording epochs at different membrane potentials from four mossy cells (different symbols). The best fit linear regression line is shown
(R ¼ 0.57621, Po0.025).

increase of population firing in the dentate. We
exploited the high levels of CA3 activity that occur
during the slow oscillatory pattern for the exam-
ination of the behavior of MCs. The transition
from DOWN to UP state was defined when a
group of dentate/hilar/CA3 neurons fired a burst
of activity. This analysis revealed that MCs be-
come periodically active and silent according to
the level of activity in the entorhinal–hippocampal
network (Fig. 8).
CA1 sharp waves
Although SPW arises in the CA3–CA1 regions,
the recurrent axon collaterals of CA3 can directly
excite hilar interneurons and even granule cells
(Li et al., 1994; Penttonen et al., 1997). Typically,
feed-forward inhibition prevails by this mechanism
as reflected by the SPW-locked hyperpolarization
of granule cells. However, this occasional failure of
inhibition and/or a concerted activation of a single
granule cell by the recurrent CA3 axons can
robustly discharge granule cells during SPWs
(Penttonen et al., 1997). Our observations, to
date, suggest that SPWs are associated with net
inhibition of MCs. Figure 9A–D show the tempo-
ral relationship between a spontaneous SPW
recorded in area CA1 (Fig. 9A, B) while record-
ing from a MC in the hilus (Fig. 9C, D). As can be
appreciated from this example, the ongoing spon-
taneous firing of the MC is inhibited in the period
overlapping with and following the SPW/ripple
complex. This inhibitory effect is probably medi-
ated via activation of chloride flux through
GABAA receptors since the change in MC mem-
brane potential associated with a CA1 SPW has a
reversal potential near 60 mV similar to GABAA
reversal potentials in vivo (Fig. 9E–G). The
potential source of this inhibition is the increased
SPW-related firing of hilar interneurons driven
either directly by the recurrent CA3 collaterals
or the rarely discharging granule cells (Penttonen
et al., 1997). It appears that, at least under an-
esthesia, the strong inhibition can prevent MCs
from discharging in response to their granule cell
inputs during SPWs. However, the situation might
be quite different in the drug-free animal and it is
213
expected that future studies will clarify whether
MCs can become active participants in SPW
events. It is worth noting here that in slices treated
with a GABAA receptor antagonist, or slices from
an epileptic rat, CA3 and MCs are engaged in
population bursts while granule cells remain
hyperpolarized, suggesting that CA3 pyramidal
cells may directly excite MCs (Scharfman, 1994a;
Scharfman et al., 2001). The failure of MCs to
discharge during SPWs in the intact brain would
imply that the hypothesized SPW-mediated con-
solidation of synaptic circuits in the CA3–CA1
networks (Buzsáki, 1989) can proceed independent
of the modification of the synapses established by
the MCs.
Conclusion
Although MCs are well-known and critical com-
ponents of the dentate circuitry, their physiological
function and exact involvement in various hyper-
excitable phenomena has remained elusive. A
major technical problem is the lack of reliable
physiological criteria that may be used in extra-
cellular recordings in freely behaving animal for
the positive identification of MCs. Our intracellu-
lar characterization of some of their biophysical
features and network-related behavior are the first
steps in this direction.
Abbreviations
CCK cholecystokinin
EPSP excitatory postsynaptic potential
GABA gamma-aminobutyric acid
GAD glutamic acid decarboxylase
MC mossy cell
NPY neuropeptide Y
References
Achermann, P. and Borbely, A.A. (1997) Low-frequency
(o1 Hz) oscillations in the human sleep electroencephalo-
gram. Neuroscience, 81: 213–222.
Acsady, L., Katona, I., Martinez-Guijarro, F.J., Buzsaki, G.
and Freund, T.F. (2000) Unusual target selectivity of
214
perisomatic inhibitory cells in the hilar region of the rat hip-
pocampus. J. Neurosci., 20: 6907–6919.
Amaral, D.G. (1978) A Golgi study of cell types in the hilar
region of the hippocampus in the rat. J. Comp. Neurol., 182:
851–914.
Amaral, D.G., Ishizuka, N. and Claiborne, B. (1990) Neurons,
numbers and the hippocampal network. Prog. Brain Res.,
83(1–11): 1–11.
Bilkey, D.K. and Schwartzkroin, P.A. (1990) Variation in
electrophysiology and morphology of hippocampal CA3 py-
ramidal cells. Brain Res., 514: 77–83.
Bragin, A., Jando, G., Nadasdy, Z., Hetke, J., Wise, K.
and Buzsaki, G. (1995a) Gamma (40–100 Hz) oscillation
in the hippocampus of the behaving rat. J. Neurosci., 15:
47–60.
Bragin, A., Jando, G., Nadasdy, Z., van Landeghem, M. and
Buzsaki, G. (1995b) Dentate EEG spikes and associated
interneuronal population bursts in the hippocampal hilar
region of the rat. J. Neurophysiol., 73: 1691–1705.
Buckmaster, P.S. and Amaral, D.G. (2001) Intracellular re-
cording and labeling of mossy cells and proximal CA3 py-
ramidal cells in macaque monkeys. J. Comp. Neurol., 430:
264–281.
Buckmaster, P.S. and Schwartzkroin, P.A. (1994) Hippocampal
mossy cell function: a speculative view. Hippocampus, 4:
393–402.
Buckmaster, P.S. and Schwartzkroin, P.A. (1995) Interneurons
and inhibition in the dentate gyrus of the rat in vivo. J.
Neurosci., 15: 774–789.
Buckmaster, P.S., Strowbridge, B.W., Kunkel, D.D., Schmiege,
D.L. and Schwartzkroin, P.A. (1992) Mossy cell axonal
projections to the dentate gyrus molecular layer in the rat
hippocampal slice. Hippocampus, 2: 349–362.
Buckmaster, P.S., Strowbridge, B.W. and Schwartzkroin,
P.A. (1993) A comparison of rat hippocampal mossy
cells and CA3c pyramidal cells. J. Neurophysiol., 70:
1281–1299.
Buckmaster, P.S., Wenzel, H.J., Kunkel, D.D. and Schwa-
rtzkroin, P.A. (1996) Axon arbors and synaptic connections
of hippocampal mossy cells in the rat in vivo. J. Comp.
Neurol., 366: 271–292.
Buhl, E.H., Han, Z.S., Lorinczi, Z., Stezhka, V.V., Karnup,
S.V. and Somogyi, P. (1994) Physiological properties of an-
atomically identified axo-axonic cells in the rat hippocampus.
J. Neurophysiol., 71: 1289–1307.
Buzsáki, G. (1989) Two-stage model of memory trace forma-
tion: a role for ‘‘noisy’’ brain states. Neuroscience, 31:
551–570.
Buzsaki, G. (2002) Theta oscillations in the hippocampus.
Neuron, 33: 325–340.
Buzsaki, G., Bickford, R.G., Armstrong, D.M., Ponomareff,
G., Chen, K.S., Ruiz, R., Thal, L.J. and Gage, F.H. (1988)
Electric activity in the neocortex of freely moving young and
aged rats. Neuroscience, 26: 735–744.
Buzsáki, G., Horvath, Z., Urioste, R., Hetke, J. and Wise, K.
(1992) High-frequency network oscillation in the hippocam-
pus. Science, 256: 1025–1027.
Buzsaki, G., Leung, L.W. and Vanderwolf, C.H. (1983) Cellu-
lar bases of hippocampal EEG in the behaving rat. Brain
Res., 287: 139–171.
Cajal, S.R. (1911) Histologie du Systeme Nerveux de l’Homme
et des Vertebres. Oxford University Press, New York.
Chicurel, M.E. and Harris, K.M. (1992) Three-dimensional
analysis of the structure and composition of CA3 branched
dendritic spines and their synaptic relationships with mossy
fiber boutons in the rat hippocampus. J. Comp. Neurol., 325:
169–182.
Csicsvari, J., Hirase, H., Czurko, A. and Buzsaki, G. (1998)
Reliability and state dependence of pyramidal cell-interneu-
ron synapses in the hippocampus: an ensemble approach in
the behaving rat. Neuron, 21: 179–189.
Csicsvari, J., Hirase, H., Czurko, A., Mamiya, A. and Buzsáki,
G. (1999) Oscillatory coupling of hippocampal pyramidal
cells and interneurons in the behaving rat. J. Neurosci., 19:
274–287.
Deller, T. (1998) The anatomical organization of the rat fascia
dentata: new aspects of laminar organization as revealed by
anterograde tracing with phaseolus vulgaris-luecoagglutinin
(PHAL). Anat. Embryol. (Berl.), 197: 89–103.
Deller, T., Katona, I., Cozzari, C., Frotscher, M. and Freund,
T.F. (1999) Cholinergic innervation of mossy cells in the rat
fascia dentata. Hippocampus, 9: 314–320.
Deller, T. and Leranth, C. (1990) Synaptic connections of ne-
uropeptide Y (NPY) immunoreactive neurons in the hilar
area of the rat hippocampus. J. Comp. Neurol., 300:
433–447.
Freund, T.F. and Buzsáki, G. (1996) Interneurons of the hip-
pocampus. Hippocampus, 6: 347–470.
Freund, T.F., Hajos, N., Acsady, L., Gorcs, T.J. and Katona, I.
(1997) Mossy cells of the rat dentate gyrus are immunore-
active for calcitonin gene-related peptide (CGRP). Eur. J.
Neurosci., 9: 1815–1830.
Frotscher, M., Seress, L., Schwerdtfeger, W.K. and Buhl, E.
(1991) The mossy cells of the fascia dentata: a comparative
study of their fine structure and synaptic connections in ro-
dents and primates. J. Comp. Neurol., 312: 145–163.
Fujise, N. and Kosaka, T. (1999) Mossy cells in the mouse
dentate gyrus: identification in the dorsal hilus and their dis-
tribution along the dorsoventral axis. Brain Res., 816:
500–511.
Goodman, J.H., Wasterlain, C.G., Massarweh, W.F., Dean, E.,
Sollas, A.L. and Sloviter, R.S. (1993) Calbindin-D28k
immunoreactivity and selective vulnerability to ischemia in
the dentate gyrus of the developing rat. Brain Res., 606:
309–314.
Gutierrez, R. (2003) The GABAergic phenotype of the
‘‘glutamatergic’’ granule cells of the dentate gyrus. Prog.
Neurobiol., 71: 337–358.
Hafting, T., Fyhn, M., Molden, S., Moser, M.B. and Moser,
E.I. (2005) Microstructure of a spatial map in the entorhinal
cortex. Nature, 436: 801–806.
Halasy, K. and Somogyi, P. (1993) Subdivisions in the multiple
GABAergic innervation of granule cells in the dentate gyrus
of the rat hippocampus. Eur. J. Neurosci., 5: 411–429.

Henze, D.A. and Buzsaki, G. (2001) Action potential threshold
of hippocampal pyramidal cells in vivo is increased by recent
spiking activity. Neuroscience, 105: 121–130.
Henze, D.A., McMahon, D.B., Harris, K.M. and Barrionuevo,
G. (2002) Giant miniature EPSCs at the hippocampal mossy
fiber to CA3 pyramidal cell synapse are monoquantal. J.
Neurophysiol., 87: 15–29.
Isomura, Y., Sirota, A., Montgomery, S., Mizuseki, K., Özen,
S., Henze, D.A. and Buzsáki, G. (2006) Integration and Seg-
regation of Activity in Entorhinal-hippocampal Subregions
by Neocortical Slow Oscillations. Neuron, 52: 871–882.
Kamondi, A., Acsady, L., Wang, X.J. and Buzsáki, G. (1998)
Theta oscillations in somata and dendrites of hippocampal
pyramidal cells in vivo: activity-dependent phase-precession
of action potentials. Hippocampus, 8: 244–261.
Kanerva, P. (1988) Sparse Distributed Memory. Bradford/MIT
Press, Cambridge, MA.
Klausberger, T., Magill, P.J., Marton, L.F., Roberts, J.D.,
Cobden, P.M., Buzsaki, G. and Somogyi, P. (2003) Brain-
state- and cell-type-specific firing of hippocampal interneu-
rons in vivo. Nature, 421: 844–848.
Klausberger, T., Marton, L.F., Baude, A., Roberts, J.D.,
Magill, P.J. and Somogyi, P. (2004) Spike timing of dendrite-
targeting bistratified cells during hippocampal network os-
cillations in vivo. Nat. Neurosci., 7: 41–47.
Klausberger, T., Marton, L.F., O’Neill, J., Huck, J.H.,
Dalezios, Y., Fuentealba, P., Suen, W.Y., Papp, E., Kaneko,
T., Watanabe, M., Csicsvari, J. and Somogyi, P. (2005)
Complementary roles of cholecystokinin- and parvalbumin-
expressing GABAergic neurons in hippocampal network os-
cillations. J. Neurosci., 25: 9782–9793.
Kohler, C. (1985) A projection from the deep layers of the
entorhinal area to the hippocampal formation in the rat
brain. Neurosci. Lett., 56: 13–19.
Leranth, C., Szeidemann, Z., Hsu, M. and Buzsaki, G. (1996)
AMPA receptors in the rat and primate hippocampus: a
possible absence of GluR2/3 subunits in most interneurons.
Neuroscience, 70: 631–652.
Li, X.G., Somogyi, P., Ylinen, A. and Buzsaki, G. (1994) The
hippocampal CA3 network: an in vivo intracellular labeling
study. J. Comp. Neurol., 339: 181–208.
Lorente de No, R. (1934) Studies on the structure of the cer-
ebral cortex. II. Continuation of the study of the ammonic
system. J. Psychol. Neurol. (Leipzig), 46: 113–177.
McNaughton, B.L. and Morris, R.G.M. (1987) Hippocampal
synaptic enhancement and information storage within a dis-
tributed memory system. TINS, 10(10): 408–415.
Mizumori, S.J., Barnes, C.A. and McNaughton, B.L. (1990)
Behavioral correlates of theta-on and theta-off cells recorded
from hippocampal formation of mature young and aged rats.
Exp. Brain Res., 80: 365–373.
Murakawa, R. and Kosaka, T. (2001) Structural features of
mossy cells in the hamster dentate gyrus, with special refer-
ence to somatic thorny excrescences. J. Comp. Neurol., 429:
113–126.
O’Keefe, J. and Nadel, L. (1978) Hippocampus as a cognitive
map. Clarendon, Oxford.
215
Penttonen, M., Kamondi, A., Acsady, L. and Buzsaki, G.
(1998) Gamma frequency oscillation in the hippocampus of
the rat: intracellular analysis in vivo. Eur. J. Neurosci., 10:
718–728.
Penttonen, M., Kamondi, A., Sik, A., Acsady, L. and Buzsaki,
G. (1997) Feed-forward and feed-back activation of the
dentate gyrus in vivo during dentate spikes and sharp wave
bursts. Hippocampus, 7: 437–450.
Petralia, R.S. and Wenthold, R.J. (1992) Light and electron
immunocytochemical localization of AMPA-selective gluta-
mate receptors in the rat brain. J. Comp. Neurol., 318:
329–354.
Ratzliff, A.H., Howard, A.L., Santhakumar, V., Osapay, I. and
Soltesz, I. (2004) Rapid deletion of mossy cells does not result
in a hyperexcitable dentate gyrus: implications for epilepto-
genesis. J. Neurosci., 24: 2259–2269.
Ratzliff, A.H., Santhakumar, V., Howard, A. and Soltesz, I.
(2002) Mossy cells in epilepsy: rigor mortis or vigor mortis?
Trends Neurosci., 25: 140–144.
Santhakumar, V., Bender, R., Frotscher, M., Ross, S.T.,
Hollrigel, G.S., Toth, Z. and Soltesz, I. (2000) Granule cell
hyperexcitability in the early post-traumatic rat dentate
gyrus: the ‘irritable mossy cell’ hypothesis. J. Physiol.,
524(Pt 1): 117–134.
Scharfman, H.E. (1991) Dentate hilar cells with dendrites in the
molecular layer have lower thresholds for synaptic activation
by perforant path than granule cells. J. Neurosci., 11:
1660–1673.
Scharfman, H.E. (1992a) Blockade of excitation reveals inhi-
bition of dentate spiny hilar neurons recorded in rat hippo-
campal slices. J. Neurophysiol., 68: 978–984.
Scharfman, H.E. (1992b) Differentiation of rat dentate neurons
by morphology and electrophysiology in hippocampal slices:
granule cells, spiny hilar cells and aspiny ‘fast-spiking’ cells.
Epilepsy Res. Suppl., 7: 93–109.
Scharfman, H.E. (1993a) Characteristics of spontaneous and
evoked EPSPs recorded from dentate spiny hilar cells in rat
hippocampal slices. J. Neurophysiol., 70: 742–757.
Scharfman, H.E. (1993b) Spiny neurons of area CA3c in
rat hippocampal slices have similar electrophysiological
characteristics and synaptic responses despite morphological
variation. Hippocampus, 3: 9–28.
Scharfman, H.E. (1994a) EPSPs of dentate gyrus granule cells
during epileptiform bursts of dentate hilar ‘‘mossy’’ cells and
area CA3 pyramidal cells in disinhibited rat hippocampal
slices. J. Neurosci., 14: 6041–6057.
Scharfman, H.E. (1994b) Evidence from simultaneous intracel-
lular recordings in rat hippocampal slices that area CA3
pyramidal cells innervate dentate hilar mossy cells. J. Ne-
urophysiol., 72: 2167–2180.
Scharfman, H.E. (1994c) Synchronization of area CA3 hippo-
campal pyramidal cells and non-granule cells of the dentate
gyrus in bicuculline-treated rat hippocampal slices. Neuro-
science, 59: 245–257.
Scharfman, H.E. (1995) Electrophysiological evidence that
dentate hilar mossy cells are excitatory and innervate both
granule cells and interneurons. J. Neurophysiol., 74: 179–194.
216
Scharfman, H.E., Kunkel, D.D. and Schwartzkroin, P.A.
(1990) Synaptic connections of dentate granule cells and
hilar neurons: results of paired intracellular recordings and
intracellular horseradish peroxidase injections. Neuroscience,
37: 693–707.
Scharfman, H.E. and Schwartzkroin, P.A. (1988) Electrophys-
iology of morphologically identified mossy cells of the dent-
ate hilus recorded in guinea pig hippocampal slices. J.
Neurosci., 8: 3812–3821.
Scharfman, H.E., Smith, K.L., Goodman, J.H. and Sollas, A.L.
(2001) Survival of dentate hilar mossy cells after pilocarpine-
induced seizures and their synchronized burst discharges with
area CA3 pyramidal cells. Neuroscience, 104: 741–759.
Sik, A., Penttonen, M. and Buzsáki, G. (1997) Interneurons in
the hippocampal dentate gyrus: an in vivo intracellular study.
Eur. J. Neurosci., 9: 573–588.
Sloviter, R.S. (1989) Calcium-binding protein (calbindin-D28k)
and parvalbumin immunocytochemistry: localization in the
rat hippocampus with specific reference to the selective
vulnerability of hippocampal neurons to seizure activity.
J. Comp. Neurol., 280: 183–196.
Sloviter, R.S. (1991a) Feedforward and feedback inhibition of
hippocampal principal cell activity evoked by perforant path
stimulation: GABA-mediated mechanisms that regulate ex-
citability in vivo. Hippocampus, 1: 31–40.
Sloviter, R.S. (1991b) Permanently altered hippocampal struc-
ture, excitability, and inhibition after experimental status ep-
ilepticus in the rat: the ‘‘dormant basket cell’’ hypothesis and
its possible relevance to temporal lobe epilepsy. Hippocam-
pus, 1: 41–66.
Sloviter, R.S. (1994) The functional organization of the hippo-
campal dentate gyrus and its relevance to the pathogenesis of
temporal lobe epilepsy. Ann. Neurol., 35: 640–654.
Sloviter, R.S., Sollas, A.L., Barbaro, N.M. and Laxer, K.D.
(1991) Calcium-binding protein (calbindin-D28k) and
parvalbumin immunocytochemistry in the normal and
epileptic human hippocampus. J. Comp. Neurol., 308:
381–396.
Sloviter, R.S., Zappone, C.A., Harvey, B.D., Bumanglag, A.V.,
Bender, R.A. and Frotscher, M. (2003) ‘‘Dormant basket
cell’’ hypothesis revisited: relative vulnerabilities of dentate
gyrus mossy cells and inhibitory interneurons after hippo-
campal status epilepticus in the rat. J. Comp. Neurol., 459:
44–76.
Soltesz, I., Bourassa, J. and Deschenes, M. (1993) The behavior
of mossy cells of the rat dentate gyrus during theta oscilla-
tions in vivo. Neuroscience, 57: 555–564.
Soriano, E. and Frotscher, M. (1994) Mossy cells of the rat
fascia dentata are glutamate-immunoreactive. Hippocampus,
4: 65–69.
Steriade, M., Contreras, D., Curro, D.R. and Nunez, A.
(1993a) The slow (o1 Hz) oscillation in reticular thalamic
and thalamocortical neurons: scenario of sleep rhythm
generation in interacting thalamic and neocortical networks.
J. Neurosci., 13: 3284–3299.
Steriade, M., Gloor, P., Llinas, R.R., Lopes de Silva, F.H. and
Mesulam, M.M. (1990) Report of IFCN committee on basic
mechanisms. Basic mechanisms of cerebral rhythmic activi-
ties. Electroencephalogr. Clin. Neurophysiol., 76: 481–508.
Steriade, M., Nunez, A. and Amzica, F. (1993b) A novel slow
(o1 Hz) oscillation of neocortical neurons in vivo: depolar-
izing and hyperpolarizing components. J. Neurosci., 13:
3252–3265.
Steriade, M., Nunez, A. and Amzica, F. (1993c) Intracellular
analysis of relations between the slow (o1 Hz) neocortical
oscillation and other sleep rhythms of the electroencephalo-
gram. J. Neurosci., 13: 3266–3283.
Strowbridge, B.W., Buckmaster, P.S. and Schwartzkroin, P.A.
(1992) Potentiation of spontaneous synaptic activity in rat
mossy cells. Neurosci. Lett., 142: 205–210.
Treves, A. and Rolls, E.T. (1992) Computational constraints
suggest the need for two distinct input systems to the hip-
pocampal CA3 network. Hippocampus, 2: 189–199.
Treves, A. and Rolls, E.T. (1994) Computational analysis of the
role of the hippocampus in memory. Hippocampus, 4:
374–391.
Walker, M.C., Ruiz, A. and Kullmann, D.M. (2001) Monosy-
naptic GABAergic signaling from dentate to CA3 with a
pharmacological and physiological profile typical of mossy
fiber synapses. Neuron, 29: 703–715.
Wenzel, H.J., Buckmaster, P.S., Anderson, N.L., Wenzel, M.E.
and Schwartzkroin, P.A. (1997) Ultrastructural localization
of neurotransmitter immunoreactivity in mossy cell axons
and their synaptic targets in the rat dentate gyrus. Hippo-
campus, 7: 559–570.
Wittner, L., Henze, D.A., Zaborszky, L. and Buzsaki, G. (2006)
Hippocampal CA3 pyramidal cells selectively innervate
aspiny interneurons. Eur. J. Neurosci., 24: 1286–1298.
Wolansky, T., Clement, E.A., Peters, S.R., Palczak, M.A. and
Dickson, C.T. (2006) Hippocampal slow oscillation: a novel
EEG state and its coordination with ongoing neocortical ac-
tivity. J. Neurosci., 26: 6213–6229.
Ylinen, A., Bragin, A., Nadasdy, Z., Jando, G., Szabo, I., Sik,
A. and Buzsáki, G. (1995) Sharp wave-associated high-fre-
quency oscillation (200 Hz) in the intact hippocampus: net-
work and intracellular mechanisms. J. Neurosci., 15: 30–46.
Zappone, C.A. and Sloviter, R.S. (2004) Translamellar disinhi-
bition in the rat hippocampal dentate gyrus after seizure-in-
duced degeneration of vulnerable hilar neurons. J. Neurosci.,
24: 853–864.
 Spontaneous Evoked
20 mV
0.5 mS
IC

EC

50 microV 40 microV
0.5ms 0.5ms

Plate 12.4. Extracellular waveforms of mossy cells during spontaneously and evoked spiking. A recording was obtained where the
intracellularly recorded mossy cell was also observed on the extracellular electrode. The extended shape of the mossy cell dendrites
allowed the extracellular signal to be observed on three shanks of the extracellular silicon probe (150 mm between shanks). The
extracellular electrode track is indicated by the white arrows. The soma of the mossy cell is indicated by the yellow arrow. There was a
difference in the shape of the action potentials, both intracellular and extracellular, suggesting differences in the site of spike initiation
for spontaneous and evoked spikes. (For B/W version, see page 206 in the volume.)
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 13

Interneurons of the dentate gyrus: an overview of cell

types, terminal fields and neurochemical identity

Carolyn R. Houser1,2,

1
Department of Neurobiology and Brain Research Institute, David Geffen School of Medicine at UCLA,
Los Angeles, CA 90095, USA
2
Research Service, VA Greater Los Angeles Healthcare System, West Los Angeles, Los Angeles, CA 90073, USA

Abstract: Interneurons of the dentate gyrus are a diverse group of neurons that use GABA as their primary
neurotransmitter. Morphological studies of these neurons have been challenging since no single neuro-
anatomical method provides a complete view of these interneurons. However, through the integration of
findings obtained from multiple methods, an interesting picture of this complex group of neurons is
emerging, and this review focuses on studies in rats and mice. In situ hybridization of mRNAs for the two
isoforms of the GABA synthesizing enzyme, glutamate decarboxylase (GAD65 and GAD67), demonstrates
the abundance of GABA neurons in the dentate gyrus and their high concentration in the hilus and along
the base of the granule cell layer. Likewise, immunohistochemical studies, particularly of GAD65, dem-
onstrate the rich fields of GABA terminals not only around the somata of granule cells but also in the
dendritic regions of the molecular layer. This broad group of GABA neurons and their terminals can be
subdivided according to their morphological characteristics, including the distribution of their axonal
plexus, and their neurochemical identity. Intracellular labeling of single interneurons has been instrumental
in demonstrating the extensiveness of their axonal plexus and the relatively specific spatial distribution of
their axonal fields. These findings have led to the broad classification of interneurons into those that
terminate primarily at perisomatic regions and those that innervate the dendrites of granule cells. The
interneurons also can be classified according to their neuropeptide and calcium-binding protein content.
These and other molecules contribute to the rich diversity of dentate interneurons and may provide
opportunities for selectively regulating specific groups of GABA neurons in the dentate gyrus in order to
enhance their function or protect vulnerable neurons from damage.

Keywords: GABA; glutamate decarboxylase (GAD); interneurons; hilus; parvalbumin; somatostatin;

calretinin

Introduction early morphological studies identified at least 21

Interneurons of the dentate gyrus are a particu-
larly fascinating and diverse group of neurons, and
Corresponding author. Tel.: +1 (310) 206-1567;
Fax: +1 (310) 825-2224; E-mail: houser@mednet.ucla.edu
cell types in the dentate hilus alone (Amaral,
1978). Despite such diversity, a unifying feature of
the dentate interneurons is their use of g-amino-
butyric acid (GABA) as their major neurotrans-
mitter. However, as in other cortical and
hippocampal regions, multiple subtypes of GABA

DOI: 10.1016/S0079-6123(07)63013-1 217

218
neurons can be distinguished, and several classifi-
cation systems have emerged. These include cate-
gorization according to the axonal distributions
and postsynaptic targets of the interneurons, their
neuropeptide or calcium-binding protein content,
and their physiological characteristics (Parra et al.,
1998; Maccaferri and Lacaille, 2003; Somogyi and
Klausberger, 2005; Jinno and Kosaka, 2006). Each
system highlights special features of the interneu-
ron population, and, while there is not a perfect
overlap between categories in the different sys-
tems, some relationships can be established. This
review will focus first on the broad distribution of
GABA neuron cell bodies and axon terminals
within the dentate gyrus, and then consider several
major categories of dentate interneurons and their
relationships to the overall distribution of GABA
neurons and their processes.
Distribution of GABA neurons
Demonstrating the entire population of GABA
neurons in the dentate gyrus in histological prep-
arations has proven to be an interesting challenge.
Immunohistochemical studies utilizing various an-
tisera to GABA and its synthesizing enzyme,
glutamate decarboxylase (GAD), have yielded
different and, occasionally, contradictory results
(Ribak et al., 1978; Sloviter and Nilaver, 1987;
Babb et al., 1988; Woodson et al., 1989; Jinno et
al., 1998). Although cell bodies in many regions of
the hippocampal formation are labeled with these
methods, several groups of neurons in the dentate
gyrus, particularly those in the hilus, are often in-
consistently labeled. Alternate methods of identi-
fying the cell bodies of GABA neurons in the
dentate gyrus are thus needed.
Currently, one of the most effective methods for
demonstrating the entire population of GABA
neurons in the dentate gyrus is in situ hybridiza-
tion of either GAD65 or GAD67 mRNA (Houser
et al., 2000, for review). Both forms of GAD can
synthesize GABA, but they have functional differ-
ences that include different interactions with the
cofactor pyridoxal phosphate (Erlander et al.,
1991; Kaufman et al., 1991; Martin et al., 1991;
Esclapez et al., 1994). However, the mRNAs for
both isoforms appear to be present in essentially
all GABAergic interneurons of the dentate gyrus
(Houser and Esclapez, 1994). Thus clear visuali-
zation of the entire population of GABA neuron
somata is possible, and the abundance of these
neurons within the relatively small region of the
dentate gyrus is striking (Fig. 1). Such in situ
hybridization methods for identifying GABA
neurons have the additional advantage of labeling
the cell bodies of these GABA neurons without
occlusion by axon terminals.
GAD mRNA-labeled neurons are distributed in
all three relatively distinct regions of the dentate
gyrus. First, GAD mRNA-labeled neurons are
highly concentrated in the hilus, and their distri-
bution often helps define the hilar region (Fig. 1A,
B). While labeled interneurons are abundant in the
deep parts of the hilus, between the tip of the hilus
and CA3, they are also numerous in the hilar
regions that are located beneath the granule cell
layer (Fig. 1A, C). Second, labeled neurons asso-
ciated with the granule cell layer are frequently
aligned along the base of this layer and may be
either highly concentrated or more dispersed along
this zone (Fig. 1A–C). Some GAD mRNA-labeled
cell bodies are also located within the granule cell
layer and along its outer border. Finally, GABA
neurons are distributed diffusely within the mo-
lecular layer, and a few are found along the hip-
pocampal fissure at the outer border of the
molecular layer (Fig. 1A, C). These basic patterns
of GABA neurons are present in both the dorsal
(septal) and ventral (temporal) regions of the dent-
ate gyrus (Fig. 1B).
The numbers of GAD mRNA-labeled neurons
in all three regions (hilus, granule cell layer and
molecular layer) generally appear to be larger than
those identified with other labeling methods, and
the numerous labeled neurons in the hilus are par-
ticularly notable. Their abundance could give the
impression that most hilar neurons are GABA
neurons. However, non-GABA neurons are also
present in the hilus (Fig. 2A), in approximately
equal numbers, and the vast majority of these are
mossy cells that use glutamate as their transmitter.
Double labeling for GAD mRNA and a general
neuronal marker, such as the neuron-specific nu-
clear protein NeuN, demonstrates the extensive
 219

Fig. 1. In situ hybridization of GAD65 in the rat dentate gyrus. (A) In a coronal section through the rostral dentate gyrus, labeled
neurons are numerous in the deep regions of the hilus (H), between the hilar tip and CA3, and in the region beneath the granule cell
layer (arrow). Labeled neurons are aligned along the base of the granule cell layer (G) and are dispersed in the molecular layer (M). (B)
In a coronal section through a more caudal level of the dentate gyrus, labeled neurons are concentrated in the dorsal and ventral hilus
and are numerous along the base of the granule cell layer (G). (C) At higher magnification, labeled neurons in the hilus (H) can be
distinguished from those at the base of the granule cell layer (G). Labeled neurons are evident at all laminar levels of the molecular
layer (M).

intermingling of GABA and non-GABA neurons
throughout the hilus (Fig. 2A). (Mossy cells are
generally considered to be principal cells and are
discussed elsewhere in this volume; see reviews by
Seress, Henze and Buszaki, and Scharfman.) Iden-
tifying the relatively large numbers of GAD
mRNA-labeled neurons in the dentate gyrus pro-
vides a base on which to consider subpopulations
of GABA neurons. Visualizing the entire group of
GABA neurons in the region also makes it clear
that only limited numbers of dentate GABA
neurons are being identified as the different sub-
groups of GABA neurons are labeled. As one
example, NADPH-diaphorase labels exclusively
GABAergic interneurons in the dentate gyrus,
and the labeled neurons are relatively abundant
(Valtschanoff et al., 1993; Czeh et al., 2005). Yet,
even though NADPH-diaphorase-labeled neurons
are present in all regions of the dentate gyrus
and are not restricted to a specific subtype of
interneuron, they represent only a portion of the
GABA neurons in each region (Fig. 2B).
220

Fig. 2. Double labeling of GAD65 mRNA and either NeuN (A) or NADPH-diaphorase (B) in the rat dentate gyrus. (A) GAD65
mRNA-labeled neurons (black) are numerous throughout the dentate hilus (H) where they are intermingled with NeuN single-labeled
(non-GABA) neurons (red; examples at arrows). Some GAD65 mRNA-labeled neurons are present in the granule cell layer (G) but are
most abundant at the base of this layer. Most neurons in the molecular layer (M) are labeled for GAD65 mRNA. Granule cells and
CA3 pyramidal cells are single-labeled for NeuN. (B) NADPH-diaphorase (dark blue) is present in only subgroups of the GAD
mRNA-labeled neurons (red) in the hilus (H), granule cell layer (G) and molecular layer (M) (examples at arrows). See Plate 13.2 in
Color Plate Section.

While it is expected that many peptides and cal-
cium-binding proteins will define specific sub-
groups of GABA neurons, as will be discussed in
later sections, it is less clear why immunohisto-
chemical labeling of GABA, as well as GAD65
and GAD67 proteins, does not label all GABA
neurons as reliably as the in situ hybridization
labeling of GAD mRNAs. One possibility is that
the actual levels of GAD and GABA in the cell
bodies of GABA neurons vary, and that this
influences the extent of cell body labeling. The
varying somal content could be related to physi-
ological conditions, including the levels of activity
of the neurons, as well as transport of GAD and
GABA from the soma to axon terminals.
As additional antibodies that recognize GAD
and GABA are obtained and new labeling meth-
ods are developed, the most effective methods for
labeling GABA neurons are likely to change.
However, currently in situ hybridization of either

GAD65 or GAD67 mRNA yields the most com-
plete and reliable labeling of cell bodies of the en-
tire population of GABA neurons in the dentate
gyrus. Immunohistochemical localization of the
GAD proteins generally provides labeling of some
but not all cell bodies, but these immunohisto-
chemical methods are essential for demonstrating
the extensive plexus of GABAergic terminals
within the dentate gyrus. The following descrip-
tions of immunohistochemical labeling will con-
sider GABA neurons in both rat and mouse. The
most detailed information on the distributions and
connections of dentate interneurons has been ob-
tained from rat, but there is considerable interest
in the characteristics of these neurons in mice, due
to the rapid increase in studies of genetically mod-
ified animals. Thus the illustrations of immuno-
labeling will be from mouse (C57/BL6) primarily.
Current findings suggest, however, that most mor-
phological and biochemical aspects of dentate
interneurons are quite similar in rat and mouse
(personal observations; Mátyás et al., 2004).
Axonal arborizations and postsynaptic targets of
GABA neurons
GABAergic terminals are abundant throughout the
dentate gyrus, but reach their highest densities in
the outer third of the molecular layer (Fig. 3A).
Slightly lower concentrations of GABAergic termi-
nals are evident in the inner two-thirds of this layer.
These high concentrations of GABAergic axon ter-
minals in the dendritic regions of the granule cells
are consistent with electron microscopic evidence
that, in quantitative terms, GABAergic synapses in
dendritic regions significantly outnumber those in
somatic regions of the dentate gyrus (75% vs. 25%)
(Halasy and Somogyi, 1993a). Within the granule
cell layer, GABAergic terminals surround the gran-
ule cell somata, and a narrow band of increased
labeling is present at the outer border of the layer
(Fig. 3A). The lowest concentration of GABAergic
terminals is found in the hilus where labeled termi-
nals are dispersed throughout the neuropil. While
the cell bodies of some hilar neurons are sur-
rounded by GABAergic terminals, others exhibit
little perisomatic labeling.
221
Several subgroups of GABA neurons contribute
to the rich fields of GABAergic terminals in the
dentate gyrus, and many of these interneurons
have extensive axonal arborizations not only
within a slice but also along the septotemporal
length of the dentate gyrus (Han et al., 1993;
Buckmaster and Schwartzkroin, 1995; Sik et al.,
1997). Indeed the axonal patterns and postsynaptic
targets of the interneurons prove to be the most
distinguishing morphological features of the differ-
ent subclasses of interneurons. The current classi-
fication of dentate interneurons is based primarily
on the studies of Somogyi and colleagues in which
interneurons were labeled intracellularly, and their
axonal plexus and targets were identified with light
and electron microscopic methods (Halasy and
Somogyi, 1993b; Han et al., 1993).
Interneurons in the dentate gyrus, as in other
cortical regions, can be divided into those that
form synapses at either perisomatic or dendritic
locations. Interneurons with synapses at periso-
matic sites can be divided further into those that
synapse on either the cell bodies and proximal
dendrites (basket cells) or the axon initial segments
(axoaxonic cells) of granule cells.
The cell bodies of many basket cells are located
along the base of the granule cell layer but are also
found within the granule cell layer and along its
outer border. While basket cells have a variety of
shapes and dendritic patterns, the pyramidal basket
cells are the most readily identified because of their
large cell bodies and long apical dendrites (Ribak
and Seress, 1983). But regardless of the dendritic
patterns, the axons are concentrated in the granule
cell layer, where each basket cell forms synaptic
contacts with the cell bodies of multiple granule cells
(Han et al., 1993). The distributions of the axon
terminals of basket cells on granule cell somata are
reflected in the numerous GAD-labeled terminals
throughout the granule cell layer (Fig. 3A). In ad-
dition, basket cells form synaptic contacts with
other GABA neurons, including other basket cells
(Freund and Buzsáki, 1996).
A second class of GABA neurons with periso-
matic targets is the axoaxonic cell. These inter-
neurons are amazingly selective in their targets
and, in the dentate gyrus, form synapses exclu-
sively on axon initial segments of granule cells
222

Fig. 3. GAD65-labeled axon terminals in the mouse dentate gyrus (A) and hypothetical cell types of origin (B). (A & B) Dense terminal
fields in the outer molecular layer (Mo) originate from hilar (H) and molecular layer GABA neurons that project to the perforant path
zone (HIPP and MOPP cells, respectively). Moderately dense terminal fields in the inner molecular layer (Mi) originate from hilar
interneurons that innervate the commissural-association projection region (HICAP cells). GABAergic terminals in the granule cell
layer (G) orginate from basket and axoaxonic cells that provide perisomatic innervation.

(Soriano and Frotscher, 1989; Soriano et al., 1990;
Han et al., 1993; DeFelipe, 1999). Their axons
branch extensively, and each branch forms a series
of synaptic contacts along an axon initial segment.
The numerous descending branches resemble the
candles or lights on a chandelier, and this led to
their early identification as chandelier cells in the
cerebral cortex (Szentágothai, 1974; Somogyi,
1977). The cell bodies of these neurons are often
located near the granule cell layer, including loca-
tions along the outer border of the layer, but can
also be found within the hilus. Because of the
proximal location of the axon terminals of basket
cells and axoaxonic cells, these neurons can exert
powerful control over the output of granule cells
and potentially contribute to synchronization of
granule cells through their contacts with multiple
granule cells (Miles et al., 1996).
The interneurons that terminate on granule cell
dendrites, and form the GABAergic plexus in the
molecular layer, can be divided into multiple
classes based on the location of their cell bodies
and the laminar distribution of their axon termi-
nals in the molecular layer (Halasy and Somogyi,
1993b; Han et al., 1993). Interestingly, the cell
bodies of neurons that innervate granule cell dend-
rites in the molecular layer are located in both the
hilus and the molecular layer itself. GABA neu-
rons with their cell bodies in the hilus and axon
projections to the outer two-thirds of the molec-
ular layer are identified as hilar perforant path-
associated (HIPP) cells. Likewise, neurons with

cell bodies in the hilus that project to the inner
third of the molecular layer are identified as
hilar commissural-association pathway-related
(HICAP) cells. Because of the location of the cell
bodies of these neurons in the hilus, they receive
considerable input from the granule cells and are
thus positioned to provide feedback inhibition to
granule cells near the location of either the exci-
tatory input from the perforant path or commis-
sural-association neurons (primarily mossy cells)
in the hilus (Han et al., 1993).
In contrast, GABA neurons with cell bodies in
the molecular layer are likely to receive excitatory
input directly from the perforant path or, poten-
tially, commissural-association fibers and thus
could provide feed-forward inhibition to granule
cell dendrites (Freund and Buzsáki, 1996). Molec-
ular layer neurons that innervate the outer molec-
ular layer have been named molecular layer
perforant path-associated (MOPP) cells (Halasy
and Somogyi, 1993b). Influences of these neurons
on granule cells could be critical for regulating the
response of the granule cells to their major exci-
tatory inputs.
As descriptions of specific cell types are ob-
tained, it is possible to relate the general locations
of the interneurons and their terminal fields to
the broad distribution of axon terminals in the
dentate gyrus (Fig. 3). Many of the cell bodies
along the base of the granule cell layer, and
within the layer, are likely to be basket cells and
axoaxonic cells that contribute to the GABAergic
terminals within the granule cell layer (Fig. 3). In
contrast, many GABA neurons in the hilus are
dendrite-innervating neurons and are a major
source of the GABAergic innervation of the mo-
lecular layer. The dense innervation in the outer
molecular layer is derived in part from the HIPP
cells whereas the more moderate densities of ter-
minals in the inner molecular region include ax-
onal projections from the HICAP cells (Fig. 3).
However, it is not possible to distinguish different
classes of hilar neurons, such as the HICAP and
HIPP cells, on the basis of their locations in the
hilus or the patterns of their proximal dendrites.
Most GABA neurons in the molecular layer pre-
sumably contribute to the GABAergic plexus
within the layer, with the highest concentration of
223
their terminals in the perforant path zone (MOPP
cells).
Two other subgroups of GABA neurons in the
dentate gyrus have unique patterns of axonal ter-
minations but cannot be readily distinguished by
their cell body locations, dendritic morphology, or
laminar distributions of their axons. The first is a
group of interneurons that selectively contact
other interneurons. Many of these GABA neu-
rons are located in the hilus and contain the cal-
cium-binding proteins calretinin and calbindin in
the rat (Gulyás et al., 1996). However, the chem-
ical identity of interneuron-specific GABA neu-
rons in the mouse is not yet known.
Another subgroup of interneurons in the hilus
projects beyond the dentate to the medial septum.
The hippocampo-septal neurons have been de-
scribed most extensively in CA1 and CA3 of the
rat (Alonso and Kohler, 1982; Toth and Freund,
1992). However, they are also found in the hilus,
and many express somatostatin (Zappone and
Sloviter, 2001; Jinno and Kosaka, 2002a; Gulyás
et al., 2003). In the septal region, these neurons
form synapses with parvalbumin (PV)-containing
GABA neurons that, in turn, project back to the
hippocampus and dentate gyrus and innervate
other GABA neurons (Freund and Antal, 1988;
Acsády et al., 2000b). This recurrent GABAergic
circuit is viewed as a potential regulator of hippo-
campal theta activity although the precise mech-
anism of action remains to be determined (Dragoi
et al., 1999).
Neurochemical identity
Interneurons of the dentate gyrus can also be clas-
sified according to their content of specific cal-
cium-binding proteins and neuropeptides, and a
number of distinct groups of interneurons have
been identified, even though overlap is found
among some subgroups. By studying the cell bod-
ies and axon terminals of these chemically defined
interneurons, it has been possible to ‘‘dissect’’ the
broad, complex pattern of GABAergic cell bodies
and terminals and relate several of the groups to
the morphologically identified cell types discussed
previously.
224

Parvalbumin dendrites that extend in the subgranular region

PV-containing interneurons have a restricted dis-
tribution of their axon terminals and form
synapses predominantly on cell bodies and axon
initial segments of granule cells, consistent with
their identity as basket and axo-axonic cells
(Ribak et al., 1990). PV-labeled cell bodies are lo-
cated primarily near the granule cell layer and are
most prominent at the base of the granule cell
layer (Fig. 4A). However, some PV-labeled inter-
neurons are also located near the junction of the
granule cell and molecular layers (Fig. 4A, C), and
smaller numbers of PV-labeled neurons are found
in the hilus and molecular layer. Many labeled
neurons along the base of the granule cell layer
resemble pyramidal basket cells, with labeled api-
cal dendrites that often branch and ascend through
the full extent of the molecular layer and basal
(Fig. 4A).
The PV-labeled neurons receive substantial in-
put from axon collaterals of the mossy fibers
through synaptic contacts on their cell bodies and
proximal dendrites within the granule cell layer
and also on their basal dendrites (Blasco-Ibánez
et al., 2000). These interneurons can thus provide
strong feedback inhibition of the granule cells.
In addition, the ascending dendrites of many
PV-labeled interneurons are positioned to receive
input from the major afferents to the dentate
gyrus, including the entorhinal cortex (Zipp et al.,
1989). Thus the PV-labeled interneurons can also
provide feed-forward inhibition to the granule
cells, in response to excitatory inputs to their
dendrites in the molecular layer.
Although PV-containing neurons are considered
to be a prominent class of interneurons in the

Fig. 4. Parvalbumin (PV)-labeled neurons in the mouse dentate gyrus. (A) Cell bodies of PV-labeled neurons are located at the base of
the granule cell layer (G), and their apical dendrites ascend through the granule cell layer and often branch as they extend through the
molecular layer (M). Basal dendrites are evident along the inner border of the granule cell layer (arrows). (B) A large PV-labeled
interneuron has the characteristics of a pyramidal basket cell and is located at the base of the granule cell layer (G). (C) A PV-labeled
multipolar interneuron in the inner molecular layer extends several dendrites (arrows) into the molecular layer (M). The axon is likely
to descend and contribute to the axonal plexus in the granule cell layer (G). (D) PV-labeled terminals are concentrated in the granule
cell layer where they surround the cell bodies and axon initial segments of granule cells.

dentate gyrus, the number of labeled cell bodies in
the dentate is often relatively low (Ribak et al.,
1990). The percentage of GAD67-labeled inter-
neurons that express PV in the granule cell layer is
estimated to be 20–25% in the mouse (Jinno and
Kosaka, 2002b). This contrasts with the consider-
ably higher 40% of interneurons that express PV
in the pyramidal cell layer of CA3 and CA1 of the
hippocampus (Jinno and Kosaka, 2002b). The
limited number of PV-labeled cell bodies could
reflect generally low numbers of PV-containing
dentate interneurons, or, possibly, a low level of
the protein in the cell bodies despite the presence
of PV in the axon terminals. Such differential in-
tracellular labeling has been suggested in some ex-
perimental and pathological conditions, including
human temporal lobe epilepsy, in which cell body
labeling of PV is substantially decreased despite
persistent terminal labeling (Sloviter et al., 1991;
Scotti et al., 1997; Wittner et al., 2001).
The precise function of PV in the interneurons
remains uncertain, but the presence of this cal-
cium-binding protein could be a physiological
adaptation related to the high levels of activity of
this group of neurons and their identification
as fast-spiking neurons in several brain regions
(Freund and Buzsáki, 1996). Interestingly, PV-
labeled interneurons express particularly high
levels of cytochrome c in the rat, and this, along
with high numbers of mitochondrial profiles in
their axon terminals, is consistent with high levels
of metabolic activity in this group of GABA neu-
rons (Gulyás et al., 2006).
Cholecystokinin
The peptide cholecystokinin (CCK) is also located
in interneurons near the granule cell layer, and these
neurons constitute a second group of GABAergic
basket cells. There appears to be little overlap
between the CCK- and PV-labeled neurons in the
hippocampus, and this suggests that there are differ-
ent groups of basket cells with potentially different
but complementary functions, as has been demon-
strated for basket cells in CA1 (Klausberger et al.,
2005). Relatively few CCK-labeled cell bodies are
found in the dentate gyrus, but their terminal fields
225
resemble the patterns of PV-containing terminals
within the granule cell layer, including a slightly
increased density of fibers and terminals along the
outer border of the granule cell layer. In contrast
to the PV-containing group, CCK-labeled neurons
are restricted to the basket cell population and
apparently do not contribute to the axoaxonic sub-
group, at least in the rat.
Somatostatin
Somatostatin neurons constitute one of the largest
chemically defined subgroups of GABA neurons in
the dentate gyrus, and virtually all somatostatin
neurons of this region are located within the
hilus (Fig. 5A). Furthermore, approximately 55%
of hilar GABA neurons express somatostatin
(Esclapez and Houser, 1995; Buckmaster and Jon-
gen-Rêlo, 1999; Jinno and Kosaka, 2003).
In both mouse and rat, somatostatin-containing
neurons are more abundant, and constitute a
slightly higher percentage of total GABA neurons
in the ventral than in the dorsal dentate gyrus
(Buckmaster et al., 1994; Esclapez and Houser,
1995; Jinno and Kosaka, 2003).
While somatostatin-labeled cell bodies are con-
fined primarily to the hilus in the dentate gyrus,
their terminals fields are most prominent in the
outer molecular layer where the perforant path
fibers terminate (Bakst et al., 1986; Katona et al.,
1999). Thus many of the somatostatin neurons in
the hilus are considered to be HIPP cells, based on
the location of their cell bodies and axon termi-
nals. The dendrites of somatostatin neurons ex-
tend for considerable distances across the hilus but
generally remain within the region (Fig. 5B).
Although the proximal dendrites of many of
the somatostatin interneurons are smooth and
sparsely spinous, as is characteristic of many
GABAergic interneurons (Ribak, 1978), the more
distal dendrites have a considerable number of
spines (Fig. 5C). These extensive dendrites and their
spines are likely targets of the granule cell mossy
fibers, and the mossy fiber contacts on GABA neu-
rons outnumber those on mossy cells by a ratio of
approximately 5:1 (Acsády et al., 1998). The som-
atostatin neurons, in turn, form a moderately dense
226

Fig. 5. Somatostatin-containing neurons in the ventral dentate gyrus of the mouse. (A) Numerous somatostatin-immunoreactive cell
bodies are present in the hilus (H), and a labeled axonal plexus is evident in the outer molecular layer (M). Little labeling is found in the
granule cell layer (G) except for axons from hilar somatostatin neurons that project through this layer. (B) Localization of green
fluorescent protein (GFP) in somatostatin neurons of a transgenic mouse reveals the extensive dendrites of these neurons. Several
dendrites extend across large extents of the hilus (arrows). (C) GFP-labeled dendrites extend from a multipolar hilar neuron, and small
spines can be detected on some dendritic segments (arrows).

axonal plexus in the outer two-thirds of the molec-
ular layer (Bakst et al., 1986). The somatostatin
terminals synapse primarily with granule cell dend-
rites and are often located adjacent to asymmetric
excitatory synapses (Katona et al., 1999). Som-
atostatin neurons are thus ideally positioned to
modulate the influence of the perforant path on
dentate granule cells in response to ongoing granule
cell activity (see chapter by Tallent in this volume).
Some hilar somatostatin neurons project beyond
the dentate gyrus to the medial septum. In both
mouse and rat, a high proportion of the hippo-
campo-septal neurons in the hilus express som-
atostatin (Zappone and Sloviter, 2001; Jinno and
Kosaka, 2002a). Furthermore, retrograde labeling
studies in the mouse indicate that a surprisingly
high percentage (44%) of somatostatin-labeled
neurons in the hilus project to the septal region
(Jinno and Kosaka, 2002a). Currently, it is unclear
whether separate groups of GABA neurons
project to either the dentate molecular layer or
the medial septum, or whether single somatostatin
neurons project to both regions. Regardless, the
high percentage of hilar somatostatin neurons that
innervate the septum suggests that somatostatin as
well as GABA could have important functional
roles in the hippocampal-septal circuitry.
Somatostatin neurons in the dentate gyrus are
highly vulnerable to excitatory and ischemic dam-
age, and loss of these neurons is commonly ob-
served in models of temporal lobe epilepsy and
forebrain ischemia (e.g. Johansen et al., 1987;
Sloviter, 1987; Buckmaster and Dudek, 1997).
Characteristics of these neurons that may contrib-
ute to their vulnerability include the high levels
of excitatory mossy fiber contacts (Acsády
et al., 1998) and the relatively low GABAergic
innervation on the soma and proximal dendrites of
these interneurons (Acsády et al., 2000a).
Calretinin
Most classes of GABAergic interneurons in the
dentate gyrus appear to be quite similar in their
patterns of neuropeptide and calcium-binding
 227

Fig. 6. Comparisons of calretinin labeling in the dorsal dentate gyrus of mouse (A) and rat (B). (A) Calretinin labeling in the mouse
includes darkly-labeled (arrows) and moderately-labeled (arrowheads) neurons in the hilus (H) that correspond to GABAergic in-
terneurons and mossy cells, respectively. In addition, small labeled cell bodies at the base of the granule cell layer (G) are likely to be
newly-generated granule cells. A dense band of labeled axon terminals (asterisk) from the mossy cells is prominent in the inner
molecular layer (M). (B) In the rat, darkly-labeled neurons (arrows) in the hilus (H) are presumed to be GABA neurons. A narrow
band of labeled terminals is evident along the outer border of the granule cell layer (G), and these may originate in the sup-
ramammillary region.

protein expression in rats and mice. However, sev-
eral species differences are found in the neurons
that express the calcium-binding protein calretinin
(Liu et al., 1996; Blasco-Ibáñez and Freund, 1997).
Most notably, in the mouse, calretinin is expressed
in mossy cells of the hilus, and such mossy cells are
particularly numerous in the ventral hilus (Blasco-
Ibáñez and Freund, 1997; Fujise et al., 1998).
Calretinin is present in axon terminals of the
mossy cells, and this creates a dense band of labe-
ling in the inner third of the molecular layer
(the major terminal field of the mossy cells) in the
mouse (Fig. 6A). In contrast, labeling is primarily
restricted to GABAergic interneurons of the dent-
ate gyrus in the rat, and only a narrow band of
calretinin labeling is present in the supragranular
region in the rat (Fig. 6B). These terminals are
not from mossy cells but may originate from the
supramammillary region and are considered to be
glutamatergic (Maglóczky et al., 1994; Kiss et al.,
2000).
Despite calretinin labeling of the mossy cells in
the mouse, some GABAergic interneurons that co-
express calretinin can be detected (Gulyás et al.,
1996). The GABA neurons are located primarily in
the hilus, near the granule cell layer, and some-
times can be distinguished by their darker labeling
compared to that of mossy cells in the region
(Fig. 6A). Another difference between the two
species is the lack of spiny calretinin-containing
neurons in the mouse (Blasco-Ibáñez and Freund,
1997; Fujise et al., 1998; Mátyás et al., 2004). Such
neurons are distinctive hilar neurons in the rat
and belong to the subgroup of GABA neurons
that contacts primarily other interneurons (Gulyás
et al., 1996). Finally, in the mouse, newborn granule
cells along the base of the granule cell layer
express calretinin (Fig. 6A) (Liu et al., 1996;
Blasco-Ibáñez and Freund, 1997). Likewise, Cajal-
Retzius cells that are presumed to use glutamate as
their neurotransmitter are labeled for calretinin in
the mouse.
Thus, calretinin labeling of cell bodies in the
dentate gyrus is largely restricted to GABA neu-
rons in the rat but is present in several groups
of glutamatergic, as well as some GABAergic,
neurons, in the mouse.
Neuropeptide Y
Within the dentate gyrus, neuropeptide Y (NPY)-
labeled cell bodies are located primarily in the
hilus (Fig. 7) and constitute one of the larger
groups of hilar GABA neurons. It has been esti-
mated that NPY-containing neurons account for
228
Fig. 7. NPY-labeled neurons in the dorsal (A) and ventral (B) hilus of the mouse dentate gyrus. (A & B) Labeled neurons are largely
confined to the hilus (H) at all levels of the dentate gyrus. Labeled neurons are relatively numerous in the hilus but are sparse in the
granule cell (G) and molecular (M) layers as well as the pyramidal cell layer of CA3.
approximately 57–82% of GAD67-labeled neu-
rons in the dorsal and ventral hilus, respectively
(Jinno and Kosaka, 2003). These relatively high
percentages suggest that there is considerable co-
localization of NPY and somatostatin in GABA
neurons in the hilus (Kohler et al., 1987; Deller
and Leranth, 1990). Small numbers of NPY-labe-
led cell bodies are also located at the base of the
granule cell layer and in the molecular layer. The
axon terminal labeling of NPY-expressing neurons
is generally limited with current immunohisto-
chemical methods, but some terminals are evident
in
the outer molecular layer, suggesting that many of
the NPY-containing neurons are HIPP cells and
resemble somatostatin neurons. Some NPY-
labeled axon terminals are also present in the
hilus, and many of their targets are likely to be
mossy cells (Deller and Leranth, 1990; Acsády
et al., 2000a).
Further characterization of dentate interneurons
The presence of specific neuropeptides and cal-
cium-binding proteins in selected interneurons has
been extremely useful for identifying major classes
of GABA neurons in the hippocampal formation.
Yet, in some regions, the number of these neurons
appears small when compared to the total number
of GABA neuron cell bodies and terminals.
GABA neurons in two locations within the dent-
ate gyrus appear to be particularly underrepre-
sented by the currently available markers for
subgroups of GABA neurons. These include the
numerous GABA neurons in the hilar or poly-
morph region that is located below the base of the
granule cell layer and GABA neurons in the mo-
lecular layer.
Other molecules however call attention to some
of these GABA neurons. For example, virtually all
of the molecular layer interneurons express high
levels of the a1 subunit of the GABAA receptor, as
do many, but not all, groups of interneurons in the
dentate gyrus (Fig. 8) (Gao and Fritschy, 1994;
Esclapez et al., 1996; Sperk et al., 1997; Peng et al.,
2004). In addition, these molecular layer interneu-
rons express the d subunit of the GABAA receptor
and, consistent with such expression, exhibit tonic
inhibition (Glykys et al., 2007). Activation of
GABAA receptors on the cell bodies of GABA
neurons could be an additional mechanism
through which the activity of GABA neurons is
modulated or limited. Inhibition of even small
numbers of GABA neurons that have extensive
axonal projections, such as those in the molecular
layer, could reduce inhibition and thus enhance
excitation of the granule cells.
Numerous other peptides and proteins are local-
ized in GABA neurons in the dentate gyrus, and
their identities and distributions highlight addi-
tional, unique characteristics of these interneurons.

Fig. 8. GABAA receptor a1 subunit labeling in the mouse
dentate gyrus. The a1 subunit is highly expressed on the surface
of cell bodies and dendritic processes of interneurons in the
granule cell (G) and molecular (M) layers (examples at arrows).
Labeled neurons at the base of the granule cell layer closely
resemble parvalbumin-labeled pyramidal basket cells (compare
with Fig. 4B).
Some expression patterns mirror the major subdi-
visions of interneurons, based on their innervation
of perisomatic or dendritic regions, for which PV
and somatostatin serve as prototypic markers,
respectively. For example, mu-opioid receptors are
preferentially localized on PV-labeled neurons
whereas delta-opioid receptors are more abundant
on somatostatin neurons (Commons and Milner,
1996; Stumm et al., 2004; Drake and Milner, 2006).
Likewise, the a1 subunit of the GABAA receptor is
highly expressed on PV-labeled neurons, as well as
other interneurons in the dentate gyrus, but shows
very little or no expression on somatostatin neurons
in the hilus (Gao and Fritschy, 1994; Esclapez et al.,
1996).
Differences between these major groups of dent-
ate interneurons are also reflected in their vulner-
ability to ischemic and seizure-induced damage.
Somatostatin interneurons are some of the most
vulnerable neurons in the dentate gyrus, whereas
PV-containing neurons are more resistant to dam-
age in some models (Sloviter, 1987; Sperk et al.,
229
1992; Buckmaster and Dudek, 1997). Continued
characterization of the vulnerable groups of
GABA neurons could eventually lead to specific
neuroprotective strategies.
Other molecules are localized primarily in inter-
neurons but are found in several subtypes of
GABA neurons in the hippocampal formation.
These includes Reelin (Pesold et al., 1998) and
nerve growth factor (Lauterborn et al., 1993;
Pascual et al., 1998). The preferential localization
of these particular substances in interneurons
suggests roles for these GABA neurons in basic
developmental processes that extend beyond their
essential roles in regulating information process-
ing, rhythm generation and excitability of the hip-
pocampal formation.
Conclusions
Despite years of study, GABA neurons in the dent-
ate gyrus continue to hold surprises and challenges.
No single morphological method provides an ade-
quate view of the richness and complexity of these
neurons. Thus the findings obtained from different
methods are pieces of a puzzle that require assem-
bly. Key features of the emerging picture are the
extensive and yet highly defined axonal plexus
of single interneurons in the dentate gyrus; the rich
repertoire of molecules that each interneuron pos-
sesses; and the remarkable functional regulation
of levels of GABA, GAD and many neuropeptides
in these interneurons. While such complexity can
be daunting, it also provides opportunities for dis-
covering new ways to enhance normal GABAergic
function and possibly protects these critical neurons
from damage in pathological conditions.
Acknowledgments
I thank Christine Huang and Siroun Tahtakran
for excellent histologic and photographic work,
and in situ hybridization labeling, respectively, and
gratefully acknowledge the numerous contribu-
tions of Drs. Monique Esclapez, Zechun Peng and
Nianhui Zhang to our studies of GABA neurons
in the hippocampus. These studies were supported
230
by Veterans Administration Medical Research
Funds and National Institutes of Health grants
NS046524 and NS051311 to C.R.H.
References
Acsády, L., Kamondi, A., Sik, A., Freund, T. and Buzsáki, G.
(1998) GABAergic cells are the major postsynaptic targets
of mossy fibers in the rat hippocampus. J. Neurosci., 18:
3386–3403.
Acsády, L., Katona, I., Martinez-Guijarro, F.J., Buzsáki, G.
and Freund, T.F. (2000a) Unusual target selectivity of
perisomatic inhibitory cells in the hilar region of the rat hip-
pocampus. J. Neurosci., 20: 6907–6919.
Acsády, L., Pascual, M., Rocamora, N., Soriano, E. and
Freund, T.F. (2000b) Nerve growth factor but not neurotro-
phin-3 is synthesized by hippocampal GABAergic neurons
that project to the medial septum. Neuroscience, 98: 23–31.
Alonso, A. and Kohler, C. (1982) Evidence for separate
projections of hippocampal pyramidal and non-pyramidal
neurons to different parts of the septum in the rat brain.
Neurosci. Lett., 31: 209–214.
Amaral, D.G. (1978) A Golgi study of cell types in the hilar
region of the hippocampus in the rat. J. Comp. Neurol., 182:
851–914.
Babb, T.L., Pretorius, J.K., Kupfer, W.R. and Brown, W.J.
(1988) Distribution of glutamate-decarboxylase-immunore-
active neurons and synapses in the rat and monkey hippo-
campus: light and electron microscopy. J. Comp. Neurol.,
278: 121–138.
Bakst, I., Avendano, C., Morrison, J.H. and Amaral, D.G.
(1986) An experimental analysis of the origins of somatosta-
tin-like immunoreactivity in the dentate gyrus of the rat. J.
Neurosci., 6: 1452–1462.
Blasco-Ibáñez, J.M. and Freund, T.F. (1997) Distribution, ul-
trastructure, and connectivity of calretinin-immunoreactive
mossy cells of the mouse dentate gyrus. Hippocampus, 7:
307–320.
Blasco-Ibánez, J.M., Martı́nez-Guijarro, F.J. and Freund, T.F.
(2000) Recurrent mossy fibers preferentially innervate parv-
albumin-immunoreactive interneurons in the granule cell
layer of the rat dentate gyrus. Neuroreport, 11: 3219–3225.
Buckmaster, P.S. and Dudek, F.E. (1997) Neuron loss, granule
cell axon reorganization, and functional changes in the dent-
ate gyrus of epileptic kainate-treated rats. J. Comp. Neurol.,
385: 385–404.
Buckmaster, P.S. and Jongen-Rêlo, A.L. (1999) Highly specific
neuron loss preserves lateral inhibitory circuits in the dentate
gyrus of kainate-induced epileptic rats. J. Neurosci., 19:
9519–9529.
Buckmaster, P.S., Kunkel, D.D., Robbins, R.J. and Schwa-
rtzkroin, P.A. (1994) Somatostatin-immunoreactivity in the
hippocampus of mouse, rat, guinea pig, and rabbit. Hippo-
campus, 4: 167–180.
Buckmaster, P.S. and Schwartzkroin, P.A. (1995) Interneurons
and inhibition in the dentate gyrus of the rat in vivo. J.
Neurosci., 15(1): 774–789.
Commons, K.G. and Milner, T.A. (1996) Cellular and subcel-
lular localization of delta opioid receptor immunoreactivity
in the rat dentate gyrus. Brain Res., 738: 181–195.
Czeh, B., Hajnal, A. and Seress, L. (2005) NADPH-diaphorase
positive neurons of the rat hippocampal formation: regional
distribution, total number and colocalization with calcium
binding proteins. Prague Med. Rep., 106: 261–274.
DeFelipe, J. (1999) Chandelier cells and epilepsy. Brain,
122(Pt 10): 1807–1822.
Deller, T. and Leranth, C. (1990) Synaptic connections of ne-
uropeptide Y (NPY) immunoreactive neurons in the hilar area
of the rat hippocampus. J. Comp. Neurol., 300: 433–447.
Dragoi, G., Carpi, D., Recce, M., Csicsvari, J. and Buzsaki, G.
(1999) Interactions between hippocampus and medial septum
during sharp waves and theta oscillation in the behaving rat.
J. Neurosci., 19: 6191–6199.
Drake, C.T. and Milner, T.A. (2006) Mu opioid receptors are
extensively co-localized with parvalbumin, but not som-
atostatin, in the dentate gyrus. Neurosci. Lett., 403: 176–180.
Erlander, M.G., Tillakaratne, N.J.K., Feldblum, S., Patel, N.
and Tobin, A.J. (1991) Two genes encode distinct glutamate
decarboxylases. Neuron, 7: 91–100.
Esclapez, M., Chang, D.K. and Houser, C.R. (1996) Sub-
populations of GABA neurons in the dentate gyrus express
high levels of the a1 subunit of the GABAA receptor.
Hippocampus, 6: 225–238.
Esclapez, M. and Houser, C.R. (1995) Somatostatin neurons
are a subpopulation of GABA neurons in the rat dentate
gyrus: evidence from co-localization of pre-prosomatostatin
and glutamate decarboxylase mRNAs. Neuroscience, 64:
339–355.
Esclapez, M., Tillakaratne, N.J.K., Kaufman, D.L., Tobin, A.J.
and Houser, C.R. (1994) Comparative localization of two
forms of glutamic acid decarboxylase and their mRNAs in
rat brain supports the concept of functional differences be-
tween the forms. J. Neurosci., 14: 1834–1855.
Freund, T.F. and Antal, M. (1988) GABA-containing neurons
in the septum control inhibitory interneurons in the hippo-
campus. Nature, 336: 170–173.
Freund, T.F. and Buzsáki, G. (1996) Interneurons of the hip-
pocampus. Hippocampus, 6: 347–470.
Fujise, N., Liu, Y., Hori, N. and Kosaka, T. (1998) Distribution
of calretinin immunoreactivity in the mouse dentate gyrus: II.
Mossy cells, with special reference to their dorsoventral
difference in calretinin immunoreactivity. Neuroscience,
82: 181–200.
Gao, B. and Fritschy, J.M. (1994) Selective allocation of
GABAA receptors containing the a1 subunit to neurochem-
ically distinct subpopulations of rat hippocampal interneu-
rons. Eur. J. Neurosci., 6: 837–853.
Glykys, J., Peng, Z., Chandra, D., Homanics, G.E., Houser,
C.R. and Mody, I. (2007) A new naturally occurring GABAA
receptor subunit partnership with high sensitivity to ethanol.
Nat. Neurosci., 10: 40–48.

Gulyás, A.I., Hajos, N., Katona, I. and Freund, T.F. (2003)
Interneurons are the local targets of hippocampal inhibitory
cells which project to the medial septum. Eur. J. Neurosci.,
17: 1861–1872.
Gulyás, A.I., Buzsáki, G., Freund, T.F. and Hirase, H. (2006)
Populations of hippocampal inhibitory neurons express
different levels of cytochrome c. Eur. J. Neurosci., 23:
2581–2594.
Gulyás, A.I., Hájos, N. and Freund, T.F. (1996) Interneurons
containing calretinin are specialized to control other inter-
neurons in the rat hippocampus. J. Neurosci., 16(10):
3397–3411.
Halasy, K. and Somogyi, P. (1993a) Distribution of
GABAergic synapses and their targets in the dentate gyrus
of rat: a quantitative immunoelectron microscopic analysis.
J. Hirnforsch, 34: 299–308.
Halasy, K. and Somogyi, P. (1993b) Subdivision in the multiple
GABAergic innervation of granule cells in the dentate gyrus of
the rat hippocampus. Eur. J. Neurosci., 5: 411–429.
Han, Z.-S., Buhl, E.H., Lorinczi, Z. and Somogyi, P. (1993)
A high degree of spatial selectivity in the axonal and dendritic
domains of physiologically identified local-circuit neurons in
the dentate gyrus of the rat hippocampus. Eur. J. Neurosci.,
5: 395–410.
Houser, C.R. and Esclapez, M. (1994) Localization of
mRNAs encoding two forms of glutamic acid decarboxy-
lase in the rat hippocampal formation. Hippocampus, 5:
530–545.
Houser, C.R., Esclapez, M. and Zhang, N. (2000) GABA Neu-
rons of the hippocampal formation: localization, vulnerability
and plasticity. In: Martin D.L. and Olsen R.W. (Eds.), GABA
in the Nervous System: The View at Fifty Years. Lippincott
Williams & Wilkins, Philadelphia, PA, pp. 337–355.
Jinno, S., Aika, Y., Fukuda, T. and Kosaka, T. (1998)
Quantitative analysis of GABAergic neurons in the mouse
hippocampus, with optical disector using confocal laser scan-
ning microscope. Brain Res., 814(1–2): 55–70.
Jinno, S. and Kosaka, T. (2002a) Immunocytochemical char-
acterization of hippocamposeptal projecting GABAergic
nonprincipal neurons in the mouse brain: a retrograde labe-
ling study. Brain Res., 945: 219–231.
Jinno, S. and Kosaka, T. (2002b) Patterns of expression of
calcium binding proteins and neuronal nitric oxide synthase
in different populations of hippocampal GABAergic neurons
in mice. J. Comp. Neurol., 449: 1–25.
Jinno, S. and Kosaka, T. (2003) Patterns of expression of
neuropeptides in GABAergic nonprincipal neurons in the
mouse hippocampus: quantitative analysis with optical
disector. J. Comp. Neurol., 461: 333–349.
Jinno, S. and Kosaka, T. (2006) Cellular architecture of the
mouse hippocampus: a quantitative aspect of chemically
defined GABAergic neurons with stereology. Neurosci. Res.,
56: 229–245.
Johansen, F.F., Zimmer, J. and Diemer, N.H. (1987) Early loss
of somatostatin neurons in dentate hilus after cerebral
ischemia in the rat precedes CA-1 pyramidal cell loss. Acta
Neuropathol. (Berl.), 73: 110–114.
231
Katona, I., Acsády, L. and Freund, T. (1999) Postsynaptic tar-
gets of somatostatin-immunoreactive interneurons in the rat
hippocampus. Neuroscience, 88: 37–55.
Kaufman, D.L., Houser, C.R. and Tobin, A.J. (1991) Two
forms of the gamma-aminobutyric acid synthetic enzyme
glutamate decarboxylase have distinct intraneuronal distri-
butions and cofactor interactions. J. Neurochem., 56:
720–723.
Kiss, J., Csaki, A., Bokor, H., Shanabrough, M. and Leranth,
C. (2000) The supramammillo-hippocampal and supramam-
millo-septal glutamatergic/aspartatergic projections in the
rat: a combined [3H]D-aspartate autoradiographic and
immunohistochemical study. Neuroscience, 97: 657–669.
Klausberger, T., Marton, L.F., O’Neill, J., Huck, J.H., Dale-
zios, Y., Fuentealba, P., Suen, W.Y., Papp, E., Kaneko, T.,
Watanabe, M., Csicsvari, J. and Somogyi, P. (2005) Com-
plementary roles of cholecystokinin and parvalbumin-ex-
pressing GABAergic neurons in hippocampal network
oscillations. J. Neurosci., 25: 9782–9793.
Kohler, C., Eriksson, L.G., Davies, S. and Chan-Palay, V.
(1987) Co-localization of neuropeptide tyrosine and som-
atostatin immunoreactivity in neurons of individual subfields
of the rat hippocampal region. Neurosci. Lett., 78: 1–6.
Lauterborn, J.C., Tran, T.M., Isackson, P.J. and Gall, C.M.
(1993) Nerve growth factor mRNA is expressed by GAB-
Aergic neurons in rat hippocampus. Neuroreport, 5:
273–276.
Liu, Y., Fujise, N. and Kosaka, T. (1996) Distribution of ca-
lretinin immunoreactivity in the mouse dentate gyrus. I.
General description. Exp. Brain Res., 108: 389–403.
Maccaferri, G. and Lacaille, J.C. (2003) Interneuron diversity
series: hippocampal interneuron classifications — making
things as simple as possible, not simpler. Trends Neurosci.,
26: 564–571.
Maglóczky, Z., Acsády, L. and Freund, T. (1994) Principal
cells are the postsynaptic targets of supramammillary
afferents in the hippocampus of the rat. Hippocampus, 4:
322–334.
Martin, D.L., Martin, S.B., Wu, S.J. and Espina, N. (1991)
Cofactor interactions and the regulation of glutamate dec-
arboxylase activity. Neurochem. Res., 16: 243–249.
Mátyás, F., Freund, T.F. and Gulyás, A.I. (2004) Immunocyto-
chemically defined interneuron populations in the hippocam-
pus of mouse strains used in transgenic technology.
Hippocampus, 14: 460–481.
Miles, R., Tóth, K., Gulyás, A.I., Hájos, N. and Freund, T.F.
(1996) Differences between somatic and dendritic inhibition
in the hippocampus. Neuron, 16: 815–823.
Parra, P., Gulyas, A.I. and Miles, R. (1998) How many sub-
types of inhibitory cells in the hippocampus? Neuron, 20:
983–993.
Pascual, M., Rocamora, N., Acsády, L., Freund, T.F. and
Soriano, E. (1998) Expression of nerve growth factor and
neurotrophin-3 mRNAs in hippocampal interneurons: mor-
phological characterization, levels of expression, and colo-
calization of nerve growth factor and neurotrophin-3. J.
Comp. Neurol., 395: 73–90.
232
Peng, Z., Huang, C.S., Stell, B.M., Mody, I. and Houser, C.R.
(2004) Altered expression of the d subunit of the GABAA
receptor in a mouse model of temporal lobe epilepsy. J. Ne-
urosci., 24: 8629–8639.
Pesold, C., Impagnatiello, F., Pisu, M.G., Uzunov, D.P., Costa,
E., Guidotti, A. and Caruncho, H.J. (1998) Reelin is pref-
erentially expressed in neurons synthesizing gamma-amino-
butyric acid in cortex and hippocampus of adult rats. Proc.
Natl. Acad. Sci. U.S.A., 95: 3221–3226.
Ribak, C.E. (1978) Aspinous and sparsely-spinous stellate neu-
rons in the visual cortex of rats contain glutamic acid dec-
arboxylase. J. Neurocytol., 7: 461–478.
Ribak, C.E., Nitsch, R. and Seress, L. (1990) Proportion of
parvalbumin-positive basket cells in the GABAergic inner-
vation of pyramidal and granule cells of the rat hippocampal
formation. J. Comp. Neurol., 300: 449–461.
Ribak, C.E. and Seress, L. (1983) Five types of basket cell in the
hippocampal dentate gyrus: a combined Golgi and electron
microscopic study. J. Neurocytol., 12: 577–597.
Ribak, C.E., Vaughn, J.E. and Saito, K. (1978) Immunocyto-
chemical localization of glutamic acid decarboxylase in neu-
ronal somata following colchicine inhibition of axonal
transport. Brain Res., 140: 315–332.
Scotti, A.L., Bollag, O., Kalt, G. and Nitsch, C. (1997) Loss of
perikaryal parvalbumin immunoreactivity from surviving
GABAergic neurons in the CA1 field of epileptic gerbils.
Hippocampus, 7: 524–535.
Sik, A., Penttonen, M. and Buzsaki, G. (1997) Interneurons in
the hippocampal dentate gyrus: an in vivo intracellular study.
Eur. J. Neurosci., 9: 573–588.
Sloviter, R.S. (1987) Decreased hippocampal inhibition and
a selective loss of interneurons in experimental epilepsy.
Science, 235: 73–76.
Sloviter, R.S. and Nilaver, G. (1987) Immunocytochemical lo-
calization of GABA-, cholecystokinin-, vasoactive intestinal
polypeptide-, and somatostatin-like immunoreactivity in the
area dentata and hippocampus of the rat. J. Comp. Neurol.,
256: 42–60.
Sloviter, R.S., Sollas, A.L., Barbaro, N.M. and Laxer, K.D.
(1991) Calcium-binding protein (calbindin-D28K) and parv-
albumin immunocytochemistry in the normal and epileptic
human hippocampus. J. Comp. Neurol., 308: 381–396.
Somogyi, P. (1977) A specific ‘axo-axonal’ interneuron in the
visual cortex of the rat. Brain Res., 136: 345–350.
Somogyi, P. and Klausberger, T. (2005) Defined types of cor-
tical interneurone structure space and spike timing in the
hippocampus. J. Physiol., 562: 9–26.
Soriano, E. and Frotscher, M. (1989) A GABAergic axo-axonic
cell in the fascia dentata controls the main excitatory hippo-
campal pathway. Brain Res., 503: 170–174.
Soriano, E., Nitsch, R. and Frotscher, M. (1990) Axo-axonic
chandelier cells in the rat fascia dentata: Golgi-electron
microscopy and immunocytochemical studies. J. Comp.
Neurol., 293: 1–25.
Sperk, G., Marksteiner, J., Gruber, B., Bellmann, R., Mahata,
M. and Ortler, M. (1992) Functional changes in neuropeptide
Y- and somatostatin-containing neurons induced by limbic
seizures in the rat. Neuroscience, 50: 831–846.
Sperk, G., Schwarzer, C., Tsunashima, K., Fuchs, J. and
Sieghart, W. (1997) GABAA receptor subunits in the rat
hippocampus I: immunocytochemical distribution of 13 sub-
units. Neuroscience, 80: 987–1000.
Stumm, R.K., Zhou, C., Schulz, S. and Hollt, V. (2004)
Neuronal types expressing mu- and delta-opioid receptor
mRNA in the rat hippocampal formation. J. Comp. Neurol.,
469: 107–118.
Szentágothai, J. (1974) Conceptual models of neural organiza-
tion. Neurosci. Res. Prog. Bull., 12: 307–510.
Toth, K. and Freund, T.F. (1992) Calbindin D28k-containing
nonpyramidal cells in the rat hippocampus: their immuno-
reactivity for GABA and projection to the medial septum.
Neuroscience, 49: 793–805.
Valtschanoff, J.G., Weinberg, R.J., Kharazia, V.N.,
Nakane, M. and Schmidt, H.H. (1993) Neurons in rat hip-
pocampus that synthesize nitric oxide. J. Comp. Neurol., 331:
111–121.
Wittner, L., Maglóczky, Z., Borhegyi, Z., Halász, P., Tóth, S.,
Eröss, L., Szabó, Z. and Freund, T.F. (2001) Preserva-
tion of perisomatic inhibitory input of granule cells in the
epileptic human dentate gyrus. Neuroscience, 108:
587–600.
Woodson, W., Nitecka, L. and Ben-Ari, Y. (1989) Organization
of the GABAergic system in the rat hippocampal formation:
a quantitative immunocytochemical study. J. Comp. Neurol.,
280: 254–271.
Zappone, C.A. and Sloviter, R.S. (2001) Commissurally
projecting inhibitory interneurons of the rat hippocampal
dentate gyrus: a colocalization study of neuronal markers
and the retrograde tracer Fluoro-gold. J. Comp. Neurol.,
441: 324–344.
Zipp, F., Nitsch, R., Soriano, E. and Frotscher, M. (1989)
Entorhinal fibers form synaptic contacts on parvalbumin-
immunoreactive neurons in the rat fascia dentata. Brain Res.,
495: 161–166.
Plate 13.2. Double labeling of GAD65 mRNA and either NeuN (A) or NADPH-diaphorase (B) in the rat dentate gyrus. (A) GAD65
mRNA-labeled neurons (black) are numerous throughout the dentate hilus (H) where they are intermingled with NeuN single-labeled
(non-GABA) neurons (red; examples at arrows). Some GAD65 mRNA-labeled neurons are present in the granule cell layer (G) but are
most abundant at the base of this layer. Most neurons in the molecular layer (M) are labeled for GAD65 mRNA. Granule cells and
CA3 pyramidal cells are single-labeled for NeuN. (B) NADPH-diaphorase (dark blue) is present in only subgroups of the GAD
mRNA-labeled neurons (red) in the hilus (H), granule cell layer (G) and molecular layer (M) (examples at arrows). (For B/W version,
see page 220 in the volume.)
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 14

Functional regulation of the dentate gyrus

by GABA-mediated inhibition

Douglas A. Coulter1,2, and Gregory C. Carlson1

1
The Children’s Hospital of Philadelphia, Abramson Pediatric Research Center, Room 410D, 3516 Civic Center
Boulevard, Philadelphia, PA 19104-4318, USA
2
Departments of Pediatrics, Neurology, and Neuroscience, University of Pennsylvania School of Medicine, Philadelphia,
PA, USA

Abstract: Dentate granule cells are characterized by their low levels of excitability, an important aspect of
hippocampal function, which distinguishes them from other principal cells of the hippocampus. This low
excitability derives in large part from the degree and nature of GABAergic inhibition evident in the dentate
gyrus. Granule cells express a unique and complex assortment of GABAA receptor subunits, found in few
areas of the brain. Associated with this receptor complexity, granule cells are endowed with both synaptic
and extrasynaptic GABAA receptors with distinctive properties. In particular, extrasynaptic GABAA re-
ceptors in granule cells exhibit high affinity for GABA and do not desensitize. This results in activation of a
tonic current by ambient levels of GABA present in the extracellular space. This tonic current contributes
significantly to the circuit properties of the dentate gyrus. Both synaptic and extrasynaptic GABAA re-
ceptors exhibit profound dysregulation in animal models of temporal lobe epilepsy, which may contribute
to the hippocampal hyperexcitability that defines this disorder.

Keywords: hippocampus; dentate gyrus; GABA receptor; inhibition; epilepsy

Introduction 1976; Wilson and McNaughton, 1993). This ‘sparse

As a part of its role in memory formation, the hip-
pocampus encodes a cognitive map of the space in
which an animal (or human) navigates. During en-
vironmental exploration, studies recording place
fields in rat hippocampal neurons have typically
demonstrated that information coding of spatial
position (i.e. degree of firing and firing patterns) is
specific but sparse in dentate granule cells (Jung and
McNaughton, 1993; Chawla et al., 2005) compared
to dorsal hippocampal pyramidal cells (O’Keefe,
Corresponding author. Tel.: +1 215-590-1937;
Fax: +1 215-590-4121; E-mail: coulterd@email.chop.edu
coding’ of dentate granule cells is theorized to be
important in information processing and memory
formation of the hippocampus (McNaughton and
Morris, 1987). It also reflects a general trend evi-
dent in most studies of dentate granule cell excit-
ability: these cells exhibit a fundamental reluctance
to fire, particularly synchronously in network
bursts. This is due, in part, to a combination of
intrinsic factors (including hyperpolarized resting
membrane potential, lack of conductances which
permit phasic firing or ‘‘bursting’’, and marked
spike frequency adaptation), but is principally a
consequence of the powerful feedforward and feed-
back GABAergic inhibition characteristic of dent-
ate gyrus circuit function.

DOI: 10.1016/S0079-6123(07)63014-3 235

236
In addition to its important role in cognitive
processing, the low excitability of the dentate gyrus
may serve to filter or ‘gate’ synchronous excitatory
activity in entorhinal cortex, preventing this type of
activity from hyperactivating and damaging the
relatively fragile hippocampal structures down-
stream, and from triggering seizure activity (Heine-
mann et al., 1992; Lothman et al., 1992). Like
sparse coding, this gating function of the dentate
gyrus is also due to powerful feedforward and feed-
back GABAergic inhibition, as well as intrinsic
properties of granule cells. Furthermore, the filter
function of the dentate gyrus may be compromised
in animals with epilepsy, or in animals in the proc-
ess of developing epilepsy, and this loss of function
may reflect alterations in GABAA receptor expres-
sion and function in dentate granule cells. The in-
tent of the present chapter is to discuss dentate
gyrus circuit excitability in context of the GABA
receptors, which granule cells express, and also to
extend this discussion to how alterations in expres-
sion and function of inhibitory synaptic receptors in
epileptic animals may disrupt normal operation of
the dentate gyrus.
Heterogeneous composition and function of GABAA
receptors in the CNS
GABAA receptors are members of the cysteine-
loop ligand-gated ion channel family, which, like
other members of this family, are pentameric as-
semblies of subunits that surround a central ion
selective pore. The properties of the pore in GABA
receptors make it primarily permeable to chloride
and bicarbonate ions. The net result of activation
of GABA receptors is therefore flow of chloride
into cells down its concentration gradient, con-
comitant hyperpolarization of the membrane po-
tential, and a large conductance increase. Both of
these events (hyperpolarization and activation of a
shunting conductance) usually result in a de-
creased propensity to fire action potentials in a
neuron following GABA receptor activation, i.e.
inhibition. To date, 19 GABAA receptor subunits
have been cloned from the mammalian CNS, en-
compassing 8 families (a1–6, b1–3, g1–3, d, e, y, p,
and r1–3). Additional complexity is conferred on
possible receptor composition by alternative splic-
ing, which is found in several subunit families (re-
viewed in Farrant and Nusser, 2005, and in
references therein). Combinatorial calculations
based on random assembly of pentameric recep-
tors would result in a million or more possible
GABAA receptors. However, this complexity is
simplified by the fact that there appears to be a
preferred receptor stoichiometry (usually two a,
two b, and a g subunit as the prevalent receptor,
with the g subunit replaced by d, e, or p subunit in
rare subsets of receptors). In GABAA receptors
isolated from brain, there may be as few as 10–20
preferred receptor subunit configurations (Wisden
et al. 1992; McKernan and Whiting, 1996; Pirker
et al. 2000). This heterogeneity in subunit compo-
sition of GABAA receptors has significant func-
tional consequences. The subunit composition of a
GABAA receptor dictates the kinetic properties,
cellular and subcellular localization, and pharma-
cology of the receptor when expressed in neurons.
GABAA receptors expressed by dentate granule
cells
In situ hybridization (Wisden et al., 1992),
immunocytochemical (Sperk et al., 1997; Pirker
et al., 2000), and single-cell combined antisense
RNA amplification profiling-functional studies
(Brooks-Kayal et al., 1998) have demonstrated
that granule cells express as many as 10 or more
GABAA receptor subunits, even when only a sin-
gle cell is measured (Brooks-Kayal et al., 1998). Of
19 possible subunits, technological and experimen-
tal issues have limited profiling to 13 subunits or
less in most cases. In these studies, granule cells
express five of six a subunits (with no expression of
a6, and a3 found in some studies, and not in oth-
ers), three of three b subunits, one of three g sub-
units (only g2), as well as d subunits (Wisden et al.,
1992; Sperk et al., 1997; Brooks-Kayal et al., 1998;
Pirker et al., 2000).
Pharmacological studies examining the actions
of subunit selective agonists and antagonists on
synaptic responses in dentate granule cells have
provided some insight into which of the above de-
scribed set of 10 subunits may comprise

subsynaptic GABAA receptors. In normal, control
rodents, zolpidem (a hypnotic that preferentially
augments responses in a1-containing GABAA
receptors markedly enhances inhibitory synaptic
responses in dentate granule cells, while zinc has
little or no antagonist effect, even at very high
concentrations (Buhl et al., 1996; Cohen et al.,
2003). This suggests that, in addition to g2 sub-
units required to anchor GABAA receptors in
synapses (Essrich et al., 1998) and an undefined set
of b subunits, a1 subunits are highly expressed in
synaptic GABAA receptors in dentate granule
cells, since GABAA receptors containing these
subunits respond preferentially to zolpidem, and
are very resistant to blockade by zinc. Other a
subunits, such as a2 and a4 may also make a con-
tribution to synaptic responses. The degree of
contribution of these other subunits has not as yet
been defined experimentally in granule cells. In
addition, this ensemble of synaptic GABAA re-
ceptors changes significantly in animals with epi-
lepsy, which will be discussed in detail below.
Tonic and phasic inhibition in dentate granule cells
In addition to synaptic GABAA receptors, which
demonstrate a significant contribution of a1 sub-
units, dentate granule cells express a relatively
unique set of GABAA receptors comprised of a4bd
subunits. Receptors with this composition are only
found to any significant extent in one other region
of the brain, the thalamus, where they are ex-
pressed in thalamocortical relay neurons (McKe-
rnan and Whiting, 1996). These two brain regions,
the dentate gyrus and the thalamus, share a func-
tional requirement to fire action potentials in dis-
crete, extremely small populations of neurons, and
to resist synchronous firing during information
transfer. These requirements are evident as ‘sparse
coding’ properties discussed above for the dentate
gyrus, and as somatosensory and visual represen-
tations evident in thalamic projections to cortex,
where synchronous firing would disrupt informa-
tion transfer. Given the shared circuit require-
ments in these two brain regions, and the selective
expression of a4bd GABA receptors in only these
regions, this suggests that receptors of this
237
composition may play a critical role in discrete
coding of information, and in opposing synchro-
nous firing.
What can the properties of a4bd receptors tell us
about how they may accomplish this role in de-
fining circuit function in the dentate gyrus? Re-
ceptors lacking the g2 subunit are not clustered in
synapses (Essrich et al., 1998), and so a4bd recep-
tors are primarily localized extra- or perisynaptic-
ally (Wei et al., 2003; Sun et al., 2004). Receptors
of this composition are endowed with a set of
pharmacological properties which make them per-
fectly suited to constitute extrasynaptic receptors:
they have extremely high affinity for GABA (in the
nM compared to the low mM range for synaptic
receptors), and they exhibit little or no desensiti-
zation (Stell and Mody, 2002; Mtchedlishvili and
Kapur, 2006; reviewed in Farrant and Nusser,
2005). Therefore, these receptors can respond to
ambient or synaptic spillover levels of GABA, and
can maintain a tonic current for sustained periods
without desensitizing, both of which define tonic
GABAA receptors (Farrant and Nusser, 2005).
Gene targeting studies in mice support this hy-
pothesis, because it was found that deletion of
either the a4 or d subunit significantly reduced
tonic current as well as accelerated decay of IPSCs
in dentate granule cells, both of which were ac-
companied by alterations in neuronal excitability
(Spiegelman et al., 2003; Wei et al., 2003; Chandra
et al., 2006). Therefore, tonic and spillover-
mediated GABAA currents derived from activity
of a4bd-containing receptors are a significant
and prevalent feature of dentate granule cell ne-
urophysiological characteristics, and are impor-
tant contributors to the function of the dentate
gyrus.
The ‘gatekeeper’ function of the dentate gyrus is
maintained by GABAergic inhibition
Assessing the role of GABAergic inhibition in reg-
ulating function of the dentate gyrus requires si-
multaneous recording in afferents to the dentate
gyrus, the dentate gyrus itself, and efferents of the
dentate gyrus during synaptic activation. This is
necessary to ascertain how the dentate gyrus
238
circuitry may filter and constrain the amplitude
and duration of afferent inputs as it passes infor-
mation on to area CA3 of the hippocampus. In
addition, these recordings need to be conducted
under conditions where the extracellular milieu
can be easily manipulated.
Several studies have been conducted examining
filtering or ‘gatekeeper’ function of the dentate
gyrus under control and pathophysiological con-
ditions (animal models of epilepsy). Behr et al.
(1998), using multiple field potential recordings in
entorhinal cortex, dentate gyrus, and area CA3,
found that, in control animals, epileptiform bursts
originating in entorhinal cortex did not propagate
through the dentate gyrus to activate CA3 with a
similar burst. Rather, the input was filtered, and
only moderate activity was recorded in CA3. This
contrasted with kindled animals, where the ent-
orhinal cortical epileptiform burst activity was
able to propagate through the dentate gyrus and
trigger similar epileptiform bursts in area CA3. No
attempt was made in the Behr et al. (1998) study to
determine what aspect of dentate gyrus circuit
physiology was responsible for the filtering func-
tion in control animals, nor what aspect of circuit
function was perturbed in kindled animals.
We (Carlson et al., 2002a; Ang et al., 2006) have
utilized voltage sensitive dye imaging techniques to
monitor function of the dentate gyrus in control
and epileptic animals, as entorhinal cortical affer-
ents to dentate gyrus are activated. Voltage sen-
sitive dyes, when combined with state-of-the-art
CCD cameras, allow monitoring of multiple sites
within a brain slice (in our case, 6400 sites), at
heretofore unprecedented temporal resolution
(1–5 kHz frame rates). This allows circuit activity
to be monitored with high temporal and spatial
resolution, which can facilitate recording of synap-
tic integration within multiple circuits. This per-
mits monitoring of afferent activation of the
dentate gyrus, processing within the dentate, and
propagation of dentate signals to efferent struc-
tures to be monitored simultaneously, while re-
cording from individual neurons with a patch
electrode (Ang et al., 2006; see Fig. 1).
Data derived from these studies clearly illustrate
both the ‘gatekeeper’ function of the dentate (lack
of hippocampal activation despite robust, syn-
chronous excitation of the dentate gyrus by ent-
orhinal cortical afferents; Fig. 1A–C; see also Ang
et al., 2006). In addition, these studies also dem-
onstrate that this dentate filtering is accomplished
by activation of GABAA receptors. Even modest
disinhibition (by picrotoxin concentrations suffi-
cient to block 20–25% of synaptic responses) re-
sults in collapse of the gate of the dentate gyrus
(Fig. 1D–F), a finding which was also evident in
earlier imaging studies of dentate gyrus function
(Iijima et al., 1996).
In the Carlson et al. (2002a) studies, picrotoxin
was used in modest concentrations to trigger dis-
inhibition. Picrotoxin is a non-competitive GABAA
antagonist, and has equivalent efficacy in blocking
both lower affinity, synaptic and higher affinity,
extrasynaptic GABAA receptors (Stell and Mody,
2002). In contrast, gabazine, a competitive GABAA
antagonist, is much more effective in blocking lower
affinity, synaptic GABAA receptors than tonic, ex-
trasynaptic receptors, since it can compete more
effectively for the GABA binding site when recep-
tors have relatively low affinity for GABA. When
applied in nanomolar concentrations, sufficient to
block 70% of synaptic GABA responses, but only a
few percent of tonic GABAA receptor responses
(Stell and Mody, 2002), gabazine had no effect on
gatekeeper function of the dentate gyrus (Carlson
et al., 2002b; Carlson and Coulter, unpublished
observations). This implicates tonic GABAA recep-
tors as critical mediators of dentate gyrus filter
function.
This concept of a critical role of tonic GABAA
receptors as important regulators of dentate gyrus
function was further supported by a recent study,
which described ovarian cycle-linked reduction in
expression of the d subunit of GABAA receptors in
dentate granule cells, together with a reduction in
tonic current, which were accompanied by an in-
crease in seizure susceptibility (Maguire et al.,
2005). Increased seizure susceptibility could be
mimicked by experimental downregulation of the d
subunit. This demonstration that reduced expres-
sion of a single subunit (d) critical for tonic current
in dentate granule cells altered dentate circuit
function in whole animals further corroborates the
 239

Fig. 1. ‘Gatekeeper’ function of the dentate gyrus is maintained by GABAergic inhibition. Simultaneous voltage sensitive dye (A
snapshot taken at the peak of the response, B trace illustrating the VSD response over time), patch clamp (C), and field potential (C)
recording of dentate gyrus response to perforant path activation in control ACSF. Note robust activation of dentate gyrus molecular
layer (red color in A, corresponding to a 10–15 mV EPSP in B), which does not result in activation of downstream structures (note lack
of response in area CA3 in A and B). This lack of CA3 activation is because dentate granule cells do not fire action potentials in
response to perforant path activation under these conditions. This is evident in both the patch (C, top trace, the neuron depolarized to
Vm of 50 mV) and field potential recording (C, bottom trace), due to powerful feedforward inhibition activated by perforant path
stimulation (C, note large IPSP in patch recording). The importance of inhibition in mediating this ‘gatekeeper’ function is illustrated
in responses in panels D, E, and F, following perfusion with 5 mM picrotoxin, a non-competitive GABAA receptor antagonist. This
concentration blocks 20–25% of inhibition (see inset [located above panel E] depicting an averaged spontaneous IPSC [sIPSC] before
and after perfusion with 5 mM picrotoxin). During 25% GABAergic blockade, perforant path activation resulted in powerful ac-
tivation of both the dentate gyrus and downstream structures (CA3 and hilus; D, E). It also triggered action potential firing in dentate
granule cells (see patch and field potential recordings in F, both of which exhibit action potential firing). A grayscale image of the slice,
with patch and field potential recording electrode location is depicted in the inset above A. From Carlson and Coulter (unpublished).
(See Color Plate 14.1 in color plate section.)
240
important role played by these small amplitude,
sustained currents in regulating function of the
dentate gyrus.
Alterations in GABAA receptor expression in
dentate granule cells, and function of the dentate
gyrus in animal models of epilepsy
Temporal lobe epilepsy is defined by seizures dis-
charges which activate the temporal lobe, includ-
ing the hippocampus. Because the dentate gyrus is
hypothesized to be a critical checkpoint regulating
excitability of the limbic system (Heinemann et al.,
1992; Lothman et al., 1992), it has been a focus of
multiple studies in animal models of epilepsy, as-
sessing whether cellular, synaptic, and circuit
properties are altered in a manner consistent with
seizure susceptibility. A primary focus of this work
has been GABAA receptor expression and func-
tion, as well as inhibitory synaptic function.
Upregulation of synaptic GABAA receptors in
granule cells of epileptic animals
If levels of GABAergic inhibition are critical in
mediating gatekeeper function of the dentate gyrus,
and compromised filter function of the dentate is a
primary contributor to seizure generation in ani-
mals with temporal lobe epilepsy, then one might
expect that overall levels of expression of GABAA
receptors might be reduced in epileptic animals. In
both kindling and post-status epilepticus models of
temporal lobe epilepsy, studies examining GABAA
receptor function in dentate granule cells have de-
scribed a paradoxical upregulation in expression of
GABAA receptors, both synaptically (Otis et al.,
1994; Buhl et al., 1996; Nusser et al., 1998; Cohen
et al., 2003), and in whole cell studies (Gibbs et al.,
1997; Brooks-Kayal et al., 1998; Mtchedlishvili et
al., 2001; Leroy et al., 2004). There is a consistent
finding of a virtual doubling in the amplitude of
quantal inhibitory synaptic responses, accompanied
by a doubling in density of GABAA receptors re-
corded in whole-cell GABA responses, evident
across multiple, divergent models of temporal lobe
epilepsy.
Alterations in pharmacology and subunit expression
of dentate granule cell GABAA receptors in epileptic
animals
This upregulation of GABAA receptors appears to
be inconsistent with the hypothesis that compro-
mised inhibition in the dentate gyrus might con-
tribute to seizure generation in epileptic animals.
However, experimental evidence provides addi-
tional clues as to how alterations in inhibition may
contribute to seizure susceptibility. Not only is
there an upregulation in density of GABAA re-
ceptors (discussed above), but the nature (subunit
composition) of the receptors themselves is altered.
This is evident as large-scale changes in the phar-
macological properties of the receptors, and
changes in the expression patterns of subunits en-
coding GABAA receptors in dentate granule cells.
Synaptic GABAA receptors exhibit the de novo
appearance of sensitivity to zinc blockade (Buhl
et al., 1996; Gibbs et al., 1997; Brooks-Kayal et al.,
1998; Cohen et al., 2003), as well as loss of sensi-
tivity to benzodiazepine site agonists (Gibbs et al.,
1997; Mtchedlishvili et al., 2001; Leroy et al., 2004).
Both of these findings (elevated zinc sensitivity, re-
duced benzodiazepine sensitivity) suggest that dent-
ate granule cell GABAA receptors exhibit altered
subunit composition following the development of
epilepsy. This has been demonstrated in individual
dentate granule cells with this pharmacological
phenotype, with a net downregulation in expression
of a1 subunits, and an upregulation in expression of
a4 subunits (Brooks-Kayal et al., 1998). This a
subunit switch, if manifest as an upregulation in a4
subunits in synapses, could explain both the en-
hanced sensitivity to zinc blockade and diminished
sensitivity to benzodiazepine site agonists, particu-
larly the a1-preferring GABA receptor modulator,
zolpidem.
Possible consequences of epilepsy-associated altered
subunit composition of GABAA receptors: zinc-
induced collapse of augmented inhibition
This altered zinc sensitivity of GABAA receptors
in animals with epilepsy may have functional sig-
nificance. In addition to alterations in inhibitory

synaptic responses, the dentate gyrus of animals
with epilepsy frequently demonstrates aberrant
sprouting of the output axons of granule cells, the
mossy fibers. These axons, perhaps in response to
loss of targets in the hilus and area CA3, sprout
and reinnervate the proximal dendritic tree of
other dentate granule cells, creating a reentrant
excitatory circuit. In addition to releasing gluta-
mate, these mossy fibers release large quantities of
zinc (reviewed in Coulter, 2000). Therefore, in the
epileptic brain, there is an apposition of the de
novo appearance of a zinc delivery system,
sprouted mossy fibers, as well as the de novo ap-
pearance of zinc-sensitive synaptic GABAA recep-
tors. The simultaneous manifestation of these two
alterations in the epileptic dentate gyrus has led to
the hypothesis that, during periods of repetitive
afferent activation of the dentate gyrus, which
triggers zinc release, there may be a zinc-induced
collapse of augmented GABAergic inhibition, fa-
cilitating seizure initiation (Buhl et al., 1996; Gibbs
et al., 1997; Coulter, 2000). Evidence supporting
this hypothesis includes the de novo appearance of
zinc-sensitivity to ‘gatekeeper’ function of the
dentate gyrus evident in voltage-sensitive dye im-
aging studies (Carlson et al., 2002a).
However, all of the above studies demonstrating
compromised circuit function and enhanced zinc
responses of GABAA receptors have examined the
effects of exogenously applied zinc. For the above
hypothesis to be operative, zinc released at exci-
tatory synapses needs to be sufficiently mobile to
diffuse to and block neighboring GABAergic
synapses. In vitro studies examining the effects of
endogenously released zinc have as yet failed to
identify effects of zinc release on inhibitory synap-
tic responses in slices prepared from epileptic an-
imals (Molnar and Nadler, 2001).
Reductions in tonic GABAA current in granule cells
of epileptic animals
A second finding has recently been described in
animal models of temporal lobe epilepsy: reduction
in tonic GABAA receptor current accompanied by
a downregulation in expression of d subunits (Peng
et al., 2004). Given that tonic GABAA receptors
241
appear critical in ‘gatekeeper’ function of the dent-
ate gyrus, this downregulation in d subunit expres-
sion could compromise function of the dentate
gyrus, despite the concomitant upregulation of
synaptic GABAA receptors. However, recent cir-
cuit studies have failed to identify compromised
‘gatekeeper’ function of the dentate gyrus in epi-
leptic animals (Ang et al., 2006), under stimulus
parameter conditions where blockade of tonic
GABAA receptors induced collapse of the dentate
gate in normal animals (Carlson et al., 2002b). This
suggests that possible dentate gate compromise me-
diated by downregulation in expression of d sub-
units (and reduction in tonic GABAA currents) in
epileptic animals may require certain patterns of
afferent activation to become operative. Perhaps
repetitive activation, as during seizure initiation,
will elevate extrasynaptic GABA concentrations
and enhance tonic current in normal animals
(suppressing seizures), and this major check on
excitability will be compromised in animals with
epilepsy.
Conclusions
The dentate gyrus is a structure characterized by
low cellular excitability of its output neurons, dent-
ate granule cells. This low excitability is important
in cognitive function of the hippocampus, and re-
sults predominantly from the high degree of inhib-
itory synaptic regulation, as well as the unique
GABAA receptor properties of these cells, including
expression of tonic GABAA receptors rarely ex-
pressed in other brain regions. Altered expression
and function of GABAA receptors in dentate gran-
ule cells is a prominent feature of temporal lobe
epilepsy, a disorder characterized by hypersynchro-
nous discharges involving the hippocampus. This
association of aberrant GABAA receptor properties
in granule cells from hyperexcitable hippocampi
further demonstrates the critical role of inhibition
in controlling dentate gyrus function.
References
Ang, C.W., Carlson, G.C. and Coulter, D.A. (2006) Massive
and specific dysregulation of direct cortical input to the
242
hippocampus in temporal lobe epilepsy. J. Neurosci., 26:
11850–11856.
Behr, J., Lyson, K.J. and Mody, I. (1998) Enhanced propaga-
tion of epileptiform activity through the kindled dentate
gyrus. J. Neurophysiol., 79: 1726–1732.
Brooks-Kayal, A.R., Shumate, M.D., Jin, H., Lin, D.D., Ri-
khter, T.Y. and Coulter, D.A. (1998) Selective changes in
single cell GABAA receptor subunit expression and function
in temporal lobe epilepsy. Nat. Med., 4: 1166–1172.
Buhl, E.H., Otis, T.S. and Mody, I. (1996) Zinc-induced col-
lapse of augmented inhibition by GABA in a temporal lobe
epilepsy model. Science, 271: 369–373.
Carlson, G., Lee, C.J. and Coulter, D.A. (2002a) Zinc facilitates
hyperexcitability in the hippocampus of epileptic rats. Epile-
psia, 43(Suppl 7): 2.031.
Carlson, G.C., Lee, C.J. and Coulter, D.A. (2002b) The role of
inhibition and dentate gatekeeper function: voltage-sensitive
dye imaging study. Soc. Neurosci. Abstr., 28: 145.9.
Chandra, D., Jia, F., Peng, Z., Suryanarayanan, A., Werner,
D.F., Spigelman, I., Houser, C.R., Olsen, R.W., Harrison,
N.L. and Homanics, G.E. (2006) GABAA receptor a4 sub-
units mediate extrasynaptic inhibition in thalamus and dent-
ate gyrus and the action of gaboxadol. PNAS, 103:
15230–15235.
Chawla, K.M., Guzowski, J.F., Ramirez-Amaya, V., Lipa, P.,
Hoffman, K.L., Marriott, L.K., Worley, P.F., McNaughton,
B.L. and Barnes, C.A. (2005) Sparse, environmentally selec-
tive expression of Arc RNA in the upper blade of the rodent
fascia dentate by brief spatial experience. Hippocampus, 15:
579–586.
Cohen, A.S., Lin, D.D., Quirk, G.L. and Coulter, D.A. (2003)
Dentate granule cell GABAA receptors in epileptic hippo-
campus: enhanced synaptic efficacy and altered pharmacol-
ogy. Eur. J. Neurosci., 17: 1607–1616.
Coulter, D.A. (2000) Mossy fiber zinc and temporal lobe ep-
ilepsy: pathological association with altered epileptic
GABAA receptors in dentate granule cells. Epilepsia, 41(Sup-
pl 6): S96–S99.
Essrich, C., Lorez, M., Benson, J.A., Fritschy, J.M. and Lusc-
her, B. (1998) Postsynaptic clustering of major GABAA re-
ceptor subtypes requires the gamma 2 subunit and gephyrin.
Nat. Neurosci., 1: 563–571.
Farrant, M. and Nusser, Z. (2005) Variations on an inhibitory
theme: phasic and tonic activation of GABAA receptors. Nat.
Rev. Neurosci., 6: 215–229.
Gibbs III, J.W., Shumate, M.D. and Coulter, D.A. (1997)
Differential epilepsy-associated alterations in postsynaptic
GABAA receptor function in dentate granule and CA1 neu-
rons. J. Neurophysiol., 77: 1924–1938.
Heinemann, U., Beck, H., Dreier, J.P., Ficker, E., Stabel, J. and
Zhang, C.L. (1992) The dentate gyrus as a regulated gate for
the propagation of epileptiform activity. Epilepsy Res. Sup-
pl., 7: 273–280.
Iijima, T., Witter, M.P., Ichikawa, M., Tominage, T., Kajiwara,
R. and Matsumoto, G. (1996) Entorhinal-hippocampal in-
teractions revealed by real-time imaging. Science, 272:
1176–1179.
Jung, M.W. and McNaughton, B.L. (1993) Spatial selectivity of
unit activity in the hippocampal granular cell layer. Hippo-
campus, 3: 165–182.
O’Keefe, J. (1976) Place units in the hippocampus of the freely
moving rat. Exp. Neurol., 51: 78–109.
Leroy, C., Poisbeau, P., Keller, A.F. and Nehlig, A. (2004)
Pharmacological plasticity of GABAA receptors at dentate
gyrus synapses in a rat model of temporal lobe epilepsy.
J. Physiol., 557: 473–487.
Lothman, E.W., Stringer, J.L. and Bertram, E.H. (1992) The
dentate gyrus as a control point for seizures in the hippo-
campus and beyond. Epilepsy Res. Suppl., 7: 301–313.
Maguire, J.L., Stell, B.M., Rafizadeh, M. and Mody, I. (2005)
Ovarian cycle-linked changes in GABA(A) receptors medi-
ating tonic inhibition alter seizure susceptibility and anxiety.
Nat. Neurosci., 8(6): 797–804.
McKernan, R.M. and Whiting, P.J. (1996) Which GABAA-
receptor subtypes really occur in the brain. Trends Neurosci.,
19: 139–143.
McNaughton, B.L. and Morris, R.G.M. (1987) Hippocampal
synaptic enhancement and information storage within a dis-
tributed memory system. Trends Neurosci., 10: 408–415.
Molnar, P. and Nadler, J.V. (2001) Lack of effect of mossy
fiber-released zinc on granule cell GABAA receptors in the
pilocarpine model of epilepsy. J. Neurophysiol., 85:
1932–1940.
Mtchedlishvili, Z., Bertram, E.H. and Kapur, J. (2001) Dimin-
ished allopregnanolone enhancement of GABAA receptor
currents in a rat model of chronic temporal lobe epilepsy.
J. Physiol., 537: 453–465.
Mtchedlishvili, Z. and Kapur, J. (2006) High-affinity, slowly
desensitizing GABAA receptors mediate tonic inhibition in
hippocampal dentate granule cells. Mol. Pharmacol., 69:
564–575.
Nusser, Z., Hajos, N., Somogyi, P. and Mody, I. (1998) In-
creased number of synaptic GABAA receptors underlies po-
tentiation at hippocampal inhibitory synapses. Nature, 395:
172–177.
Otis, T.S., De Koninck, Y. and Mody, I. (1994) Lasting po-
tentiation of inhibition is associated with an increased
number of gamma-aminobutyric acid type A receptors acti-
vated during miniature inhibitory postsynaptic currents.
PNAS, 91: 7698–7702.
Peng, Z., Huang, C.S., Stell, B.S., Mody, I. and Houser, C.R.
(2004) Altered expression of the d subunit of the GABAA
receptor in a mouse model of temporal lobe epilepsy. J. Ne-
urosci., 24: 8629–8639.
Pirker, S., Schwarzer, C., Wieselthaler, A., Sieghart, W. and
Sperk, G. (2000) GABAA receptors: immunocytochemical
distribution of 13 subunits in the adult rat brain. Neurosci-
ence, 101: 815–850.
Sperk, G., Schwarzer, C., Tsunaskima, K., Fuchs, K. and Si-
eghart, W. (1997) GABA(A) receptor subunits in the rat
hippocampus. I: immunocytochemical distribution of 13 sub-
units. Neuroscience, 80: 987–1000.
Spiegelman, I., Li, Z., Liang, J., Cagetti, E., Sanzadeh, S., Mi-
halek, R.M., Homanics, G.E. and Olsen, R.W. (2003)

Reduced inhibition and sensitivity to neurosteroids in hip-
pocampus of mice lacking the GABA(A) receptor delta sub-
unit. J. Neurophysiol., 90: 903–910.
Stell, B.M. and Mody, I. (2002) Receptors with different
affinities mediate phasic and tonic GABAA conduc-
tances in hippocampal neurons. J. Neurosci., 22(1–5):
RC223.
Sun, C., Sieghart, W. and Kapur, J. (2004) Distribution of al-
pha1, alpha4, gamma2, and delta subunits of GABAA re-
ceptors in hippocampal dentate granule cells. Brain Res.,
1029: 207–216.
243
Wei, W., Zhang, N., Peng, Z., Houser, C.R. and Mody, I.
(2003) Perisynaptic localization of d-containing GABAA re-
ceptors and their activation by GABA spillover in the mouse
dentate gyrus. J. Neurosci., 23: 10650–10661.
Wilson, M.A. and McNaughton, B.L. (1993) Dynamics of the
hippocampal ensemble code for space. Science, 261(5124):
1055–1058.
Wisden, W., Laurie, D.J., Monyer, H. and Seeburg, P.H. (1992)
The distribution of 13 GABAA receptor subunit mRNAs in
the rat brain. I. Telencephalon, diencephalon, mesencepha-
lon. J. Neurosci., 12: 1040–1062.
Plate 14.1. ‘Gatekeeper’ function of the dentate gyrus is maintained by GABAergic inhibition. Simultaneous voltage sensitive dye (A
snapshot taken at the peak of the response, B trace illustrating the VSD response over time), patch clamp (C), and field potential (C)
recording of dentate gyrus response to perforant path activation in control ACSF. Note robust activation of dentate gyrus molecular
layer (red color in A, corresponding to a 10–15 mV EPSP in B), which does not result in activation of downstream structures (note lack
of response in area CA3 in A and B). This lack of CA3 activation is because dentate granule cells do not fire action potentials in
response to perforant path activation under these conditions. This is evident in both the patch (C, top trace, the neuron depolarized to
Vm of 50 mV) and field potential recording (C, bottom trace), due to powerful feedforward inhibition activated by perforant path
stimulation (C, note large IPSP in patch recording). The importance of inhibition in mediating this ‘gatekeeper’ function is illustrated
in responses in panels D, E, and F, following perfusion with 5 mM picrotoxin, a non-competitive GABAA receptor antagonist. This
concentration blocks 20–25% of inhibition (see inset [located above panel E] depicting an averaged spontaneous IPSC [sIPSC] before
and after perfusion with 5 mM picrotoxin). During 25% GABAergic blockade, perforant path activation resulted in powerful ac-
tivation of both the dentate gyrus and downstream structures (CA3 and hilus; D, E). It also triggered action potential firing in dentate
granule cells (see patch and field potential recordings in F, both of which exhibit action potential firing). A grayscale image of the slice,
with patch and field potential recording electrode location is depicted in the inset above A. From Carlson and Coulter (unpublished).
(For B/W version, see page 239 in the volume.)
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 15

Opioid systems in the dentate gyrus

Carrie T. Drake1, Charles Chavkin2 and Teresa A. Milner1,

1
Division of Neurobiology, Department of Neurology and Neuroscience, Weill-Cornell Medical College,
411 East 69th Street, New York, NY 10021, USA
2
Department of Pharmacology, University of Washington, Seattle, WA 98195, USA

Abstract: Opiate drugs alter cognitive performance and influence hippocampal excitability, including long-
term potentiation (LTP) and seizure activity. The dentate gyrus (DG) contains two major opioid peptides,
enkephalins and dynorphins, which have opposing effects on excitability. Enkephalins preferentially bind
to delta- and mu-opioid receptors (DORs and MORs) while dynorphins preferentially bind to kappa-
opioid receptors (KORs). Opioid receptors can also be activated by exogenous opiate drugs such as the
MOR agonist morphine. Enkephalins are contained in the mossy fiber pathway, in the lateral perforant
path (PP) and in scattered GABAergic interneurons. MORs and DORs are predominantly in distinct
subpopulations of GABAergic interneurons known to inhibit granule cells, and are present at low levels
within granule cells. MOR and DOR agonists increase excitability and facilitate LTP in the molecular layer.
Anatomical and physiological evidence is consistent with somatodendritic and axon terminal targeting of
both MORs and DORs. Dynorphins are in the granule cells, most abundantly in mossy fibers but also in
dendrites. KORs have been localized to granule cell mossy fibers, supramammillary afferents to granule
cells, and PP terminals. KOR agonists, including endogenous dynorphins, diminish the induction of LTP.
Recent evidence indicates that opiates and opioids also modulate other processes in the hippocampal
formation, including adult neurogenesis, the actions of gonadal hormones, and development of neonatal
transmitter systems.

Keywords: enkephalin; dynorphin; opioid receptors; GABA interneuron; disinhibition; neurogenesis;

seizure; long-term potentiation

Introduction Endogenous opioid peptides also modulate hippo-

Application of opiate drugs alters cognitive per-
formance and influences hippocampal excitability,
including long-term potentiation (LTP) (Guerra
et al., 1987; Bramham, 1992; Wagner and Chavkin,
1992; Pu et al., 2002; Robbins and Everitt,
2002) and seizure activity (Hong et al., 1993).
Corresponding author. Tel.: +1 (212) 570 2900;
Fax: +1 (212) 988 3672; E-mail: tmilner@med.cornell.edu
campal excitability (reviewed by Simmons and
Chavkin, 1996) and the endogenous hippocampal
opioid systems are implicated in learning (Sandin
et al., 1998), including that associated with drug
use (Nestler, 2001). Recent evidence indicates that
opiates and opioids also modulate other processes
in the hippocampal formation, including adult
neurogenesis, the actions of gonadal hormones,
and development of neonatal transmitter systems
(Eisch et al., 2000; Slamberová et al., 2003;
Schindler et al., 2004). The dentate gyrus (DG)

DOI: 10.1016/S0079-6123(07)63015-5 245

246
contains several types of opioid peptides, which
have varying degrees of receptor selectivity, and
opioid receptors, which can be activated by both
endogenous opioid peptides and exogenous opiate
drugs. Most studies of opioid systems have been
conducted on rat, and although other species stud-
ied are generally similar, there are some notable
species differences.
Opioid peptides
The enkephalins and dynorphins are the most
abundant opioids in the DG. They are derived
from two genes with distinct but overlapping cel-
lular distributions. Proenkephalin-derived opioids
include [Met5]-enkephalin, [Leu5]-enkephalin, and
other C-terminally extended forms of [Met5]-
enkephalin (Gubler et al., 1982; Noda et al.,
1982). Prodynorphin-derived opioids include
[Leu5]-enkephalin, dynorphin A(1–17), dynorphin
A(1–8), dynorphin B(1–13), leumorphin [a.k.a.
dynorphin B(1–29)], alpha-neo-endorphin, and
beta-neo-endorphin (Kakidani et al., 1982). The
first five amino acids of dynorphin peptides
are identical to the (entire) sequence of [Leu5]-
enkephalin. At times this has complicated the
localization and study of the individual opioids,
however, fairly selective agents have been
developed.
The recently discovered opioid endomorphin
is extremely rare in the rat DG (Pierce and
Wessendorf, 2000). Another opioid, beta-endorphin
(a product of the pro-opiomelanocortin precur-
sor), was reported in the hippocampal formation
(Zakarian and Smyth, 1979, 1982) based on radio-
immunoassay and immunofluorescence; however
later chromatographic work showed that the
immunoreactivity did not correspond to authentic
beta-endorphin (Chavkin et al., 1985b). Finally,
nociceptin/orphanin FQ is a hectadecapeptide sim-
ilar in structure to dynorphin A (Meunier et al.,
1995; Reinscheid et al., 1995). Although it shares
structural similarity to opioids, its actions and
pharmacology are distinct in that it acts through
the naloxone-insensitive ORL1 (a.k.a NOP) re-
ceptor (Heinricher, 2003).
Opioid receptors
Three major types of opioid receptors have been
characterized; MORs, DORs, and KORs (reviewed
by Martin, 1983). These receptor types have vary-
ing preferences for endogenous opioid peptides and
exogenous ligands. MORs have a high affinity for
endomorphin and the opiate alkaloid morphine
(Martin, 1983; Zadina et al., 1997). MORs and
DORs preferentially interact with enkephalins over
dynorphins, and DORs have a 10-fold higher affin-
ity than MORs for enkephalins (Corbett et al.,
1993). KORs have a high affinity for dynorphins
but not enkephalins (Chavkin et al., 1982, 1985b;
Corbett et al., 1982) and also bind tightly to the
exogenous agent ethylketocyclazocine (Corbett
et al., 1993). As described below, there is also ev-
idence that dynorphins, although primarily KOR
agonists, have some affinity for MORS and can
modulate N-methyl-D-aspartic acid (NMDA) re-
ceptors (Chavkin et al., 1985a; Caudle et al., 1994;
Caudle and Dubner, 1998).
The complexity of opioid receptors goes far
beyond three straightforward types. Additional
opioid receptor types and subtypes have been pro-
posed based on receptor-binding studies, receptor
autoradiography, and behavioral studies. These
include mu1 and mu2 (Wolozin and Pasternak,
1981), mu/delta (Rothman et al., 1983, 1987),
delta1 and delta2 (Jiang et al., 1991; Mattia et al.,
1991), and kappa1, kappa2, and kappa3 (Nock
et al., 1988; Zukin et al., 1988; Clark et al., 1989).
MOR, DOR, and KOR subtypes may arise from
several sources. The MOR, DOR, and KOR have
been cloned (Evans et al., 1992; Kieffer et al., 1992;
Chen et al., 1993; Minami et al., 1993; Yasuda
et al., 1993), and although to date only one gene
has been found for each receptor type (Slowe et al.,
1999), other genes may exist. Another potential
source of subtypes is alternatively spliced opioid
receptor mRNA (Zimprich et al., 1995; Pan et al.,
1999). In both rat and mouse DG, splice variants
derived from the cloned MOR1-gene are found,
and have slightly different distributions, implying
region-specific mRNA processing (Abbadie et al.,
2000). Yet another source of opioid receptor
diversity is functional association with other pro-
teins. Like other G protein-coupled receptors,

opioid receptors form stable associations with one
another (termed ‘‘dimers’’) and with accessory
proteins. These interactions may serve to increase
pharmacological complexity, to aid in transport or
maturation of the receptors, and to regulate signal
transduction. Both homodimers (e.g., DOR–DOR)
and heterodimers (e.g., KOR–DOR) have been
demonstrated (see Bouvier, 2001; Rios et al., 2001).
A novel opioid-like receptor, termed ORL1, was
identified based on homology to opioid receptors,
particularly KORs. This receptor has a high affin-
ity for OFQ/nociceptin, but not opioids (Bunzow
et al., 1994). The ORL1 and its endogenous ligand
OFQ/nociceptin are present in the DG and modify
many of the same functions as do opioids; the
history and details of this system are reviewed in
detail elsewhere (Mogil and Pasternak, 2001).
As we describe in the following review of the
anatomy and physiology of opioid systems in the
DG, the different opioid systems share at least two
features. One, in the hippocampal opioid system
there is frequently a mismatch between opioid re-
ceptors and endogenous opioid peptides. Mis-
matches at the regional level (e.g., tens or hundreds
of micrometers) raise the question of how much of
a given pool of opioid peptides reaches opioid
receptors (see McLean et al., 1987) and predict dis-
tinct effects of endogenously released opioids com-
pared to exogenous opiates. Mismatches at the
ultrastructural level indicate that opioids often are
not active at synapses, but instead function through
nonsynaptic diffusion to nearby cellular targets
(e.g., Drake et al., 1994, 2002). Two, although dif-
ferent opioids have differing effects on excitability,
they all share the property of being modulators of
fast neurotransmission. Their modulatory influ-
ences have regionally selective effects on plasticity
and sometimes exert complex or subtle influences
on dentate circuits. Furthermore, physiologically
significant functional effects may occur in regions
with relatively few receptors (Weisskopf et al., 1993).
Anatomical distribution of opioids and
opioid receptors
The distribution of opioid peptides and their
receptors is summarized schematically in Figs. 1
247
and 2 and discussed in detail below. The present
description of opioid peptides is restricted to
enkephalins and dynorphins, since the endogenous
MOR ligand endomorphin is present only in
rare scattered fibers in the rat DG (Pierce and
Wessendorf, 2000).
Enkephalins
Both [met5]-enkephalin and [leu5]-enkephalin are
present in the DG (Hong et al., 1980; Gall et al.,
1981). Enkephalins are present in afferents to the
DG and in DG neurons in a wide range of species
including mice, rats, guinea pigs, hamsters, and
humans (Gall, 1988b; Herkenham and McLean,
1988; Rees et al., 1994). Enkephalin is present in
a subset of mossy fibers (Gall et al., 1981; Gall,
1988a), and is expressed by
15% of granule cells
(Johnston and Morris, 1994). Leu-enkephalin
immunoreactivity is present in both types of
terminals formed by mossy fibers, that is, small
terminals in the hilus and the large complex mossy
terminals (Commons and Milner, 1995).
Enkephalins also are found in a portion of the
lateral perforant path (PP) (Gall et al., 1981;
Fredens et al., 1984; Gall, 1988a), particularly the
temporal portion of this pathway. At the subcel-
lular level, leu-enkephalin immunoreactivity in the
rat is primarily associated with distinctive organ-
elles known as large dense-core vesicles (described
in more detail in Opioid Peptide Release) in both
the small hilar and large mossy terminals. Addi-
tionally, a few interneuron-like cells in the hilus are
leu-enkephalin-immunoreactive (-ir) (Gall et al.,
1981; Commons and Milner, 1995). Analysis of
proenkephalin mRNA expression indicates addi-
tional proenkephalin-expressing interneurons in
the granule cell layer (GCL) and molecular layer
(Stumm et al., 2004).
There are some relevant species differences in
dentate enkephalins. Compared to the rat, the
mouse appears to have sparse enkephalin
immunoreactivity in the hilus (Khalid et al., per-
sonal observation). In the guinea pig there is a
prominent bulge of enkephalin immunoreactivity
in stratum lucidum (SLu) at the CA3/CA2 border.
Enkephalin-ir fibers can been seen to course along
248

SO CA1
p
SP
SR
SUB

SLM
TAT
PP

p HIL m
SLu g
CA3
DG

GCL
ML

ENK

DYN

Fig. 1. Location of enkephalins and dynorphins in the hippocampal formation. Opioid peptides are found in all regions of the
hippocampal formation: the hippocampus proper (fields CA1 and CA3), the dentate gyrus (DG), and the subiculum (SUB). The
glutamatergic neurons [pyramidal cells (p), granule cells (g), mossy cells (m)] and their projections are shown. GABAergic interneurons
are scattered in all laminae. The perforant path (PP) from entorhinal cortex provides the major excitatory innervation. Subcortical
afferents and contralateral hippocampal projections enter via the fimbria/fornix (not shown) adjacent to CA3. Enkephalins (green) are
in some granule cells, particularly their mossy fiber terminals in hilus and stratum lucidum (SLu). Mossy fibers extensively innervate
hilar mossy cells and CA3 pyramidal cells. Enkephalins are in a portion of the projection from entorhinal cortex; the lateral portion of
the perforant path (PP) in the DG and the temporal-ammonic tract (TAT) in CA1. Enkephalins are in a few scattered interneurons,
which are relatively abundant on the border of stratum radiatum (SR) and stratum lacunosum-moleculare (SLM). Dynorphins (yellow)
are in the granule cells. Dynorphin peptides are most striking in the mossy fiber terminals, and are also in the mossy fiber axons,
granule cell somata, and granule cell dendrites. Abbreviations: SO, stratum oriens; SP, stratum pyramidale; GCL, granule cell layer.
(See Color Plate 15.1 in color plate section.)

the longitudinal axis of the hippocampal forma-
tion at this point and are almost certainly in the
thick end-bulb formed by the septotemporal pro-
jections of mossy fibers (Tielen et al., 1982;
McLean et al., 1987). In the rat DG, the
enkephalin-containing PP fibers are restricted to
the outer molecular layer, that is, the terminal
zone of the lateral PP (Gall et al., 1981). In con-
trast, two bands of enkephalin immunoreactivity
are seen in molecular layer of guinea pig, one in
the lateral PP terminal zone and the other in the
medial PP terminal zone (Tielen et al., 1982). Ad-
ditionally, while the rat contains dense enkephalin
immunoreactivity in the PP terminal zone, in the
guinea pig this immunoreactivity is considerably
lighter (McLean et al., 1987). Consistent with
studies described below, this observation indicates
that prodynorphin-derived peptides play a larger
role in the DG of the guinea pig than the rat.
Anatomical data are consistent with the sugges-
tion that enkephalins nonsynaptically modulate
inhibitory transmission. In the dentate molecular
layer, leu-enkephalin-ir terminals occasionally
appose GABA-ir terminals, and in the hilus leu-
enkephalin-ir terminals often contact gamma
amino butyric acid (GABA)-containing perikarya
and dendrites (Commons and Milner, 1996b).
These data suggest that leu-enkephalin may have
an important role in regulating inhibition of
GABA-containing neurons.
One interesting aspect of enkephalin biology is
that the production may be regulated by highly
 249

+
Molecular ENK
Layer PP
_ GABA

?
?

+
SubP ACh

Granule Granule ?
Cell Cell
Layer

GABA
Hilus PARV

+
ENK/DYN
+

ACh GABA
NPY
SOM

MOR

to CA3 DOR

KOR
+

Fig. 2. Subcellular locations of opioid peptides and receptors in the dentate gyrus. Enkephalins (green shading) are in: (1) som-
atodendritic, axon, and terminal portions of granule cells; (2) lateral perforant path axons and terminals in the outer molecular layer;
and (3) occasional GABAergic interneurons (not shown). Mu-opioid receptors (MORs; blue) are most frequently in all portions of
parvalbumin (PARV)-containing basket cells that innervate granule cell somata and proximal dendrites. MORs are also in cholinergic
and GABAergic afferents from the lateral septum/diagonal band. Delta-opioid receptors (DORs; pink) are in many hilar interneurons
that contain somatostatin or neuropeptide Y (SOM/NPY) and inhibit the distal dendrites of granule cells. PARV-containing neurons
express DOR mRNA. DORs, and to a lesser extent MORs, are occasionally present in granule cell dendrites. Kappa-opioid receptors
(KORs) in guinea pigs are in substance P containing afferents to granule cells, in granule cell mossy fibers, and most likely on perforant
path terminals in the outer molecular layer. (See Color Plate 15.2 in color plate section.)

region-specific mechanisms. For example, the Dynorphin

biosynthesis and posttranslational proteolytic
processing of proenkephalin in the mossy fiber
system is distinct from other brain regions, sug-
gesting that particular functions of the enkephalins
may be elicited depending on which met-enkepha-
lin-containing peptides are produced in a given
region or subregion (White et al., 1986).
Dynorphin has a similar distribution in rats, mice,
guinea pigs, squirrels, hamsters, and humans
(Chavkin et al., 1985b; McLean et al., 1987; Gall,
1988a; Houser et al., 1990). The vast majority of
dynorphin in the hippocampal formation is made
by the dentate granule cells, and dynorphin
250
immunoreactivity is abundant in the mossy fiber
pathway in the hilus and CA3 SLu (McGinty
et al., 1983; Khachaturian et al., 1993). In the hi-
lus, both the large mossy terminals and the smaller
mossy fiber collateral terminals contain dynorphin
immunoreactivity (Pierce et al., 1999), suggesting
nonselective release near all types of synaptic tar-
gets of these terminals, that is, CA3 pyramidal
cells, mossy cells, and interneurons. Released
dynorphin is more stable than enkephalin
(Wagner et al., 1991) and diffusion may play a
large role in the actions of dynorphin.
Some dynorphin may be contributed by affer-
ents. For example, in situ hybridization histo-
chemistry in rat identified preprodynorphin
mRNA-containing perikarya in hippocampally
projecting regions of the septum (Merchenthaler
et al., 1997). Also, at the extreme ventral pole of
the guinea pig hippocampal formation, a diffuse
band of varicose dynorphin-ir processes is found
in the molecular layer (Drake et al., 1994).
The granule cell dendrites may also be a relevant
source of dynorphin. As discussed in more detail
below, dynorphin, like enkephalins and other pep-
tides, is stored in dense-core vesicles that are read-
ily visible by electron microscopy (Drake et al.,
1994; Pierce et al., 1999). In the molecular layer of
the guinea pig, midway along the septotemporal
axis, electron microscopic analysis showed that al-
though some dynorphin-immunoreactivity is pre-
sent in unmyelinated axons, the majority (74%) of
dynorphin-labeled dense-core vesicles are found in
granule cell dendrites (Drake et al., 1994). Den-
dritic dynorphin has also been observed in samples
from epileptic human DG (Zhang and Houser,
1999). The functional relevance of dynorphin re-
lease in guinea pig outer molecular layer was sug-
gested by physiological data showing that the time
course of released dynorphin actions was more
consistent with local release than with release from
hilar mossy fibers (Drake et al., 1994). Release
of dendritic dynorphin in the molecular layer
involves calcium channel types and mechanisms
distinct from axonal dynorphin release (Simmons
et al., 1995).
There is an interesting compartmentalization of
dynorphin within guinea pig granule cell dendrites.
The dendritic spines contain more dynorphin
immunoreactivity than do dendritic shafts, and
dynorphin-containing dense-core vesicles in spines
are usually near the extrasynaptic plasma mem-
brane and seldom near synapses (Drake et al.,
1994). In contrast, dynorphin-containing dense-
core vesicles in dendritic shafts are often in the
center of the profile, appearing to be in transit.
This localization may be relevant to synaptic plas-
ticity. In the outer two-thirds of the molecular
layer, spines receive virtually all of the excitatory
input to granule cells. Individual spines in some
ways act as discrete functional units, with different
calcium dynamics and chemical events than the
adjacent dendritic shaft (reviewed by Yuste and
Tank, 1996). Such compartmentalization is
thought to be critical for synapse-specific plastic-
ity (e.g., LTP) (Yuste and Tank, 1996). Only 26%
of dynorphin-labeled dense-core vesicles in the
guinea pig molecular layer are in axons or axon
terminals; of these, most are in small terminals
that may have originated in entorhinal cortex or
are rare collaterals of granule cells. Consistent with
the latter possibility, the inner molecular layer
(which is almost completely devoid of entorhinal
afferents) contains the largest proportion of ax-
onal dense-core vesicles (Drake et al., 1994). In the
inner molecular layer, dynorphin-containing ter-
minals form asymmetric synapses on granule cell
dendrites. These may be mossy fiber collaterals,
consistent with physiological evidence for direct
granule cell-to-granule cell connections (Terman
et al., 2000).
Interestingly, at the extreme ventral pole of the
guinea pig hippocampal formation, the number of
axonal dynorphin-labeled dense-core vesicles was
higher than at other levels, with no change in the
number of dendritic dense-core vesicles (Drake
et al., 1994), suggesting that some innervation
differences are present. The most likely possibility
seems to be the mossy fiber collaterals, which as
described above are more common at the ventral
pole. This dorsal–ventral difference in dynorphin
distribution has not been further studied, but is
intriguing in light of other functional and struc-
tural differences in dynorphin systems of the
dorsal compared to the ventral hippocampal
formation (McDaniel et al., 1990; Pierce et al.,
1999).

Opioid peptide release
Like many other neuropeptides, dynorphin and
enkephalins are stored in large dense-core vesicles.
These are larger (80–120 nm diameter) than small
synaptic vesicles, the storage vesicles for fast trans-
mitters such as GABA or glutamate. Dense-core
vesicles are synthesized, transported, and released
differently from small synaptic vesicles (reviewed
in De Camilli and Jahn, 1990). In particular,
dense-core vesicles require a more prolonged, in-
tense stimulation of the neuron to fuse with the
membrane and release their contents (Seward
et al., 1995). This is linked to differential calcium
coupling: dense-core vesicle release is facilitated by
a small, widespread elevation of calcium while small
synaptic vesicle release requires robust, localized
calcium elevation (Verhage et al., 1991). Consistent
with this, endogenous opioid release in the hippo-
campal formation requires specific forms of high-
intensity stimulation (Wagner et al., 1990, 1991,
1993). Dense-core vesicles often fuse with, and re-
lease their contents from, portions of the membrane
that are distant from classical synaptic specializa-
tions (Thureson-Klein and Klein, 1990). Such non-
synaptic or extrasynaptic sites of release can be the
preferential location of dense-core vesicle fusion,
as in Aplysia californica (Karhunen et al., 2001).
In rat and guinea pig mossy fiber terminals,
dense-core vesicles containing enkephalin- or
dynorphin-immunoreactivities are distributed
along both extrasynaptic and synaptic portions
of the membrane (Drake et al., 1994; Commons
and Milner, 1995; Pierce et al., 1999). The majority
of these dense-core vesicles are extrasynaptic, and
changes in the distribution of membrane-associ-
ated dense-core vesicles following seizure suggest
that dynorphin-containing dense-core vesicles are
specifically targeted to particular portions of the
terminals (Pierce et al., 1999).
Opioid receptors
Delta-opioid receptors (DORs)
Functional DOR binding, as indicated by receptor
autoradiography, is diffusely distributed throughout
251
the DG of mice and rats (Goodman et al., 1980;
Crain et al., 1986; Gulya et al., 1986; Mansour et al.,
1987; McLean et al., 1987; Tempel and Zukin, 1987;
Mansour et al., 1988). Immunocytochemical labe-
ling of DORs in the rat (Mansour et al., 1993) and
mouse (Bausch et al., 1995a) revealed that DOR-ir
neurons and processes are scattered in the DG, in a
distribution that overlaps with, but is not restricted
to, regions known to contain enkephalin. Interest-
ingly, neurons with intense DOR immunoreactivity
(Commons and Milner, 1996a) or high levels of
DOR mRNA (Stumm et al., 2004) are found in the
central hilus, while neurons with light DOR-labeling
are in the GCL. Close examination with electron
microscopy (Commons and Milner, 1996a) further
identified DOR immunoreactivity in dendritic spines
of granule cells, as well as in perikarya and dendrites
of granule cells and nongranule cells. In DOR-ir
dendrites, DOR labeling is most often affiliated with
the plasmalemmal surface near excitatory-type
synapses, consistent with a role in modulating ex-
citatory neurotransmission. Additionally, presynap-
tic modulation by DORs is suggested by DOR
immunoreactivity in axon terminals of both mouse
(Bausch et al., 1995a) and rat (Commons and
Milner, 1996a). DOR-labeled terminals exhibit a
range of excitatory and inhibitory-type morph-
ologies and are in all dentate lamina, consistent
with a variety of neuronal sources (Commons and
Milner, 1996a). However, DOR immunoreactivity is
found on the plasmalemmal surface in axon termi-
nals less frequently than in dendrites.
Most DOR-labeled perikarya contain GABA
(Bausch et al., 1995a; Commons and Milner,
1996a; Stumm et al., 2004). DOR has been identi-
fied within several interneuron subtypes, although
variations in prevalence have been reported de-
pending on the technique used. For example, with
dual-labeling immunocytochemistry 75% of DOR-
ir neurons sampled in the hilus contained neuro-
peptide Y (NPY) immunoreactivity and 39% con-
tained somatostatin (SOM) immunoreactivity
(Commons and Milner, 1996a). Conversely, 75%
of NPY neurons and 93% of SOM neurons con-
tained DOR immunoreactivity (Commons and
Milner, 1996a). This suggests that almost all of
the somatostatinergic cells in the hilus, which are a
subpopulation of NPY neurons, contain DOR. In
252
this study, few DOR-ir neurons were found in the
infragranular zone (home to many of the parval-
bumin (PARV)-containing basket cells). On the
other hand, using a combination of immunocyto-
chemistry and in situ hybridization, DOR mRNA
was observed in more than 90% of the PARV-
positive neurons sampled in the GCL and hilus, in

11% of hilar SOM-ir neurons, and in none of the
calretinin-ir neurons (Stumm et al., 2004). The rea-
son for the discrepancy in prevalence with the ear-
lier study (Commons and Milner, 1996a) is not
clear. Possibilities include translation-induced
differences (e.g., low production of the DOR pro-
tein in PARV-containing neurons or low levels of
SOM production in some of the neurons containing
SOM mRNA), or perhaps a dimerization of
MORs and DORs (George et al., 2000; Gomes
et al., 2000) in PARV-containing neurons that leads
to lower immunocytochemical recognition of DOR
protein.
None of the DOR mRNA-containing neurons
in the dentate were observed to contain pro-
enkephalin mRNA (Stumm et al., 2004), suggest-
ing little autoregulation of DOR-bearing interneu-
rons. However, since enkephalins and modest
DOR levels have both been observed in granule
cells, autocrine regulation of mossy fibers or gran-
ule cell dendrites remains a possibility.
In summary, DORs are primarily in several in-
terneuron populations that inhibit granule cells,
are present at low levels in granule cells, and are
absent from the calretinin-containing interneurons
that inhibit other interneurons. Since the targets of
the identified DOR-containing interneuron popu-
lations are relatively well understood (for review
see (Freund and Buzsáki, 1996), a functional link
can be postulated. The presence of DORs in SOM/
NPY-containing interneurons suggests a role in
the inhibition of granule cell distal dendrites, and
in PARV neurons (if the DOR protein is indeed
made by these cells) suggests a role in inhibiting
granule cell output.
Mu-opioid receptors (MORs)
Receptor autoradiography has shown that MOR
binding in the DG is more prominent than DOR
binding, although both are similarly distributed in
a diffuse pattern (Mansour et al., 1988). By light
microscopy, MOR immunoreactivity is present in
scattered neurons in the rat and mouse DG
(Arvidsson et al., 1995; Bausch et al., 1995b;
Mansour et al., 1995; Ding et al., 1996; Bausch
and Chavkin, 1997; Kalyuzhny and Wessendorf,
1997; Drake and Milner, 1999). Dual-labeling
immunocytochemistry and ultrastructural mor-
phological characterization showed that MOR-
labeling is in both the somatodendritic and axonal
compartments of GABAergic interneurons (Drake
and Milner, 1999). Notably, MOR is common in
presumed basket cell terminals that form inhibi-
tory-type synapses with granule cell somata and
dendrites (Drake and Milner, 1999). Analysis of
the neurochemical content of MOR-labeled cells
indicates that MOR mRNA (Stumm et al., 2004)
and immunoreactivity (Drake and Milner, 2006)
are most frequent in PARV-containing interneu-
rons, which comprise one group of basket cells
(see Freund and Buzsáki, 1996 for review). MOR
immunoreactivity (Drake and Milner, 2006) and
mRNA (Stumm et al., 2004) are also a modest
number of the SOM-containing hilar interneurons
(the ‘‘HIPP’’ neurons that project to the outer
molecular layer). In the outer molecular layer,
MOR-ir axons and terminals form inhibitory-type
synapses with dendrites containing NMDA recep-
tor immunoreactivity (Milner and Drake, 2001).
These results are consistent with physiological
studies described below, and suggest that opioids
could inhibit GABAergic terminals that modulate
granule cell dendrites, boosting depolarizing
events in granule cells and facilitating the activa-
tion of NMDA receptors. Finally, MOR mRNA is
in a small proportion (10%) of calretinin-contain-
ing interneurons in the GCL (Stumm et al., 2004).
This is in contrast to DOR mRNA, which is
absent from calretinin-ir neurons (Stumm et al.,
2004). Together the above morphological and
dual-labeling studies suggest that MORs are most
frequent in interneurons specialized to inhibit
granule cell output (i.e., PARV-containing basket
cells), are in a limited number of interneurons
that inhibit granule cell distal dendrites, and are
occasionally in interneurons that innervate other
interneurons.

The subcellular location of MOR immunoreac-
tivity is consistent with extrasynaptic activation by
diffusing endogenous opioids. In somata and
dendrites of GABAergic neurons, MOR immuno-
reactivity is often along the plasma membrane but
rather distant from synapses (Drake and Milner,
1999). Moreover, in the rat hilus, profiles bearing
MORs are not contacted by many leu-enkephalin-
labeled profiles, but about one-third of MOR-ir
axons, terminals, and dendrites are within 3 mm of
leu-enkephalin-ir profiles. An estimation of the
effective range of one leu-enkephalin-containing
vesicle can be made by calculating how far leu-
enkephalin released from a single dense-core ves-
icle would diffuse to reach a concentration that
would activate half of the available MORs. By
knowing the estimated affinity of leu-enkephalin
for MORs, the concentration of leu-enkephalin in
a single dense-core vesicle, and the extracellular
volume fraction of the hilus, we have calculated
that the approximate volume occupied by relevant
concentration of peptide from one dense-core ves-
icle is 35–104 fl, with a radius of 2.03–2.92 mm (for
more detail see Drake et al., 2002). Thus, many
MORs are likely to be within a functionally rel-
evant distance of released enkephalin (Drake et al.,
2002). The lack of synaptic associations suggests
that activation of MORs by leu-enkephalin re-
quires volume transmission, and furthermore that
dendritic, axonal, and terminal MORs are equally
likely to be activated by leu-enkephalin released
from mossy fibers. There is also a possibility that
some MORs are activated in an autocrine fashion
by enkephalins. One group observed that 75% of
hilar cells with MOR mRNA also label for pro-
enkephalin mRNA (Stumm et al., 2004). However,
these neurons may not translate both MORs and
enkephalin, as proenkephalin- and MOR- mRNAs
are not seen in the same transmitter-identified hilar
populations (Stumm et al., 2004) and an electron
microsopic study of the hilus showed little colo-
calization of immunoreactivities for MOR and leu-
enkephalin in dendrite or axon terminal profiles
(Drake et al., 2002). This suggests that, at least in
the hilus, autoregulation of neurotransmitter re-
lease by MORs may be minor. Rather, enkephalins
released from the mossy fibers or other interneu-
rons appear more likely to activate hilar MORs.
253
MORs are in a few other interesting locations in
the DG. First, MOR immunoreactivity is seen in
occasional dendrites of granule cells (Drake and
Milner, 1999). This location is consistent with
other evidence that MOR agonists can act directly
on granule cells (Piguet and North, 1993; Stumm
et al., 2004), and contrasts with the hippocampus
proper, where MORs are absent from principal
cells (Drake and Milner, 1999). Second, MORs are
on some afferents to the DG. MOR labeling is in
cholinergic septal terminals in the hilus (Kaplan
et al., 2004) and in a subset of GABAergic septal
afferents (Alreja et al., 2000). MOR-ir terminals
also contact the same hilar neuronal targets as do
cholinergic terminals (Kaplan et al., 2004). These
findings suggest the existence of numerous inter-
actions between opioid systems and septal projec-
tions in the DG.
Kappa-opioid receptors (KORs)
The rat DG contains little KOR binding and vir-
tually no KOR immunoreactivity (Mansour et al.,
1988, 1996; Unterwald et al., 1991). [However, see
Maderspach et al., 1991, 1995)] Considerably more
KOR binding is found in the guinea pig DG
(McLean et al., 1987; Wagner et al., 1991). In
guinea pig, KOR binding is most intense in the
dentate outer molecular layer, with a thin band in
the inner molecular layer (McLean et al., 1987).
The outer molecular layer KORs disappear fol-
lowing entorhinal cortex lesions, consistent with a
presynaptic localization in PP afferents (Simmons
et al., 1994). At least two subtypes of pharmaco-
logically identified KORs (K1 and K2) have been
identified in the guinea pig DG (Zukin et al., 1988;
Unterwald et al., 1991). Two groups, using differ-
ent carefully characterized antibodies, have local-
ized KOR immunoreactivity in the guinea pig
brain, and found a pattern that largely overlaps
with KOR binding but has some notable differ-
ences (Arvidsson et al., 1994; Drake et al., 1996).
The light microscopic distribution of KOR
immunoreactivity overlaps the distribution of
KOR binding in the inner molecular layer (Drake
et al., 1996), but is much less intense in the outer
molecular layer. This initially surprising result
254
suggests that in these latter regions KOR-binding
may reflect receptors that are less accessible to the
antibodies (due to alternative conformations or
associations with other proteins), or are different
subtypes. KOR immunoreactivity in the inner
molecular layer is near (within 100 mm) large stores
of dynorphin, but these immunoreactivities do not
overlap, suggesting diffusion must occur if the
KORs are to be activated (Drake et al., 1996). The
reliance on diffusion is consistent with other
strong evidence that released dynorphin can
diffuse a sufficient distance to activate dentate
KORs (Wagner et al., 1991, 1993; Drake et al.,
1994).
KOR immunoreactivity in the granule cell and
inner molecular layers is restricted to unmyelin-
ated axons and axon terminals. Many KOR
immunolabeled terminals colocalize substance P
(SubP) and form asymmetric synapses with gran-
ule cell perikarya and large unlabeled dendrites
concentrated in the dentate inner molecular layer
(Drake et al., 1997). Following fornix lesions,
KOR- and SubP-ir processes dramatically de-
crease in the molecular layer. These data suggest
a role for KORs in presynaptically modulating
supramammillary afferents to granule cells, thus
helping to regulate information flow into the DG.
KORs in the DG have an interesting subcellular
location. There is an association with the inner
surfaces of plasma membranes (Drake et al.,
1996), as expected for the cytoplasmic portion of
functional cell-surface receptors. KOR immuno-
reactivity is also found in large dense-core vesicles
(Drake et al., 1996). This localization suggests the
possibility that KORs are incorporated into the
presynaptic membrane when the dense-core vesi-
cle’s contents are released. Since strong stimula-
tion is required to release dense-core vesicles
(Thureson-Klein and Klein, 1990), the implication
is that these KORs may be a reserve pool of
receptors that undergo excitation-dependent pres-
entation to the plasma membrane. Such a mech-
anism is of obvious relevance for a role in LTP or
in response to hyperstimulation. A presence in
dense-core vesicles may be a common property
of KORs, since KORs have been localized to
dense-core vesicles in other brain regions (Shuster
et al., 1999; Svingos et al., 1999). Moreover, in
hypothalamus, translocation of KORs to the
plasma membrane is observed following the stim-
ulation of peptide release (Shuster et al., 1999),
providing direct evidence for activity-dependent
presentation of KORs to the membrane.
Opioid physiology
Mu- and delta-opioid actions in the hippocampus
are largely disinhibitory (Zieglgänsberger et al.,
1979). For example, enkephalin was found to
hyperpolarize GABAergic interneurons in the rat
hippocampus (Madison and Nicoll, 1988), and
MOR and DOR activation blocked GABAergic
input and induced disinhibitory effects in the rat
DG (Neumaier et al., 1988). Whole-cell voltage
clamp recordings of granule cells also showed that
mu-receptor activation reduced monosynaptic
stimulation-evoked inhibitory post-synaptic cur-
rents (IPSCs) (Xie et al., 1992). The reduction in
GABAergic tone generally increases the excitabil-
ity of the hippocampus in ways that affect seizure
susceptibility and synaptic plasticity [i.e., LTP and
long-term depression (LTD)] (Cohen et al., 1992;
Morris and Johnston, 1995). LTP induction by
high-frequency stimulation of the PP input to the
DG in rat hippocampal slices was facilitated by
MOR and DOR activation (Bramham and Sarvey,
1996) and this was found to be caused by disin-
hibition.
The inhibition of GABAergic neurons by
MORs and DORs is due to membrane hyperpo-
larization, and the ionic basis has been character-
ized by several groups using voltage clamp
electrophysiology. Mu opioids activate both in-
ward-rectifying and voltage-gated potassium chan-
nels in hippocampal interneurons (Wimpey and
Chavkin, 1991). Opioids can also increase the
M-type potassium channel conductance in hippo-
campus (Moore et al., 1994). G protein-coupled
inwardly rectifying potassium channels (GIRK or
Kir3) can be activated by the MOR agonist DAM-
GO in slices from mice (Luscher et al., 1997).
These multiple hyperpolarizing mechanisms re-
duce inhibitory input within the DG and indirectly
increase granule cell excitability.

Kappa opioids can directly affect presynaptic
release of excitatory amino acids in the guinea pig
DG, leading to inhibition of excitatory transmis-
sion. (However, in rat, disinhibition resembling
MOR and DOR activation has been reported
(Neumaier et al., 1988).) Intracellular recordings
of dentate granule cells demonstrated that KOR
activation significantly reduced the amplitude of
glutamatergic excitatory post-synaptic potentials
(EPSPs) while having no direct effects on mem-
brane conductance (Wagner et al., 1992). Provid-
ing further evidence for presynaptic inhibition, a
KOR-selective agonist inhibited excitatory re-
sponses to PP stimulation but not to applied
glutamate, and potentiated paired-pulse facilita-
tion (Simmons et al., 1994). Kappa receptor acti-
vation blocked excitatory amino acid release from
PP and recurrent hilar input to granule cells in the
DG, and this inhibition reduced LTP induction of
these two pathways (Terman et al., 1994, 2000).
Much of the evidence for KOR agonist inhibition
of LTP in the DG (and in region CA3) has been
reviewed by (Morris and Johnston, 1995).
Opioids cause presynaptic inhibition of certain
subcortical afferents to the DG. Neurochemical
measures of transmitter release showed that the
stimulated release of serotonin (Yoshioka et al.,
1993), norepinephrine (Jackisch et al., 1984;
Matsumoto et al., 1994), and acetylcholine
(Jackisch et al., 1986) were each directly inhibited
by opioid receptor activation. The ionic mechanisms
of these inhibitory presynaptic actions have not
been directly defined because it is difficult to electro-
physiologically record from hippocampal nerve ter-
minals, but molecular mechanisms that have been
suggested include activation of delayed rectifying
potassium channels (Wimpey and Chavkin, 1991),
opioid inhibition of voltage-sensitive calcium chan-
nels by Gbg binding (Herlitze et al., 1996; Ikeda,
1996), or direct inhibition of the vesicular fusion
machinery by opioid receptor activation (Scholz
and Miller, 1992; Capogna et al., 1996).
Opioids and seizures
Seizures modulate, and are modulated by, hippo-
campal opioid systems. The topic of seizures and
255
dentate opioids has been reviewed elsewhere
(Tortella et al., 1988; Simmons and Chavkin,
1996; Simonato and Romualdi, 1996), and will be
discussed only briefly here. Levels of enkephalins
and dynorphins in the hippocampal formation are
altered in humans with temporal lobe epilepsy
(Houser et al., 1990; Rees et al., 1994), as well as in
rodent seizure models (Hong et al., 1993). Dentate
levels of enkephalin and dynorphin are altered in
rodents by seizure-inducing treatments, including
electroconvulsive shock, amygdala kindling
(Vindrola et al., 1981; McGinty et al., 1986), and
administration of excitatory chemical agents
(Hong et al., 1980, 1988; Pierce et al., 1999; Pierce
and Milner, 2001). The time course and direction
of peptide changes varies; in general enkephalin
mRNA and immunoreactivity increase within
hours of seizure while dynorphin decreases rap-
idly but then recovers (for reviews, see Hong et al.,
1993; Pierce et al., 1999). Opioid receptors in the
DG also change following seizures, with both
MORs and DORs showing a shift in distribution
that is in line with anatomical reorganization of
particular neuronal circuits, such as mossy fiber
sprouting (Bausch and Chavkin, 1997; Skyers et
al., 2003). Electrophysiological responses to MOR
and DOR agonists also are altered after such sei-
zure-associated circuit reorganization (Bausch and
Chavkin, 1997), consistent with a persistent func-
tional change in the dentate opioid systems.
Opioid regulation of excitability within the DG
and hippocampus also has profound effects on
seizure susceptibility. Iontophoretic application of
opioids produced epileptiform increases in excita-
bility (French and Siggins, 1980). Morphine-
pretreated rats show enhanced susceptibility to
amydala kindling and morphological changes in
the DG (Rocha et al., 1996). Seizure development
is thought to involve MOR activation, since MOR
agonists, but not DOR agonists, produce convul-
sions and wet-dog shakes (Lee et al., 1989; Hong
et al., 1993). However, in some reports, MOR
agonists administered systemically had anti-
convulsant actions that may reflect their inhibi-
tory actions on other brain regions (reviewed by
Simmons and Chavkin, 1996). KOR activation
has been consistently reported to dampen seizure
(reviewed by Tortella et al., 1988). For example,
256
KOR activation in rats reduces excitability of the
dentate and reduces pilocarpine-induced seizures
(Bausch et al., 1998). This anti-convulsant action is
consistent with the inhibition of excitatory neuro-
transmission in the DG and CA3 produced by
exogenous and endogenous KOR agonists.
Opioids and gonadal steroids
There is ample precedent for interactions between
opioids and gonadal steroids in other brain regions
(e.g., hypothalamus). Because of this, and because
both opioids (for details see Opioid Physiology)
and gonadal steroids (for review see Woolley and
Schwartzkroin, 1998) modulate hippocampal plas-
ticity, it has been hypothesized that these two sys-
tems may interact in the hippocampal formation.
A few studies have examined gonadal steroid
effects on opioid receptors. Estradiol treatment
has been reported to increase MOR binding in
hippocampal homogenates (Piva et al., 1995).
Since estrogens do not appear to alter levels of
hippocampal MOR mRNA (Quinones-Jenab
et al., 1997), the increased MOR binding presum-
ably involves another mechanism, such as un-
masking, receptor dimerization, or increased
translation efficiency.
Evidence also suggests that sex hormones may
affect levels of dentate opioid peptides. One group
(Roman et al., 2006) used radioimmunoassay to
determine whether opioid peptide levels changed
across the estrous cycle. They found no significant
changes in the levels of dynorphins or pro-
enkephalin-derived peptides in homogenates of
the hippocampal formation, however, the ratio of
dynorphins to their conversion products (including
perhaps leu-enkephalin) fluctuated. This suggests
that enzymatic processing of dynorphin peptides
changes across the estrous cycle, and may affect
the availability of dynorphin and its derivatives.
Using anatomical methods and focusing on par-
ticular subregions, we recently found evidence that
estrogen affects the levels of leu-enkephalin and
dynorphin immunoreactivity in the DG and the
CA3 SLu (Torres-Reveron et al., 2006a, b). Rats
in proestrus (high estrogen) have significantly
higher levels of leu-enkephalin in the hilus and in
CA3 compared to rats in diestrus. Rats with 24 h
(but not 6 or 72 h) of estrogen replacement showed
increased levels of leu-enkephalin. Dynorphin-
immunoreactivity was elevated only in the apex
of the hilus in proestrus rats, but showed signifi-
cant increases in the entire hilus and CA3 SLu 24 h
after estrogen replacement in ovariectomized rats.
These data suggest that during the estrogen peaks
of the ovarian cycle, a greater amount of
enkephalin is available for release from mossy
fibers and perhaps enkephalinergic interneurons,
suggesting greater excitability in hippocampal cir-
cuits. We also found that immunoreactivity for
the estrogen receptor ERb was colocalized with
enkephalin- or dynorphin-labeling in mossy fiber
terminals and smaller terminals. Thus, estrogen
may directly influence enkephalin and/or dynor-
phin release through activation of estrogen recep-
tors on opioidergic neurons.
Opioids and neurogenesis
The DG is one of the few brain regions that un-
dergoes neurogenesis throughout adulthood (see
Gould and Gross, 2002). New granule cells, and
interneurons (Liu et al., 2003), are generated in the
subgranular zone of the hilus by progenitor cells
(Seri et al., 2001; Turlejski and Djavadian, 2002).
As noted in earlier sections, granule cells express
low levels of MORs and DORs (Commons and
Milner, 1996a; Drake and Milner, 1999), and con-
tain enkephalin (Gall et al., 1981) and dynorphin
(Chavkin, 2000). Some hilar interneurons likewise
contain opioid receptors and/or enkephalin
(Gall et al., 1981; Commons and Milner, 1996a).
Neurogenesis has been linked to certain forms
of hippocampal-dependent associative learning
(Shors et al., 2001, 2002) and is influenced by
many factors; it is enhanced by exercise and anti-
depressant drugs, inhibited by stress, and fluctu-
ates across the female estrous cycle (Tanapat et al.,
1999; Eisch et al., 2000; Gould and Gross, 2002).
Since cognitive impairment is one of the problems
associated with chronic opiate use, and opiates can
impair neurogenesis in the prenatal brain, Eisch
et al. (2000) examined whether opiates also may
affect neurogenesis in adult rats. Although an

acute exposure to morphine did not affect neuro-
genesis, neurogenesis was inhibited by chronic
morphine treatment or heroin self-administration.
The mechanism requires opioid-receptor activa-
tion, as evidenced by sensitivity to the opioid re-
ceptor antagonist naltrexone, and does not involve
stress hormones (Eisch et al., 2000). Newly born
neurons contain MORs (Eisch and Harburg,
2006), consistent with the possibility that mor-
phine may directly act on newly born neurons.
These findings suggest that inhibition of neuro-
genesis is one mechanism by which opiates may
exert long-lasting effects on dentate circuits.
Prenatal morphine and opioid system development
Prenatal exposure to opiates has many conse-
quences for the developing brain, including alter-
ation of the dentate opioid system and of
phenomena known to be influenced by opioids,
such as LTP and seizure threshold. In a study of
the DG of adult male rats exposed to morphine
prenatally, proenkephalin mRNA and met-enke-
phalin peptides were decreased, while prodynor-
phin mRNA and dynorphin B peptides were
increased (Schindler et al., 2004). Increases in the
density of MOR, but not DOR, binding were re-
ported in some strata of hippocampal subfields in
morphine-exposed males (Schindler et al., 2004).
In an investigation of the interactions between
prenatal morphine exposure and adult hormone
status (Slamberová et al., 2003), MOR density in
the DG was reduced in prenatal, morphine-
exposed males that received gonadal hormone re-
placement compared to gonadectomized males.
Controls exposed to prenatal saline did not show
this difference. There was a lower density of
MORs in the DG of females after hormone re-
placement, but this occurred with either prenatal
morphine treatment or with prenatal saline expo-
sure, suggesting that interactions between prenatal
morphine and adult hormone status were less rel-
evant in the female. Interestingly, the hippocam-
pus proper was affected differently than the DG,
with prenatal exposure to morphine altering the
density of MORs in adult female but not adult
male rats (Slamberová et al., 2003).
257
Prenatal morphine exposure produces functional
consequences for the hippocampal formation. Al-
though rats that are exposed to morphine prenatally
did not have a reduced seizure threshold to MOR
agonists, they showed an opioid-receptor dependent
decrease in sensitivity to bicuculline-induced seizures
(Schindler et al., 2004). Similarly, in slices from
adult male rats, prenatal morphine exposure was
reported to enhance susceptibility to epileptiform
activity in the outer molecular layer, but to shift
long-term plasticity of PP-granule cell synapses in
favor of LTD (Velı́sek et al., 2000). Clearly the
effects of prenatal morphine are complex, and
involve both afferents and intrinsic dentate circuits.
Summary
Opiate drugs alter cognitive performance and in-
fluence hippocampal excitability, including LTP
and seizure activity. In the DG, the two major
opioid peptides, enkephalins, and dynorphins,
have opposing effects on excitability. Enkephalins
preferentially bind to DORs and MORs while
dynorphins preferentially bind to KORs. Opioid
receptors can also be activated by exogenous opi-
ate drugs such as morphine. Enkephalins are con-
tained in the mossy fiber pathway, in the lateral PP
and in scattered GABAergic interneurons. MORs
and DORs are predominantly in distinct sub-
populations of GABAergic interneurons known to
inhibit granule cells, and are present at low levels
within granule cells. MOR and DOR agonists
increase excitability and facilitate LTP in the mo-
lecular layer. Anatomical and physiological evi-
dence is consistent with somatodendritic and axon
terminal targeting of both MORs and DORs.
Dynorphins are most abundant in mossy fibers but
also are in the granule cells and their dendrites.
KORs have been localized to granule cell mossy
fibers, supramammillary afferents to granule cells,
and PP terminals. KOR agonists, including en-
dogenous dynorphins, diminish the induction of
LTP. Opiates and opioids also modulate other
processes in the hippocampal formation, including
adult neurogenesis, the actions of gonadal hor-
mones, and development of neonatal transmitter
systems.
258

Abbreviations Arvidsson, U., Riedl, M., Chakrabarti, S., Lee, J.H., Nakano,
A.H., Dado, R.J., Loh, H.H., Law, P.Y., Wessendorf, M.W.
DG dentate gyrus and Elde, R. (1995) Distribution and targeting of a mu-
opioid receptor (MOR1) in brain and spinal cord. J. Neuro-
DOR delta-opioid receptor sci., 15: 3328–3341.
EPSPs excitatory post-synaptic potentials Arvidsson, U., Risling, M., Frisén, J., Piehl, F., Fried, K.,
GABA gamma amino butyric acid Hökfelt, T. and Cullheim, S. (1994) trkC-like immunoreac-
GCL granule cell layer tivity in the primate descending serotoninergic system. Eur. J.
GIRK G protein-coupled inwardly recti- Neurosci., 6: 230–236.
Bausch, S.B. and Chavkin, C. (1997) Changes in hippocampal
fying potassium channels circuitry after pilocarpine-induced seizures as revealed by
ir immunoreactive opioid receptor distribution and activation. J. Neurosci., 17:
KOR kappa-opioid receptor 477–492.
LTD long-term depression Bausch, S.B., Esteb, T.M., Terman, G.W. and Chavkin, C.
LTP long-term potentiation (1998) Administered and endogenously released kappa op-
ioids decrease pilocarpine-induced seizures and seizure-
MOR mu-opioid receptor
induced histopathology. J. Pharmacol. Exp. Ther., 284:
ORL1 opioid-like receptor 1147–1155.
PARV parvalbumin Bausch, S.B., Patterson, T.A., Appleyard, S.M. and Chavkin,
PP perforant path C. (1995a) Immunocytochemical localization of delta
SLM stratum lacunosum-moleculare opioid receptors in mouse brain. J. Chem. Neuroanat., 8:
175–189.
SLu stratum lucidum
Bausch, S.B., Patterson, T.A., Ehrengruber, M., Lester, H.A.,
SO stratum oriens Davidson, N. and Chavkin, C. (1995b) Colocalization of mu
SOM/NPY somatostatin/neuropeptide Y opioid receptors with GIRK1 potassium channels in rat
SP stratum pyramidale brain: an immunocytochemical study. Receptors Channels, 3:
SubP substance P 221–241.
Bouvier, M. (2001) Oligomerization of G-protein-coupled
SR stratum radiatum
transmitter receptors. Nat. Rev. Neurosci., 2: 274–286.
SUB subiculum Bramham, C.R. (1992) Opioid receptor dependent long-term
TAT temporal-ammonic tract potentiation: peptidergic regulation of synaptic plasticity in
the hippocampus. Neurochem. Int., 20: 441–455.
Bramham, C.R. and Sarvey, J.M. (1996) Endogenous activa-
tion of m and d-1 opioid receptors is required for long-
Acknowledgments term potentiation induction in the lateral perforant path:
Dependence on GABAergic inhibition. J. Neurosci., 16:
We thank Ms. Katherine Mitterling and Mr. 8123–8131.
Bunzow, J.R., Saez, C., Mortrud, M., Bouvier, C., Williams,
Bradley Graustein for technical assistance. Sup-
J.T., Low, M. and Grandy, D.K. (1994) Molecular cloning
ported by NIH grants: DA 08259 (T.A.M; and tissue distribution of a putative member of the rat opioid
C.T.D.; C.C.) and HL 18974 (T.A.M.; CTD) and receptor gene family that is not a mu, delta or kappa opioid
DA04123 (C.C.). receptor type. FEBS Lett., 347: 284–288.
Capogna, M., Gahwiler, B.H. and Thompson, S.M. (1996)
Presynaptic inhibition of calcium-dependent and -independ-
ent release elicited with ionomycin, gadolinium, and alpha-
References latrotoxin in the hippocampus. J. Neurophysiol., 75:
2017–2028.
Abbadie, C., Pan, Y.X., Drake, C.T. and Pasternak, G.W. Caudle, R.M., Chavkin, C. and Dubner, R. (1994) Kappa2
(2000) Comparative immunohistochemical distributions of opioid receptors inhibit NMDA receptor-mediated synaptic
carboxy terminus epitopes from the mu-opioid receptor currents in guinea pig CA3 pyramidal cells. J. Neurosci., 14:
splice variants MOR-1D, MOR-1 and MOR-1C in the mouse 5580–5589.
and rat CNS. Neuroscience, 100: 141–153. Caudle, R.M. and Dubner, R. (1998) Ifenprodil blocks the ex-
Alreja, M., Shanabrough, M., Liu, W.M. and Leranth, C. citatory effects of the opioid peptide dynorphin 1–17 on
(2000) Opioids suppress IPSCs in neurons of the rat medial NMDA receptor mediated currents in the CA3 region of the
septum/diagonal band of Broca: Involvement of m-opioid guinea pig hippocampus. Neuropeptides, 32: 87–95.
receptors and septohippocampal GABAergic neurons. Chavkin, C. (2000) Dynorphins are endogenous opioid peptides
J. Neurosci., 20: 1179–1189. released from granule cells to act neurohumorly and inhibit

excitatory neurotransmission in the hippocampus. Prog.
Brain Res., 125: 363–367.
Chavkin, C., Henriksen, S.J., Siggins, G.R. and Bloom, F.E.
(1985a) Selective inactivation of opioid receptors in rat hip-
pocampus demonstrates that dynorphin-A and -B may act on
mu-receptors in the CA1 region. Brain Res., 331: 366–370.
Chavkin, C., James, I.F. and Goldstein, A. (1982) Dynorphin is
a specific endogenous ligand of the kappa opioid receptor.
Science, 215: 413–415.
Chavkin, C., Shoemaker, W.J., McGinty, J.F., Bayon, A. and
Bloom, F.E. (1985b) Characterization of the prodynorphin
and proenkephalin neuropeptide systems in rat hippocam-
pus. J. Neurosci., 5: 808–816.
Chen, Y., Mestek, A., Liu, J., Hurley, J.A. and Yu, L. (1993)
Molecular cloning and functional expression of a mu-opioid
receptor from rat brain. Mol. Pharmacol., 44: 8–12.
Clark, J.A., Liu, L., Price, M., Hersh, B., Edelson, M. and
Pasternak, G.W. (1989) Kappa opiate receptor multiplicity:
evidence for two U50,488-sensitive kappa 1 subtypes and a
novel kappa 3 subtype. J. Pharmacol. Exp. Ther., 251:
461–468.
Cohen, G.A., Doze, V.A. and Madison, D.V. (1992) Opioid
inhibition of GABA release from presynaptic terminals of rat
hippocampal interneurons. Neuron, 9: 325–335.
Commons, K.G. and Milner, T.A. (1995) Ultrastructural het-
erogeneity of enkephalin-containing neurons in the rat hip-
pocampal formation. J. Comp. Neurol., 358: 324–342.
Commons, K.G. and Milner, T.A. (1996a) Cellular and sub-
cellular localization of delta opiate receptor immunoreactiv-
ity in the rat dentate gyrus. Brain Res., 738: 181–195.
Commons, K.G. and Milner, T.A. (1996b) The ultrastructural
relationships between leu-enkephalin and GABA containing
neurons differ within the hippocampal formation. Brain Res.,
724: 1–15.
Corbett, A.D., Paterson, S.J. and Kosterlitz, H.W. (1993) Se-
lectivity of ligands for opioid receptors. In: Herz A. (Ed.),
Opioids I. Springer-Verlag, Berlin, pp. 645–680.
Corbett, A.D., Paterson, S.J., McKnight, A.T., Magnan, J. and
Kosterlitz, H.W. (1982) Dynorphin and dynorphin are lig-
ands for the kappa-subtype of opiate receptor. Nature, 299:
79–81.
Crain, B.J., Chang, K.J. and McNamara, J.O. (1986) Quanti-
tative autoradiographic analysis of mu and delta opioid
binding sites in the rat hippocampal formation. J. Comp.
Neurol., 246: 170–180.
De Camilli, P. and Jahn, R. (1990) Pathways to regulated ex-
ocytosis in neurons. Annu. Rev. Physiol., 52: 625–645.
Ding, Y.Q., Kaneko, T., Nomura, S. and Mizuno, N. (1996)
Immunohistochemical localization of m-opioid receptors in
the central nervous system of the rat. J. Comp. Neurol., 367:
375–402.
Drake, C.T., Chang, P.C., Harris, J.A. and Milner, T.A. (2002)
Neurons with mu opioid receptors interact indirectly with
enkephalin-containing neurons in the rat dentate gyrus. Exp.
Neurol., 176: 254–261.
Drake, C.T., Chavkin, C. and Milner, T.A. (1997) Kappa op-
ioid receptor-like immunoreactivity is present in substance
259
P-containing subcortical afferents in guinea pig dentate
gyrus. Hippocampus, 7: 36–47.
Drake, C.T. and Milner, T.A. (1999) Mu opioid receptors are in
somatodendritic and axonal compartments of GABAergic
neurons in rat hippocampal formation. Brain Res., 849:
203–215.
Drake, C.T. and Milner, T.A. (2006) Mu opioid receptors are
extensively co-localized with parvalbumin, but not som-
atostatin, in the dentate gyrus. Neurosci. Lett., 403: 176–180.
Drake, C.T., Patterson, T.A., Simmons, M.L., Chavkin, C. and
Milner, T.A. (1996) Kappa opioid receptor-like immunore-
activity in guinea pig brain: ultrastructural localization in
presynaptic terminals in hippocampal formation. J. Comp.
Neurol., 370: 377–395.
Drake, C.T., Terman, G.W., Simmons, M.L., Milner, T.A.,
Kunkel, D.D., Schwartzkroin, P.A. and Chavkin, C. (1994)
Dynorphin opioids present in dentate granule cells may func-
tion as retrograde inhibitory neurotransmitters. J. Neurosci.,
14: 3736–3750.
Eisch, A.J., Barrot, M., Schad, C.A., Self, D.W. and Nestler,
E.J. (2000) Opiates inhibit neurogenesis in the adult rat hip-
pocampus. Proc. Natl. Acad. Sci. U.S.A., 97: 7579–7584.
Eisch, A.J. and Harburg, G.C. (2006) Opiates, psychostimu-
lants, and adult hippocampal neurogenesis: Insights for
addiction and stem cell biology. Hippocampus, 16: 271–286.
Evans, C.J., Keith, D.E., Morrison, H., Magendzo, K. and
Edwards, R.H. (1992) Cloning of a delta opioid receptor by
functional expression. Science, 258: 1952–1955.
Fredens, K., Stengaard Pedersen, K. and Larsson, L.I. (1984)
Localization of enkephalin and cholecystokinin immunore-
activities in the perforant path terminal fields of the rat hip-
pocampal formation. Brain Res., 304: 255–263.
French, E.D. and Siggins, G.R. (1980) An iontophoretic
survey of opioid peptide actions in the rat limbic system: in
search of opiate epileptogenic mechanisms. Regul. Pept., 1:
127–146.
Freund, T.F. and Buzsáki, G. (1996) Interneurons of the hip-
pocampus. Hippocampus, 6: 347–470.
Gall, C. (1988a) Seizures induce dramatic and distinctly differ-
ent changes in enkephalin, dynorphin, and CCK immunore-
activities in mouse hippocampal mossy fibers. J. Neurosci., 8:
1852–1862.
Gall, C.M. (1988b) Localization and seizure-induced alterations
of opioid peptides and CCK in the hippocampus. Natl. Inst.
Drug Abuse Res. Monogr. Ser., 82: 12–32.
Gall, C., Brecha, N., Karten, H.J. and Chang, K.-J. (1981)
Localization of enkephalin-like immunoreactivity to identi-
fied axonal and neuronal populations of the rat hippocam-
pus. J. Comp. Neurol., 198: 335–350.
George, S.R., Fan, T., Xie, Z., Tse, R., Tam, V., Varghese, G.
and O’Dowd, B.F. (2000) Oligomerization of mu- and
delta-opioid receptors, generation of novel functional prop-
erties. J. Biol. Chem., 275: 26128–26135.
Gomes, I., Jordan, B.A., Gupta, A., Trapaidze, N., Nagy, V.
and Devi, L.A. (2000) Heterodimerization of m and d opioid
receptors: a role in opiate synergy. J. Neurosci., 20:
NIL22–NIL26.
260
Goodman, R.R., Snyder, S.H., Kuhar, M.J. and Young III,
W.S. (1980) Differentiation of delta and mu opiate receptor
localizations by light microscopic autoradiography. Proc.
Natl. Acad. Sci. U.S.A., 77: 6239–6243.
Gould, E. and Gross, C.G. (2002) Neurogenesis in adult
mammals: some progress and problems. J. Neurosci., 22:
619–623.
Gubler, U., Seeburg, P., Hoffman, B.J., Gage, L.P. and Uden-
friend, S. (1982) Molecular cloning establishes proenkephalin
as precursor of enkephalin-containing peptides. Nature, 295:
206–208.
Guerra, D., Sole, A., Cami, J. and Tobena, A. (1987) Neuro-
physiological performance in opiate addicts after rapid
detoxification. Drug Alcohol Depend., 20: 261–270.
Gulya, K., Gehlert, D.R., Wamsley, J.K., Mosberg, H., Hruby,
V.J. and Yamamura, H.I. (1986) Light microscopic autora-
diographic localization of delta opioid receptors in the rat
brain using a highly selective bis-penicillamine cyclic enke-
phalin analog. J. Pharmacol. Exp. Ther., 238: 720–726.
Heinricher, M.M. (2003) Orphanin FQ/nociceptin: from neural
circuitry to behavior. Life Sci., 73: 813–822.
Herkenham, M. and McLean, S. (1988) The anatomical rela-
tionship of opioid peptides and opiate receptors in the hip-
pocampi of four rodent species. Natl. Inst. Drug Abuse Res.
Monogr. Ser., 82: 33–47.
Herlitze, S., Garcia, D.E., Mackie, K., Hille, B., Scheuer, T.
and Catterall, W.A. (1996) Modulation of Ca2+ channels by
G-protein beta gamma subunits. Nature, 380: 258–262.
Hong, J.S., McGinty, J.F., Grimes, L., Kanamatsu, T., Obie, J.
and Mitchell, C.L. (1988) Seizure-induced alterations in the
metabolism of hippocampal opioid peptides suggest opioid
modulation of seizure-related behaviors. Natl. Inst. Drug
Abuse Res. Monogr. Ser., 82: 48–66.
Hong, J.S., McGinty, J.F., Lee, P.H.K., Xie, C.W. and Mitc-
hell, C.L. (1993) Relationship between hippocampal opioid
peptides and seizures. Prog. Neurobiol., 40: 507–528.
Hong, J.S., Wood, P.L., Gillin, J.C., Yang, H.Y. and Costa, E.
(1980) Changes of hippocampal Met-enkephalin content af-
ter recurrent motor seizures. Nature, 285: 231–232.
Houser, C.R., Miyashiro, J.E., Swartz, B.E., Walsh, G.O.,
Rich, J.R. and Delgado-Escueta, A.V. (1990) Altered pat-
terns of dynorphin immunoreactivity suggest mossy fiber re-
organization in human hippocampal epilepsy. J. Neurosci.,
10: 267–282.
Ikeda, S.R. (1996) Voltage-dependent modulation of N-type
calcium channels by G-protein beta gamma subunits. Nature,
380: 255–258.
Jackisch, R., Geppert, M., Brenner, A.S. and Illes, P. (1986)
Presynaptic opioid receptors modulating acetylcholine re-
lease in the hippocampus of the rabbit. Naunyn Schmiedeb-
ergs Arch. Pharmacol., 332: 156–162.
Jackisch, R., Werle, E. and Hertting, G. (1984) Identification of
mechanisms involved in the modulation of release of nor-
adrenaline in the hippocampus of the rabbit in vitro.
Neuropharmacology, 23: 1363–1371.
Jiang, Q., Takemori, A.E., Sultana, M., Portoghese, P.S.,
Bowen, W.D., Mosberg, H.I. and Porreca, F. (1991)
Differential antagonism of opioid delta antinociception by
[D-Ala2,Leu5,Cys6]enkephalin and naltrindole 50 -isothiocyan-
ate: evidence for delta receptor subtypes. J. Pharmacol. Exp.
Ther., 257: 1069–1075.
Johnston, H.M. and Morris, B.J. (1994) Nitric oxide alters
proenkephalin and prodynorphin gene expression in hippo-
campal granule cells. Neuroscience, 61: 435–439.
Kakidani, H., Furutani, Y., Takahashi, H., Noda, M., Mori-
moto, Y., Hirose, T., Asai, M., Inayama, S., Nakanishi, S.
and Numa, S. (1982) Cloning and sequence analysis of
cDNA for porcine beta-neo-endorphin/dynorphin precursor.
Nature, 298: 245–249.
Kalyuzhny, A.E. and Wessendorf, M.W. (1997) CNS GABA
neurons express the mu-opioid receptor: immunocytochem-
ical studies. Neuroreport, 20(8): 3367–3372.
Kaplan, T.J., Skyers, P.R., Tabori, N.E., Drake, C.T. and
Milner, T.A. (2004) Ultrastructural evidence for mu-opioid
modulation of cholinergic pathways in rat dentate gyrus.
Brain Res., 1019: 28–38.
Karhunen, T., Vilim, F.S., Alexeeva, V., Weiss, K.R. and
Church, P.J. (2001) Targeting of peptidergic vesicles in
cotransmitting terminals. J. Neurosci., 21: RC127.
Khachaturian, H., Schaefer, M.K.H. and Lewis, M.E. (1993)
Anatomy and function of the endogenous opioid system.
In: Herz A. (Ed.), Opioid I. Springer-Verlag, Berlin,
pp. 471–497.
Kieffer, B.L., Befort, K., Gaveriaux-Ruff, C. and Hirth, C.G.
(1992) The delta opioid receptor: isolation of a cDNA by
expression cloning and pharmacological characterization.
Proc. Natl. Acad. Sci. U.S.A., 89: 12048–12052.
Lee, P.H., Obie, J. and Hong, J.S. (1989) Opioids induce con-
vulsions and wet dog shakes in rats: mediation by hippo-
campal mu, but not delta or kappa opioid receptors. J.
Neurosci., 9: 692–697.
Liu, S., Wang, J., Zhu, D.Y., Fu, Y., Lukowiak, K. and Lu,
Y.M. (2003) Generation of functional inhibitory neurons in
the adult rat hippocampus. J. Neurosci., 23: 732–736.
Luscher, C., Jan, L.Y., Stoffel, M., Malenka, R.C. and Nicoll,
R.A. (1997) G protein-coupled inwardly rectifying K+ chan-
nels (GIRKs) mediate postsynaptic but not presynaptic trans-
mitter actions in hippocampal neurons. Neuron, 19: 687–695.
Maderspach, K., Nemeth, K., Simon, J., Benyhe, S., Szucs, M.
and Wollemann, M. (1991) A monoclonal antibody recog-
nizing kappa—but not mu—and delta-opioid receptors.
J. Neurochem., 56: 1897–1904.
Maderspach, K., Takacs, J., Niewiadomska, G. and Csillag, A.
(1995) Postsynaptic and extrasynaptic localization of kappa-
opioid receptor in selected brain areas of young rat and chick
using an anti-receptor monoclonal antibody. J. Neurocytol.,
24: 478–486.
Madison, D.V. and Nicoll, R.A. (1988) Enkephalin hyperpo-
larizes interneurons in the rat hippocampus. J. Physiol., 398:
123–130.
Mansour, A., Burke, S., Pavlic, R.J., Akil, H. and Watson, S.J.
(1996) Immunohistochemical localization of the cloned
kappa 1, receptor in the rat CNS and pituitary. Neuro-
science, 71: 671–690.

Mansour, A., Fox, C.A., Burke, S., Akil, H. and Watson, S.J.
(1995) Immunohistochemical localization of the cloned mu
opioid receptor in the rat CNS. J. Chem. Neuroanat., 8:
283–305.
Mansour, A., Khachaturian, H., Lewis, M.E., Akil, H. and
Watson, S.J. (1987) Autoradiographic differentiation of mu,
delta, and kappa opioid receptors in the rat forebrain and
midbrain. J. Neurosci., 7: 2445–2464.
Mansour, A., Khachaturian, H., Lewis, M.E., Akil, H. and
Watson, S.J. (1988) Anatomy of CNS opioid receptors.
Trends Neurosci., 11: 308–314.
Mansour, A., Thompson, R.C., Akil, H. and Watson, S.J.
(1993) Delta opioid receptor mRNA distribution in the brain:
comparison to delta receptor binding and proenkephalin
mRNA. J. Chem. Neuroanat., 6: 351–362.
Martin, W.R. (1983) Pharmacology of opioids. Pharmacol.
Rev., 35: 283–323.
Matsumoto, M., Yoshioka, M., Togashi, H., Hirokami, M.,
Tochihara, M., Ikeda, T., Smith, C.B. and Saito, H. (1994)
mu-Opioid receptors modulate noradrenaline release from
the rat hippocampus as measured by brain microdialysis.
Brain Res., 636: 1–8.
Mattia, A., Vanderah, T., Mosberg, H.I. and Porreca, F. (1991)
Lack of antinociceptive cross-tolerance between [D-Pen2,
D-Pen5]enkephalin and [D-Ala2]deltorphin II in mice: evi-
dence for delta receptor subtypes. J. Pharmacol. Exp. Ther.,
258: 583–587.
McDaniel, K.L., Mundy, W.R. and Tilsom, H.A. (1990)
Microinjection of dynorphin into the hippocampal forma-
tion impairs spatial learning in rats. Pharmacol. Biochem.
Behav., 35: 429–435.
McGinty, J.F., Henriksen, S.J., Goldstein, A., Terenius, L.
and Bloom, F.E. (1983) Dynorphin is contained within
hippocampal mossy fibers: immunochemical altera-
tions after Kainic acid administration and colchicine-
induced neurotoxicity. Proc. Natl. Acad. Sci. U.S.A., 80:
589–593.
McGinty, J.F., Kanamatsu, T., Obie, J. and Hong, J.S. (1986)
Modulation of opioid peptide metabolism by seizures: differ-
entiation of opioid subclasses. Natl. Inst. Drug Abuse Res.
Monogr. Ser., 71: 89–101.
McLean, S., Rothman, R.B., Jacobson, A.E., Rice, K.C. and
Herkenham, M. (1987) Distribution of opiate receptor sub-
types and enkephalin and dynorphin immunoreactivity in the
hippocampus of squirrel, guinea pig, rat, and hamster.
J. Comp. Neurol., 255: 497–510.
Merchenthaler, I., Maderdrut, J.L., Cianchetta, P., Shughrue,
P.J. and Bronstein, D. (1997) In situ hybridization histo-
chemical localization of prodynorphin messenger RNA in the
central nervous system of the rat. J. Comp. Neurol., 384:
211–232.
Meunier, J.-C., Mollereau, C., Toll, L., Suaudeau, C., Moisand,
C., Alvinerie, P., Butour, J.-L., Guillemot, J.-C., Ferrara, P.,
Monsarrat, B., Mazarguil, H., Vassart, G., Parmentier, M.
and Costentin, J. (1995) Isolation and structure of the en-
dogenous agonist of opioid receptor-like ORL1 receptor.
Nature, 377: 532–536.
261
Milner, T.A. and Drake, C.T. (2001) Ultrastructural evidence
for presynaptic m opioid receptor modulation of synaptic
plasticity in NMDA-receptor-containing dendrites in the
dentate gyrus. Brain Res. Bull., 54: 131–140.
Minami, M., Hosoi, Y., Toya, T., Katao, Y., Maekawa, K.,
Katsumata, S., Yabuuchi, K., Onogi, T. and Satoh, M.
(1993) In situ hybridization study of kappa-opioid receptor
mRNA in the rat brain. Neurosci. Lett., 162: 161–164.
Mogil, J.S. and Pasternak, G.W. (2001) The molecular
and behavioral pharmacology of the orphanin FQ/nocicep-
tin peptide and receptor family. Pharmacol. Rev., 53:
381–415.
Moore, S.D., Madamba, S.G., Schweitzer, P. and Siggins, G.R.
(1994) Voltage-dependent effects of opioid peptides on hip-
pocampal CA3 pyramidal neurons in vitro. J. Neurosci., 14:
809–820.
Morris, B.J. and Johnston, H.M. (1995) A role for hippocampal
opioids in long-term functional plasticity. Trends Neurosci.,
18: 350–355.
Nestler, E.J. (2001) Molecular basis of long-term plasticity un-
derlying addiction. Nat. Rev. Neurosci., 2: 119–128.
Neumaier, J.F., Mailheau, S. and Chavkin, C. (1988) Opioid
receptor-mediated responses in the dentate gyrus and CA1
region of the rat hippocampus. J. Pharmacol. Exp. Ther.,
244: 564–570.
Nock, B., Rajpara, A., O’Connor, L.H. and Cicero, T.J. (1988)
Autoradiography of [3 H]U-69593 binding sites in rat brain:
evidence for kappa opioid receptor subtypes. Eur. J. Pharma-
col., 154: 27–34.
Noda, M., Furutani, Y., Takahashi, H., Toyosato, M., Hirose,
T., Inayama, S., Nakanishi, S. and Numa, S. (1982) Cloning
and sequence analysis of cDNA for bovine adrenal prep-
roenkephalin. Nature, 295: 202–206.
Pan, Y.X., Xu, J., Bolan, E., Abbadie, C., Chang, A., Zucker-
man, A., Rossi, G. and Pasternak, G.W. (1999) Identification
and characterization of three new alternatively spliced mu-
opioid receptor isoforms. Mol. Pharmacol., 56: 396–403.
Pierce, J.P., Kurucz, O. and Milner, T.A. (1999) The morph-
ometry of a peptidergic transmitter system before and after
seizure. I. Dynorphin B-like immunoreactivity in the hippo-
campal mossy fiber system. Hippocampus, 9: 255–276.
Pierce, J.P. and Milner, T.A. (2001) Parallel increases in the
synaptic and surface areas of mossy fiber terminals following
seizure induction. Synapse, 39: 249–256.
Pierce, T.L. and Wessendorf, M.W. (2000) Immunocytochem-
ical mapping of endomorphin-2-immunoreactivity in rat
brain. J. Chem. Neuroanat., 18: 181–207.
Piguet, P. and North, R.A. (1993) Opioid actions at mu and
delta receptors in the rat dentate gyrus in vitro. J. Pharmacol.
Exp. Ther., 266: 1139–1146.
Piva, F., Limonta, P., Dondi, D., Pimpinelli, F., Martini, L.
and Maggi, R. (1995) Effects of steroids on the brain opioid
system. J. Steroid Biochem. Mol. Biol., 53: 343–348.
Pu, L., Bao, G.B., Xu, N.J., Ma, L. and Pei, G. (2002)
Hippocampal long-term potentiation is reduced by chronic
opiate treatment and can be restored by re-exposure to opi-
ates. J. Neurosci., 22: 1914–1921.
262
Quinones-Jenab, V., Jenab, S., Ogawa, S., Inturrisi, C. and
Pfaff, D.W. (1997) Estrogen regulation of mu-opioid receptor
mRNA in the forebrain of female rats. Brain Res., 47: 1.
Rees, H., Ang, L.C., Shul, D.D., George, D.H., Begley, H. and
McConnell, T. (1994) Increase in enkephalin-like immuno-
reactivity in hippocampi of adults with generalized epilepsy.
Brain Res., 652: 113–119.
Reinscheid, R.K., Nothacker, H.P., Bourson, A., Ardati, A.,
Henningsen, R.A., Bunzow, J.R., Grandy, D.K., Langen, H.,
Monsma Jr., F.J. and Civelli, O. (1995) Orphanin FQ: a ne-
uropeptide that activates an opioid-like G protein-coupled
receptor. Science, 270: 792–794.
Rios, C.D., Jordan, B.A., Gomes, I. and Devi, L.A. (2001)
G-protein-coupled receptor dimerization: modulation of re-
ceptor function. Pharmacol. Ther., 92: 71–87.
Robbins, T.W. and Everitt, B.J. (2002) Limbic-striatal memory
systems and drug addiction. Neurobiol. Learn. Mem., 78:
625–636.
Rocha, L., Ackermann, R.F. and Engel Jr., J. (1996) Effects of
chronic morphine pretreatment on amygdaloid kindling de-
velopment, postictal seizure and suppression and benzodiaze-
pine receptor binding in rats. Epilepsy Res., 23: 225–233.
Roman, E., Ploj, K., Gustafsson, L., Meyerson, B.J. and
Nylander, I. (2006) Variations in opioid peptide levels during
the estrous cycle in Sprague-Dawley rats. Neuropeptides, 40:
195–206.
Rothman, R.B., Bowen, W.D., Schumacher, U.K. and Pert,
C.B. (1983) Effect of beta-FNA on opiate receptor binding:
preliminary evidence for two types of mu receptors. Eur. J.
Pharmacol., 95: 147–148.
Rothman, R.B., Jacobson, A.E., Rice, K.C. and Herkenham,
M. (1987) Autoradiographic evidence for two classes of mu
opioid binding sites in rat brain using [125I]FK33824. Pep-
tides, 8: 1015–1021.
Sandin, J., Nylander, I., Georgieva, J., Schött, P.A., Ögren,
S.O. and Terenius, L. (1998) Hippocampal dynorphin B in-
jections impair spatial learning in rats: a kappa-opioid re-
ceptor-mediated effect. Neuroscience, 85: 375–382.
Schindler, C.J., Slamberova, R., Rimanoczy, A., Hnactzuk,
O.C., Riley, M.A. and Vathy, I. (2004) Field-specific changes
in hippocampal opioid mRNA, peptides, and receptors due
to prenatal morphine exposure in adult male rats. Neurosci-
ence, 126: 355–364.
Scholz, K.P. and Miller, R.J. (1992) Inhibition of quantal
transmitter release in the absence of calcium influx by a G
protein-linked adenosine receptor at hippocampal synapses.
Neuron, 8: 1139–1150.
Seri, B., Garcia-Verdugo, J.M., McEwen, B.S. and Alvarez-
Buylla, A. (2001) Astrocytes give rise to new neurons in the
adult mammalian hippocampus. J. Neurosci., 21: 7153–7160.
Seward, E.P., Chernevskaya, N.I. and Nowycky, M.C. (1995)
Exocytosis in peptidergic nerve terminals exhibits two cal-
cium-sensitive phases during pulsatile calcium entry. J. Ne-
urosci., 15: 3390–3399.
Shors, T.J., Miesegaes, G., Beylin, A., Zhao, M., Rydel, T. and
Gould, E. (2001) Neurogenesis in the adult is involved in the
formation of trace memories. Nature, 410: 372–376.
Shors, T.J., Townsend, D.A., Zhao, M., Kozorovitskiy, Y. and
Gould, E. (2002) Neurogenesis may relate to some but not all
types of hippocampal-dependent learning. Hippocampus, 12:
578–584.
Shuster, S.J., Riedl, M., Li, X.R., Vulchanova, L. and Elde, R.
(1999) Stimulus-dependent translocation of kappa opioid re-
ceptors to the plasma membrane. J. Neurosci., 19: 2658–2664.
Simmons, M.L. and Chavkin, C. (1996) Endogenous opioid
regulation of hippocampal function. Int. Rev. Neurobiol., 39:
145–196.
Simmons, M.L., Terman, G.W., Drake, C.T. and Chavkin, C.
(1994) Inhibition of glutamate release by presynaptic kappa
1—opioid receptors in the guinea pig dentate gyrus. J. Ne-
urophysiol., 72: 1697–1705.
Simmons, M.L., Terman, G.W., Gibbs, S.M. and Chavkin, C.
(1995) L-type calcium channels mediate dynorphin neuro-
peptide release from dendrites but not axons of hippocampal
granule cells. Neuron, 14: 1265–1272.
Simonato, M. and Romualdi, P. (1996) Dynorphin and epi-
lepsy. Prog. Neurobiol., 50: 557–583.
Skyers, P.S., Einheber, S., Pierce, J.P. and Milner, T.A. (2003)
Increased mu-opioid receptor labeling is found on inner mo-
lecular layer terminals of the dentate gyrus following seizures.
Exp. Neurol., 179: 200–209.
Slamberová, R., Rimanoczy, A., Bar, N., Schindler, C.J. and
Vathy, I. (2003) Density of mu-opioid receptors in the hip-
pocampus of adult male and female rats is altered by prenatal
morphine exposure and gonadal hormone treatment. Hippo-
campus, 13: 461–471.
Slowe, S.J., Simonin, F., Kieffer, B. and Kitchen, I. (1999)
Quantitative autoradiography of mu-, delta- and kappa 1
opioid receptors in kappa-opioid receptor knockout mice.
Brain Res., 818: 335–345.
Stumm, R.K., Zhou, C., Schulz, S. and Hollt, V. (2004) Neuronal
types expressing mu- and delta-opioid receptor mRNA in the
rat hippocampal formation. J. Comp. Neurol., 469: 107–118.
Svingos, A.L., Colago, E.E.O. and Pickel, V.M. (1999) Cellular
sites for dynorphin activation of kappa-opioid receptors in
the rat nucleus accumbens shell. J. Neurosci., 19: 1804–1813.
Tanapat, P., Hastings, N.B., Reeves, A.J. and Gould, E. (1999)
Estrogen stimulates a transient increase in the number of new
neurons in the dentate gyrus of the adult female rat. J.
Neurosci., 19: 5792–5801.
Tempel, A. and Zukin, R.S. (1987) Neuroanatomical patterns
of the mu, delta, and kappa opioid receptors of rat brain as
determined by quantitative in vitro autoradiography. Proc.
Natl. Acad. Sci. U.S.A., 84: 4308–4312.
Terman, G.W., Drake, C.T., Simmons, M.L., Milner, T.A. and
Chavkin, C. (2000) Opioid modulation of recurrent excitation
in the hippocampal dentate gyrus. J. Neurosci., 20: 4379–4388.
Terman, G.W., Wagner, J.J. and Chavkin, C. (1994) Kappa
opioids inhibit induction of long-term potentiation in the
dentate gyrus of the guinea pig hippocampus. J. Neurosci.,
14: 4740–4747.
Thureson-Klein, A.K. and Klein, R.L. (1990) Exocytosis from
neuronal large dense-core vesicles. Int. Rev. Cytol., 121:
67–126.

Tielen, A.M., van Leeuwen, F.W. and Lopes da Silva, F.H.
(1982) The localization of leucine-enkephalin immunoreac-
tivity within the guinea pig hippocampus. Exp. Brain Res.,
48: 288–295.
Torres-Reveron, A., Khalid, S., Drake, C.T. and Milner, T.A.
(2006a) Enkephalin and dynorphin immunoreactivities in-
crease in the dorsal hippocampus in response to estrogens
and colocalize with estrogen receptor beta in mossy fiber
terminals. Program No. 724.6 2006. Neuroscience Meeting
Planner Atlanta GA: Society for Neuroscience, 2006. Online.
Torres-Reveron, A., Khalid, S., Drake, C.T. and Milner, T.A.
(2006b) Enkephalin-immunoreactivity colocalizes with estro-
gen receptor immunoreactivity in terminals and increases in
proestrus and estrogen replaced female rats in the dorsal
hippocampus. Program No. Tn-26 2006. Poster Session Ab-
stracts St. Paul Minnesota. International Narcotics Research
Conference.
Tortella, F.C., Echevarria, E., Robles, L., Mosberg, H.I. and
Holaday, J.W. (1988) Anticonvulsant effects of mu (DAGO)
and delta (DPDPE) enkephalins in rats. Peptides, 9:
1177–1181.
Turlejski, K. and Djavadian, R. (2002) Life-long stability of
neurons: a century of research on neurogenesis, neuronal
death and neuron quantification in adult CNS. Prog. Brain
Res., 136: 39–65.
Unterwald, E.M., Knapp, C. and Zukin, R.S. (1991) Neuro-
anatomical localization of kappa 1 and kappa 2 opioid
receptors in rat and guinea pig brain. Brain Res., 562:
57–65.
Velı́sek, L., Stanton, P.K., Moshé, S.L. and Vathy, I. (2000)
Prenatal morphine exposure enhances seizure susceptibility
but suppresses long-term potentiation in the limbic system of
adult male rats. Brain Res., 869: 186–193.
Verhage, M., McMahon, H.T., Ghijsen, W.E.J.M., Boomsma,
F., Scholten, G., Wiegant, V.M. and Nicholls, D.G. (1991)
Differential release of amino acids, neuropeptides, and cat-
echolamines from isolated nerve terminals. Neuron, 6:
517–524.
Vindrola, O., Briones, R., Asai, M. and Fernandez Guardiola,
A. (1981) Amygdaloid kindling enhances the enkephalin
content in rat brain. Neurosci. Lett., 21: 39–44.
Wagner, J.J., Caudle, R.M. and Chavkin, C. (1992) Kappa-
opioids decrease excitatory transmission in the dentate gyrus
of the guinea pig hippocampus. J. Neurosci., 12: 132–141.
Wagner, J.J., Caudle, R.M., Neumaier, J.F. and Chavkin, C.
(1990) Stimulation of endogenous opioid release displaces mu
receptor binding in rat hippocampus. Neuroscience, 37: 45–53.
Wagner, J.J. and Chavkin, C. (1992) Endogenous opioids and
LTP: another view. Neurochem. Int., 20: 457–460.
Wagner, J.J., Evans, C.J. and Chavkin, C. (1991) Focal stim-
ulation of the mossy fibers releases endogenous dynorphins
that bind kappa 1-opioid receptors in guinea pig hippocam-
pus. J. Neurochem., 57: 333–343.
Wagner, J.J., Terman, G.W. and Chavkin, C. (1993) Endog-
enous dynorphins inhibit excitatory neurotransmission and
block LTP induction in the hippocampus. Nature, 363:
451–454.
263
Weisskopf, M.G., Zalutsky, R.A. and Nicoll, R.A. (1993) The
opioid peptide dynorphin mediates heterosynaptic depression
of hippocampal mossy fibre synapses and modulates long-
term potentiation. Nature, 362: 423–427.
White, J.D., Gall, C.M. and McKelvy, J.F. (1986) Pro-
enkephalin is processed in a projection-specific manner in
the rat central nervous system. Proc. Natl. Acad. Sci. U.S.A.,
83: 7099–7103.
Wimpey, T.L. and Chavkin, C. (1991) Opioids activate both an
inward rectifier and a novel voltage-gated potassium con-
ductance in the hippocampal formation. Neuron, 6: 281–289.
Wolozin, B.L. and Pasternak, G.W. (1981) Classification of
multiple morphine and enkephalin binding sites in the central
nervous system. Proc. Natl. Acad. Sci. U.S.A., 78:
6181–6185.
Woolley, C.S. and Schwartzkroin, P.A. (1998) Hormonal
effects on the brain. Epilepsia, 39(Suppl 8): S2–S8.
Xie, C.W., Morrisett, R.A. and Lewis, D.V. (1992) Mu opioid
receptor-mediated modulation of synaptic currents in dentate
granule cells of rat hippocampus. J. Neurophysiol., 68:
1113–1120.
Yasuda, K., Raynor, H., Kong, C.D., Breder, J., Takeda, J.,
Reisine, T. and Bell, G.I. (1993) Cloning and functional
comparison of kappa and delta opioid receptors from mouse
brain. Proc. Natl. Acad. Sci. U.S.A., 90: 6736–6740.
Yoshioka, M., Matsumoto, M., Togashi, H., Smith, C.B. and
Saito, H. (1993) Opioid receptor regulation of 5-hydroxytry-
ptamine release from the rat hippocampus measured by in
vivo microdialysis. Brain Res., 613: 74–79.
Yuste, R. and Tank, D.W. (1996) Dendritic integration in
mammalian neurons, a century after Cajal. Neuron, 16:
701–716.
Zadina, J.E., Hackler, L., Ge, L.J. and Kastin, A.J. (1997) A
potent and selective endogenous agonist for the m-opiate re-
ceptor. Nature, 386: 499–502.
Zakarian, S. and Smyth, D. (1979) Distribution of active and
inactive forms of endorphins in rat pituitary and brain. Proc.
Natl. Acad. Sci. U.S.A., 76: 5972–5976.
Zakarian, S. and Smyth, D.G. (1982) Distribution of beta-
endorphin-related peptides in rat pituitary and brain. Bio-
chem. J., 202: 561–571.
Zieglgänsberger, W., French, E.D., Siggins, G.R. and Bloom,
F.E. (1979) Opioid peptides may excite hippocampal pyram-
idal neurons by inhibiting adjacent inhibitory interneurons.
Science, 205: 415–417.
Zimprich, A., Simon, T. and Hollt, V. (1995) Cloning and ex-
pression of an isoform of the rat mu opioid receptor
(rMOR1B) which differs in agonist induced desensitization
from rMOR1. FEBS Lett., 359: 142–146.
Zukin, R.S., Eghbali, M., Olive, D., Unterwald, E.M. and
Tempel, A. (1988) Characterization and visualization of rat
and guinea pig brain kappa opioid receptors: evidence for
kappa 1 and kappa 2 opioid receptors. Proc. Natl. Acad. Sci.
U.S.A., 85: 4061–4065. RU:1 do you mean KORs too? above
you mention KORs do opposite things( and see below com-
ment 8. RU:2 could you provide a reference - to a review -
because yours is the only opioids chapter in the book.
 SO CA1
p
SP
SR
SUB

SLM
TAT
PP

p HIL m
SLu g
CA3
DG

GCL
ML

ENK

DYN

Plate 15.1. Location of enkephalins and dynorphins in the hippocampal formation. Opioid peptides are found in all regions of the
hippocampal formation: the hippocampus proper (fields CA1 and CA3), the dentate gyrus (DG), and the subiculum (SUB). The
glutamatergic neurons [pyramidal cells (p), granule cells (g), mossy cells (m)] and their projections are shown. GABAergic interneurons
are scattered in all laminae. The perforant path (PP) from entorhinal cortex provides the major excitatory innervation. Subcortical
afferents and contralateral hippocampal projections enter via the fimbria/fornix (not shown) adjacent to CA3. Enkephalins (green) are
in some granule cells, particularly their mossy fiber terminals in hilus and stratum lucidum (SLu). Mossy fibers extensively innervate
hilar mossy cells and CA3 pyramidal cells. Enkephalins are in a portion of the projection from entorhinal cortex; the lateral portion of
the perforant path (PP) in the DG and the temporal-ammonic tract (TAT) in CA1. Enkephalins are in a few scattered interneurons,
which are relatively abundant on the border of stratum radiatum (SR) and stratum lacunosum-moleculare (SLM). Dynorphins (yellow)
are in the granule cells. Dynorphin peptides are most striking in the mossy fiber terminals, and are also in the mossy fiber axons,
granule cell somata, and granule cell dendrites. Abbreviations: SO, stratum oriens; SP, stratum pyramidale; GCL, granule cell layer.
(For B/W version, see page 248 in the volume.)
 +
Molecular ENK
Layer PP
_ GABA

?
?

+
SubP ACh

Granule Granule ?
Cell Cell
Layer

GABA
Hilus PARV

+
ENK/DYN
+

ACh GABA
NPY
SOM

MOR

to CA3 DOR

KOR
+

Plate 15.2. Subcellular locations of opioid peptides and receptors in the dentate gyrus. Enkephalins (green shading) are in: (1)
somatodendritic, axon, and terminal portions of granule cells; (2) lateral perforant path axons and terminals in the outer molecular
layer; and (3) occasional GABAergic interneurons (not shown). Mu-opioid receptors (MORs; blue) are most frequently in all portions
of parvalbumin (PARV)-containing basket cells that innervate granule cell somata and proximal dendrites. MORs are also in
cholinergic and GABAergic afferents from the lateral septum/diagonal band. Delta-opioid receptors (DORs; pink) are in many hilar
interneurons that contain somatostatin or neuropeptide Y (SOM/NPY) and inhibit the distal dendrites of granule cells. PARV-
containing neurons express DOR mRNA. DORs, and to a lesser extent MORs, are occasionally present in granule cell dendrites.
Kappa-opioid receptors (KORs) in guinea pigs are in substance P containing afferents to granule cells, in granule cell mossy fibers, and
most likely on perforant path terminals in the outer molecular layer. (For B/W version, see page 249 in the volume.)
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 16

Somatostatin in the dentate gyrus

Melanie K. Tallent

Department of Pharmacology and Physiology, Drexel University College of Medicine, 245 N. 15 St., Philadelphia,
PA 19102, USA

Abstract: The neuropeptide somatostatin (SST) is expressed in a discrete population of interneurons in the
dentate gyrus. These interneurons have their soma in the hilus and project to the outer molecular layer onto
dendrites of dentate granule cells, adjacent to perforant path input. SST-containing interneurons are very
sensitive to excitotoxicty, and thus are vulnerable to a variety of neurological diseases and insults, including
epilepsy, Alzheimer’s disease, traumatic brain injury, and ischemia. The SST gene contains a prototypical
cyclic AMP response element (CRE) site. Such a regulatory site confers activity-dependence to the gene,
such that it is turned on when neuronal activity is high. Thus SST expression is increased by pathological
conditions such as seizures and by natural stimulation such as environmental enrichment. SST may play
an important role in cognition by modulating the response of neurons to synaptic input. In the dentate,
SST and the related peptide cortistatin (CST) reduce the likelihood of generating long-term potentiation,
a cellular process involved in learning and memory. Thus these neuropeptides would increase the threshold
of input required for acquisition of new memories, increasing ‘‘signal to noise’’ to filter out irrelevant
environmental cues. The major mechanism through which SST inhibits LTP is likely through inhibition
of voltage-gated Ca2+ channels on dentate granule cell dendrites. Transgenic overexpression of CST in the
dentate leads to profound deficits in spatial learning and memory, validating its role in cognitive process-
ing. A reduction of synaptic potentiation by SST and CST in dentate may also contribute to the
well-characterized antiepileptic properties of these neuropeptides. Thus SST and CST are important
neuromodulators in the dentate gyrus, and disruption of this signaling system may have major impact on
hippocampal function.

Keywords: somatostatin; cortistatin; neuropeptide; electrophysiology; long-term potentiation; interneuron;

cognition; epilepsy

Introduction neurons or HIPP cells), however, its physiological

Somatostatin (SST) is a neuropeptide abundantly
expressed throughout the brain and periphery. In
the dentate gyrus, SST has been extensively used as
a chemical marker of a subset of interneurons lo-
cated in the hilus (hilar perforant path associated
Corresponding author. Tel.: +1 215 762 8208;
Fax: +1 215 762 2299; E-mail: tallent@drexel.edu
role here is relatively unknown. The selective tar-
geting of SST-containing interneuron terminals to
distal dendrites of principal neurons in both dent-
ate and cornu ammonis suggest that this peptide
may play a similar role throughout the hippocam-
pus. Interest in the role of SST in the dentate has
been heightened by studies showing it is upregu-
lated by mild seizures, but SST-containing inter-
neurons are vulnerable to severe seizures. What

DOI: 10.1016/S0079-6123(07)63016-7 265

266
role, if any, the loss of the SSTergic system plays in
postseizure excitability in the dentate is still under
investigation in our lab and others. Recent evi-
dence suggests that SST may be a critical regulator
of synaptic plasticity that underlies learning and
memory in the dentate (Baratta et al., 2002).
SST was first discovered in 1972 by Vale and col-
leagues as a hypothalamic inhibitor of growth hor-
mone release from the anterior pituitary (Brazeau
et al., 1972). In the periphery, SST generally acts as
an inhibitor of hormone release from the gut and
pancreas (Reisine, 1995). In the brain, SST also
generally has inhibitory actions at the cellular
level, but can increase neurotransmitter release in
some instances, likely through disinhibitory mech-
anisms [i.e., inhibition of inhibitory interneurons,
Chesselet and Reisine (1983)]. SST distribution in
the brain is extensive, with virtually every region
showing some localization at some point in devel-
opment. SST expression in forebrain is prominent,
containing perhaps the highest brain levels outside
the hypothalamus. More recently, a related
peptide from a distinct gene was discovered by
de Lecea and colleagues, cortistatin (CST). The
distribution of CST is more regionally limited
than SST, being largely restricted to cortex and
hippocampus (de Lecea et al., 1996).
SST activates at least three types of K+ chan-
nels, KIR, KM, and KLK (Moore et al., 1988;
Velimirovic et al., 1995; Schweitzer et al., 1998).
These actions of SST hyperpolarize neurons,
moving them away from their threshold for firing.
Activation of K+ channels also decreases input
resistance of neurons, lessening the depolarizing
actions of excitatory synaptic inputs. SST inhibits
voltage-sensitive Ca2+ channels in neurons as well
(Wang et al., 1990b; Boehm and Betz, 1997). This
action could also limit excitability, either through
postsynaptic decrease of Ca2+-induced depolari-
zations and/or signaling, or through presynaptic
inhibition of neurotransmitter release. However,
increasing evidence suggests that the major mech-
anism for inhibition of neurotransmitter release by
Gi/Go coupled receptors is by direct inhibition of
the synaptic protein SNAP-25 by bg subunits
(Blackmer et al., 2005; Gerachshenko et al., 2005).
SST inhibition of glutamate release in cultured
hippocampal neurons was shown to require Gai2,
Gai3, or Gao1 (Straiker et al., 2002). Since Ga
subunits show preference in coupling to distinct bg
subunits (Robishaw and Berlot, 2004), specificity
in Ga coupling likely confers specificity in bg
subunits that that interact with SNAP-25.
SST and CST mediate their actions via the same
family of receptors. Five SST receptors have been
cloned, SST1–SST5. These are prototypical Gi/Go
G-protein coupled receptors most closely related
to the opioid family of receptors. Only SST2 is
alternatively spliced, with two described variants,
SST2a and SST2b (Schindler et al., 1999). Expres-
sion of SST1–SST4 in brain has been confirmed;
SST5 expression in brain is still somewhat contro-
versial (Stroh et al., 1999; Schulz et al., 2000).
SST2, SST3, and SST4 are all expressed in hippo-
campus (Dournaud et al., 1996a; Handel et al.,
1999; Schreff et al., 2000; Schulz et al., 2000); SST1
expression in forebrain has not been validated
using antibodies that have had their specificity
confirmed in SST1 receptor knockout mice
(Hervieu and Emson, 1998; Schulz et al., 2000).
Interesting distinct subcellular localization of the
different SST receptor subtypes has been reported
on the principal neurons of hippocampus. SST2
expression is largely restricted to soma and prox-
imal dendrites (Dournaud et al., 1996a), SST4 to
medial and distal dendrites (Schreff et al., 2000),
and SST3 has a pattern of expression unique from
any other G protein-coupled receptor in that it is
exclusively expressed on neuronal cilia (Handel
et al., 1999). Thus, although these three receptors
are expressed in the same neurons, there is little
spatial overlap (Schreff et al., 2000). Excepting
axons (which may contain SST1, Hervieu and
Emson, 1998), the principal projecting neurons in
hippocampus therefore are virtually coated with
SST receptors, suggesting they are exquisitely
sensitive to this neuropeptide, and that SST is
an extremely important signaling molecule in this
region.
SST receptors activate multiple signaling systems
Studies in expression systems have shown SST to
couple to multiple signaling systems, including
adenylyl cyclase, protein kinase C, and MAP

kinases. Signaling of these receptors in brain
has been less studied, besides their regulation of
voltage-dependent ion channels. Most of the stud-
ies examining signaling pathways activated by
SST receptors have used recombinant receptors in
expression systems. Such studies have shown that
SST receptors can activate or inhibit a multitude
of effectors, most of which are typical for the
Gi/Go family of G proteins (Neves et al., 2002).
For example, activation of all cloned SST recep-
tors leads to the inhibition of adenylyl cyclase
(Moller et al., 2003). Activation of tyrosine phos-
phates and phospholipase C (PLC) has also been
reported for all of the cloned receptors (Lahlou
et al., 2004), although one study in which all
five human receptors were expressed in the same
cell line showed only SST3 and SST5 significantly
activated PLC and increased intracellular Ca2+
(Siehler and Hoyer, 1999). Activation of MAP
kinases has been observed for the majority of the
receptors (Florio et al., 1999), specifically ERK
(Lahlou et al., 2003), and p38 (Moller et al., 2003),
although inhibition of ERK has also been reported
(Todisco et al., 1994; Pola et al., 2003; Lahlou
et al., 2004). SST has also been shown to activate
PI3 kinase (Lahlou et al., 2003).
Several studies have addressed signaling mech-
anisms of endogenous SST receptors in cell lines
and in situ; interest has been raised because of the
finding that SST has antiproliferative actions on
cancer cells. Since activation of SST receptors in-
hibits adenylyl cyclase activity, inhibition of
cAMP-dependent signaling pathways would be a
likely mechanism for the antiproliferative effects of
SST. One such molecule is the cAMP response el-
ement binding protein (CREB), a transcription
factor that regulates protein expression by binding
to CRE sites in their promoter region. Indeed, the
SST gene itself has such a CRE site (as does SST2,
Kimura et al., 2001), suggesting a mechanism of
reciprocal regulation (Montminy and Bilezikjian,
1987). SST was shown to inhibit CREB phos-
phorylation in GH4 somatotrophic cells via inhi-
bition of PKA (Tentler et al., 1997). SST has been
shown to regulate other CREB-regulated signaling
molecules in several cell lines. In the mouse insu-
linoma cell line MIN6, that expresses only SST3,
SST transiently increases and then decreases the
267
CREB-regulated immediate early gene c-fos,
and MAP kinase activity, an activator of CREB
(Yoshitomi et al., 1997). This suggests SST recep-
tor activation can inhibit CREB independently of
adenylyl cyclase, although this was not directly
assessed in this study. Interestingly, although
SST inhibition of c-fos was blocked by pertussis
toxin (PTX), activation of c-fos expression was
PTX-insensitive, suggesting a non-Gi/Go mecha-
nism (Yoshitomi et al., 1997). In the pituitary
adenoma cell line GH3, SST inhibited c-fos
mRNA expression through a PTX-sensitive mech-
anism that appeared to involve inhibition of ERK
activity and Elk-1 phosphorylation (Todisco et al.,
1994). This same group also found SST inhibited
binding of the AP-1 transcription complex to its
target DNA, which is necessary for c-fos activity.
This action of SST was blocked by PTX and the
phosphatase inhibitors sodium orthovanadate
and okadaic acid (Todisco et al., 1994, 1995).
In the pancreatic cell line AR42J and in human
leukemia cell lines, SST also inhibited c-fos
expression in a PTX-sensitive fashion (Ishihara
et al., 1999; Cowles et al., 2002). Thus most studies
suggest that SST inhibits cAMP-mediated gene
expression, which is in keeping with its antiprolif-
erative actions and its Gi/Go coupling. However,
robust activation of c-fos expression by SST3 has
been reported (Yoshitomi et al., 1997). This action
may be mediated by activation of MAP kinases
via the small GTPase Ras (Florio et al., 1999;
Lahlou et al., 2003), which is a prototypical Gi-
mediated signaling pathway (Neves et al., 2002).
SST has also been found to stimulate adenylyl
cyclase in somatotrophic cells (Ramirez et al.,
2002), suggesting atypical G-protein coupling can
occur in some cell types.
It is interesting that SST receptors inhibit aden-
ylyl cyclase, leading to a decrease in cAMP levels
and decreased activity of PKA, while these same
receptors can activate the MAP kinase ERK.
Thus, although most studies suggest that SST in-
hibits CREB-mediated transcription, activation
has also been reported. These contrasting actions
may be receptor subtype and tissue specific,
although this issue has not been adequately
addressed in situ or in vivo. This complexity in
signaling is of course not unique to SST receptors,
268
and simply because SST receptors have been
shown that they can couple to a pathway in an
artificial system, it does not necessarily follow that
they do couple to these pathways to mediate their
important physiological effects.
Surprisingly few studies have addressed modu-
lation of signaling pathways by SST in the brain,
aside from its well-known actions on K+ and
Ca2+ channels (Mihara et al., 1987; Inoue et al.,
1988; Ikeda and Schofield, 1989; Wang et al.,
1990a, b; Meriney et al., 1994; Connor et al., 1997;
Sodickson and Bean, 1998). Inhibition of cAMP
by SST in membranes from several brain regions
has been observed, including cortex and hippo-
campus (Raynor and Reisine, 1992; Blake, 2001).
Inhibition of c-fos mRNA expression has been
reported in the trigeminal nucleus and arcuate
nucleus (Bereiter, 1997; Dickson et al., 1997;
Zheng et al., 1997), however, it is unclear whether
this action is through activation of signaling path-
ways or through inhibition of neuronal firing, a
widespread action of SST due to its augmentation
of K+ currents.
Decrease in SST levels are found in Alzheimer’s
brain
Davies et al. (1980) first reported reduced SST
levels in postmortem brains of Alzheimer’s disease
patients. Since then there have been dozens of
studies confirming and expanding on this initial
observation. SST-containing interneurons were
shown to be selectively associated with neuronal
tangles in Alzheimer’s brain, suggesting a mecha-
nism by which these neurons are vulnerable
(Roberts et al., 1985). However, studies in rats
suggest that lesioning cholinergic neurons also de-
creases SST neurons in cortex, suggesting the SST
reduction is downstream of the well-characterized
cholinergic deficits (Zhang et al., 1998). Other
studies have examined regional changes in SST
reductions, showing a loss of SST-containing neu-
rons in frontal and parahippocampal cortex and
hippocampus (Grouselle et al., 1998). A reduction
in SST mRNA levels per surviving neuron was
also observed in hippocampus but not cortex
(Dournaud et al., 1994). A strong negative
correlation was found between levels of SST in
cerebrospinal fluid (CSF) and degree of dementia
in patients with Alzheimer’s disease (Edvinsson et
al., 1993; Minthon et al., 1997). Thus a reduction
in the CNS SST system is consistently observed in
brains from Alzheimer’s patients; however, it is
currently unknown how or whether this loss con-
tributes to Alzheimer’s dementia, as the function
of SST in cognition is still unclear. Interestingly,
SST receptors appear to be largely preserved in
Alzheimer’s disease (Kumar, 2005), suggesting
them as potential targets for therapeutics. In
mouse models of Alzheimer’s, decreases in brain
SST levels have also been observed (Tomidokoro
et al., 2000; Ramos et al., 2006), suggesting they
could be appropriate models to study the role of
SST in this disease.
A novel role for SST with direct implications
on Alzheimer’s disease has recently been re-
ported. SST was shown to increase degradation
of b-amyloid by upregulating the activity of ne-
prilysin, an Ab degrading peptidase (Saito et al.,
2005). Interestingly, in SST knockout mice,
immunoreactive neprilysin was reduced by 78%
in the molecular layer of the dentate gyrus. A
corresponding 1.5-fold increase in Ab 42, the b
amyloid fragment most closely associated with
Alzheimer’s disease and amyloid plaques, was
found in brains of SST knockout mice (Saito
et al., 2005). Like SST, neprilysin decreases with
aging and in Alzheimer’s disease (Yasojima et al.,
2001; Iwata et al., 2002). An Alzheimer’s trans-
genic mouse model (APP23) crossed with a ne-
prilysin knockout mouse showed increased
oligomeric forms of Ab at synapses, and cognitive
impairment, including deficits in spatial learning
and object recognition memory at young ages
prior to accumulation of amyloid plaques (Huang
et al., 2006). Thus one functional consequence
of reduced SST levels in brain may be increased
accumulation of Ab, and upregulation of SST
could be protective.
Changes in CNS SST levels have been reported
in other diseases as well, in particular those asso-
ciated with cognitive dysfunction. Whether these
changes influence the disease state is unknown.
Reduced CSF SST levels were found in patients
with depression (Agren and Lundqvist, 1984) or

multiple sclerosis (Vecsei et al., 1990), although
regional changes in brain have not been explored
in these diseases. In bipolar disorder, transcription
analysis using microarrays found that SST was one
of eight genes that were significantly correlated
with the severity of the disease in postmortem
tissue. In this study, SST was upregulated, and
was the only gene to show significant association
with bipolarism when genetic polymorphisms were
analyzed (Nakatani et al., 2006). Increases in CSF
SST was also found in the related disorder, mania
(Sharma et al., 1995). Elevated SST levels are also
detected in basal ganglia of patients with Hunt-
ington’s disease (Aronin et al., 1983; Nemeroff
et al., 1983). In Parkinson’s disease, a reduction of
SST in hippocampus and cortex was found only in
patients showing cognitive deficits, suggesting that
changes occur in the later stages of this disease
(Epelbaum et al., 1983). Reduction in cortical SST
is also found in schizophrenia (Gabriel et al.,
1996); interestingly, in rat chronic haloperidol
treatment increases cortical SST levels (Sakai
et al., 1995).
SST is expressed in a subset of hilar interneurons in
dentate gyrus
In rat dentate gyrus, 16% of GAD-containing in-
terneurons express SST (Kosaka et al., 1988). This
neuropeptide has a distinct pattern of expression;
it is co-localized with gamma amino butyric acid
(GABA) in a specific subset of hilar neurons, the
HIPP cells that project to the outer molecular
layer. These neurons also have an axonal plexus
within the hilus, suggesting they can modulate
hilar neurons in addition to their projection to
the outer molecular layer, where they presumably
influence granule cells (Leranth et al., 1990; Lubke
et al., 1998). Some SST neurons have axons that
cross the hippocampal fissure into CA1, suggesting
interactions of SST terminals from the dentate
with SST terminals from the analogs population of
SST neurons in CA1, which terminate nearby in
the stratum lanunosum moleculare. The dendrites
of HIPP cells are spiny and appear to be restricted
to the hilus. They express mGluR1 and substance
P receptors (Freund and Buzsaki, 1996).
269
Approximately 30% of SST interneurons co-
express neuropeptide Y. This is a common theme
throughout the brain, and co-localization can be
as common among SST and neuropeptide Y cells
as 100% (see Chronwall et al., 1984; McDonald,
1989; Figueredo-Cardenas et al., 1996). The func-
tional implication of this co-expression has
not been examined in the dentate or elsewhere,
but since both of these neuropeptides have gener-
ally inhibitory actions, they are likely to act
synergistically when co-released.
In rat and primate, including humans, immuno-
staining reveals a dense plexus of SST immunore-
activity in the outer molecular layer, in agreement
with the projection pattern of HIPP cells (Amaral
et al., 1988; Milner and Bacon, 1989; Leranth
et al., 1990; Austin and Buckmaster, 2004;
Csaba et al., 2005). Similar findings have been
reported in other species, including pig (Holm
et al., 1992), guinea pig, and rabbit (Buckmaster
et al., 1994). In mice, however, this SST staining
profile is absent (Buckmaster et al., 1994; Jinno
and Kosaka, 2000). This does not appear to be due
to a difference in the projection pattern of HIPP
cells, since enhanced green fluorescent protein
(EGFP) staining is present in the outer molecular
layer in mice engineered to express this marker in
SST-containing neurons (see below Oliva et al.,
2000). Whether this difference in SST density in
the outer molecular layer reflects a functional
difference of SST between rats and mice is
unknown.
Electron microscopy has shown that HIPP cell
axons terminate on spines of dentate granule
cell (DGC) dendrites, forming symmetrical (inhib-
itory) synapses. Interestingly, asymmetrical (exci-
tatory) synapses are frequently observed on the
same spines, suggesting entorhinal afferents and
HIPP axons form synapses on the same spines
(Milner and Bacon, 1989; Leranth et al., 1990).
Thus, based on the anatomical data, one would
predict that HIPP interneurons would mediate
feedback inhibition of granule cells, with their
dendrites remaining in the hilus and receiving in-
put from mossy fibers, and their axons terminating
in the outer molecular layer adjacent to excitatory
perforant path input onto granule cell dendrites.
Definitive proof of this feedback role is lacking,
270
in fact, enhanced feedback inhibition has been
consistently observed after seizure-induced death
of HIPP neurons (Swanson et al., 1998; Sokal and
Large, 2001; Kobayashi and Buckmaster, 2003).
HIPP-like neurons have been characterized
electrophysiologically in rat (Scharfman, 1992;
Mott et al., 1997; Lubke et al., 1998), however,
identification in these studies were based on lim-
ited morphological analysis and there are some
discrepancies in the findings. An interesting model
to study HIPP neuron properties has been devel-
oped. A subset of interneurons was engineered
to express EGFP in a transgenic mouse, the
GIN mice (GFP-expressing inhibitory neurons),
using the GAD67 promoter (Oliva et al., 2000).
SST-containing neurons account for 90–98% of
the EGFP expressing cells in hippocampus and
cortex. In the hilus, 95% of the EGFP expressing
neurons were found to contain SST, although only
16% of the SST-containing neurons expressed
EGFP (Oliva et al., 2000). Anatomically, the hilar
EGFP interneurons were similar to the HIPP cells
described by Freund and Buszaki (1996) in rat,
with terminals in the outer two-thirds of the
molecular layer. These neurons also were positive
for mGluR1, also a characteristic of HIPP cells
(Freund and Buzsaki, 1996). These mice have been
used to characterize electrophysiological proper-
ties of SST interneurons in sensorimotor cortex,
which revealed distinct subpopulations of these
neurons (Halabisky et al., 2006). These mice pro-
vide an excellent model system for a rigorous study
of electrophysiological and network properties of
SST-containing HIPP cells in mice.
SST expression in the hippocampus is regulated
during development
Expression of SST in hippocampus changes during
development. A surge in SST immunoreactivity
occurs between days E12 and E15, with expression
beginning in the subiculum and progressing ‘‘back-
ward’’ though the hippocampus, so that hilar ex-
pression appears last (Rapp and Amaral, 1988).
After maturity, a decline in SST levels occurs
throughout rat hippocampus, with hilar expres-
sion showing the largest decrease in aged animals
(Vela et al., 2003). In aged primates, a dramatic
decline in hippocampal SST mRNA is also
observed, although no regional analysis was
performed in this study (Hayashi et al., 1997).
Interestingly, a significant negative correlation
has been found in cognitive decline in aged rats
and SST levels (Dournaud et al., 1996b). That a
similar negative relationship has been observed
between SST levels and dementia in both Parkin-
son’s (Epelbaum et al., 1983) and Alzheimer’s
disease in humans (Davies et al., 1980) suggests an
intriguing potential role for SST in cognition.
However, in most cases it is unclear whether this
decrease in SST expression is due to changes in
SST gene or transcript regulation, or due to the
death of SST-containing interneurons (see below).
If the latter is true, the relationship between SST
levels and cognitive decline becomes more indirect.
CST expression in the dentate gyrus is postna-
tally regulated as well, with a pattern of expression
that is distinct from SST. At P0 in rat,
CST mRNA is apparent in hippocampal stratum
oriens but not the dentate. By P10, expression
throughout the hippocampus is increasing, and
SST-positive neurons appear in the hilus. Expres-
sion of CST mRNA in hilar neurons peaks at
approximately postnatal day 16 and then declines,
with very few labeled neurons present in adult
hilus (de Lecea et al., 1997a). Whether CST has
a specific function during development is un-
known, although developmental differences in its
electrophysiological actions have been reported
(de Lecea et al., 1997a). Interestingly, in aged mice
(24 months), CST levels have increased in hilar
interneurons (Winsky-Sommerer et al., 2004).
SST expression is activity-dependent
The SST gene promoter has a prototypical CRE
site that has been studied extensively to under-
stand mechanisms of action of CREB (Montminy
and Bilezikjian, 1987; Powers et al., 1989; Mont-
miny et al., 1996). The transcription factor CREB
is a major transducer of activity-dependent gene
expression and thus mediates transformation of
electrical to genetic signaling. CREB mediates
its actions by interacting with CRE sites on

promoters. Thus, SST mRNA and ultimately pep-
tide levels are upregulated by neuronal activity.
SST in hilar interneurons appears to be especially
sensitive to upregulation by seizure-like activity.
This could be due to dense innervation of these
neurons by mossy fibers, the axons of DGCs, be-
cause granule cells and other principal neurons
of hippocampus are active during many types of
seizures, particularly limbic seizures. Studies using
c-fos, a downstream target of CREB, indicate that
hilar interneurons are ‘‘turned on’’ transcription-
ally shortly after DGCs and before CA3 or CA1
pyramidal neurons (Le Gal La Salle, 1988; Peng
and Houser, 2005).
Thus, it is not surprising that the intense
neuronal activation underlying seizures results in
upregulation of SST expression. Hippocampal
kindling upregulates hilar SST mRNA and pep-
tide once stage II seizures (mild seizures) are
reached (Bendotti et al., 1993; Vezzani et al.,
1996). This effect was transient, lasting
1 week.
Repeated, but not single electroconvulsive shock
(ECS)-induced seizures increases SST mRNA in
the hilus (Passarelli and Orzi, 1993; Mikkelsen and
Woldbye, 2006). This regulation is specific to dent-
ate, as no significant changes were detected in CA3
and CA1 (Passarelli and Orzi, 1993). Another
study showed repeated ECS increased the density
of SST immunoreactivity in the outer molecular
layer as well as in hilar cell bodies. This effect was
additive over 36 days of repeated ECS. Interest-
ingly, lesioning the perforant path input to the
dentate did not alter ECS-induced upregulation of
SST (Kragh et al., 1994). Kainate-induced seizures
also induce expression of SST mRNA in the hilus,
and elsewhere in hippocampus, beginning as early
as 3 h after seizures (Hashimoto and Obata, 1991).
Ectopic expression in DGCs and pyramidal cells
has also been reported after kainate-induced sei-
zures, suggesting SST expression can be activated
in neurons that normally do not express the ne-
uropeptide (Hashimoto and Obata, 1991; Calbet
et al., 1999). Expression of CST, with a promoter
distinct from SST, is not induced by kainate,
demonstrating that these two related peptides are
independently regulated (Calbet et al., 1999).
Considering that neuropeptides are released
when peptidergic neurons are activated at high
271
frequencies, it seems logical that during seizures,
SST release would be robust. This has proven to be
the case. In vivo voltammetry showed detectable
SST release in hippocampus 48 min following a
seizure, with increased release continuing for the
3 h recording period after the seizure (Manfridi
et al., 1991). In hippocampal slices, both baseline
and K+-induced SST release was significantly
higher in kindled rats than in control rats (Vezzani
et al., 1992). An in vivo study showed similar
findings using microdialysis, that is, an increase
in both basal and K+-stimulated release in
hippocampus after kindling (Marti et al., 2000).
That SST release is enhanced after kindling
suggests that upregulation of SST after seizures
has functional implications.
Interestingly, SST has antiepileptic properties in
hippocampus. This was first shown using in vivo
models in rats, where SST was injected intrahip-
pocampally or ICV. For example, Vezzani and
colleagues demonstrated that SST or SST2 selec-
tive ligands could decrease the severity of seizures
evoked by pentylenetetrazol (Perez et al., 1995),
kainate (Vezzani et al., 2000), or quinolic acid
(Vezzani et al., 1991). This group also showed that
continuous infusion of anti-SST antibody into the
hippocampus decreased the latency, and dramat-
ically reduced the number of kindling stimulations
required to reach stage V seizures, suggesting
endogenous SST plays an important role in main-
taining a seizure-free state. This was confirmed by
a different group, who also demonstrated that
anti-SST antibody injected into the hippocampus
blocked the protective effects of SST, administered
ICV, on picrotoxin-kindled rats (Mazarati and
Telegdy, 1992). This study supports the hypothesis
that hippocampal SST is critical in controlling
epileptogenesis and seizures. We used an in vitro
model system to examine whether SST could
suppress epileptiform activity in the rat hippo-
campal slice (Tallent and Siggins, 1999). We found
a robust effect of SST to decrease epileptiform
events generated either by increasing excitability
(by removing extracellular Mg++) or by decreas-
ing inhibition (by blocking GABAA receptors).
Although this study focused on epileptiform
events in CA1 and CA3, our later studies in
mouse hippocampal slices showed that SST has
272
robust effects on synaptic potentiation in dentate
gyrus (Baratta et al., 2002, discussed below), a
mechanism that likely contributes to epileptogen-
esis in this region (Sutula and Steward, 1986, 1987;
Wasterlain et al., 1999). Thus, SST has potential
sites of action throughout the hippocampus to
control seizure activity.
SST expression in the dentate is also regulated
by pathological conditions that do not involve
seizures. For example, SST expression increases
72 h following traumatic brain injury (Cook et al.,
1998). Interestingly, this upregulation of endog-
enous SST could be protective, since ICV admin-
istration of SST or CST protected against
ischemia-induced neuronal death (Rauca et al.,
1999). Stress also increases expression and release
of SST in the hilus (Arancibia et al., 2001).
Expression of SST receptors in dentate gyrus
Prior to the cloning of the SST receptors, autora-
diographic studies showed dense, high-affinity
binding of radiolabeled SST analogs throughout
the molecular layer, with much less binding in the
granule cell layer and diffuse binding in the hilus
(Leroux et al., 1993). These data suggested that
SST released from terminals could bind to recep-
tors throughout the dendritic layer, if it could
diffuse far enough without proteolysis. Character-
ization of mRNA expression of the different re-
ceptor subtypes extended these earlier studies by
providing evidence for SST1–SST4 in DGCs
(Bruno et al., 1992; Kong et al., 1994).
Immunohistochemical detection has confirmed
the presence of SST2, SST3, and SST4 on DGCs
(Dournaud et al., 1996a; Handel et al., 1999;
Schreff et al., 2000). As described above for all
hippocampal principal neurons, SST2 is expressed
predominately on DGCs soma and proximal
dendrites, SST3 on neuronal cilia, and SST4 is
almost exclusively dendritic.
Developmental changes in SST receptor mRNA
expression have been reported (Thoss et al., 1995,
1996). Only SST2 and SST3 mRNA are present in
significant amounts at E18, suggesting these two
receptors could play a role in development. Robust
high-affinity binding of SST ligands is present in
dentate at E18, validating the presence of func-
tional SST receptors (Thoss et al., 1996). At P5,
SST2 and SST3 expression levels remain flat,
while SST4 significantly increases, so that mRNA
for SST2, SST3, and SST4 are expressed at similar
levels. Likewise, an increase in labeling of SST
receptors using radioligands is apparent in dentate
from E18 to P5 (Thoss et al., 1996). From P5 to
adult (no intermediate ages have been reported),
SST1 levels have increased so that it is significantly
expressed for the first time, SST2 levels modestly
increase, SST4 levels modestly decrease, while
SST3 levels robustly increase (Thoss et al., 1996).
Regulation of SST receptor subtypes during aging
has not been reported, although one study showed
expression of SST1–SST5 in cortex of aged humans
(Kumar, 2005).
Plasticity in expression of SST receptors
Activity-dependent changes in expression of SST
receptors in the dentate following seizures have
been observed, but do not appear to involve
changes in gene expression. After kainate-induced
seizures, no change was observed in mRNA for
SST1–SST4 in dentate gyrus, although a decrease
for SST3 and SST4 mRNA was found in CA1,
corresponding to the death of pyramidal cells in
this region (Perez et al., 1995). Kindling also did
not cause changes in mRNA for SST receptors
in the dentate, although a 40% reduction in SST
receptor binding was observed in the molecular
layer immediately, and 1 week following, kindling-
induced seizures. Binding returned to normal
30 days following seizures. Decreased SST bind-
ing without changes in mRNA expression could
have been due to receptor downregulation as a
consequence of increased release of SST, although
it is unclear why reductions in binding were still
apparent 1 week following seizures unless recep-
tor-containing neurons were damaged or were lost
(Piwko et al., 1996). Kindling also reduced
immunoreactivity for SST2 in the outer molecular
layer; this was associated with an increase in SST
immunoreactivity, again suggesting downregula-
tion on binding of the receptor to its ligand (Csaba
et al., 2004). In dentate gyrus from humans with

epilepsy, SST2 receptor is strongly regulated
(Csaba et al., 2005). In hippocampal tissue re-
sected from patients with temporal lobe seizures,
SST2 expression was decreased in CA1 and CA3,
likely reflecting the neuronal loss found in this re-
gions in patients exhibiting hippocampal sclerosis
(De Lanerolle et al., 2003). In contrast, in DGCs,
which are more resistant to seizure-induced death,
SST2 mRNA was strongly upregulated. However,
SST2 receptor binding and immunoreactivity
showed a strong decrease in the outer molecular
layer, suggesting SST2 may be redistributed.
Hilar SST interneurons are vulnerable to
excitotoxicity
A hallmark of epileptic hippocampus is the death
of the SST-containing hilar interneurons. This was
first shown by Sloviter (1987) in a rat model after
induction of seizures by perforant path stimula-
tion. In hippocampal tissue surgically removed
from humans with intractable temporal lobe
seizures, a similar pattern of loss was seen
(De Lanerolle et al., 1989; Robbins et al., 1991).
Of all subsets of GABAergic neurons, this class of
interneurons appears to be selectively vulnerable,
since other interneuronal markers, such as chole-
cystokinin and vasoactive intestinal peptide,
appear unchanged (Robbins et al., 1991). This
vulnerability of SST interneurons to seizure-
induced death has since been confirmed in other
animal models, including kainate- and pilocarpine-
induced seizures (Buckmaster and Dudek, 1997;
Kobayashi and Buckmaster, 2003). Interestingly,
in electrically kindled rats, a model in which spon-
taneous seizures do not generally develop, most
studies show that the hilar SST neurons are spared
(Bendotti et al., 1993; Schwarzer et al., 1996;
Vezzani et al., 1996; Simonato et al., 1998). These
studies suggest that a functional consequence of
this interneuronal loss is the development of spon-
taneous seizures, and that the hilar SST neurons or
SST itself is important in preventing epileptogen-
esis. However, rate of epileptogenesis is not strictly
correlated with degree of hilar neuron loss (e.g.,
Lahteinen et al., 2003), suggesting other factors are
likely to contribute.
273
Interestingly, changes in the distribution of SST-
immunoreactive axons are also observed in the
dentate of hippocampal tissue resected from
patients with temporal lobe epilepsy (Mathern
et al., 1995; Csaba et al., 2005). In control post-
mortem tissue or in hippocampus of patients
without hippocampal sclerosis, a fine, diffuse
network of SST-containing processes was
observed in the outer two-thirds of the molecular
layer. However, in patients with hippocampal
sclerosis, accompanying the dramatic loss of
SST-containing hilar neurons was a change in
the SST-immunoreactive fibers in the dentate.
SST-containing axons in the molecular layer were
more numerous, thicker, longer, and more
strongly immunoreactive for SST. Large terminal
varicosities were also present. Similar, though less
dramatic changes were observed in fibers in the
hilus (Csaba et al., 2005). These results suggest
changes in the remaining SST-containing neurons
that are consistent with axon sprouting (Mathern
et al., 1995).
Thus, SST interneurons are activated by
seizures, and SST expression initially increases.
This is followed by death of the SST neurons,
presumably via excitotoxic mechanisms. Interest-
ingly, this group of neurons is selectively vulner-
able to other insults as well. Ischemia, for example,
also causes a reduction in SST-containing hilar
neurons (Johansen et al., 1987; Bering et al., 1997),
as does traumatic brain injury (Lowenstein et al.,
1992). Both ethanol ingestion and withdrawal
induces the death of these neurons (Andrade
et al., 1992). Neurological diseases in which a
reduction in hilar SST has been found to include
Alzheimer’s disease (Chan-Palay, 1987) and an
animal model of HIV dementia (Mitchell et al.,
1999).
What causes this subset of interneurons to be
selectively vulnerable to multiple types of insults?
Two major theories have been advanced; that
these interneurons are ‘‘overstimulated’’ during
hyperexcitability because they receive input from
both mossy fibers and perforant path, or that
they lack appropriate Ca2+ binding proteins so
that they are vulnerable to Ca2+-induced excito-
toxicity. The first hypothesis assumes that HIPP
interneurons have a unique innervation pattern
274
compared with other, less vulnerable, hilar inter-
neurons. However, the majority of anatomical
data suggests that their main excitatory input
comes exclusively from mossy fibers of DGCs (see
above). In fact, when activity-dependent c-fos
activation is measured following seizures, HIPP
neurons are activated after DGCs (Le Gal La
Salle, 1988; Peng and Houser, 2005), supporting
the anatomical data.
As for Ca2+ binding proteins, SST containing
interneurons in the hilus do not contain parval-
bumin, calretinin, or calbindin, although most
CA1 and CA3 SST interneurons contain one of
these Ca2+ binding proteins (Bouilleret et al.,
2000). Thus, it is an appealing hypothesis that the
underlying vulnerability of hilar SST interneurons
vs. SST interneurons in the rest of the hippocam-
pus is due to their lack of Ca2+ binding proteins.
Indeed, a study using Ca2+ binding protein
knockout mice suggests that this may be a con-
tributing factor. This study found that in wild
type mice, intrahippocampal injection of kainate
resulted in a reduction of SST containing inter-
neurons in CA1 to 43 and 29% of control 1 and
30 days after injection, respectively. However, a
dramatic decrease in SST-containing interneurons
after kainate injection was observed in mice with
both parvalbumin and calbindin knocked out,
but not with a knockout of parvalbumin alone.
In the double knockout, only 12 and 6% of the
CA1 SST-containing neurons remained 1 and
30 days after kainate injection, respectively. In
mice with both parvalbumin and calretinin
knocked out, a larger decrease in SST immuno-
reactivity in CA1 was observed 30 days but not 1
day following kainate injections (Bouilleret et al.,
2000). Although changes in network excitability
cannot be ruled out (although no differences in
behavioral seizures was observed between the
groups), this study suggests that knockout of cal-
bindin in particular leads to increased suscepti-
bility of SST interneurons to seizure-induced
death. Thus, the lack of any of these Ca2+ bind-
ing proteins in hilar SST interneurons may con-
tribute to their vulnerability to seizures and other
insults.
Electrophysiological effects of SST in dentate gyrus
On the basis of projection profile of SST-contain-
ing HIPP cells, in that their major target is the
distal dendrites of DGCs, and that SST-containing
terminals are often found adjacent to presumed
lateral perforant path synapses, it seems likely that
a major role for SST would be to regulate perfo-
rant path input onto DGCs. Unlike CA1, and to a
lesser extent CA3, the electrophysiological effects
of SST in the dentate gyrus have not been well
characterized, especially in rat. In rat CA1 pyram-
idal neurons, the major effects of SST appear to be
postsynaptic. SST and CST robustly enhance the
M-current (Moore et al., 1988; Schweitzer et al.,
1990; de Lecea et al., 1996), a subthreshold non-
inactivating K+ current that is important in set-
ting the excitatory tone of neurons. SST has sim-
ilar actions on the M-current in mouse CA1
pyramidal neurons (Qiu and Tallent, unpublished
observations). SST and CST also augment another
K+ current in these neurons, a voltage-independ-
ent leak current (Schweitzer et al., 1998, 2003). By
acting on these two K+ currents, application of
SST causes a several millivolt hyperpolarization of
the resting membrane potential and a decrease in
input resistance. Thus, incoming excitatory synap-
tic input would be less likely to trigger action po-
tentials. SST also has presynaptic actions in rat
CA1 pyramidal neurons, inhibiting EPSCs but not
IPSCs (Tallent and Siggins, 1997). Interestingly,
we have observed no presynaptic effect of SST in
mouse CA1 pyramidal neurons, even though in
side-by-side studies SST reduced EPSC amplitudes
in rat CA1 pyramidal neurons (Qiu and Tallent,
unpublished observations). In rat CA3 pyramidal
neurons, SST also causes membrane hyperpolari-
zation and reduces associational/commissural
EPSCs presynaptically (Tallent and Siggins,
1999). In cultured immature rat hippocampal py-
ramidal neurons, SST presynaptically inhibits
glutamate but not GABA release, and inhibits
Ca2+ currents (Boehm and Betz, 1997).
In vivo studies by de Lecea and colleagues (de
Lecea et al., 1997a) have shown that CST de-
presses population spike amplitude in DGCs in

immature (P15) but not adult rats. Interestingly,
CST is significantly expressed in dentate gyrus (in
hilar interneurons) only at early stages of devel-
opment but not in adults (see above; de Lecea et
al., 1997a). These results suggest developmental
changes in some SST receptor that mediates this
CST effect. The developmental regulation of SST
receptors in the dentate gyrus has not been exten-
sively examined, although only SST4 mRNA ex-
pression is higher in immature than adult dentate
gyrus (Thoss et al., 1995).
The only detailed characterization of the elect-
rophysiological effects of SST in dentate was done
by our group in adult mouse (Baratta et al., 2002).
In mouse DGCs, no action of SST on postsynaptic
K+ currents was detected, and the peptide did not
alter membrane potential. SST also did not affect
the firing properties of DGCs. We also examined
whether SST modulated excitatory synaptic input
onto DGCs. We found that SST did not affect
field EPSPs recorded in the outer molecular
layer and evoked by low frequency stimulation
the lateral perforant path (entorhinal cortical
input). However, in spite of these negative find-
ings on pre- and postsynaptic actions, SST
robustly depressed long-term potentiation (LTP)
at lateral perforant path synapses. Thus, although
not acting on low-frequency input, SST depressed
synaptic plasticity generated by high frequency
input. This finding corroborates the speculated
role of SST based on the close association between
SST-containing terminals and lateral perforant
path input. We further examined the mechanism
by which SST inhibited LTP. Since LTP at this
synapse is n-methyl-D-aspartate (NMDA)-receptor
dependent, we determined whether SST could re-
duce pharmacologically isolated NMDA receptor-
mediated synaptic responses. SST did not act on
NMDA receptor-mediated EPSCs recorded in
voltage-clamp, but modestly depressed NMDA
receptor-mediated field EPSPs recorded extra-
cellularly. Interestingly, the SST effect was blocked
when recordings were made in the presence of o-
conotoxin, a specific blocker of N-type Ca2+
channels. These results indicated that SST does
not directly inhibit NMDA receptors but acts on
275
Ca2+ channels activated during the NMDA-
receptor mediated depolarization. We next re-
corded Ca2+ spikes generated in DGCs by
depolarization. SST inhibited the Ca2+ spikes;
this inhibition remained in the presence of the
L-type Ca2+ blocker nifedipine and the T-type
Ca2+ channel blocker nickel. However, SST did
not inhibit Ca2+ spikes generated in the presence
of o-contotoxin, indicating that the peptide
was acting specifically on N-type Ca2+ channels.
In the presence of o-conotoxin, no LTP could be
generated at lateral perforant path synapses,
suggesting this mechanism alone could account
for SST inhibition of LTP (Baratta et al., 2002).
We have also demonstrated that transgenic
overexpression of CST prevents generation of
LTP at lateral perforant path synapses (Tallent
et al., 2005a). In adult wild type mice, CST is not
expressed in dentate (de Lecea et al., 1997a, b). We
generated a CST overexpressing transgenic mouse
in which the transgene was driven by the neuron-
specific enolase promoter. CST transgene expres-
sion was detected in the hippocampus, cortex, and
reticular thalamus, with especially high expression
in the dentate. These mice had normal baseline
synaptic transmission at lateral perforant path
synapses, with the exception of reduced paired-
pulse potentiation at interstimulus intervals less
than 100 ms. However, no LTP could be generated
at these synapses, both in vivo and in vitro (Fig. 1).
CA1 LTP in the mice was not significantly differ-
ent from wild type 60 min following induction.
A correlated deficit in spatial learning was present
in the CST transgenic mice (Fig. 2), suggesting
synaptic plasticity in dentate gyrus may be critical
for this type of learning, as previous studies had
suggested (Kauer et al., 1988; Moser et al., 1998;
Nakao et al., 2002).
These results and our results with exogenous
SST (Baratta et al., 2002) indicate SST receptors
may mediate important regulatory control over
synaptic plasticity in dentate gyrus. It is interesting
that SST appears to have a somewhat different
physiological role in dentate gyrus than in
CA1 and CA3 hippocampus. In cornu ammonis,
high frequency activation of somatostatinergic
276

A 260 B
Control CST Transgenic 220
Baseline Post-HFT
220 0.25 mV
% Control fEPSP slope

5ms
180

180

140
140

100
100

CST
60 60
-10 0 10 20 30 40 50 60 -10 0 10 20 30 40 50 60
Time (min) Time (min)

C
260
Baseline CST/Post-HFT

220

180

140

CST
100

60
-10 0 10 20 30 40 50 60
Time (min)

Fig. 1. Transgenic overexpression or exogenous application of CST prevents induction of LTP at lateral perforant path/dentate
granule cell synapses. (A) Mean data from control and CST transgenic slices. Initial slope is plotted over time 10 min prior to 60 min
following high frequency trains (double arrows). Reduced short-term potentiation and no long-term potentiation is seen in CST
transgenics (open circles) relative to controls (filled triangles). Inset: Representative fEPSPs from a control and a transgenic mouse
before (gray) and 60 min following two 1 s 100 Hz trains (black). No potentiation of the fEPSP is seen in slices from dentate of the CST
transgenic mouse. Scale bar labels are the same for all subsequent traces. (B) Exogenous CST also reduces LTP. CST (1 mM; open
circles) was superfused beginning 7 min prior to HFTs for a total of 8 min (black bar). A small decrease in the fEPSP slope was
observed prior to HFTs (maximal inhibition was 1174% 7 min following the beginning of the superfusion). No significant potent-
iation is present 60 min following the trains when CST was superfused during the induction protocol. Inset: Representative fEPSPs
before and 60 min following HFTs with CST superfusion. (C) Exogenous CST does not affect LTP maintenance. 1 mM CST was
applied from 30 to 45 min (black bar) following HFTs (open circles). No significant effect on magnitude of LTP was observed. Inset:
Representative baseline and 60 min post-train fEPSP from experiment where CST was applied 30–45 min following train. (Reprinted
from Tallent et al., 2005a, copyright 2005, with permission from Elsevier.)

Fig. 2. CST transgenic mice show a deficit in spatial learning in
the Barnes platform maze. In this mouse strain, overexpression
of CST was largely restricted to dentate and reticular thalamus.
The percent of mice in each group of males (top) and females
(bottom) meeting the criterion for learning in this task (three
errors or less on seven of eight consecutive trials) across the 40
days of testing. Both male and female CST transgenic mice
showed impairment in this task as determined by significantly
fewer mice in these groups learning this task. (Reprinted from
Tallent et al., 2005a, copyright 2005, with permission from
Elsevier.)
interneurons would release SST to hyperpolarize
pyramidal neurons, decreasing their likelihood of
firing during intense activation. Presynaptic inhi-
bition of glutamate release by SST (in rat but not
mouse) would act synergistically to inhibit high
frequency activity. Thus, SST has robust effects on
epileptiform activity in both CA1 and CA3 hip-
pocampus, in both rats (Tallent and Siggins, 1999)
and mice (Qiu et al., 2005). Interestingly, though,
SST and CST has less robust actions on synaptic
plasticity in CA1, reducing but not blocking LTP
(Qiu et al., 2003; Tallent et al., 2005b). In dentate
gyrus, SST has no observable actions on mem-
brane potential or firing rate, thus this peptide
would not regulate responsivity of DGCs to nor-
mal, low-frequency synaptic input. Perhaps this
regulatory mechanism, important in pyramidal
277
neurons, would be less effective mechanism in
DGCs, with their already quite hyperpolarized
membrane potential (
78 to 80 mV). Certainly
regulation of the M-current in particular would be
an ineffective mechanism to control output of
granule cells, since, unlike in CA1, this current
would not be active at the resting membrane
potential of DGCs (in fact we do not observe this
current in granule cells, even though immunohisto-
chemical studies suggest its component subunits
are present: Cooper et al., 2000; Yus-Najera et al.,
2003). We also did not observe presynaptic effects
of SST or CST in mouse dentate gyrus (Baratta
et al., 2002; Tallent et al., 2005b). Thus in spite of
the rather dense localization of SST receptors on
granule cell dendrites, they do not appear to be
functionally important during standard synaptic
transmission even when SST or CST is exogenou-
sly applied. However, the effect of these peptides
on synaptic plasticity in dentate gyrus is profound.
No LTP is generated at lateral perforant path
synapses when nanomolar concentrations of SST
is applied prior to induction. Likewise, transgenic
overexpression of CST prevents generation of any
perforant path LTP both in vivo and in vitro
(Tallent et al., 2005b). Therefore, in spite of the
similar projection pattern of somatostatinergic in-
terneurons to distal dendrites of principle neurons,
and of the similar localization of SST receptors on
principle cell soma and dendrites in both regions,
SST may have different physiological roles in
the cornu ammonis vs. dentate gyrus. In dentate,
release of SST during high frequency events would
act to increase the threshold for inducing LTP.
This mechanism may be important in increasing
the signal to noise properties of synaptic input into
the dentate, as well as in preventing invasion of
seizures into the hippocampus.
Conclusions
SST in the dentate gyrus is expressed primarily
in a subset of interneurons of the hilus which co-
express GABA and project to the dendritic layer of
DGCs. This places them in a position to modulate
granule cell activity, and the striking plasticity of
these neurons, as well as their vulnerability,
278
suggests that they could play an important role in
normal and pathological conditions. Indeed, SST
levels in different brain regions are altered in many
diseases associated with cognitive deficits. It is
tempting to speculate that these peptides are im-
portant in regulation of synaptic changes that un-
derlie learning and memory. Although our studies
in CST overexpressing mice support this hypoth-
esis, further characterization of the role of SST
and CST in cognitive processing is needed.
List of abbreviations
CREB cAMP response element binding
protein
CSF cerebrospinal fluid
CST cortistatin
DGC dentate granule cell
EGFP enhanced green fluorescent pro-
tein
GABA gamma amino butyric acid
HIPP hilar perforant path associated
LTP long-term potentiation
NMDA n-methyl-D-aspartate
PTX pertussis toxin
SST somatostatin
References
Agren, H. and Lundqvist, G. (1984) Low levels of somatostatin
in human CSF mark depressive episodes. Psycho-
neuroendocrinology, 9(3): 233–248.
Amaral, D.G., Insausti, R. and Campbell, M.J. (1988) Distri-
bution of somatostatin immunoreactivity in the human dent-
ate gyrus. J. Neurosci., 8(9): 3306–3316.
Andrade, J.P., Fernando, P.M., Madeira, M.D., Paula-Barb-
osa, M.M., Cadete-Leite, A. and Zimmer, J. (1992) Effects of
chronic alcohol consumption and withdrawal on the som-
atostatin-immunoreactive neurons of the rat hippocampal
dentate hilus. Hippocampus, 2(1): 65–71.
Arancibia, S., Payet, O., Givalois, L. and Tapia-Arancibia, L.
(2001) Acute stress and dexamethasone rapidly increase hip-
pocampal somatostatin synthesis and release from the dent-
ate gyrus hilus. Hippocampus, 11(4): 469–477.
Aronin, N., Cooper, P.E., Lorenz, L.J., Bird, E.D., Sagar,
S.M., Leeman, S.E. and Martin, J.B. (1983) Somatostatin is
increased in the basal ganglia in Huntington disease. Ann.
Neurol., 13(5): 519–526.
Austin, J.E. and Buckmaster, P.S. (2004) Recurrent excitation
of granule cells with basal dendrites and low interneuron
density and inhibitory postsynaptic current frequency in the
dentate gyrus of macaque monkeys. J. Comp. Neurol.,
476(3): 205–218.
Baratta, M.V., Lamp, T. and Tallent, M.K. (2002) Som-
atostatin depresses long-term potentiation and Ca2+ signa-
ling in mouse dentate gyrus. J. Neurophysiol., 88(6):
3078–3086.
Bendotti, C., Vezzani, A., Tarizzo, G. and Samanin, R. (1993)
Increased expression of GAP-43, somatostatin and neuro-
peptide Y mRNA in the hippocampus during development of
hippocampal kindling in rats. Eur. J. Neurosci., 5(10):
1312–1320.
Bereiter, D.A. (1997) Morphine and somatostatin analogue
reduce c-fos expression in trigeminal subnucleus caudalis
produced by corneal stimulation in the rat. Neuroscience,
77(3): 863–874.
Bering, R., Draguhn, A., Diemer, N.H. and Johansen, F.F.
(1997) Ischemia changes the coexpression of somatostatin
and neuropeptide Y in hippocampal interneurons. Exp.
Brain. Res., 115(3): 423–429.
Blackmer, T., Larsen, E.C., Bartleson, C., Kowalchyk, J.A.,
Yoon, E.-J., Preininger, A.M., Alford, S., Hamm, H.E.
and Martin, T.F.J. (2005) G protein [beta][gamma] directly
regulates SNARE protein fusion machinery for secretory
granule exocytosis. Nat. Neurosci., 8(4): 421–425.
Blake, A.D. (2001) Somatostatin receptor subtype 1 (sst(1))
regulates intracellular 30 ,50 -cyclic adenosine monophosphate
accumulation in rat embryonic cortical neurons: evidence
with L-797,591, an sst(1)-subtype-selective nonpeptidyl
agonist. Neuropharmacology, 40(4): 590–596.
Boehm, S. and Betz, H. (1997) Somatostatin inhibits excitatory
transmission at rat hippocampal synapses via presynaptic
receptors. J. Neurosci., 17(11): 4066–4075.
Bouilleret, V., Schwaller, B., Schurmans, S., Celio, M.R. and
Fritschy, J.M. (2000) Neurodegenerative and morphogenic
changes in a mouse model of temporal lobe epilepsy do not
depend on the expression of the calcium-binding proteins
parvalbumin, calbindin, or calretinin. Neuroscience, 97(1):
47–58.
Brazeau, P., Vale, W., Burgus, R., Ling, N., Rivier, J. and
Guillemin, R. (1972) Hypothalamic polypeptide that inhibits
the secretion of immunoreactive pituitary growth hormone.
Science, 129: 77–79.
Bruno, J., Yu, Y., Song, J. and Berlowitz, M. (1992) Molecular
cloning and functional expression of a novel brain som-
atostatin receptor. PNAS, 89: 11151–11155.
Buckmaster, P.S. and Dudek, F.E. (1997) Neuron loss, granule
cell axon reorganization, and functional changes in the dent-
ate gyrus of epileptic kainate-treated rats. J. Comp. Neurol.,
385(3): 385–404.
Buckmaster, P.S., Kunkel, D.D., Robbins, R.J. and Schwa-
rtzkroin, P.A. (1994) Somatostatin-immunoreactivity in the
hippocampus of mouse, rat, guinea pig, and rabbit. Hippo-
campus, 4(2): 167–180.
Calbet, M., Guadano-Ferraz, A., Spier, A.D., Maj, M.,
Sutcliffe, J.G., Przewlocki, R. and de Lecea, L. (1999) Co-
rtistatin and somatostatin mRNAs are differentially

regulated in response to kainate. Brain Res. Mol. Brain Res.,
72(1): 55–64.
Chan-Palay, V. (1987) Somatostatin immunoreactive neurons
in the human hippocampus and cortex shown by immuno-
gold/silver intensification on vibratome sections: coexistence
with neuropeptide Y neurons, and effects in Alzheimer-type
dementia. J. Comp. Neurol., 260(2): 201–223.
Chesselet, M.F. and Reisine, T.D. (1983) Somatostatin regu-
lates dopamine release in rat striatal slices and cat caudate
nuclei. J. Neurosci., 3(1): 232–236.
Chronwall, B.M., Chase, T.N. and O’donohue, T.L. (1984)
Coexistence of neuropeptide Y and somatostatin in rat and
human cortical and rat hypothalamic neurons. Neurosci.
Lett., 52(3): 213–217.
Connor, M., Ingram, S.L. and Christie, M.J. (1997) Cortistatin
increase of a potassium conductance in rat locus coeruleus in
vitro. Br. J. Pharmacol., 122(8): 1567–1572.
Cook, J.L., Marcheselli, V., Alam, J., Deininger, P.L. and
Bazan, N.G. (1998) Temporal changes in gene expression
following cryogenic rat brain injury. Brain Res. Mol. Brain
Res., 55(1): 9–19.
Cooper, E.C., Aldape, K.D., Abosch, A., Barbaro, N.M.,
Berger, M.S., Peacock, W.S., Jan, Y.N. and Jan, L.Y. (2000)
Colocalization and coassembly of two human brain M-type
potassium channel subunits that are mutated in epilepsy.
Proc. Natl. Acad. Sci. U.S.A., 97(9): 4914–4919.
Cowles, R.A., Segura, B.J. and Mulholland, M.W. (2002)
Regulation of carbachol-induced c-fos mRNA expression in
AR42J cells by somatostatin receptor subtypes 1, 2, and 3.
Pancreas, 25(3): 239–244.
Csaba, Z., Pirker, S., Lelouvier, B., Simon, A., Videau, C.,
Epelbaum, J., Czech, T., Baumgartner, C., Sperk, G. and
Dournaud, P. (2005) Somatostatin receptor type 2 undergoes
plastic changes in the human epileptic dentate gyrus. J.
Neuropathol. Exp. Neurol., 64(11): 956–969.
Csaba, Z., Richichi, C., Bernard, V., Epelbaum, J., Vezzani, A.
and Dournaud, P. (2004) Plasticity of somatostatin and
somatostatin sst2A receptors in the rat dentate gyrus during
kindling epileptogenesis. Eur. J. Neurosci., 19(9): 2531–2538.
Davies, P., Katzman, R. and Terry, R.D. (1980) Reduced som-
atostatin-like immunoreactivity in cerebral cortex from cases
of Alzheimer disease and Alzheimer senile dementia. Nature,
288(5788): 279–280.
De Lanerolle, N.C., Kim, J.H., Robbins, R.J. and Spencer,
D.D. (1989) Hippocampal interneuron loss and plasticity
in human temporal lobe epilepsy. Brain Res., 495(2):
387–395.
De Lanerolle, N.C., Kim, J.H., Williamson, A., Spencer, S.S.,
Zaveri, H.P., Eid, T. and Spencer, D.D. (2003) A retrospec-
tive analysis of hippocampal pathology in human temporal
lobe epilepsy: evidence for distinctive patient subcategories.
Epilepsia, 44(5): 677–687.
de Lecea, L., Criado, J.R., Prospero-Garcia, O., Gautivik,
K.M., Schweitzer, P., Danielson, P.E., Dunlop, C.L.M., Sig-
gins, G.R., Henriksen, S.J. and Sutcliffe, J.G. (1996) A cor-
tical neuropeptide with neuronal depressant and sleep-
modulating properties. Nature, 381: 242–245.
279
de Lecea, L., Del Rio, J.A., Criado, J.R., Alcantara, S., Mo-
rales, M., Danielson, P.E., Henriksen, S.J., Soriano, E. and
Sutcliffe, J.G. (1997a) Cortistatin is expressed in a distinct
subset of cortical interneurons. J. Neurosci., 17(15):
5868–5880.
de Lecea, L., Ruiz-Lozano, P., Danielson, P.E., Peelle-Kirley,
J., Foye, P.E., Frankel, W.N. and Sutcliffe, J.G. (1997b)
Cloning, mRNA expression, and chromosomal mapping of
mouse and human preprocortistatin. Genomics, 42(3):
499–506.
Dickson, S.L., Viltart, O., Bailey, A.R. and Leng, G. (1997)
Attenuation of the growth hormone secretagogue induction
of Fos protein in the rat arcuate nucleus by central som-
atostatin action. Neuroendocrinology, 66(3): 188–194.
Dournaud, P., Cervera-Pierot, P., Hirsch, E., Javoy-Agid, F.,
Kordon, C., Agid, Y. and Epelbaum, J. (1994) Somatostatin
messenger RNA-containing neurons in Alzheimer’s disease:
an in situ hybridization study in hippocampus, parahippo-
campal cortex and frontal cortex. Neuroscience, 61(4):
755–764.
Dournaud, P., Gu, Y.Z., Schonbrunn, A., Mazella, J., Tan-
nenbaum, G.S. and Beaudet, A. (1996a) Localization of the
somatostatin receptor SST2A in rat brain using a specific anti-
peptide antibody. J. Neurosci., 16(14): 4468–4478.
Dournaud, P., Jazat-Poindessous, F., Slama, A., Lamour, Y.
and Epelbaum, J. (1996b) Correlations between water maze
performance and cortical somatostatin mRNA and high-
affinity binding sites during ageing in rats. Eur. J. Neurosci.,
8(3): 476–485.
Edvinsson, L., Minthon, L., Ekman, R. and Gustafson, L.
(1993) Neuropeptides in cerebrospinal fluid of patients with
Alzheimer’s disease and dementia with frontotemporal lobe
degeneration. Dementia, 4(3–4): 167–171.
Epelbaum, J., Ruberg, M., Moyse, E., Javoy-Agid, F., Dubois,
B. and Agid, Y. (1983) Somatostatin and dementia in Par-
kinson’s disease. Brain Res., 278(1–2): 376–379.
Figueredo-Cardenas, G., Morello, M., Sancesario, G.,
Bernardi, G. and Reiner, A. (1996) Colocalization of som-
atostatin, neuropeptide Y, neuronal nitric oxide synthase and
NADPH-diaphorase in striatal interneurons in rats. Brain
Res., 735(2): 317–324.
Florio, T., Yao, H., Carey, K.D., Dillon, T.J. and Stork, P.J.
(1999) Somatostatin activation of mitogen-activated
protein kinase via somatostatin receptor 1 (SSTR1). Mol.
Endocrinol., 13(1): 24–37.
Freund, T.F. and Buzsaki, G. (1996) Interneurons of the
hippocampus. Hippocampus, 6(4): 347–470.
Gabriel, S.M., Davidson, M., Haroutunian, V., Powchik, P.,
Bierer, L.M., Purohit, D.P., Perl, D.P. and Davis, K.L.
(1996) Neuropeptide deficits in schizophrenia vs. Alzheimer’s
disease cerebral cortex. Biol. Psychiatry, 39(2): 82–91.
Gerachshenko, T., Blackmer, T., Yoon, E.-J., Bartleson, C.,
Hamm, H.E. and Alford, S. (2005) G[beta][gamma] acts at
the C terminus of SNAP-25 to mediate presynaptic inhibi-
tion. Nat. Neurosci., 8(5): 597–605.
Grouselle, D., Winsky-Sommerer, R., David, J.P., Delacourte,
A., Dournaud, P. and Epelbaum, J. (1998) Loss of
280
somatostatin-like immunoreactivity in the frontal cortex
of Alzheimer patients carrying the apolipoprotein epsilon 4
allele. Neurosci. Lett., 255(1): 21–24.
Halabisky, B., Shen, F., Huguenard, J.R. and Prince, D.A.
(2006) Electrophysiological classification of somatostatin-
positive interneurons in mouse sensorimotor cortex. J. Ne-
urophysiol., 96(2): 834–845.
Handel, M., Schulz, S., Stanarius, A., Schreff, M., Erdtmann-
Vourliotis, M., Schmidt, H., Wolf, G. and Hollt, V. (1999)
Selective targeting of somatostatin receptor 3 to neuronal
cilia. Neuroscience, 89(3): 909–926.
Hashimoto, T. and Obata, K. (1991) Induction of somatostatin
by kainic acid in pyramidal and granule cells of the rat hip-
pocampus. Neurosci. Res., 12(4): 514–527.
Hayashi, M., Yamashita, A. and Shimizu, K. (1997) Som-
atostatin and brain-derived neurotrophic factor mRNA
expression in the primate brain: decreased levels of mRNAs
during aging. Brain Res., 749(2): 283–289.
Hervieu, G. and Emson, P.C. (1998) The localization of som-
atostatin receptor 1 (sst1) immunoreactivity in the rat brain
using an N-terminal specific antibody. Neuroscience, 85(4):
1263–1284.
Holm, I.E., Geneser, F.A. and Zimmer, J. (1992) Somatostatin-
and neuropeptide Y-like immunoreactivity in the dentate
area, hippocampus, and subiculum of the domestic pig.
J. Comp. Neurol., 322(3): 390–408.
Huang, S.M., Mouri, A., Kokubo, H., Nakajima, R., Suemoto,
T., Higuchi, M., Staufenbiel, M., Noda, Y., Yamaguchi, H.,
Nabeshima, T., Saido, T.C. and Iwata, N. (2006) Neprilysin-
sensitive synapse-associated Abeta oligomers impair neuron-
al plasticity and cognitive function. J. Biol. Chem., 281(26):
17941–17951.
Ikeda, S.R. and Schofield, G.G. (1989) Somatostatin blocks
a calcium current in rat sympathetic ganglion neurons.
J. Physiol., 409: 221–240.
Inoue, M., Nakajima, S. and Nakajima, Y. (1988) Somatostatin
induces an inward rectification in rat locus coeruleus neurons
through a pertussis toxin sensitive mechanism. J. Physiol.,
407: 177–198.
Ishihara, S., Hassan, S., Kinoshita, Y., Moriyama, N.,
Fukuda, R., Maekawa, T., Okada, A. and Chiba, T. (1999)
Growth inhibitory effects of somatostatin on human le-
ukemia cell lines mediated by somatostatin receptor subtype
1. Peptides, 20(3): 313–318.
Iwata, N., Takaki, Y., Fukami, S., Tsubuki, S. and Saido, T.C.
(2002) Region-specific reduction of A beta-degrading end-
opeptidase, neprilysin, in mouse hippocampus upon aging.
J. Neurosci. Res., 70(3): 493–500.
Jinno, S. and Kosaka, T. (2000) Colocalization of parvalbumin
and somatostatin-like immunoreactivity in the mouse hippo-
campus: quantitative analysis with optical dissector.
J. Comp. Neurol., 428(3): 377–388.
Johansen, F.F., Zimmer, J. and Diemer, N.H. (1987) Early loss
of somatostatin neurons in dentate hilus after cerebral is-
chemia in the rat precedes CA-1 pyramidal cell loss. Acta
Neuropathol., 73(2): 110–114.
Kauer, J.A., Malenka, R.C. and Nicoll, R.A. (1988) A persist-
ent postsynaptic modification mediates long-term potentiat-
ion in the hippocampus. Neuron, 1(10): 911–917.
Kimura, N., Tomizawa, S., Arai, K.N. and Osamura, R.Y.
(2001) Characterization of 50 -flanking region of rat
somatostatin receptor sst2 gene: transcriptional regulatory
elements and activation by Pitx1 and estrogen. Endocrino-
logy, 142(4): 1427–1441.
Kobayashi, M. and Buckmaster, P.S. (2003) Reduced inhibition
of dentate granule cells in a model of temporal lobe epilepsy.
J. Neurosci., 23(6): 2440–2452.
Kong, H., Depaoli, A.M., Breder, C.D., Yasuda, K., Bell, G.I.
and Reisine, T. (1994) Differential expression of messenger
RNAs for somatostatin receptor subtypes SSTR1, SSTR2,
and SSTR3 in adult rat brain: analysis by RNA blotting and
in situ hybridization histochemistry. Neuroscience, 59(1):
175–184.
Kosaka, T., Wu, J.-Y. and Benoit, R. (1988) GABAergic
neurons containing somatostatin-like immunoreactivity in
the rat hippocampus and dentate gyrus. Exp. Brain Res., 71:
388–398.
Kragh, J., Tonder, N., Finsen, B.R., Zimmer, J. and Bolwig,
T.G. (1994) Repeated electroconvulsive shocks cause tran-
sient changes in rat hippocampal somatostatin and neuro-
peptide Y immunoreactivity and mRNA in situ hybridization
signals. Exp. Brain Res., 98(2): 305–313.
Kumar, U. (2005) Expression of somatostatin receptor subtypes
(SSTR1-5) in Alzheimer’s disease brain: an immunohisto-
chemical analysis. Neuroscience, 134(2): 525–538.
Lahlou, H., Guillermet, J., Hortala, M., Vernejoul, F., Pyron-
net, S., Bousquet, C. and Susini, C. (2004) Molecular signa-
ling of somatostatin receptors. Ann. N.Y. Acad. Sci.,
1014(1): 121–131.
Lahlou, H., Saint-Laurent, N., Esteve, J.P., Eychene, A.,
Pradayrol, L., Pyronnet, S. and Susini, C. (2003) sst2
Somatostatin receptor inhibits cell proliferation through
Ras-, Rap1-, and B-Raf-dependent ERK2 activation. J. Biol.
Chem., 278(41): 39356–39371.
Lahteinen, S., Pitkanen, A., Koponen, E., Saarelainen, T. and
Castren, E. (2003) Exacerbated status epilepticus and acute
cell loss, but no changes in epileptogenesis, in mice
with increased brain-derived neurotrophic factor signaling.
Neuroscience, 122(4): 1081–1092.
Le Gal La Salle, G. (1988) Long-lasting and sequential increase
of c-fos oncoprotein expression in kainic acid-induced status
epilepticus. Neurosci. Lett., 88(2): 127–130.
Leranth, C., Malcolm, A.J. and Frotscher, M. (1990) Affer-
ent and efferent synaptic connections of somatostatin-
immunoreactive neurons in the rat fascia dentata. J. Comp.
Neurol., 295(1): 111–122.
Leroux, P., Weissmann, D., Pujol, J.-F. and Vaudry, H. (1993)
Quantitative autoradiography of somatostatin receptors in
the rat limbic system. J. Comp. Neurol., 331: 389–401.
Lowenstein, D.H., Thomas, M.J., Smith, D.H. and Mcintosh,
T.K. (1992) Selective vulnerability of dentate hilar neurons
following traumatic brain injury: a potential mechanistic link

between head trauma and disorders of the hippocampus.
J. Neurosci., 12(12): 4846–4853.
Lubke, J., Frotscher, M. and Spruston, N. (1998) Specialized
electrophysiological properties of anatomically identified
neurons in the hilar region of the rat fascia dentata. J.
Neurophysiol., 79(3): 1518–1534.
Manfridi, A., Forloni, G.L., Vezzani, A., Fodritto, F. and De
Simoni, M.G. (1991) Functional and histological consequen-
ces of quinolinic and kainic acid-induced seizures on hippo-
campal somatostatin neurons. Neuroscience, 41(1): 127–135.
Marti, M., Bregola, G., Morari, M., Gemignani, A. and
Simonato, M. (2000) Somatostatin release in the hippocam-
pus in the kindling model of epilepsy: a microdialysis study.
J. Neurochem., 74(6): 2497–2503.
Mathern, G.W., Babb, T.L., Pretorius, J.K. and Leite, J.P.
(1995) Reactive synaptogenesis and neuron densities for
neuropeptide Y, somatostatin, and glutamate decarboxylase
immunoreactivity in the epileptogenic human fascia dentata.
J. Neurosci., 15(5 pt 2): 3990–4004.
Mazarati, A.M. and Telegdy, G. (1992) Effects of somatostatin
and anti-somatostatin serum on picrotoxin-kindled seizures.
Neuropharmacology, 31(8): 793–797.
Mcdonald, A.J. (1989) Coexistence of somatostatin with
neuropeptide Y, but not with cholecystokinin or vasoactive
intestinal peptide, in neurons of the rat amygdala. Brain Res.,
500(1–2): 37–45.
Meriney, S.D., Gray, D.B. and Pilar, G.R. (1994) Somatostatin-
induced inhibition of neuronal calcium current modulated by
cGMP-dependent protein kinase. Nature, 369(6478): 336–339.
Mihara, S., North, R.A. and Surprenant, A. (1987) Somatosta-
tin increases an inwardly rectifying potassium conductance in
guinea pig submucosa plexus neurones. J. Physiol., 390:
335–355.
Mikkelsen, J.D. and Woldbye, D.P. (2006) Accumulated in-
crease in neuropeptide Y and somatostatin gene expression of
the rat in response to repeated electroconvulsive stimulation.
J. Psychiatr. Res., 40(2): 153–159.
Milner, T.A. and Bacon, C.E. (1989) Ultrastructural localiza-
tion of somatostatin-like immunoreactivity in the rat dentate
gyrus. J. Comp. Neurol., 290(4): 544–560.
Minthon, L., Edvinsson, L. and Gustafson, L. (1997) Som-
atostatin and neuropeptide Y in cerebrospinal fluid: correla-
tions with severity of disease and clinical signs in Alzheimer’s
disease and frontotemporal dementia. Dement. Geriatr.
Cogn. Disord., 8(4): 232–239.
Mitchell, T.W., Buckmaster, P.S., Hoover, E.A., Whalen, L.R.
and Dudek, F.E. (1999) Neuron loss and axon reorganization
in the dentate gyrus of cats infected with the feline immuno-
deficiency virus. J. Comp. Neurol., 411(4): 563–577.
Moller, L.N., Stidsen, C.E., Hartmann, B. and Holst, J.J.
(2003) Somatostatin receptors. Biochim. Biophys. Acta,
1616(1): 1–84.
Montminy, M., Brindle, P., Arias, J., Ferreri, K. and Arm-
strong, R. (1996) Regulation of somatostatin gene transcrip-
tion by cyclic adenosine monophosphate. Metabolism, 45(8
Suppl 1): 4–7.
281
Montminy, M.R. and Bilezikjian, L.M. (1987) Binding of a
nuclear protein to the cyclic-AMP response element of the
somatostatin gene. Nature, 328(6126): 175–178.
Moore, S.D., Madamba, S.G., Joels, M. and Siggins, G.R.
(1988) Somatostatin augments the M-current in hippocampal
neurons. Science, 239(4837): 278–280.
Moser, E.I., Krobert, K.A., Moser, M.B. and Morris, R.G.
(1998) Impaired spatial learning after saturation of long-term
potentiation. Science, 281(5385): 2038–2042.
Mott, D.D., Turner, D.A., Okazaki, M.M. and Lewis, D.V.
(1997) Interneurons of the dentate-hilus border of the rat
dentate gyrus: morphological and electrophysiological heter-
ogeneity. J. Neurosci., 17(11): 3990–4005.
Nakao, K., Ikegaya, Y., Yamada, M.K., Nishiyama, N. and
Matsuki, N. (2002) Hippocampal long-term depression as an
index of spatial working memory. Eur. J. Neurosci., 16(5):
970–974.
Nakatani, N., Hattori, E., Ohnishi, T., Dean, B., Iwayama, Y.,
Matsumoto, I., Kato, T., Osumi, N., Higuchi, T., Niwa, S.I.
and Yoshikawa, T. (2006) Genome-wide expression analysis
detects eight genes with robust alterations specific to bipolar I
disorder: relevance to neuronal network perturbation. Hum.
Mol. Genet., 15(12): 1949–1962.
Nemeroff, C.B., Youngblood, W.W., Manberg, P.J., Prange
Jr., A.J. and Kizer, J.S. (1983) Regional brain concentrations
of neuropeptides in Huntington’s chorea and schizophrenia.
Science, 221(4614): 972–975.
Neves, S.R., Ram, P.T. and Iyengar, R. (2002) G protein path-
ways. Science, 296(5573): 1636–1639.
Oliva Jr., A.A., Jiang, M., Lam, T., Smith, K.L. and Swann,
J.W. (2000) Novel hippocampal interneuronal subtypes iden-
tified using transgenic mice that express green fluorescent
protein in GABAergic interneurons. J. Neurosci., 20(9):
3354–3368.
Passarelli, F. and Orzi, F. (1993) Somatostatin mRNA in the
hippocampal formation following electroconvulsive shock in
the rat. Neurosci. Lett., 153(2): 197–201.
Peng, Z. and Houser, C.R. (2005) Temporal patterns of fos
expression in the dentate gyrus after spontaneous seizures in
a mouse model of temporal lobe epilepsy. J. Neurosci.,
25(31): 7210–7220.
Perez, J., Vezzani, A., Civenni, G., Tutka, P., Rizzi, M.,
Schuepbach, E. and Hoyer, D. (1995) Functional effects of D-
Phe-c[Cys-Tyr-D-Trp-Lys-Val-Cys]-Trp-NH2 and differential
changes in somatostatin receptor messenger RNAs, binding
sites and somatostatin release in kainic acid-treated rats.
Neuroscience, 65(4): 1087–1097.
Piwko, C., Thoss, V.S., Samanin, R., Hoyer, D. and Vezzani,
A. (1996) Status of somatostatin receptor messenger RNAs
and binding sites in rat brain during kindling epileptogenesis.
Neuroscience, 75(3): 857–868.
Pola, S., Cattaneo, M.G. and Vicentini, L.M. (2003) Anti-
migratory and anti-invasive effect of somatostatin in
human neuroblastoma cells: involvement of Rac and
MAP kinase activity. J. Biol. Chem., 278(42):
40601–40606.
282
Powers, A.C., Tedeschi, F., Wright, K.E., Chan, J.S. and Ha-
bener, J.F. (1989) Somatostatin gene expression in pancreatic
islet cells is directed by cell-specific DNA control elements and
DNA-binding proteins. J. Biol. Chem., 264(17): 10048–10056.
Qiu, C., Suntheimer, R.J. and Tallent, M.K. (2003) Somatosta-
tin depresses CA1 long term potentiation. Abstract Viewer/
Itinerary Planner. Washington, DC: Society for Neurosci-
ence, Program No. 889.1.
Qiu, C., Suzuki, C., Zeyda, T., Hochgeschwender, U., de Lecea,
L. and Tallent, M.K. (2005) Increased seizure severity and
reduced somatostatin effects in hippocampus of somatostatin
receptor subtype 4 (SST4) knockout mice. 2005 Abstract
Viewer/Itinerary Planner. Washington, DC: Society for
Neuroscience, Program No. 607.11.
Ramirez, J.L., Gracia-Navarro, F., Garcia-Navarro, S.,
Torronteras, R., Malagon, M.M. and Castano, J.P. (2002)
Somatostatin stimulates GH secretion in two porcine soma-
totrope subpopulations through a cAMP-dependent path-
way. Endocrinology, 143(3): 889–897.
Ramos, B., Baglietto-Vargas, D., Del Rio, J.C., Moreno-
Gonzalez, I., Santa-Maria, C., Jimenez, S., Caballero, C.,
Lopez-Tellez, J.F., Khan, Z.U., Ruano, D., Gutierrez, A.
and Vitorica, J. (2006) Early neuropathology of somatosta-
tin/NPY GABAergic cells in the hippocampus of a
PS1  APP transgenic model of Alzheimer’s disease.
Neurobiol. Aging, 27(11): 1658–1672.
Rapp, P.R. and Amaral, D.G. (1988) The time of origin of
somatostatin-immunoreactive neurons in the rat hippocam-
pal formation. Brain Res., 469(1–2): 231–239.
Rauca, C., Schafer, K. and Hollt, V. (1999) Effects of som-
atostatin, octreotide and cortistatin on ischaemic neuronal
damage following permanent middle cerebral artery occlu-
sion in the rat. Naunyn. Schmiedebergs Arch. Pharmacol.,
360(6): 633–638.
Raynor, K. and Reisine, T. (1992) Differential coupling of
somatostatin1 receptors to adenylyl cyclase in the rat stria-
tum vs. the pituitary and other regions of the rat brain.
J. Pharmacol. Exp. Ther., 260: 841–848.
Reisine, T. (1995) Somatostatin. Cell Mol. Neurobiol., 15(6):
597–614.
Robbins, R.J., Brines, M.L., Kim, J.H., Adrian, T., De Lane-
rolle, N., Welsh, S. and Spencer, D.D. (1991) A selective loss
of somatostatin in the hippocampus of patients with tempo-
ral lobe epilepsy. Ann. Neurol., 29(3): 325–332.
Roberts, G.W., Crow, T.J. and Polak, J.M. (1985) Location
of neuronal tangles in somatostatin neurons in Alzheimer’s
disease. Nature, 314: 92–94.
Robishaw, J.D. and Berlot, C.H. (2004) Translating G protein
subunit diversity into functional specificity. Curr. Opin. Cell
Biol., 16(2): 206–209.
Saito, T., Iwata, N., Tsubuki, S., Takaki, Y., Takano, J.,
Huang, S.M., Suemoto, T., Higuchi, M. and Saido, T.C.
(2005) Somatostatin regulates brain amyloid beta peptide
Abeta42 through modulation of proteolytic degradation.
Nat. Med., 11(4): 434–439.
Sakai, K., Maeda, K., Chihara, K. and Kaneda, H. (1995) In-
creases in cortical neuropeptide Y and somatostatin
concentrations following haloperidol-depot treatment in rats.
Neuropeptides, 29(3): 157–161.
Scharfman, H.E. (1992) Differentiation of rat dentate neurons
by morphology and electrophysiology in hippocampal slices:
granule cells, spiny hilar cells and aspiny ‘fast-spiking’ cells.
Epilepsy Res. Suppl., 7: 93–109.
Schindler, M., Humphrey, P.P., Lohrke, S. and Friauf, E.
(1999) Immunohistochemical localization of the somatostatin
sst2(b) receptor splice variant in the rat central nervous sys-
tem. Neuroscience, 90(3): 859–874.
Schreff, M., Schulz, S., Handel, M., Keilhoff, G., Braun, H.,
Pereira, G., Klutzny, M., Schmidt, H., Wolf, G. and Hollt, V.
(2000) Distribution, targeting, and internalization of the sst4
somatostatin receptor in rat brain. J. Neurosci., 20(10):
3785–3797.
Schulz, S., Handel, M., Schreff, M., Schmidt, H. and Hollt, V.
(2000) Localization of five somatostatin receptors in the rat
central nervous system using subtype-specific antibodies. J.
Physiol. Paris, 94(3–4): 259–264.
Schwarzer, C., Sperk, G., Samanin, R., Rizzi, M., Gariboldi,
M. and Vezzani, A. (1996) Neuropeptides-immunoreactivity
and their mRNA expression in kindling: functional implica-
tions for limbic epileptogenesis. Brain Res. Brain Res. Rev.,
22(1): 27–50.
Schweitzer, P., Madamba, S. and Siggins, G.R. (1990)
Arachidonic acid metabolites as mediators of somatostatin-
induced increase of neuronal M-current. Nature, 346(6283):
464–467.
Schweitzer, P., Madamba, S.G. and Siggins, G.R. (1998) Som-
atostatin increases a voltage-insensitive K+ conductance in
rat CA1 hippocampal neurons. J. Neurophysiol., 79(3):
1230–1238.
Schweitzer, P., Madamba, S.G. and Siggins, G.R. (2003) The
sleep-modulating peptide cortistatin augments the h-current
in hippocampal neurons. J. Neurosci., 23(34): 10884–10891.
Sharma, R.P., Bissette, G., Janicak, P.G., Davis, J.M. and Ne-
meroff, C.B. (1995) Elevation of CSF somatostatin concen-
trations in mania. Am. J. Psychiatry, 152(12): 1807–1809.
Siehler, S. and Hoyer, D. (1999) Characterisation of human
recombinant somatostatin receptors. 4. Modulation of
phospholipase C activity. Naunyn Schmiedebergs Arch.
Pharmacol., 360(5): 522–532.
Simonato, M., Bregola, G., Beani, L., Vezzani, A., Sala, R.,
Raiteri, M. and Bonanno, G. (1998) Time- and region-spe-
cific variations in somatostatin release following amygdala
kindling in the rat. J. Neurochem., 70(1): 252–259.
Sloviter, R.S. (1987) Decreased hippocampal inhibition and a se-
lective loss of interneurons in experimental epilepsy. Science,
235(4784): 73–76.
Sodickson, D.L. and Bean, B.P. (1998) Neurotransmitter acti-
vation of inwardly rectifying potassium current in dissociated
hippocampal CA3 neurons: interactions among multiple re-
ceptors. J. Neurosci., 18(20): 8153–8162.
Sokal, D.M. and Large, C.H. (2001) The effects of GABA(B)
receptor activation on spontaneous and evoked activity in
the dentate gyrus of kainic acid-treated rats. Neuropharma-
cology, 40(2): 193–202.

Straiker, A.J., Borden, C.R. and Sullivan, J.M. (2002) G-pro-
tein alpha subunit isoforms couple differentially to receptors
that mediate presynaptic inhibition at rat hippocampal
synapses. J. Neurosci., 22(7): 2460–2468.
Stroh, T., Kreienkamp, H.J. and Beaudet, A. (1999)
Immunohistochemical distribution of the somatostatin re-
ceptor subtype 5 in the adult rat brain: predominant expres-
sion in the basal forebrain. J. Comp. Neurol., 412(1): 69–82.
Sutula, T. and Steward, O. (1986) Quantitative analysis of
synaptic potentiation during kindling of the perforant path.
J. Neurophysiol., 56(3): 732–746.
Sutula, T. and Steward, O. (1987) Facilitation of kindling by
prior induction of long-term potentiation in the perforant
path. Brain Res., 420(1): 109–117.
Swanson, T.H., Sperling, M.R. and O’connor, M.J. (1998)
Strong paired pulse depression of dentate granule cells in
slices from patients with temporal lobe epilepsy. J. Neural
Transm., 105(6–7): 613–625.
Tallent, M.K., Fabre, V., Qiu, C., Calbet, M., Lamp, T., Bar-
atta, M.V., Suzuki, C., Levy, C.L., Siggins, G.R. and Hen-
riksen, S.J. (2005a) Cortistatin overexpression in transgenic
mice produces deficits in synaptic plasticity and learning.
Mol. Cell. Neurosci., 30(3): 465–475.
Tallent, M.K., Patterson, C., Hon, B., Qiu, C., Zeyda, T., Ho-
chgeschwender, U. and de Lecea, L. (2005b) Somatostatin
receptor subtype 3 is critical for hippocampal-dependent
memory: studies in SST3 receptor knockout mice. Abstract
Viewer/Itinerary Planner. Washington, DC: Society for
Neuroscience, Program No. 607.10.
Tallent, M.K. and Siggins, G.R. (1997) Somatostatin depresses
excitatory but not inhibitory neurotransmission in rat CA1
hippocampus. J. Neurophysiol., 78(6): 3008–3018.
Tallent, M.K. and Siggins, G.R. (1999) Somatostatin acts in
CA1 and CA3 to reduce hippocampal epileptiform activity.
J. Neurophysiol., 81(4): 1626–1635.
Tentler, J.J., Hadcock, J.R. and Gutierrez-Hartmann, A. (1997)
Somatostatin acts by inhibiting the cyclic 30 ,50 -adenosine
monophosphate (cAMP)/protein kinase A pathway, cAMP
response element-binding protein (CREB) phosphorylation,
and CREB transcription potency. Mol. Endocrinol., 11(7):
859–866.
Thoss, V.S., Duc, D. and Hoyer, D. (1996) Somatostatin re-
ceptors in the developing rat brain. Eur. J. Pharmacol.,
297(1–2): 145–155.
Thoss, V.S., Perez, J., Duc, D. and Hoyer, D. (1995) Embryonic
and postnatal mRNA distribution of five somatostatin re-
ceptor subtypes in the rat brain. Neuropharmacology, 34(12):
1673–1688.
Todisco, A., Campbell, V., Dickinson, C.J., Delvalle, J.
and Yamada, T. (1994) Molecular basis for somatostatin
action: inhibition of c-fos expression and AP-1 binding.
Am. J. Physiol. Gastrointest. Liver Physiol., 267(2):
G245–G253.
Todisco, A., Seva, C., Takeuchi, Y., Dickinson, C.J. and Ya-
mada, T. (1995) Somatostatin inhibits AP-1 function via
multiple protein phosphatases. Am. J. Physiol., 269(1 Pt 1):
G160–G166.
283
Tomidokoro, Y., Harigaya, Y., Matsubara, E., Ikeda, M.,
Kawarabayashi, T., Okamoto, K. and Shoji, M. (2000) Im-
paired neurotransmitter systems by Abeta amyloidosis in
APPsw transgenic mice overexpressing amyloid beta protein
precursor. Neurosci. Lett., 292(3): 155–158.
Vecsei, L., Csala, B., Widerlov, E., Ekman, R., Czopf, J. and
Palffy, G. (1990) Lumbar cerebrospinal fluid concentrations
of somatostatin and neuropeptide Y in multiple sclerosis.
Brain Res. Bull., 25(3): 411–413.
Vela, J., Gutierrez, A., Vitorica, J. and Ruano, D. (2003) Rat
hippocampal GABAergic molecular markers are differen-
tially affected by ageing. J. Neurochem., 85(2): 368–377.
Velimirovic, B.M., Koyano, K., Nakajima, S. and Nakajima,
Y. (1995) Opposing mechanisms of regulation of a G-pro-
tein-coupled inward rectifier K+ channel in rat brain neu-
rons. Proc. Natl. Acad. Sci. U.S.A., 92(5): 1590–1594.
Vezzani, A., Monno, A., Rizzi, M., Galli, A., Barrios, M. and
Samanin, R. (1992) Somatostatin release is enhanced in the
hippocampus of partially and fully kindled rats. Neurosci-
ence, 51(1): 41–46.
Vezzani, A., Rizzi, M., Conti, M. and Samanin, R. (2000)
Modulatory role of neuropeptides in seizures induced in
rats by stimulation of glutamate receptors. J. Nutr., 130(4S
Suppl): 1046S–1048S.
Vezzani, A., Schwarzer, C., Lothman, E.W., Williamson, J. and
Sperk, G. (1996) Functional changes in somatostatin and
neuropeptide Y containing neurons in the rat hippocampus
in chronic models of limbic seizures. Epilepsy Res., 26(1):
267–279.
Vezzani, A., Serafini, R., Stasi, M.A., Vigano, G., Rizzi, M. and
Samanin, R. (1991) A peptidase-resistant cyclic octapeptide
analogue of somatostatin (SMS 201-995) modulates seizures
induced by quinolic acid and kainic acid differently in the rat
hippocampus. Neuropharmacology, 30(4): 345–352.
Wang, H.-L., Dichter, M.A. and Reisine, T. (1990a) Lack of
cross-desensitization of somatostatin-14 and somatostatin-28
receptors coupled to potassium channels in rat neocortical
neurons. Mol. Pharm., 38(3): 357–361.
Wang, H.-L., Reisine, T. and Dichter, M.A. (1990b) Som-
atostatin-14 and somatostatin-28 inhibit calcium currents in
rat neocortical neurons. Neuroscience, 38(2): 335–342.
Wasterlain, C.G., Mazarati, A.M., Shirasaka, Y., Thompson,
K.W., Penix, L., Liu, H. and Katsumori, H. (1999) Seizure-
induced hippocampal damage and chronic epilepsy: a he-
bbian theory of epileptogenesis. Adv. Neurol., 79: 829–843.
Winsky-Sommerer, R., Spier, A.D., Fabre, V., de Lecea, L. and
Criado, J.R. (2004) Overexpression of the human beta-am-
yloid precursor protein downregulates cortistatin mRNA in
PDAPP mice. Brain Res., 1023(1): 157–162.
Yoshitomi, H., Fujii, Y., Miyazaki, M., Nakajima, N., Inagaki,
N. and Seino, S. (1997) Involvement of MAP kinase and
c-fos signaling in the inhibition of cell growth by somatosta-
tin. Am. J. Physiol., 272(5 Pt 1): E769–E774.
Yasojima, K., Akiyama, H., Mcgeer, E.G. and Mcgeer, P.L.
(2001) Reduced neprilysin in high plaque areas of Alzheimer
brain: a possible relationship to deficient degradation of beta-
amyloid peptide. Neurosci. Lett., 297(2): 97–100.
284
Yus-Najera, E., Munoz, A., Salvador, N., Jensen, B.S.,
Rasmussen, H.B., Defelipe, J. and Villarroel, A. (2003)
Localization of KCNQ5 in the normal and epileptic human
temporal neocortex and hippocampal formation. Neurosci-
ence, 120(2): 353–364.
Zhang, Z.J., Lappi, D.A., Wrenn, C.C., Milner, T.A.
and Wiley, R.G. (1998) Selective lesion of the
cholinergic basal forebrain causes a loss of cortical
neuropeptide Y and somatostatin neurons. Brain Res.,
800(2): 198–206.
Zheng, H., Bailey, A., Jiang, M.H., Honda, K., Chen, H.Y.,
Trumbauer, M.E., Van Der Ploeg, L.H., Schaeffer, J.M.,
Leng, G. and Smith, R.G. (1997) Somatostatin receptor sub-
type 2 knockout mice are refractory to growth hormone-
negative feedback on arcuate neurons. Mol. Endocrinol.,
11(11): 1709–1717.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 17

Neuropeptide Y in the dentate gyrus

Günther Sperk1,, Trevor Hamilton2 and William F. Colmers2,

1
Department of Pharmacology, Medical University Innsbruck, Peter-Mayr-Str. 1a, 6020 Innsbruck, Austria
2
Department of Pharmacology, University of Alberta, 9– 36 Medical Sciences Building, Edmonton, AB, T6G 2H7, Canada

Abstract: Neuropeptide Y (NPY) is contained in at least four types of GABAergic interneurons in the
dentate gyrus, many of which also contain somatostatin and give rise to the dense NPY innervation of the
dentate outer molecular layer. In humans but not rats, minute amounts of NPY are also normally expressed
in dentate granule cells, while seizure activity in rats induces robust NPY expression in granule cells. Y1 and
Y2 receptors are the most abundant NPY receptors expressed in the dentate gyrus. Y1 receptors are
postsynaptic receptors, primarily located on granule cell dendrites in the molecular layer and some in-
terneurons, while Y2 receptors are presynaptic receptors mediating inhibition of glutamate release, and
potentially that of NPY and GABA depending on their presynaptic localization, and may also be expressed
on some hilar interneurons. In humans, monkeys and mice, Y2 receptors are also present on mossy fibers,
but not in most rat species, though functional evidence suggests their presence. Hilar interneurons con-
taining NPY degenerate in temporal lobe epilepsy and in Alzheimer’s disease and reduced levels of NPY in
dentate hilus are associated with depression. By activating Y1 receptors, NPY also exerts powerful ne-
uroproliferative effects on subgranular zone progenitor cells, increasing the number of newly born granule
cells in the adult dentate gyrus. Functionally, NPY exerts anticonvulsive actions mediated by Y2 receptors
at mossy fiber terminals, but there are no presynaptic responses to NPY at perforant path inputs to dentate
granule cells in rats or mice. NPY also has potentially complicated actions on NPY-containing interneu-
rons. Elevated expression of NPY in mossy fibers of the rat, sprouting of NPY interneurons in the human
dentate, and over-expression of Y2 receptors in mossy fibers indicate an anticonvulsive role of endogenous
NPY in epilepsy. However, the physiological role of NPY in the healthy dentate gyrus remains unclear.

Keywords: neuropeptide Y; Y1 receptors; Y2 receptors; epilepsy

Neuropeptide Y (NPY) is a 36 amino acid peptide
first isolated from porcine brain extracts (Tatemoto
et al., 1982). It is so named because it possesses
tyrosine residues at both the C- and the N-terminal
ends and three other tyrosine residues within its
amino acid sequence. It was discovered and
Corresponding author. Tel.: +43-512-9003-71210;
E-mail: guenther.sperk@i-med.ac.at (G. Sperk)
Corresponding author.Tel.: 780-492-1933;
E-mail: William.colmers@ualberta.ca (W.F. Colmers)
isolated due to its C-terminal amidation common
for many neuropeptides. NPY belongs to a larger
family of pancreatic peptides that in addition
includes peptide YY (PYY) and pancreatic poly-
peptide (PP). In the periphery, NPY is primarily
expressed in sympathetic neurons and in the
adrenal gland, and also in neurons of the enteric
nervous system including the submucosal ganglion
(Lundberg et al., 1983; Sundler et al., 1983). Unlike
the other members of the pancreatic PP family,
NPY is abundant in the central nervous system

DOI: 10.1016/S0079-6123(07)63017-9 285

286
(de Quidt and Emson, 1986; Morris, 1989). It is
preferentially located in subpopulations of GABA-
ergic interneurons in the telencephalon (Hendry
et al., 1984) in noradrenergic neurons originating
from the locus coeruleus (Everitt et al., 1984;
Harfstrand et al., 1987; Smialowska, 1988), and in
hypothalamic neurons of the arcuate and other nu-
clei (Smith and Parent, 1986; Elmquist et al., 2005).
Neurons containing NPY in the dentate gyrus and
their synaptic contacts
NPY-immunoreactive (IR) neurons, all of which
express GABA, are abundant in the dentate gyrus
of humans (Chan-Palay et al., 1986; Furtinger
et al., 2001), monkeys (Smith et al., 1985; Smith and
Parent, 1986; Kohler et al., 1986), rodents (Kohler
et al., 1986; Haas et al., 1987; Morris, 1989; Deller
and Leranth, 1990) and pigs (Holm et al., 1992).
NPY-containing perikarya are present in all layers
of the fascia dentata. The largest population of
hippocampal NPY-IR neurons is contained in the
polymorphic region of the dentate gyrus, the dent-
ate hilus (Kohler et al., 1986). In their careful ex-
amination of NPY-IR neurons in the rat dentate
gyrus, Deller and Leranth (1990) classified four
different types of neurons: The majority (>60%,
type 2 cells) were identified as medium-sized mul-
tipolar and fusiform hilar neurons with dendrites
occasionally reaching the outer molecular layer;
the next most common (20%, type 3 cells) were
pyramidal-shaped cells in the granule cell layer with
long apical dendrites reaching the outer molecular
layer. Comprising the remaining 20% were large
multipolar cells located in the deep hilus (type 1);
and small multipolar NPY-IR cells located in the
molecular layer (type 4). About half of the hilar
NPY-IR neurons (types 1 and 2) are also IR for
somatostatin and the projection of these neurons
has a high density of somatostatin-IR axon termi-
nals (Kohler et al., 1986; Freund and Buzsaki,
1996).
Light and electron microscopic studies show
that the majority of NPY-IR terminals are located
in the outer molecular layer of the rat dentate
gyrus, where they establish symmetric (Gray type
2) synaptic contacts on dendritic shafts and
occasionally on spines of granule cells (Kohler et
al., 1986; Deller and Leranth, 1990). While most of
these axons pass through the granule cell layer and
do not form contacts with somata of granule cells,
a moderate number of NPY-IR synapses are also
present on dendrites in the inner molecular layer
and on the cell body of granule cells. The few
synaptic contacts formed are symmetric, consistent
with the GABAergic nature of these interneurons.
Furthermore, numerous symmetric NPY-IR
synapses are found on dendrites and somata of
neurons in the hilar area, which support the phar-
macological results reported below. Dendrites of
some NPY-IR neurons in the hilar area receive
input from granule cell axons, the mossy fibers
axon collaterals. From transection studies, there is
also evidence that NPY-IR dendrites are inner-
vated by perforant pathway axon terminals in the
outer molecular layer (Deller and Leranth, 1990).
Commissurotomy revealed direct commissural in-
put to NPY-IR dendrites at the border of the inner
and outer molecular layer and in the hilus.
Although the vast majority of NPY-IR neurons
appear to be local circuit neurons, 2% of hilar
NPY-IR neurons seem to project to the contra-
lateral hippocampus, because they become retro-
gradely labeled when horseradish peroxidase
(coupled to wheat germ agglutinin) is injected into
the contralateral hippocampus (Deller and Lera-
nth, 1990). These authors also report that sym-
metric NPY-IR synapses are found on the cell
bodies and dendrites of hilar neurons.
In primates (humans and monkeys) the distri-
bution of NPY-IR neurons is similar to the rat
(Chan-Palay et al., 1986; Kohler et al., 1986). The
innervation of the outer molecular layer by hilar
neurons appears to be, however, more prominent
than in the rat. In the monkey, a small portion of
NPY-IR hilar interneurons contains parvalbumin
(Nitsch and Leranth, 1991). In humans, unlike
naı̈ve rats, faint NPY-IR is also observed in mossy
fiber terminals (Chan-Palay et al., 1986; Furtinger
et al., 2001).
Neuropeptide Y receptors in the fascia dentata
Several types of NPY receptors (Y1, Y2, Y4, Y5)
have been identified in the brain (Dumont et al.,

1992; Parker and Herzog, 1999). In the hippo-
campus of rats, mice, and humans, Y1 and Y2
receptors are the most abundant. Y1 receptor
mRNA is prominently expressed in granule cells
and correspond to Y1 receptor binding sites on
granule cell dendrites in the molecular layer,
where they receive their input from hilar NPY
interneurons (Dumont et al., 1998; Gackenheimer
et al., 2001). Expression of Y1 receptors on other
cells of the dentate gyrus (e.g., hilar interneurons)
has not unequivocally been demonstrated so far,
although electrophysiological results strongly
suggest their presence (below). Interestingly
NPY seems to augment proliferation of progen-
itor cells located in the subgranular zone (SGZ)
by activation of Y1 receptors, a mechanism re-
duced in Y1 receptor knock out mice (Howell
et al., 2005). This suggests that Y1 receptors
are also present on dentate granule cell (DGC)
progenitors.
The site of Y2 receptor expression seems to be in
part species-dependent (Dumont et al., 1998). In
all species investigated so far, Y2 receptor mRNA
is highly expressed in pyramidal cells. Whereas in
most rat strains, notably in Sprague-Dawley rats,
no Y2 message was demonstrated in DGCs,
whereas Wistar rats and most mouse strains seem
to express Y2 mRNA there (Parker and Herzog,
1999; Wolak et al., 2003; Kishi et al., 2005). Cu-
riously, there is strong evidence that mossy fiber
terminals in Sprague-Dawley rats express highly
functional Y2 receptors (McQuiston and Colmers,
1996), and some evidence exists for Y2 receptor-
mediated postsynaptic actions in some DGCs
(McQuiston et al., 1996). In humans, prominent
expression of Y2 mRNA was demonstrated in
granule cells, and Y2 receptor binding was found
in the terminal areas of mossy fibers in the dentate
hilus and stratum lucidum of humans but not rats.
Recent studies using antibodies to Y1 and Y2 re-
ceptors support the findings using receptor auto-
radiography (Stanic et al., 2006). In addition,
Stanic et al. (2006) provided evidence that Y2 re-
ceptors may also be located on hilar interneurons,
indicating that these receptors may also mediate
inhibition of GABA release from interneurons in
addition to suppressing glutamate release at mossy
fibers.
287
Electrophysiological effects of NPY on dentate
neurons
Mossy fiber terminals
As it does in area CA1 (Colmers et al., 1987, 1988;
Colmers and Bleakman, 1994), NPY also nega-
tively modulates synaptic transmission at the
mossy fiber-to-CA3 synapse. Synaptic excitation,
evoked by stimulation of the mossy fiber pathway,
is potently and reversibly inhibited by bath appli-
cation of NPY in a concentration-dependent man-
ner (Klapstein and Colmers, 1993; Guo et al.,
2002; El Bahh et al., 2005). The mechanism
underlying this NPY-mediated inhibition of gluta-
mate release from mossy fibers was determined by
quantal synaptic analysis. Measurements of the
frequency and amplitude of spontaneous (sEPSC)
and miniature (mEPSC) excitatory postsynaptic
currents were made in CA3 pyramidal neurons in
the whole-cell patch configuration. Application of
NPY and the Y2-preferring agonist [ahx5 24]NPY
increased the interval and decreased the amplitude
of sEPSCs (Fig. 1) but had no effect on either pa-
rameter when mEPSCs were measured in the pres-
ence of the voltage-dependent Na+ channel
blocker, tetrodotoxin (McQuiston and Colmers,
1996). This indicates that Y2 receptors on mossy
fiber terminals inhibit glutamate release and
decrease excitatory input into CA3 pyramidal
neurons. Subsequent experiments in area CA1
determined that NPY inhibits glutamate release
via a suppression of presynaptic calcium influx
(Qian et al., 1997). Controversy developed because
it was unclear whether the Y5 receptor also played
a role in this inhibition (Guo et al., 2002). How-
ever, in experiments at the mossy fiber-CA3 and
stratum radiatum-CA1 synapse in hippocampal
slices, presynaptic inhibitory responses to the
application either of the Y2-preferring agonist
[ahx5 24]NPY, or the Y5-preferring agonist,
AlaAib NPY, were both blocked entirely by pre-
treatment with the potent and selective Y2 recep-
tor antagonist, BIIE0246 (El Bahh et al., 2005). As
this antagonist has no activity at the Y5 receptor
(Doods et al., 1999), this indicates that the Y2
receptor is the only one mediating this action
of NPY, and that at high concentrations, the
288

a.
Control NPY (1 µM) Washout

50 pA

500 msec

b. c.

NPY
Control
300 pA

Control NPY13-36

20 msec

d. e.

[Leu31Pro34] - NPY
PYY
Control
Control

Fig. 1. Y2- and Y1-mediated effects in the dentate gyrus. (a) Spontaneous synaptic currents recorded in CA3 pyramidal neuron.
Sweeps in each column represent consecutive traces from neuron in control (left), in the presence of 1 mM NPY (center) and after
60 min washout (right). Note the reduction in frequency of excitatory postsynaptic currents (downward deflections) with NPY, and
increase upon washout. Subsequent experiments with selective agonists in this paper demonstrate the Y2 nature of this response
(adapted from McQuiston and Colmers, 1996, with permission). (b–e) Ca2+currents in acutely isolated rat dentate granule cells are
inhibited by agonists at Y1 receptors. Shown superimposed are current traces evoked in this neuron in control and in the presence of
the respective agonist. NPY and PYY are pan-agonists, while Leu31, Pro34 NPY is a Y1/Y5-preferring agonist, and NPY13 36 is a Y2/
Y5-preferring agonist. In 30% of neurons tested, a Y2 receptor agonist also inhibited the Ca2+current, while the Y1 agonist was
effective in every neuron tested (adapted from McQuiston et al., 1996, with permission. r 1996, by the Society for Neuroscience).

Y5-preferring agonist has some activity at Y2 re-
ceptors (El Bahh et al., 2005). This conclusion is
further supported by the observation that in hip-
pocampal slices prepared from mutant mice lack-
ing functional Y2 receptors, there is no inhibition
of either the mossy fiber or stratum radiatum-
evoked field excitatory postsynaptic potentials
(EPSPs) by NPY or the Y5-preferring agonist
AlaAib NPY (El Bahh et al., 2005). The Y2
receptor is therefore responsible for NPY-medi-
ated inhibition of glutamate release from mossy
fiber terminals, most likely by suppression of volt-
age-dependent Ca2+ influx through presynaptic
calcium channels.
Actions of NPY at somata and dendrites of dentate
granule cells (DGCs)
In the dentate molecular layer, the effect of NPY
on EPSPs is, if at all detectable, certainly far less
pronounced. Stimulation of the perforant path or
commissural inputs evokes EPSPs in the molecular
layer that are either unaltered (Klapstein and Col-
mers, 1993), or show very minor inhibition (Bijak
and Smialowska, 1995) with bath application of
NPY. This apparent absence of NPY actions
within the dentate gyrus itself was puzzling, given
the significant levels of NPY receptor expression
previously reported in this region (Chan-Palay
et al., 1986; Kohler et al., 1986), and prompted
further investigations.
Using patch clamp recording and simultaneous
calcium imaging, McQuiston et al. (1996) evoked
action potentials in rat DGCs in brain slices. Bath
application of NPY did not alter resting Ca2+lev-
els in the soma or dendrites of the DGCs, but did
significantly and reversibly decrease the depolari-
zation-induced Ca2+ influx in the soma and dend-
rites of these same DGCs. Because voltage clamp
conditions are less than ideal in neurons with their
dendritic trees intact, as occurs in brain slices,
mechanistic studies of this postsynaptic NPY
action were undertaken in acutely isolated DGCs,
which are electrically more tractable. Somatic
Ca2+ currents were isolated pharmacologically
and NPY and receptor subtype-preferring agonists
were applied, as were selective blockers of different
289
Ca2+channel subtypes. Calcium currents were
inhibited by the NPY Y1- and Y5-preferring
agonist, [Leu31Pro34]NPY but were much less
commonly inhibited by the NPY Y2- (and Y5)-
preferring agonist, NPY13-36 (Fig. 1). These ob-
servations were consistent with the actions of NPY
on Ca2+ influx preferentially occurring via Y1
receptor activation. Selective blockade of N-type
Ca2+ channels occluded all actions of NPY,
consistent with an action of Y1 receptors to selec-
tively suppress current through this subtype of
voltage-dependent Ca2+channel (VDCC)
(McQuiston et al., 1996). Functionally, the inhibi-
tion of calcium currents by NPY was postulated to
regulate the release of dynorphin from DGC dend-
rites, which is mediated by L- and N-type VDCCs
(Simmons et al., 1995), but this has not been
investigated further. Certainly, the Y1 receptor
does mediate the inhibition of N-type VDCCs in
DGCs. Given the remarkable amounts of NPY
in the dentate, particularly in the molecular layer,
it is reasonable to presume that there will be some
significant physiological consequences of Y1 re-
ceptor activation on the dendrites and soma of
DGCs.
NPY effects on hilar interneurons
The presence of NPY receptors in the hilus is
controversial, with some studies showing no
immunoreactivity (Dumont et al., 1998; Gack-
enheimer et al., 2001) and others demonstrating
their existence (Kopp et al., 2002; Paredes et al.,
2003). Electrophysiological recordings of hilar in-
terneurons in the mouse hippocampus performed
by Paredes et al. (2003) demonstrated the pres-
ence of a G-protein-coupled inwardly rectifying
potassium current (GIRK) that was activated by
NPY in 8/17 dentate hilar interneurons studied.
Furthermore, the Y1- and Y5-preferring agonist,
[Leu31Pro34]NPY mimicked the effects of NPY,
also in about half the neurons tested, consistent
with a Y1 receptor-mediated modulation of
GIRK currents. This is similar to a Y1 recep-
tor-mediated action described in reticular and
ventrobasal nucleus neurons of the thalamus (Sun
et al., 2001a, b). Paredes et al. (2003) also used
290
immunocytochemisty to determine that NPY-
containing interneurons of the dentate hilus
contain Y1 receptors, consistent with their elect-
rophysiological findings. These observations sug-
gest that a significant subpopulation of hilar
interneurons, similar in percentage to that ex-
pressing NPY itself, are inhibited by NPY. This
inhibition of inhibitory interneurons is a potential
explanation for the proconvulsant actions of Y1
agonists in the dentate in vivo (Gariboldi et al.,
1998).
NPY effects on long-term potentiation (LTP)
LTP is reliably evoked with high frequency stim-
ulation of the perforant pathway in the dentate
gyrus in vivo. Induction of perforant pathway
LTP occurs as a result of DGC dendrites under-
going an NMDA receptor-mediated AMPA re-
ceptor insertion that is responsible for the
transformation of ‘silent synapses’ to active ones
(Bi and Poo, 1998; Lin et al., 2006; Moga et al.,
2006). Whittaker et al. (1999) injected NPY in-
tracerebroventricularly (ICV) prior to LTP in-
duction in vivo. This resulted in an inhibition of
the induction and maintenance of perforant path
LTP. The previously described Y1 receptor-me-
diated inhibition of dendritic calcium currents
(McQuiston et al., 1996) must be considered a
potential candidate postsynaptic mechanism for
the observed inhibition of LTP. Alternatively, Y1
receptor activation may also alter downstream
effectors necessary for LTP induction, such as
calcium-calmodulin kinase II, protein kinase C or
mitogen activated/extracellular –signal-regulated
kinase (Lin et al., 2006). An alternative interpre-
tation is that NPY inhibits perforant pathway
glutamate release on to dendrites of DGCs (Whit-
taker et al., 1999). However, there is scant evi-
dence for any presynaptic actions of NPY on
perforant path inputs, as discussed above (Klap-
stein and Colmers, 1993). Therefore, the possibil-
ity exists that NPY may inhibit perforant
pathway LTP by reducing postsynaptic Ca2+in-
flux through its documented action at N-type
VDCCs, or by inhibiting a downstream signal
effector.
NPY effects on synaptosomes
Experiments using synaptosomes to examine NPY
function have produced interesting results and in-
terpretations. Synaptosomes, a biochemical prep-
aration of isolated synaptic terminals, are
commonly used for the analysis of synaptic re-
lease. The application of a high extracellular K+
solution stimulates synaptosomal glutamate re-
lease, which can be quantitated fluorimetrically.
This preparation can also be used to study presy-
naptic calcium influx, using fluorescent calcium
indicators. Preincubation of synaptosomes with
the compound of interest allows testing of possible
effects on transmitter release. The release of gluta-
mate from synaptosomes prepared from the dent-
ate gyrus is significantly reduced by preincubation
of the Y1- and Y5-preferring agonist [Leu31-
Pro34]NPY and by the Y2- and Y5-preferring ago-
nist NPY13-36 (Silva et al., 2001). These effects
occurred both in the presence of normal and very
low levels of extracellular Ca2+, indicating that the
mechanism is Ca2+-independent. Thus, in syna-
ptosomes prepared from the rat dentate gyrus, Y1
and Y2 receptors appear to mediate suppression of
glutamate release. With the recent discovery of the
Y5 receptor and Y5 selective agonists, the Y5 re-
ceptor has also been investigated in the dentate
gyrus. As is the case with Y1 and Y2 agonists, the
Y5 agonist NPY (19-23)-(Gly(1),Ser(3),Gln(4),-
Thr(6),Ala(31), Aib(32),Gln(34))-pancreatic PP
also inhibits P/Q-type VDCC-mediated glutamate
release from synaptosomes (Silva et al., 2003).
Interestingly, despite the significant actions of Y1-
and Y5-preferring agonists in this preparation,
neither the Y1 receptor antagonist BIBP3226 nor
the Y5 receptor antagonist L-152,804 are able to
block the inhibition mediated by these agonists
The Y2 antagonist BIIE0246, on the other hand,
potently blocks the inhibition of glutamate release,
which has lead to the suggestion that Y2 receptors
are functionally coupled to Y1 and Y5 receptors
and can override their modulatory effects (Silva et
al., 2003). However, an alternate interpretation is
suggested by the way the experiments in this paper
were conducted: the investigators added the ago-
nist and antagonists at the same time, spun the

synaptosomes, then reapplied the mixture when
resuspending them. Because the agonist peptides
are much larger and ‘‘stickier’’ molecules than are
the antagonists, it is possible that more of the
agonists remained behind when the supernatant
was removed, and preferentially occupied the re-
ceptors, giving an impression that the antagonists
were ineffective. This would not be the case for the
Y2 receptor antagonist, BIIE0246, which exhibits
an irreversible form of antagonism upon pro-
longed exposure to the receptor (Dautzenberg &
Neysari, 2005; El Bahh et al., 2005). It might be
possible to test this hypothesis by pretreating the
synaptosomes with the antagonist alone, then re-
suspending them after centrifugation in the pres-
ence of both antagonist and agonist.
Notwithstanding this possibility, it appears that
in synaptosomes prepared from the dentate gyrus,
Y1, Y2, and Y5 agonists can inhibit glutamate re-
lease. Moreover, the most prominent inhibition
occurs with the Y2 agonist that likely mediates the
inhibitory effect (Silva et al., 2003). These results
differ from those from the electrophysiological
studies which suggest that NPY-mediated inhibi-
tion of glutamate release does not occur in the
dentate gyrus (Klapstein and Colmers, 1993;
Mcquiston and Colmers, 1996; Mcquiston et al.,
1996; El Bahh et al., 2002, 2005). A number of
potential differences in methods may explain these
differences. First, while the electrophysiological
studies represent the evoked or spontaneous ac-
tivity in defined neurons and synaptic pathways,
synaptosome preparations contain all terminals
that survive isolation, including inputs to inter-
neurons, terminals from other pathways, etc. in
addition to the perforant path inputs studied. Fur-
thermore, mossy fiber terminals, which have mul-
tiple release sites (Lawrence and McBain, 2003)
appear to respond to Y2 receptor agonists
(McQuiston and Colmers, 1996), and although
there is some evidence for the presence of Y5
receptors in the dentate gyrus (Guo et al., 2002),
this is disputed (El Bahh et al., 2005). However,
the presence of Y2 receptors on the terminals
of mossy fibers may be sufficient to explain
the observed inhibition of glutamate release in
synaptosomes.
291
NPY effects on neurogenesis
In the last decade, it has become clear that neural
precursor cells in the SGZ of the dentate gyrus
supply new granule cells on an ongoing basis (see
chapter by Parent, this volume). In the SGZ, neu-
ronal precursors proliferate and continually mi-
grate into the granule cell layer where they mature
and become functional granule cells. The specific
factors that influence dentate gyrus neurogenesis
are, therefore, of significant interest because their
impact will modify the memory formation process
(Aimone et al., 2006). Exercise, growth factors,
environmental enrichment, aging, hormones such
as estrogen and prolactin, glutamatergic neuro-
transmission, adrenal hormones, LTP, lesions, sei-
zure activity, and the presence of NPY are all
associated with a regulation of dentate neurogen-
esis (cf. Parent, this volume). In NPY knockout
mice, it was fortuitously observed that prolifera-
tion of olfactory precursor cells was impaired, and
NPY was seen to promote neurogenesis there, via
a Y1 receptor mechanism (Hansel et al., 2001).
Howell et al. (2005) examined whether NPY might
affect proliferation of neural progenitors in the
SGZ by culturing neuroblast precursor cells from
the early postnatal dentate gyrus in the absence or
presence of NPY in the culture medium. NPY
treatment significantly increased the number of
neurons and neuroblasts that incorporated Brd-U,
consistent with a proliferative action of NPY. The
strongly Y1-preferring agonist, F7,P34 NPY mim-
icked the effect of NPY application in vitro, and
the effect of NPY was blocked by application of
the Y1 antagonist BIBP3226 in the presence of
NPY. The neuroproliferative action of NPY was
also abolished by inhibiting the extracellular sig-
nal-regulated kinase (ERK)1/2, a subgroup of mi-
togen activated kinases (Scharfman et al., 2000).
Therefore, it is likely that a downstream effect of
the Y1 receptor is the activation of ERK1/2. These
authors then validated the observations in Y1
receptor knockout mice in vivo. Y1 / mice had a
significantly lower SGZ proliferation rate and a de-
creased number of neurons expressing doublecor-
tin than in wild types. Thus, data from in vitro and
in vivo models supports a Y1 receptor-mediated
292
neuroproliferative action of NPY in the hilus of
the dentate gyrus. This may also play a role in
depression (below).
Neuropeptide Y and its receptors in disease
Epilepsy
In animal models of epilepsy and in human
temporal lobe epilepsy, the NPY system under-
goes considerable changes in the dentate gyrus,
both in the expression of the peptide and of its
receptors (Vezzani et al., 1999). With repeated,
mildly convulsive stimuli, such as kindling,
expression of NPY and its Y2 receptors become
up-regulated in hilar interneurons (Fig. 2) (Gruber
et al., 1994; Vezzani et al., 1999), and NPY ex-
pression is up-regulated in DGCs (Marksteiner
et al., 1990; Sperk et al., 1992; Goodman and
Sloviter, 1993; Causing et al., 1996). Viral vectors
inducing overexpression of NPY have clear anti-
convulsive effects when infused locally into the
hippocampus of rats (Richichi et al., 2004). On the
other hand, prolonged seizures (e.g., during status
epilepticus) cause degeneration of hilar interneu-
rons (Sloviter, 1983; Sperk et al., 1992). However,
surviving neurons may still up-regulate NPY
expression. In rat models of temporal lobe epi-
lepsy, expression of Y2 receptors is up-regulated
simultaneously in granule cells and mossy fibers
(Fig. 2) (Schwarzer et al., 1998). Over-expression
of NPY and its Y2 receptors seems to have a
protective and anticonvulsive role caused by the
Y2-mediated presynaptic inhibition of glutamate
release from mossy fibers (Klapstein and Colmers,
1993; Greber et al., 1994; El Bahh et al., 2005).
Conversely, Y1 receptors located on granule cell
dendrites are down-regulated with seizure activity
(Kofler et al., 1997). Because Gariboldi et al.
(1998) demonstrated that Y1 receptor agonists in-
jected into the dentate gyrus have proconvulsant
action in vivo, this change could ameliorate seizure
activity or compensate for the increased excitation
(see also below). The prominent anticonvulsive
action of NPY has been suggested to be mediated
by Y5 receptor stimulation (Woldbye et al., 1997).
This has been debated, however, especially since
expression of Y5 receptors seems to be negligible,
at least in the mouse hippocampus and a crucial
role of Y2 receptors has been shown in the anti-
convulsive action of NPY (El Bahh et al., 2005;
see also below).
In hippocampal tissue surgically removed from
patients for treatment of temporal lobe epilepsy,
enhanced expression of NPY mRNA has been
observed in interneurons of the dentate hilus,
together with a markedly increased total length of
NPY-positive fibers in the inner and outer dentate
molecular layer, and the stratum lacunosum-mole-
culare (Furtinger et al., 2001). Interestingly, the
pattern of interneuronal axon sprouting over-
lapped in part with terminal areas both of normal
mossy fibers and the collaterals that sprout into
the inner molecular layer in the epileptic brain
(mossy fiber sprouting). In dentate gyrus of epi-
lepsy patients, Y2 receptors were up-regulated and
Y1 receptors down-regulated in mossy fibers and
in the molecular layer, respectively (Furtinger
et al., 2001). This indicates that NPY, released
from sprouted interneurons could also reach Y2
receptors located on mossy fibers. In epileptic rats,
NPY has been shown to have tonic inhibitory ac-
tions, mediated via Y2 receptors (Tu et al., 2005).
These changes in NPY-related neuronal circuitry
may be the basis for the anticonvulsive action of
NPY demonstrated ex vivo in hippocampal tissue
obtained from patients with temporal lobe epilepsy
(Patrylo et al., 1999).
Alzheimer’s disease
Chan-Palay et al. (1986) observed a marked
reduction in hilar NPY neurons in Alzheimer’s
Disease (AD) patients at autopsy. Assuming that
the degeneration of the septo-hippocampal path-
way may be related to the cognitive impairment in
AD patients, it is interesting that lesions of the
septo-hippocampal pathway result in an initial in-
crease in NPY-IR in the dentate hilus (Hortnagl
et al., 1990; Bayer and Milner, 1993) and a sub-
sequent loss of hilar NPY neurons (Milner et al.,
1997). It is possible that the initial increase NPY-
IR may reflect sustained stimulation of these
neurons leading then to their loss through
 293

Fig. 2. NPY and its receptors in control and epileptic rats. NPY-IR is present in numerous interneurons of the dentate gyrus that
project to the outer molecular layer and have collaterals within the dentate hilus (see the respective labeling in a, representing a high
magnification image of NPY-IR in the dorsal dentate gyrus). In epileptic rats (b; 30 days after kainic acid-induced seizures; lower
magnification image of the dorsal hippocampus) varying portions of hilar NPY neurons degenerate (note the loss in NPY-IR in the
outer molecular layer). After seizures mossy fibers strongly express NPY-IR unlike in naı̈ve animals. The arrow in b marks sprouted
mossy fiber terminals in the inner molecular layer also containing NPY. (c) Y1 receptors are postsynaptic receptors located mainly on
dendrites of granule cells (here labeled with the Y1/Y5 ligand 125I-Pro34-PYY) and become reduced after kainic acid-induced seizures
(arrow in d; 8 days after application of kainic acid injection). Panels e and f depict brain sections labeled with the Y2/Y5 specific ligand
125
I-PYY(3 36). Note the strong Y2-specific labeling of strata oriens and radiatum in control rats (e). In kainic acid treated rats strong
labeling of mossy fibers can be seen representing presynaptic Y2 receptors (arrow in f; 8 days after application of kainic acid injection).

excitotoxic mechanisms (Chan-Palay et al., 1986;
Milner et al., 1997).
Depression
In ‘‘depressed’’ Flinder’s Sensitive Line (FSL) rats,
basal expression of NPY and Y1 mRNA in the
dentate gyrus is significantly lower than in control
animals (Jimenez-Vasquez et al., 2006). An up-
regulation of Y1 receptor binding in the dentate
gyrus was also observed, most likely due to an
increased affinity for and/or decreased internali-
zation of NPY receptors (Jimenez-Vasquez et al.,
2006). Electroconvulsive therapy (ECT), a com-
mon treatment for depressive disorders, raises
NPY mRNA levels in the hilus (Mikkelsen et al.,
1994), and Y1 receptor mRNA expression in
DGCs (Madsen et al., 2000) of naı̈ve rats. In
FSL rats, ECT also increases NPY mRNA and Y1
mRNA expression, and reduces Y1 receptor bind-
ing in the dentate gyrus (Jimenez-Vasquez et al.,
2006). This altered NPY expression was measured
after an ECT-induced reduction in depressive
behavior, measured with the Porsolt swim test.
Interestingly, ECT increases seizure threshold in
epileptic rats, which correlates with a reduction in
NPY binding sites in the dentate gyrus (Bolwig
et al., 1999). Fluoxetine, a serotonin-selective re-
uptake inhibitor used in the treatment of depres-
sion, has also been observed to increase NPY
mRNA in the dentate gyrus of Flinders Sensitive
Line rats (Caberlotto et al., 1999), although this
was not replicated in a different lab using tricyclic
294
antidepressants in Sprague-Dawley rats (Bellmann
and Sperk, 1993). Thus, a downregulation of Y1
receptor mRNA is intimately related to the patho-
physiology of depression and is altered by treat-
ments that have been efficacious in treating
depressive disorders, possibly via an increase in
NPY receptor expression.
These effects are not only limited to animal
models. Depressed humans have a decreased
NPY-like immunoreactivity (NPY-LI) in their
cerebrospinal fluid compared to non-depressed
control subjects (Widerlov et al., 1988; Gjerris et
al., 1992). Consistent with the hypothesis that a
hypofunction of the NPY system is related to de-
pression, depressed human subjects exhibit in-
creased NPY-LI after ECT (Mathe, 1999). Lastly,
it appears that neurogenesis in the dentate gyrus
is necessary for the actions of antidepressants
(Dranovsky and Hen, 2006). It is therefore tempt-
ing to speculate that some of the antidepressant
actions of the NPY system may be mediated
through neuroproliferative actions.
Summary
NPY is an extraordinarily abundant peptide in the
dentate gyrus, but the roles that it plays remain
somewhat obscure. It is clear that the dentate
gyrus NPY peptide-receptor system is extremely
sensitive to ongoing activity, and is poised to reg-
ulate many important aspects of the plasticity of
dentate gyrus circuitry. Certainly, the links be-
tween NPY and excitability, memory formation
and depression are important and warrant inten-
sive study.
Abbreviations
AD Alzheimer’s disease
DGC dentate granule cell
ECT electroconvulsive therapy
EPSP excitatory postsynaptic potential
GABA gamma-aminobutyric acid
GIRK G-protein-coupled inwardly-rec-
tifying potassium current
ICV intracerebroventricular
LTP long-term potentiation
MEPSC miniature excitatory postsynaptic
current
NPY neuropeptide Y
NPY-IR neuropeptide Y immunoreactive
NPY-LI neuropeptide Y-like immunore-
activity
PYY peptide YY
PP pancreatic polypeptide
SEPSC spontaneous excitatory postsy-
naptic current
SGZ subgranular zone
VDCC voltage dependent Ca2+ channel
References
Aimone, J.B., Wiles, J. and Gage, F.H. (2006) Potential role for
adult neurogenesis in the encoding of time in new memories.
Nat. Neurosci., 9(6): 723–727.
Bayer, L.E. and Milner, T.A. (1993) Transient increases in ne-
uropeptide Y-like immunoreactivity in dentate hilar neurons
following fimbria/fornix transection. J. Neurosci. Res., 34(4):
434–441.
Bellmann, R. and Sperk, G. (1993) Effects of antidepressant
drug treatment on levels of NPY or prepro-NPY-mRNA in
the rat brain. Neurochem. Int., 22(2): 183–187.
Bi, G.Q. and Poo, M.M. (1998) Synaptic modifications in cul-
tured hippocampal neurons: dependence on spike timing,
synaptic strength, and postsynaptic cell type. J. Neurosci.,
18(24): 10464–10472.
Bijak, M. and Smialowska, M. (1995) Effects of neuropeptide Y
on evoked potentials in the CA1 region and the dentate gyrus
of the rat hippocampal slice. Pol. J. Pharmacol., 47(4):
333–338.
Bolwig, T.G., Woldbye, D.P. and Mikkelsen, J.D. (1999) Elec-
troconvulsive therapy as an anticonvulsant: a possible role of
neuropeptide Y (NPY). J. ECT, 15(1): 93–101.
Caberlotto, L., Jimenez, P., Overstreet, D.H., Hurd, Y.L., Ma-
the, A.A. and Fuxe, K. (1999) Alterations in neuropeptide Y
levels and Y1 binding sites in the Flinders Sensitive Line rats,
a genetic animal model of depression. Neurosci. Lett., 265(3):
191–194.
Causing, C.G., Makus, K., Ma, Y., Miller, F.D. and Colmers,
W.F. (1996) Selective upregulation of Ta1 a-tubulin and Ne-
uropeptide Y mRNAs after intermittent excitatory stimula-
tion in adult rat hippocampus in vivo. J. Comp. Neurol.,
367(1): 132–146.
Chan-Palay, V., Kohler, C., Haesler, U., Lang, W. and
Yasargil, G. (1986) Distribution of neurons and axons
immunoreactive with antisera against neuropeptide Y in the
normal human hippocampus. J. Comp. Neurol., 248(3):
360–375.
Colmers, W.F. and Bleakman, D. (1994) Neuropeptide Y
effects on the electrical properties of neurons. Trends Ne-
urosci., 17(9): 373–379.

Colmers, W.F., Lukowiak, K.D. and Pittman, Q.J. (1987) Pre-
synaptic action of neuropeptide Y in area CA1 of the rat
hippocampal slice. J. Physiol., 383: 285–299.
Colmers, W.F., Lukowiak, K. and Pittman, Q.J. (1988) Ne-
uropeptide Y action in the rat hippocampal slice: site and
mechanism of presynaptic inhibition. J. Neurosci., 8(10):
3827–3837.
Dautzenberg, F.M. and Neysari, S. (2005) Irreversible binding
kinetics of neuropeptide Y ligands to Y2 but not to Y1 and
Y5 receptors. Pharmacology, 75(1): 21–29.
Deller, T. and Leranth, C. (1990) Synaptic connections of ne-
uropeptide Y (NPY) immunoreactive neurons in the hilar
area of the rat hippocampus. J. Comp. Neurol., 300(3):
433–447.
Doods, H., Gaida, W., Wieland, H.A., Dollinger, H.,
Schnorrenberg, G., Esser, F., Engel, W., Eberlein, W. and
Rudolf, K. (1999) BIIE0246: a selective and high affinity ne-
uropeptide Y Y(2) receptor antagonist. Eur. J. Pharmacol.,
384(2-3): R3–R5.
Dranovsky, A. and Hen, R. (2006) Hippocampal neurogenesis:
regulation by stress and antidepressants. Biol. Psychiatry,
59(12): 1136–1143.
Dumont, Y., Jacques, D., Bouchard, P. and Quirion, R. (1998)
Species differences in the expression and distribution of the
neuropeptide Y Y1, Y2, Y4, and Y5 receptors in rodents,
guinea pig, and primates brains. J. Comp. Neurol., 402(3):
372–384.
Dumont, Y., Martel, J.C., Fournier, A., St-Pierre, S. and Quir-
ion, R. (1992) Neuropeptide Y and neuropeptide Y receptor
subtypes in brain and peripheral tissues. Prog. Neurobiol.,
38(2): 125–167.
El Bahh, B., Balosso, S., Hamilton, T., Herzog, H., Beck-Sick-
inger, A.G., Sperk, G., Gehlert, D.R., Vezzani, A. and
Colmers, W.F. (2005) The anti-epileptic actions of neuro-
peptide Y in the hippocampus are mediated by Y and not Y
receptors. Eur. J. Neurosci., 22(6): 1417–1430.
El Bahh, B., Cao, J.Q., Beck-Sickinger, A.G. and Colmers, W.F.
(2002) Blockade of neuropeptide Y(2) receptors and suppres-
sion of NPY’s anti-epileptic actions in the rat hippocampal
slice by BIIE0246. Br. J. Pharmacol., 136(4): 502–509.
Elmquist, J.K., Coppari, R., Balthasar, N., Ichinose, M. and
Lowell, B.B. (2005) Identifying hypothalamic pathways con-
trolling food intake, body weight, and glucose homeostasis. J.
Comp. Neurol., 493(1): 63–71.
Everitt, B.J., Hokfelt, T., Terenius, L., Tatemoto, K., Mutt, V.
and Goldstein, M. (1984) Differential co-existence of neuro-
peptide Y (NPY)-like immunoreactivity with catecholamines
in the central nervous system of the rat. Neuroscience, 11(2):
443–462.
Freund, T.F. and Buzsaki, G. (1996) Interneurons of the hip-
pocampus. Hippocampus, 6(4): 347–470.
Furtinger, S., Pirker, S., Czech, T., Baumgartner, C.,
Ransmayr, G. and Sperk, G. (2001) Plasticity of Y1 and
Y2 receptors and neuropeptide Y fibers in patients with tem-
poral lobe epilepsy. J. Neurosci., 21(15): 5804–5812.
Gackenheimer, S.L., Schober, D.A. and Gehlert, D.R. (2001)
Characterization of neuropeptide Y Y1-like and Y2-like
295
receptor subtypes in the mouse brain. Peptides, 22(3):
335–341.
Gariboldi, M., Conti, M., Cavaleri, D., Samanin, R. and
Vezzani, A. (1998) Anticonvulsant properties of BIBP3226, a
non-peptide selective antagonist at neuropeptide Y Y1 re-
ceptors. Eur. J. Neurosci., 10(2): 757–759.
Gjerris, A., Widerlov, E., Werdelin, L. and Ekman, R. (1992)
Cerebrospinal fluid concentrations of neuropeptide Y in de-
pressed patients and in controls. J. Psychiatry Neurosci.,
17(1): 23–27.
Goodman, J.H. and Sloviter, R.S. (1993) Cocaine neurotoxicity
and altered neuropeptide Y immunoreactivity in the rat hip-
pocampus; a silver degeneration and immunocytochemical
study. Brain Res., 616(1–2): 263–272.
Greber, S., Schwarzer, C. and Sperk, G. (1994) Neuropeptide Y
inhibits potassium-stimulated glutamate release through Y2
receptors in rat hippocampal slices in vitro. Br. J. Pharma-
col., 113(3): 737–740.
Gruber, B., Greber, S., Rupp, E. and Sperk, G. (1994) Differ-
ential NPY mRNA expression in granule cells and interneu-
rons of the rat dentate gyrus after kainic acid injection.
Hippocampus, 4(4): 474–482.
Guo, H., Castro, P.A., Palmiter, R.D. and Baraban, S.C. (2002)
Y5 receptors mediate neuropeptide Y actions at excitatory
synapses in area CA3 of the mouse hippocampus. J. Ne-
urophysiol., 87(1): 558–566.
Haas, H.L., Hermann, A., Greene, R.W. and Chan-Palay, V.
(1987) Action and location of neuropeptide tyrosine (Y) on
hippocampal neurons of the rat in slice preparations. J.
Comp. Neurol., 257(2): 208–215.
Hansel, D.E., Eipper, B.A. and Ronnett, G.V. (2001) Neuro-
peptide Y functions as a neuroproliferative factor. Nature,
410(6831): 940–944.
Harfstrand, A., Fuxe, K., Terenius, L. and Kalia, M. (1987)
Neuropeptide Y-immunoreactive perikarya and nerve termi-
nals in the rat medulla oblongata: relationship to cytoarchi-
tecture and catecholaminergic cell groups. J. Comp. Neurol.,
260(1): 20–35.
Hendry, S.H., Jones, E.G., DeFelipe, J., Schmechel, D., Bran-
don, C. and Emson, P.C. (1984) Neuropeptide-containing
neurons of the cerebral cortex are also GABAergic. Proc.
Natl. Acad. Sci. USA, 81(20): 6526–6530.
Holm, I.E., Geneser, F.A. and Zimmer, J. (1992) Somatostatin-
and neuropeptide Y-like immunoreactivity in the dentate
area, hippocampus, and subiculum of the domestic pig. J.
Comp. Neurol., 322(2): 390–408.
Hortnagl, H., Sperk, G., Sobal, G. and Maas, D. (1990)
Cholinergic deficit induced by ethylcholine aziridinium
(AF64A) transiently affects somatostatin and neuropeptide
Y levels in rat brain. J. Neurochem., 54(5): 1608–1613.
Howell, O.W., Doyle, K., Goodman, J.H., Scharfman, H.E.,
Herzog, H., Pringle, A., Beck-Sickinger, A.G. and Gray,
W.P. (2005) Neuropeptide Y stimulates neuronal precursor
proliferation in the post-natal and adult dentate gyrus. J.
Neurochem., 93(3): 560–570.
Jimenez-Vasquez, P.A., Diaz-Cabiale, Z., Caberlotto, L., Bel-
lido, I., Overstreet, D., Fuxe, K. and Mathe, A.A. (2006)
296
Electroconvulsive stimuli selectively affect behavior and ne-
uropeptide Y (NPY) and NPY Y(1) receptor gene expres-
sions in hippocampus and hypothalamus of Flinders
Sensitive Line rat model of depression. Eur. Neuropsychop-
harmacol., 17(4): 298–308.
Kishi, T., Aschkenasi, C.J., Choi, B.J., Lopez, M.E., Lee, C.E.,
Liu, H., Hollenberg, A.N., Friedman, J.M. and Elmquist,
J.K. (2005) Neuropeptide Y Y1 receptor mRNA in rodent
brain: distribution and colocalization with melanocortin-4
receptor. J. Comp. Neurol., 482(3): 217–243.
Klapstein, G.J. and Colmers, W.F. (1993) On the sites of pre-
synaptic inhibition by neuropeptide Y in rat hippocampus in
vitro. Hippocampus, 3(1): 103–111.
Kofler, N., Kirchmair, E., Schwarzer, C. and Sperk, G. (1997)
Altered expression of NPY-Y1 receptors in kainic acid in-
duced epilepsy in rats. Neurosci. Lett., 230(2): 129–132.
Kohler, C., Eriksson, L., Davies, S. and Chan-Palay, V. (1986)
Neuropeptide Y innervation of the hippocampal region in the
rat and monkey brain. J. Comp. Neurol., 244(3): 384–400.
Kopp, J., Xu, Z.Q., Zhang, X., Pedrazzini, T., Herzog, H.,
Kresse, A., Wong, H., Walsh, J.H. and Hokfelt, T. (2002)
Expression of the neuropeptide Y Y1 receptor in the CNS of
rat and of wild-type and Y1 receptor knock-out mice. Focus
on immunohistochemical localization. Neuroscience, 111(3):
443–532.
Lawrence, J.J. and McBain, C.J. (2003) Interneuron diversity
series: containing the detonation–feedforward inhibition in
the CA3 hippocampus. Trends Neurosci., 26(11): 631–640.
Lin, Y.W., Yang, H.W., Wang, H.J., Gong, C.L., Chiu, T.H.
and Min, M.Y. (2006) Spike-timing-dependent plasticity at
resting and conditioned lateral perforant path synapses on
granule cells in the dentate gyrus: different roles of N-methyl-
D-aspartate and group I metabotropic glutamate receptors.
Eur. J. Neurosci., 23(9): 2362–2374.
Lundberg, J.M., Terenius, L., Hokfelt, T. and Goldstein, M.
(1983) High levels of neuropeptide Y in peripheral nor-
adrenergic neurons in various mammals including man. Ne-
urosci. Lett., 42(2): 167–172.
Madsen, T.M., Greisen, M.H., Nielsen, S.M., Bolwig, T.G. and
Mikkelsen, J.D. (2000) Electroconvulsive stimuli enhance
both neuropeptide Y receptor Y1 and Y2 messenger RNA
expression and levels of binding in the rat hippocampus.
Neuroscience, 98(1): 33–39.
Mathe, A.A. (1999) Neuropeptides and electroconvulsive treat-
ment. J. ECT, 15(1): 60–75.
Marksteiner, J., Ortler, M., Bellmann, R. and Sperk, G. (1990)
Neuropeptide Y biosynthesis is markedly induced in mossy
fibers during temporal lobe epilepsy of the rat. Neurosci.
Lett., 112(2-3): 143–148.
McQuiston, A.R. and Colmers, W.F. (1996) Neuropeptide Y2
receptors inhibit the frequency of spontaneous but not min-
iature EPSCs in CA3 pyramidal cells of rat hippocampus. J.
Neurophysiol., 76(5): 3159–3168.
McQuiston, A.R., Petrozzino, J.J., Connor, J.A. and Colmers,
W.F. (1996) Neuropeptide Y1 receptors inhibit N-type cal-
cium currents and reduce transient calcium increases in rat
dentate granule cells. J. Neurosci., 16(4): 1422–1429.
Mikkelsen, J.D., Woldbye, D., Kragh, J., Larsen, P.J. and
Bolwig, T.G. (1994) Electroconvulsive shocks increase the
expression of neuropeptide Y (NPY) mRNA in the piriform
cortex and the dentate gyrus. Brain Res. Mol. Brain Res.,
23(4): 317–322.
Milner, T.A., Wiley, R.G., Kurucz, O.S., Prince, S.R. and
Pierce, J.P. (1997) Selective changes in hippocampal neuro-
peptide Y neurons following removal of the cholinergic septal
inputs. J. Comp. Neurol., 386(1): 46–59.
Moga, D.E., Shapiro, M.L. and Morrison, J.H. (2006) Bidi-
rectional redistribution of AMPA but not NMDA receptors
after perforant path simulation in the adult rat hippocampus
in vivo. Hippocampus, 16(11): 990–1003.
Morris, B.J. (1989) Neuronal localisation of neuropeptide Y
gene expression in rat brain. J. Comp. Neurol., 290(3):
358–368.
Nitsch, R. and Leranth, C. (1991) Neuropeptide Y (NPY)-
immunoreactive neurons in the primate fascia dentata; occa-
sional coexistence with calcium-binding proteins: a light and
electron microscopic study. J. Comp. Neurol., 309(4):
430–444.
Paredes, M.F., Greenwood, J. and Baraban, S.C. (2003) Ne-
uropeptide Y modulates a G protein-coupled inwardly rec-
tifying potassium current in the mouse hippocampus.
Neurosci. Lett., 340(1): 9–12.
Parker, R.M. and Herzog, H. (1999) Regional distribution of
Y-receptor subtype mRNAs in rat brain. Eur. J. Neurosci.,
11(4): 1431–1448.
Patrylo, P.R., van den Pol, A.N., Spencer, D.D. and William-
son, A. (1999) NPY inhibits glutamatergic excitation in the
epileptic human dentate gyrus. J. Neurophysiol., 82(1):
478–483.
Qian, J., Colmers, W.F. and Saggau, P. (1997) Inhibition of
synaptic transmission by neuropeptide Y in rat hippocampal
area CA1: modulation of presynaptic Ca2+ entry. J. Neuro-
sci., 17(21): 8169–8177.
de Quidt, M.E. and Emson, P.C. (1986) Distribution of neuro-
peptide Y-like immunoreactivity in the rat central nervous
system–II. Immunohistochemical analysis. Neuroscience,
18(3): 545–618.
Richichi, C., Lin, E.J., Stefanin, D., Colella, D., Ravizza, T.,
Grignaschi, G., Veglianese, P., Sperk, G., During, M.J. and
Vezzani, A. (2004) Anticonvulsant and antiepileptogenic
effects mediated by adeno-associated virus vector neuropep-
tide Y expression in the rat hippocampus. J. Neurosci.,
24(12): 3051–3059.
Scharfman, H.E., Goodman, J.H. and Sollas, A.L. (2000)
Granule-like neurons at the hilar/CA3 border after status
epilepticus and their synchrony with area CA3 pyramidal
cells: functional implications of seizure-induced neurogenesis.
J. Neurosci., 20(16): 6144–6158.
Schwarzer, C., Kofler, N. and Sperk, G. (1998) Up-regulation
of neuropeptide Y-Y2 in an animal model of temporal lobe
epilepsy. Mol. Pharmacol., 53(1): 6–13.
Silva, A.P., Carvalho, A.P., Carvalho, C.M. and Malva, J.O.
(2001) Modulation of intracellular calcium changes and
glutamate release by neuropeptide Y1 and Y2 receptors in the

rat hippocampus: differential effects in CA1, CA3 and dent-
ate gyrus. J. Neurochem., 79(2): 286–296.
Silva, A.P., Carvalho, A.P., Carvalho, C.M. and Malva, J.O.
(2003) Functional interaction between neuropeptide Y re-
ceptors and modulation of calcium channels in the rat hip-
pocampus. Neuropharmacology, 44(2): 282–292.
Simmons, M.L., Terman, G.W., Gibbs, S.M. and Chavkin, C.
(1995) L-type calcium channels mediate dynorphin neuro-
peptide release from dendrites but not axons of hippocampal
granule cells. Neuron, 14(6): 1265–1272.
Sloviter, R.S. (1983) ‘‘Epileptic’’ brain damage in rats induced
by sustained electrical stimulation of the perforant path. I.
Acute electrophysiological and light microscopic studies.
Brain Res. Bull., 10(5): 675–697.
Smialowska, M. (1988) Neuropeptide Y immunoreactivity in
the locus coeruleus of the rat brain. Neuroscience, 25(1):
123–131.
Smith, Y. and Parent, A. (1986) Neuropeptide Y-immunoreac-
tive neurons in the striatum of cat and monkey: morpholog-
ical characteristics, intrinsic organization and co-localization
with somatostatin. Brain Res., 372(2): 241–252.
Smith, Y., Parent, A., Kerkerian, L. and Pelletier, G. (1985)
Distribution of neuropeptide Y immunoreactivity in the ba-
sal forebrain and upper brainstem of the squirrel monkey
(Saimiri sciureus). J. Comp. Neurol., 236(1): 71–89.
Sperk, G., Marksteiner, J., Gruber, B., Bellmann, R., Mahata,
M. and Ortler, M. (1992) Functional changes in neuropeptide
Y- and somatostatin-containing neurons induced by limbic
seizures in the rat. Neuroscience, 50(4): 831–846.
Stanic, D., Brumovsky, P., Fetissov, S., Shuster, S., Herzog, H.
and Hokfelt, T. (2006) Characterization of neuropeptide Y2
receptor protein expression in the mouse brain. I. Distribu-
tion in cell bodies and nerve terminals. J. Comp. Neurol.,
499(3): 357–390.
Sun, Q.Q., Akk, G., Huguenard, J.R. and Prince, D.A. (2001a)
Differential regulation of GABA release and neuronal excit-
ability mediated by neuropeptide Y1 and Y2 receptors in rat
thalamic neurons. J. Physiol., 531(1): 81–94.
297
Sun, Q.Q., Huguenard, J.R. and Prince, D.A. (2001b) Neuro-
peptide Y receptors differentially modulate G-protein-acti-
vated inwardly rectifying K+ channels and high-voltage-
activated Ca2+ channels in rat thalamic neurons. J. Physiol.,
531(1): 67–79.
Sundler, F., Moghimzadeh, E., Hakanson, R., Ekelund, M.
and Emson, P. (1983) Nerve fibers in the gut and pancreas
of the rat displaying neuropeptide-Y immunoreactivity.
Intrinsic and extrinsic origin. Cell Tissue Res., 230(3):
487–493.
Tatemoto, K., Carlquist, M. and Mutt, V. (1982) Neuropeptide
Y—a novel brain peptide with structural similarities to pep-
tide YY and pancreatic polypeptide. Nature, 296(5858):
659–660.
Tu, B., Timofeeva, O., Jiao, Y. and Nadler, J.V. (2005) Spon-
taneous release of neuropeptide Y tonically inhibits recurrent
mossy fiber synaptic transmission in epileptic brain. J. Ne-
urosci., 25(7): 1718–1729.
Vezzani, A., Sperk, G. and Colmers, W.F. (1999) Neuropeptide
Y: emerging evidence for a functional role in seizure mod-
ulation. Trends Neurosci., 22(1): 25–30.
Widerlov, E., Lindstrom, L.H., Wahlestedt, C. and Ekman, R.
(1988) Neuropeptide Y and peptide YY as possible cerebro-
spinal fluid markers for major depression and schizophrenia,
respectively. J. Psychiatr. Res., 22(1): 69–79.
Whittaker, E., Vereker, E. and Lynch, M.A. (1999)
Neuropeptide Y inhibits glutamate release and long-term
potentiation in rat dentate gyrus. Brain Res., 827(1-2):
229–233.
Wolak, M.L., DeJoseph, M.R., Cator, A.D., Mokashi, A.S.,
Brownfield, M.S. and Urban, J.H. (2003) Comparative dis-
tribution of neuropeptide Y Y1 and Y5 receptors in the rat
brain by using immunohistochemistry. J. Comp. Neurol.,
464(3): 285–311.
Woldbye, D.P., Larsen, P.J., Mikkelsen, J.D., Klemp, K.,
Madsen, T.M. and Bolwig, T.G. (1997) Powerful inhibition
of kainic acid seizures by neuropeptide Y via Y5-like recep-
tors. Nat. Med., 3(7): 761–764.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 18

Norepinephrine and the dentate gyrus

Carolyn W. Harley

Department of Psychology, Memorial University of Newfoundland, St. John’s, NL, A1B 3X9, Canada

Abstract: Norepinephrine’s role in the dentate gyrus is assessed based on a review of what is known about
its innervation and receptor patterns and its functional effects at both cellular and behavioral levels. The
data support seven hypotheses: (1) Norepinephrine’s functional actions are primarily mediated by b ad-
renoceptors and include electrophysiological enhancement of responses to excitatory input and glycogen-
olytic metabolic support of excitatory synaptic activity. (2) At the cellular level, locus coeruleus burst
release of norepinephrine transiently inhibits feedforward interneurons and either excites or inhibits sub-
populations of feedback interneurons. Consistent with reduced feedforward inhibition, granule cell firing is
transiently increased. Concomitant EEG effects include transient increases in theta power and decreases in
beta and gamma power. (3) Norepinephrine selectively promotes the processing of medial perforant path
spatial input. This effect is mediated both through short- and long-term potentiation of cell excitability and
through delayed potentiation of synaptic input. A critical level of norepinephrine release is required for
long-term effects to norepinephrine alone. Norepinephrine release switches early phase frequency-induced
long-term potentiation of perforant path input to an enduring late phase form and can reinstate decayed
long-term potentiation. Norepinephrine also promotes frequency-induced potentiation of granule cell out-
put at the mossy fiber to CA3 connection. (4) Local increases in norepinephrine accompany glutamate
release and release of other neurotransmitters providing a mechanism for norepinephrine enhancement
effects independent of locus coeruleus firing. (5) Stimuli, such as novelty and reward and punishment,
which activate locus coeruleus neurons, enhance responses to medial perforant path input and engage late
phase frequency-induced long-term potentiation through b adrenoceptor activation. (6) Behavioral studies
are consistent with the mechanistic evidence for a norepinephrine role in promoting learning and memory
and assisting retrieval. (7) The overall profile suggests lower levels of norepinephrine may facilitate pattern
completion or memory retrieval while higher levels would recruit global remapping and promote long-term
episodic memory.

Keywords: locus coeruleus; LTP; LTD; novelty; glycogen metabolism; theta EEG; gamma EEG; global
remapping; feedback inhibition; feedforward inhibition; alpha adrenoceptors; beta adrenoceptors; medial
perforant path; lateral perforant path; synaptic plasticity

The present chapter examines norepinephrine DG cells, and reviews NE’s role in promoting
(NE) innervation and receptor patterns in the long-term plasticity, based primarily on rodent
dentate gyrus (DG), considers NE’s effects on data.

Corresponding author. Tel.: +1 709 737 7974;

Fax: +1 709 737 4000; E-mail: charley@mun.ca

DOI: 10.1016/S0079-6123(07)63018-0 299

300
NE innervation
Blackstad first characterized NE innervation of
the rat hippocampus, reporting the densest NE
innervation in the hilus of the DG, particularly
the subgranular zone (Blackstad et al., 1967). NE
fibers were less common in cell body layers, sug-
gesting axodendritic contacts. Within the molec-
ular layer, the dendritic zone of DG granule cells,
NE fibers were denser in the middle molecular
layer, a target of medial perforant path fibers from
the medial entorhinal cortex, as compared to the
outer molecular layer, the target of lateral ent-
orhinal fibers.
With retrograde tracing, the origin of the DG
NE innervation was identified as multipolar and
fusiform cells in the dorsal part of the dorsal
pontine locus coeruleus (LC) (Haring and Davis,
1985). Electron microscope (EM) images of the
NE fibers in DG reveal small granular vesicles,
indicating NE content in
0.4–1.2 mm diameter
varicosities. Each varicosity contains
20 vesi-
cles, of which 55% can be classified as small
granular vesicles. Large granular vesicles (6%)
and a few mitochondria are also present (Koda
and Bloom, 1977; Milner and Bacon, 1989b). In-
tervaricosity axons are 0.1–0.15 mm in diameter
(Milner and Bacon, 1989b). In 6600 mm2 samples,
the hilus has 1 NE bouton (varicosity) for every
400 boutons of other types; while the granule cell
layer has 1/500, and the molecular layer has 1/
4000 NE boutons. Total NE bouton density
differs among regions. An average of 1500 and
2000 boutons are located in hilar and molecular
layer samples, respectively, while only 500 bou-
tons are present in each granule cell layer sample
(Koda and Bloom, 1977). These estimates of NE
innervation density may be compared to estimates
of 1/8800–1/14,500 boutons for the neocortex
(Lapierre et al., 1973).
The majority of NE axon contacts in DG end on
small dendrites in the subgranular layer where a
third make symmetric synapses, a third form
asymmetric synapses, and a third type make close
associations without specializations. Terminals in
the molecular layer have a similar typology. Thirty
percent of terminals end at spines and do not have
specialized synaptic profiles. On large dendrites
and granule cell somata, symmetric synapses and
appositions are typical (Milner and Bacon, 1989b).
Biochemically, NE content in the hippocampus
is highest in DG (about twice that in CA1) with
higher levels in ventral (600 ng/g) compared to
dorsal DG (360 ng/g) (Loy et al., 1980) or similar
values (500 ng/g) in both regions (Hortnagl et al.,
1991) depending on the study. Quantification of
varicosities in DG (2.4 million/cubic mm) also sug-
gest twice as many varicosities there as in CA1
(Oleskevich et al., 1989); however when divided by
cell number, the ratio of NE varicosities to cells
(20–40/1) is smaller in DG than CA1 (180/1). The
ratio is lowest in the granule cell layer itself (2–4/
1), consistent with the low total number of vari-
cosities in this layer. Seventy percent of the input
from the LC to DG courses through the cingulum,
with the fornix and ventral amygdaloid pathways
each accounting for
15% (Loy et al., 1980).
Developmentally, the NE input, as character-
ized by dopamine-b-hydroxylase immunocyto-
chemistry, is very sparse in the DG at four days
postnatal, and is heaviest at this time point in
CA1. By 10 days postnatal, a band of fibers in
the subgranular zone becomes detectable (Moudy
et al., 1993). By 21 days, a mainly adult-like
pattern is developed.
Adrenoceptor distribution
Both a and b adrenoceptors occur in DG. In an
early study using several methods to assess ad-
renoceptors, Crutcher and Davis (1980) reported b
adrenoreceptors at uniform levels in the dentate
(146 fmol/mg) and hippocampal gyri (178 fmol/
mg), while a1 receptors were more concentrated in
the DG (269 fmol/mg). The uniformity of b ad-
renoceptor distribution led the authors to suggest
these receptors might occur separately from LC
terminals, while a1 adrenoreceptors showed better
correspondence with LC innervation. Both recep-
tors occurred most densely in the synaptosomal/
mitochondrial fractions. The a1 receptor popula-
tion was not changed by 6-OHDA lesions (U’Pri-
chard et al., 1980) suggesting a postsynaptic
localization.

b Adrenoceptors
Milner’s more recent work (Milner et al., 2000)
using both EM and light microscope methodol-
ogies, suggests NE adrenoceptor distribution is
best identified using EM. Her quantified EM
receptor counts show laminar differences in b
adrenoceptors that were not evident in light
microscope studies. Approximately 500 b ad-
renoceptor-immunoreactive profiles per 6000 mm2
sample occur in the middle molecular layer, while
only
250 profiles are seen in the inner and outer
molecular layers, and in the subgranular and hilar
zones;
125 b adrenoceptor-immunoreactive pro-
files/sample occur in the granule cell layer. The
molecular layer distribution shows interesting
parallels with b adrenoceptor physiological
effects (see Physiology, below) and is consistent
with Blackstad’s initial description of NE fiber
innervation.
Granule cell somata express b adrenoceptor
immunoreactivity in the endoplasmic reticulum,
consistent with granule cell production of b
adrenoceptors. A few hilar interneurons, includ-
ing those reactive for parvalbumin, which marks
basket cells (Kosaka et al., 1987; Ribak et al.,
1990, 1992), are also immunoreactive for b ad-
renoceptors.
In the molecular layers, 40–50% of the b ad-
renoceptor reactive profiles are associated with
dendrites, and 50–60% with astrocytes. In the
granule cell layer, 30% are somatic, dendritic and
astrocytic, respectively. In the subgranular region
(an
55 mm zone below the granule cells), 65%
are astrocytic, while 25% are dendritic. Below
the subgranular zone, in the hilar region, 30% of
the b adrenoceptor reactive sites are in axons or
axon terminals and over 50% are in astrocytes.
Some b adrenoceptors are present within parval-
bumin-immunoreactive terminals. Reactive sites
in axons are less common outside the hilar re-
gion, comprising only 2–10% of the b adrenocep-
tor-reactive sites in other layers (Milner et al.,
2000).
Dendritic receptors are associated with postsy-
naptic densities on both large (inner molecular
layer) and small (middle and outer molecular lay-
ers) dendritic profiles. Immunoreactive sites also
301
occur in spines. Axonal b adrenoceptors occur in
both axons and in axon terminals ending on spines
(Milner et al., 2000).
The b adrenoreceptors on astrocytes are usu-
ally next to terminals that make asymmetric
synapses on dendrites and which are thought
to be excitatory inputs. Many of the receptors
on astrocytes are closely apposed to the terminals
that form synapses on spines (Milner et al.,
2000). The astrocytic adrenoceptor distribution,
in close association with synapses, may support
glycogen breakdown at active synaptic sites
(see Physiology, below). Astrocytic b adrenocep-
tors also occur around blood vessels.
NE axons, identified by tryrosine hydroxylase
immunoreactivity, occur close (1–2 mm) to both
astrocytic and dendritic b adrenoceptors, but
direct contacts were not observed (Milner et al.,
2000). The b adrenoceptors in DG are well-
placed to modify granule cell function, and some
interneurons, either directly through postsynap-
tic receptors or indirectly, through effects on
glial processes near synapses. The common
postsynaptic location of b adrenoceptors on the
dendrites of granule cells near, or with, asym-
metric synaptic input and their predominance in
the middle molecular layer, is consistent with a
selective modulation of responses to glut-
amatergic input from the medial entorhinal cor-
tex to the DG (see Physiology, below). Activation
of presynaptic axonal b adrenoceptors and,
possibly, astrocytic b adrenoreceptors, may play
a role in the increase in glutamate release also
reported with b-adrenoceptor activation (see
Physiology, below).
Molecular and binding studies suggest the
majority of b adrenoceptors in the DG are b1
adrenoceptors (Minneman et al., 1979; Rainbow
et al., 1984), but both b1 and b2 adrenoceptors
occur in the molecular layer and hilus (Duncan
et al., 1991; Booze et al., 1993) and granule
cells produce mRNA for both b1 and b2 ad-
renoceptors with a stronger signal from the b2
mRNA probe (Nicholas et al., 1993b). The hip-
pocampus has a high proportion of high-affinity
b adrenoceptors, consistent with a strong func-
tional role for these receptors (Arango et al.,
1990).
302
a Adrenoceptors
Three a1 receptor subtypes are recognized. Specific
probes for each (a1a, a1b, a1d) reveal that granule
cells have an intense mRNA signal for the a1d
subtype (Pieribone et al., 1994; Day et al., 1997),
but do not contain the a1b subtype mRNA. The
a1a subtype is restricted to polymorph cells of the
hilar region (Day et al., 1997). The a1 receptor is
described as half as dense in hippocampus as in
neocortex, despite the fact that the hippocampus
has twice the NE innervation density of neocortex
(Zilles et al., 1991). Zilles reports that a1 receptors
are more concentrated in the inner 2/3rd of the
molecular layer, but that overall a1 receptor den-
sity is similar in molecular, granule and hilar zones
and does not correspond with NE innervation
(both a1a and a1b receptors are represented at
similar levels in contrast to the mRNA data).
Zilles’ a1 pattern differs from earlier work using a
different a1-ligand in which a1-receptor density
closely followed NE fiber density (Jones et al.,
1985). Since neither a2a nor b adrenoceptors are in
clear synaptic association with NE terminals, a1
adrenoceptors are candidates, by default, for sites
of synaptic specialization for tyrosine hydroxylase
fibers. No a1 adrenoceptor subtype distribution
has been described at the EM level in DG.
Milner (Milner et al., 1998) characterized the
a2a receptor at the light microscope and EM levels
and describes its distribution as complementary to
the b adrenoceptors in that the majority of recep-
tors are presynaptic — in axons and axon termi-
nals — rather than postsynaptic. Nonetheless, a2a
receptors exist both pre- and postsynaptically, and
while a2a receptors are implicated in the negative
feedback regulation of NE release, most presy-
naptic profiles are in unmyelinated non-nor-
adrenergic axons.
Granule cells demonstrate immunoreactivity for
a2a-like receptors in association with endoplasmic
reticulum, suggesting granule cell synthesis of a2a
receptors. The a2a-immunoreactive sites are
evenly distributed in the middle and outer molec-
ular layers with about 40% on axons, 30% on as-
trocytes and another 30% on dendrites or spines.
There are about 1/3rd fewer immunoreactive a2a
receptors in the inner molecular layer with only
10% on axons, 50% on dendrites and 40% on glia.
Throughout the molecular layer, a2a receptors on
dendrites are primarily on spines at asymmetric
synapses. Astroctyic a2a receptors are also near
asymmetric synapses. No direct synaptic contacts
between the postsynaptic a2a receptors and tyro-
sine hyroxylase reactive axons are observed. The
granule cell layer has the fewest a2a profiles (2/3rd
less than the molecular layer) and 50% are somatic
profiles. In the subgranular zone, there is a paucity
of dendritic profiles, with mainly axonal (44%)
and glial (46%) profiles. In the hilar region 60%
are in axonal elements. The remaining a2a profiles
are in glia. The axonal profiles throughout are
predominantly in non-noradrenergic axons. A
small subpopulation of hilar interneurons, with
characteristics of the somatostatin interneuron
subtype, contained a2a receptors (Milner et al.,
1998).
In situ studies examining mRNA for the three
a2 receptor subtypes report more of a2c subtype in
DG granule cells than a2a (Nicholas et al., 1993a;
Scheinin et al., 1994) and none of a2b subtype. A
mouse study did find evidence of granule cell a2b
receptor expression, with the most signal in more
mature granule cells (located at the border with the
inner molecular layer) and none in the subgranular
zone (Wang et al., 2002). Developmental studies
report that the a2a mRNA signal is moderate at
1–3 days postnatal and then becomes weak after
maturity (Winzer-Serhan et al., 1997a), while the
a2c signal intensity is high from 11 to 14 days
postnatal, and remains a moderately-intense signal
through adulthood (Winzer-Serhan et al., 1997b).
It would be of interest to know the localization of
a2c receptors.
Physiology
Glycogen metabolism
The astrocytic metabolic enzyme, glycogen phos-
phorylase, which breaks down glycogen to glu-
cose, is strongly activated by NE in DG, as
demonstrated pharmacologically using an a2 ad-
renoceptor antagonist (Fara-On et al., 2005), or
physiologically, using glutamate activation of the

LC (unpublished observations). In 1971, Phelps
found high levels of glycogen accumulation in as-
trocytic processes near synapses in the DG in bar-
biturate-treated mice (Phelps, 1972). Earlier
evidence that NE release depleted glycogen, while
suppression of NE release increased glycogen, led
Phelps to hypothesize that NE controlled glycogen
breakdown in the DG at sites of synaptic activa-
tion through b adrenoceptors. This hypothesis is
consistent with the common association of as-
trocytic b adrenoceptors and asymmetric synaptic
contacts.
At the light microscope level, a patchy, modular
distribution of glycogen phosphorylase, which co-
exists with glycogen, is evident in the DG. Patches
are more numerous during the dark phase of the
daily cycle, when rodents are aroused, and pre-
dominate (
60%) in the middle molecular layer,
with the remaining patches equally divided be-
tween inner and outer layers (Harley and Rusak,
1993). Glycogen phosphorylase activity is reduced
during the theta EEG state, possibly in relation to
a higher overall inhibition in the DG during theta
rhythm (Uecker et al., 1997). NE activation of
glycogen phosphorylase and glycogenolysis in
vitro requires b adrenoceptor coupled cAMP ac-
tivation (Edwards et al., 1974; Nahorski et al.,
1975; Quach et al., 1978). NE also increases as-
trocyte metabolism through a adrenoceptor acti-
vation (Subbarao and Hertz, 1991).
Intracellular recording
Granule cells
Three studies have examined NE effects in vitro.
Haas and Rose found that NE and the b ad-
renoceptor agonist, isoproterenol, produced a small
depolarization of the membrane potential, an at-
tenuated afterhyperpolarization, and a decrease in
accommodation (Haas and Rose, 1987). The at-
tenuated afterhyperpolarization was attributed to a
cAMP-mediated block of Ca++-activated K+ cur-
rents. In a few granule cells, a b adrenoceptor in-
itiated block of the transient outward K+ current
(A-type current) was reported. Blocking the slow
Ca++-activated afterhyperpolarization slows the
303
decay of excitatory postsynaptic currents (EPSCs)
and increases temporal summation in other hippo-
campal principal cells (Lancaster et al., 2001).
Gray and Johnston blocked K+ currents and
showed that b adrenoceptor activation increased
a voltage-dependent Ca++ current, which is ac-
tivated when cells are depolarized to at least
15 mV, and is therefore likely to be L-type cur-
rent (Gray and Johnston, 1987). They showed an
increase in Ca++ channel open times and a
dependence on cAMP elevation. No changes in
the N-type Ca++ current were reported. The a2
agonist, clonidine, was without effect. NE also
produced an increase in the amplitude and dura-
tion of Ca++-dependent action potentials, which
contribute to back-propagating action potentials
and spike-timing dependent plasticity (Dan and
Poo, 2006). Thus, NE might be expected to en-
hance this form of plasticity. Intracellular calcium
is also increased in granule cells by dose-depend-
ent activation of a1 adrenoceptors (Kusaka et al.,
2004).
Lacaille and Schwartzkroin used focal applica-
tion of a b adrenoceptor agonist in hippocampal
slices to evaluate the effects of NE. They replicated
the depolarization of the granule cell membrane
potential (Lacaille and Schwartzkroin, 1988). They
also showed that the depolarization was accom-
panied by an increased input resistance, which
they attributed to a blockade of K+ channels that
are open under normal conditions.
Collectively, one would predict that these b ad-
renoceptor effects would enhance the excitability
and responsivity of granule cells to afferent input.
No other adrenoceptor subtype has been impli-
cated in direct modulation of granule cell mem-
brane characteristics.
Hilar interneurons
Misgeld and Bijak sampled hilar interneurons
in the subgranular zone (Bijak and Misgeld, 1995).
Aspiny interneurons, which are GABAergic and
inhibitory, have strong afterhyperpolarizations
and little spike accommodation. EM studies show
they receive NE input (Milner and Bacon, 1989a).
Spiny hilar neurons, which are glutamatergic and
304
excitatory (mossy cells), have strong accommoda-
tion and little afterhyperpolarization. NE blocked
both aspiny neuron afterhyperpolarizations and
spiny neuron accommodation through b ad-
renoceptors. NE, or a b adrenoceptor agonist,
increased the firing frequency of the excitatory
spiny neurons to current injection, but did not
alter the response of the aspiny inhibitory inter-
neurons to current injection. Spontaneous firing of
both interneuron types increased with NE or b
adrenoceptor activation, as did the frequency of
excitatory postsynaptic potentials (EPSPs) and
inhibitory postsynaptic potentials (IPSPs).
IPSPs increased in neighboring granule cells
with increased activity in inhibitory interneurons.
With glutamate release blocked, GABA-A recep-
tor-mediated IPSPs still increased in granule cells,
but NE no longer increased inhibitory interneuron
firing. a adrenoceptor agonists suppressed
inhibitory interneuron activity. Thus, NE activa-
tion of both b adrenoceptors and a adrenoceptors
either did not change or else suppressed, inhibitory
interneurons. The enhancement of spontaneous
IPSPs, without increased inhibitory interneuron
firing, suggests presynaptic b adrenoceptors
enhance GABA release. Presynaptic b adrenocep-
tors also appear to enhance glutamate release. The
outcome of natural NE release on inhibitory
interneurons, and network inhibition, will reflect
the balance of a and b adrenoceptor activation on
all elements of the circuit. Extracellular unit
recording (see Extracellular unit recording, below)
shows that the effects of NE naturally released
include decreases and increases in inhibitory
interneuron firing, depending on the subtype of
the interneuron.
Extracellular unit recording
Segal and Bloom were the first to record changes
in unit activity in DG in response to electrical
stimulation of the LC (Segal and Bloom, 1976a).
The cells they monitored had relatively high firing
rates and therefore were likely to be interneurons
(Jung and McNaughton, 1993). LC stimulation
inhibited this cell type in anesthetized rats. In
awake rats, a loud tone also inhibited these DG
cells, but when paired with sweet milk it evoked
excitation; LC stimulation enhanced the cell exci-
tation as well as inhibition (Segal and Bloom,
1976b). This suggests a circuit-dependent LC mod-
ulation of interneuron responses.
The inhibition appeared to be a direct effect of
NE release, since LC cells transplanted to a den-
ervated hippocampus also inhibited spontaneously
active units in the DG. A brief stimulus train to the
LC transplant inhibited DG cell firing for
30 s
(Bjorklund et al., 1979). This inhibition was an-
tagonized by a b adrenoceptor antagonist.
In two pharmacological studies, theta inter-
neurons (putative inhibitory interneurons) in the
hilus and near the granule cell layer were excited
by NE and by a2 and b agonists, while granule
cells were inhibited by NE and by a1 agonists, but
excited by a2 and b agonists (Pang and Rose,
1987; Rose and Pang, 1989). The excitatory
effects of a2 receptors appeared to be postsynap-
tic, since they occurred in the presence of high
concentrations of magnesium in the extracellular
buffer.
A more recent study using glutamatergic acti-
vation of LC neurons demonstrates both inhibi-
tory and excitatory effects on subgranular and
hilar interneurons with natural NE release (Brown
et al., 2005). Using NE release from LC activation
(natural NE release) has the advantage of reveal-
ing in situ actions of NE rather than pharmaco-
logical actions, but non-NE effects may also be
recruited. The nature of the LC-NE effect de-
pended on the physiological identity of the inter-
neurons. Feedforward interneurons, activated by
perforant path input from the entorhinal cortex
and firing with a lower threshold than granule
cells, were consistently inhibited (
60 s) by brief
LC activation (see Fig. 1). Feedback interneurons,
recruited only after granule cells discharged, were
either excited or inhibited, suggesting differences
between subpopulations. Consistent with inhibi-
tion of feedforward interneurons was the observa-
tion of a strongly increased granule cell response
after LC activation. These results reveal that nat-
ural release of NE produces selective modulation
of DG circuitry. More work is needed to classify
the feedback interneurons that respond differen-
tially to NE release.

Fig. 1. Normalized firing rate of 16 feedforward interneurons in the dentate gyrus following ejection of glutamate in the locus
coeruleus demonstrated the robust effects on this cell type. Glutamate was ejected at 0 min. Adapted with permission from The Society
for Neuroscience, 2005.
EEG recording
In the same study (Brown et al., 2005), LC acti-
vation enhanced DG EEG theta in the 4–8 Hz
range while suppressing beta (12–20 Hz) and
gamma (20–40 Hz) frequencies (see Fig. 2). This
result appears consistent with concomitant inhibi-
tion of the feedforward interneuron population,
which potentially mediates the more distant influ-
ences implicated in beta oscillations (Kopell et al.,
2000), and excitation of a feedback subpopulation
enhancing theta amplitude as reported for 5-HT
modulation (Nitz and McNaughton, 1999).
Gray originally proposed that LC activation in-
creases hippocampal theta power specifically in a
7.5–7.8 Hz range in awake rats (Gray and Ball,
1970). He observed this increase when rats were
responding to novelty or a disconfirmation of ex-
pectation. The low driving threshold for this fre-
quency range was lost following NE depletion or
blockade (Gray et al., 1975). We have observed
that this theta frequency (7.5–7.8 Hz) increases
with LC activation in awake rats (unpublished
observations).
Segal reported that electrical LC stimulation in-
duces theta burst firing in medial septal neurons
(Segal, 1976). Berridge and Foote later showed
that tonic cholinergic LC activation produces a
dramatic increase in hippocampal theta in anest-
hetized rats (Berridge and Foote, 1991), which de-
pends on b adrenoceptors in medial septum
(Berridge et al., 1996).
305
A pattern of enhanced theta and suppressed
beta/gamma activity is consistent with the hypoth-
esis that disengagement from established represen-
tations and enhancement of processes that
promote incorporation of new information is an
effect of LC activation (Brown et al., 2005). Con-
sistent with this view, Bouret and Sara have pro-
posed ‘network resetting’ as a primary LC
function (Bouret and Sara, 2005).
Evoked potential recording
NE potentiates the perforant path-evoked popu-
lation spike both transiently, and in a sustained
manner, depending on the degree of NE exposure
in the DG. Transient potentiation of the perforant
path-evoked population spike amplitude, which is
b adrenoceptor-dependent, has been reported in
numerous in vivo experiments after either ex-
ogenous NE application (Neuman and Harley,
1983; Winson and Dahl, 1985; Babstock and
Harley, 1993; Harley et al., 1996; Chaulk and
Harley, 1998) or activation of LC (Dahl and
Winson, 1985; Harley and Milway, 1986; Harley et
al., 1989; Washburn and Moises, 1989; Babstock
and Harley, 1992; Frizzell and Harley, 1994;
Klukowski and Harley, 1994). NE or b adrenocep-
tor agonists in vitro increase the perforant path-
evoked EPSP slope, as well as the population
spike, although enhanced E-S coupling is also ob-
served (Lacaille and Harley, 1985). The duration
306

Fig. 2. Frequency-dependent increases and decreases in dentate gyrus EEG in the 4–40 Hz range following ejection of glutamate in the
locus coeruleus. Asterisks indicate significant effects at po0.05. Upward arrow denotes time of ejection. Adapted with permission from
The Society for Neuroscience, 2005.

of in vitro effects are transient when agonist con- NE-induced plasticity

centrations are low or briefly applied (Lacaille and
Harley, 1985; Stanton and Sarvey, 1985c).
The discovery of sustained NE-induced in-
creases in the perforant path-evoked potential or
NE-induced long-term potentiation (NE-LTP)
(Neuman and Harley, 1983) was the first direct
evidence that LC-NE could provide a heterosy-
naptic signal to initiate long-lasting increases in
DG responses to glutamate-mediated information.
NE-LTP is consistent with Kety’s early hypothe-
sis, based on pharmacological and behavioral ev-
idence, that LC-NE strengthens brain responses to
significant stimuli (Harley, 1987), and NE-LTP
will be examined in detail in the next section.
Spike potentiation in anesthetized rats
While studies at the cellular and EEG level suggest
transient LC-NE modulation of the DG network,
studies of the perforant path-evoked potential, as
noted, reveal a sustained component of LC-NE
modulation of the DG response to entorhinal in-
put. Iontophoresed NE was first reported in 1983
to produce a long-lasting potentiation of the per-
forant path-evoked DG population spike in anest-
hetized rats that endured for hours (Neuman and
Harley, 1983). NE-induced LTP of the population
spike has been repeatedly confirmed in vivo using

direct NE application (Winson and Dahl, 1985;
Harley et al., 1996; Chaulk and Harley, 1998), LC
activation by glutamate (Harley and Milway,
1986; Harley and Sara, 1992; Klukowski and
Harley, 1994; Walling and Harley, 2004) and LC
activation by orexin (Walling et al., 2004). Re-
peated LC electrical stimulation can also induce
LTP of the perforant path-evoked potential
(Harley et al., 1989). LC electrical stimulation-
induced potentiation is controversial, however,
with respect to dependence on b adrenoceptor
activation. There is evidence both for (Washburn
and Moises, 1989) and against (Harley et al., 1989)
such dependence. DG evoked response potentiat-
ion associated with glutamate- or orexin-induced
activation of LC neurons is consistently blocked
by b adrenoceptor antagonists, either systemically-
administered (Harley and Milway, 1986; Harley et
al., 1989; Babstock and Harley, 1992; Walling and
Harley, 2004; Walling et al., 2004) or locally-de-
livered to the DG (Harley and Evans, 1988). The
requirement for DG b adrenoceptor activation is
consistent with the intracellular evidence that b
adrenoceptor activation increases the excitability
of granule cells to exogenous input. NE-LTP does
not depend, however, on continued exposure to
NE, as demonstrated by the efficacy of brief ion-
tophoretic applications (Neuman and Harley,
1983; Winson and Dahl, 1985) and by measure-
ment of hippocampal NE during NE-LTP induced
by exogenous NE application (Harley et al., 1996)
or LC activation (Walling et al., 2004). (For fur-
ther discussion, see LC firing and NE release pat-
terns, below).
Synaptic and spike potentiation in vitro
In vitro studies of NE and perforant path stimu-
lation differ from in vivo studies because potent-
iation of both synaptic and population spike
components of the perforant path-evoked poten-
tial occurred consistently in vitro but not in vivo
(Lacaille and Harley, 1985; Stanton and Sarvey,
1987). In vitro NE-LTP depends on concomitant
NMDA receptor activation (Burgard et al., 1989;
Sarvey et al., 1989), as well as on b1 adrenoceptor
activation and protein synthesis (Stanton and
307
Sarvey, 1985c), strongly suggesting an associative
component to NE-LTP. However in vitro studies
failed to find evidence that pairing electrical
stimulation of perforant path input with NE bath-
application was critical for NE-LTP (Lacaille and
Harley, 1985; Dahl and Sarvey, 1990) (but see Spike
potentiation and pairing requirements, below).
Frizzell and Harley found an NMDA receptor-
independent potentiation of EPSP slope and pop-
ulation spike using LC-NE activation and keta-
mine anesthesia in vivo, but potentiation was
typically short-lived (Frizzell and Harley, 1994). In
vitro, Dahl reported that two applications of a low
b adrenoceptor agonist dose (75 nM), spaced by
30 min, which, individually, elicit only weak tran-
sient potentiation, induce NE-LTP of the perfo-
rant path population spike only, and NMDA
receptor activation was not required (Dahl and Li,
1994). This effect of a spaced NE agonist suggests
a mechanism by which LC-NE might contribute to
a stronger memory with spaced training.
Pathway selectivity
The in vitro NE-LTP of EPSP slope occurred
only for medial perforant path input, while the
same application of NE with an a adrenoceptor
antagonist, or of a b adrenoceptor agonist alone,
produced a long-term depression (LTD) of the
lateral perforant path EPSP slope (Dahl and
Sarvey, 1989). It may be relevant for this selective
effect that b adrenoceptors are densest in the ter-
minal zone of the medial perforant path (Milner
et al., 2000) and that enhancement of back-prop-
agating calcium spikes by NE would be more
likely to induce LTP in the adjacent medial per-
forant path target than the more distal lateral
perforant path target (Dan and Poo, 2006).
A possible explanation for the failure to induce
NE-LTP of the field EPSP in vivo could be that
stimulation in vivo, particularly the stronger
stimulation that is required to evoke a DG pop-
ulation spike, would activate a mixture of medial
and lateral perforant path fibers. NE-LTP in
vivo occurs for short-latency population spikes,
which reflect medial perforant path activation.
In one attempt to examine the selectivity of the
308
NE effect in vivo, lateral perforant path input was
activated by stimulating the lateral olfactory
tract. The synaptic input produced by lateral
olfactory stimulation was decreased by pairing
with paragigantocellularis stimulation, a major
glutamatergic afferent pathway to the LC
(Babstock and Harley, 1993). Depression of the
lateral perforant path input depended on b
adrenoceptor activation. Since paragiganto-
cellularis stimulation enhances the medial perfo-
rant path population spike through a b
adrenoceptor-dependent mechanism (Babstock
and Harley, 1992), the combined results are con-
sistent with enhanced medial perforant path input
and depressed lateral perforant path input with
LC-NE activation in vivo.
One functional interpretation of selectivity is
suggested by a new perspective on the entorhinal-
hippocampal complex, which suggests two forms
of spatial mapping occur in hippocampus. Global
remapping, reflecting a completely new context, is
associated with changes in medial perforant path
‘grid cell’ input; while rate remapping, which
reflects firing changes in the same place cell map, is
ascribed to alterations in nonspatial sensory
input mediated by the lateral perforant path
(McNaughton et al., 2006). If strong LC-NE ac-
tivation potentiates medial, and depresses lateral,
perforant path inputs, it would favor global re-
mapping in DG over rate remapping. Global re-
mapping would provide a new context for memory
storage.
It should be noted that high frequency-induced
LTP of either the medial or lateral perforant path
requires b adrenoceptor activation (Bramham
et al., 1997). Thus, while b adrenoceptors may
gate weaker inputs to favor the medial perforant
path, when signals are strong, plasticity should be
promoted by b adrenoceptor activation in both
pathways similarly.
Spike potentiation and pairing requirements
In vivo experiments provide evidence that LC-NE
activation induces long-lasting increases in the
medial perforant path-evoked population spike,
suggesting that excitability in the postsynaptic
granule cells increases. Although mixed medial
and lateral perforant path input is a possible
explanation of the lack of EPSP potentiation
observed in vivo, the ubiquity of studies that see
no change in EPSP slope suggests an increase in
E-S coupling, or granule cell excitability, is a
prominent feature of LC-NE modulation. Recent
evidence provides support for the hypothesis
that an increase in intrinsic excitability, through
such mechanisms as reduced K+ conductance in
activated dendrites (Frick et al., 2004), is an im-
portant component of associative plasticity func-
tioning similar to, and independent of, increases in
synaptic drive per se (Daoudal and Debanne,
2003).
Evidence for associative plasticity in LC-NE ac-
tivation effects on DG has been lacking. In vitro
studies have not found that NE-LTP requires con-
current pairing of bath-applied NE and perforant
path stimulation (Lacaille and Harley, 1985; Dahl
and Sarvey, 1990). However, we have recently
found that pairing of LC glutamate activation and
perforant path input is critical for NE-LTP in vivo
(Reid and Harley, 2005). With LC activation
10 min after cessation of perforant path stimula-
tion and 10 min prior to resumption of perforant
path stimulation, no NE-LTP occurs. The same
LC activation during concurrent perforant path
stimulation produces reliable NE-LTP. Both EPSP
slope and spike potentiation occurred in our in
vivo study (Reid and Harley, 2005). The impor-
tance of timing in LC-NE activation for potent-
iation effects is consistent with evidence that
conduction velocity in LC-NE axons varies across
species to maintain a constant delay-to-signal
arrival in forebrain structures (Aston-Jones et al.,
1985).
Two differences may be noted between the in
vitro and in vivo pairing experiments. First, there
was no delay between the offset of perforant path
stimulation and the beginning of NE perfusion in
the in vitro studies. Second, in vitro bath applica-
tion of NE is associated with an extended elevation
of cAMP levels in DG, permitting perforant path
interaction with cAMP effects even at late time
points (Stanton and Sarvey, 1985b). This is un-
likely to be the case with LC activation (Siggins
et al., 1973).

Delayed long-term synaptic potentiation in awake
rats
In contrast to the evidence that NE induces rapid
potentiation of perforant path synaptic input from
in vitro studies is new in vivo data from awake rats
in which potentiation of perforant path synaptic
input is first observed 24 h after LC activation
(Walling and Harley, 2004). In the Walling and
Harley study, LC burst activation initially had no
effect on synaptic input, but induced the usual
rapid and enduring potentiation of the perforant
path-evoked population spike. Twenty-four hours
later, both the perforant path field EPSP and
the population spike were strongly potentiated
(Fig. 3). The 24 h increase in field EPSP slope pre-
dicted the 24 h increase in population spike am-
plitude. This delayed NE-LTP of EPSP slope and
population spike amplitude depended on b ad-
renoceptor activation at the time of LC activation,
and on protein synthesis.
Aplysia also demonstrates a 24 h delayed potent-
iation of synaptic input in response to heterosy-
naptic activation of a cAMP-dependent cascade
(Brunelli et al., 1976; Schacher et al., 1988). Wall-
ing and Harley suggest mammalian delayed synap-
tic potentiation implies a special role for DG NE
modulation in long-term memory. Such an associ-
ation is likely to require relatively strong NE
release. NE applied intracerebroventricularly and
Fig. 3. Delayed potentiation of the perforant path-evoked field EPSP with immediate and delayed potentiation of the perforant path-
evoked population spike in awake rats following infusion of glutamate in the locus coeruleus at the arrow. Adapted with permission
from The Society for Neuroscience, 2005.
309
measured by microdialysis in the DG is only asso-
ciated with NE-LTP in the DG when the NE in-
crease exceeds a critical threshold, estimated to be

400 nM, intrasynaptically (Harley et al., 1996).
LC firing and NE release patterns
LC neurons fire tonically at rates from less than 1
to 5 Hz (Aston-Jones and Bloom, 1981b) as a
function of level of arousal. They also fire phasic-
ally in bursts in which 2 or 3 spikes occur in rapid
succession (10–15 Hz) followed by 200–500 ms
pauses (Aston-Jones and Bloom, 1981a). Burst
events are typically associated with responses to
environmental stimuli (Aston-Jones and Bloom,
1981a). Harley and Sara have shown that gluta-
mate activation of LC produces a strong burst
followed by a pause, lasting minutes, in LC firing
(Harley and Sara, 1992). Higher levels of NE re-
lease are associated with phasic rather than tonic
firing patterns (Florin-Lechner et al., 1996).
Glutamate administration to the LC in vivo in-
creases NE release in hippocampus by
200%,
measured by microdialysis, in the first 20 min after
LC activation, subsequently, NE levels return to
baseline (Walling et al., 2004). Orexin A is also a
potent activator of LC neurons, and its adminis-
tration to the LC led to a similar increase in NE,
restricted to the first 20 min (Walling et al., 2004).
310
Both orexin A and glutamate activation of LC
produce a b adrenoceptor-dependent increase in
population spike amplitude that lasts for hours
and is thus triggered, but not maintained by, hip-
pocampal NE release. High frequency electrical
stimulation of the perforant path also increases
NE release in the DG (Bronzino et al., 2001), and
may contribute to the level and duration of LTP
induced (Bronzino et al., 1999).
Voltammetry provides better temporal resolu-
tion of NE release. Electrical stimulation of the LC
for 2 s at 50 Hz induces a 10 fold increase in ox-
idation signal in the DG (maximal extracellular
signal
0.18 mM NE), appearing immediately after
stimulation and returning to baseline within
10–15 s (Yavich et al., 2005). Repeated stimula-
tion at 5 min intervals yields stable NE signals,
while stimulation at 30 s intervals shows decreas-
ing responses, suggesting a slow rate of vesicle re-
filling. Natural stimuli that activate LC (tail pinch
or vibrissae stimulation) do not produce a visible
NE signal without enhancement of NE release us-
ing the a2 receptor-antagonist, idazoxan.
Glutamate activation of LC produces an aver-
age increase of
0.1 mM in hippocampal NE be-
ginning 30 s after glutamate infusion, peaking at
1.5 min and returning to baseline by 7 min in ur-
ethane-anesthetized rats (Palamarchouk et al.,
2000). NE patterns in awake rats with glutamate
infusion were similar but peaked at 5 min and re-
turned to baseline within 20 min (Palamarchouk
et al., 2002).
NE release modulation by glutamate and vice-versa:
local effects
Both NMDA and AMPA receptors are present
presynaptically on NE axon terminals in hippo-
campus, and when activated by glutamate or gluta-
mate agonists, induce an increase in NE release
(Pittaluga and Raiteri, 1992; Raiteri et al., 1992).
NMDA-induced NE increases are greatest in DG
microslices (Andres et al., 1993). Nictotine and
somatostatin receptors are also present on NE ter-
minals in hippocampus and induce NE release
alone (nicotine) or in concert with NMDA receptor
stimulation (nicotine and somtatostatin) in the
presence of normal magnesium levels and without
membrane depolarization (Pittaluga et al., 2000,
2005; Risso et al., 2004). This provides a mechanism
for local NE release in the absence of direct termi-
nal activation. Finally, cognitive enhancers signifi-
cantly facilitate NE release in response to
glutamate, even in the presence of the endogenous
NMDA receptor antagonist, kynurenate (Pittaluga
et al., 1997). This facilitation is suggested to be an
important mediator of cognitive enhancement
(Desai et al., 1995; Pittaluga et al., 2001).
Conversely, NE enhances glutamate release in
DG, an effect which depends on presynaptic b
adrenoceptors (Lynch and Bliss, 1986). NE can
also enhance GABA release in hippocampus
through a2 adrenoceptor activation (Pittaluga
and Raiteri, 1987; Maura et al., 1988). Reversible
EPSP slope potentiation of perforant path input
to the DG has been related to b adrenoceptor-
mediated phosphorylation of the presynaptic pro-
teins synapsins I and II (Parfitt et al., 1991, 1992).
The positive feedback between NE and gluta-
mate release further suggests local increases in NE
could occur independently of LC activation. The
release of NE by glutamate and other neurotrans-
mitters may explain extended limbic NE elevation
when rats receive shock in a novel context (McIn-
tyre et al., 2002). Suggestively, memory for the
context correlates positively with NE levels at the
time of acquisition. LTP of perforant path fibers
also evokes longer lasting increases in NE release
than LC activation alone (Bronzino et al., 1999,
2001).
LC-NE modulation of DG in relation to
environmental events
Novel objects investigated by a rat using a board
with objects contained in holes (holeboard) elicit
LC bursts (Vankov et al., 1995). Perforant path-
evoked population spikes are potentiated for

1 min after a rat places its nose in a hole in re-
sponse to a novel object (nose poke). Spike po-
tentiation depends on b adrenoceptor activation
(Kitchigina et al., 1997); as does the enhanced ex-
ploration of novelty (Sara et al., 1995). Prolonged
spike enhancement occurs upon first exposure to

the novel holeboard environment, possibly be-
cause the novel environment elicits a larger LC
response (Kitchigina et al., 1997). These effects
appear to model the attentional (Berridge and
Waterhouse, 2003), rather than the memory,
effects of NE.
Novel environment exposure can promote mem-
ory-like effects when a weak, high-frequency LTP
stimulus is also presented (Straube et al., 2003).
Placement in a novel environment converts early-
phase LTP to late-phase LTP in the DG. This
conversion depends on b adrenoceptor activation,
suggesting LC-NE activation by novelty induces
an enhancement of the weak LTP stimulus. NE in
the ventricles mediates a similar conversion of
early to late-LTP in the DG (Almaguer-Melian et
al., 2005). b adrenoceptor activation is critical for
late DG LTP both in vitro (Stanton and Sarvey,
1985a) and in vivo (Straube and Frey, 2003). Only
the use of strong LTP protocols (for example, 15
0.25 ms pulses at 200 Hz every 10 s for 200 s) with
repeated 15 pulse protocols twice within 5 min can
produce LTP in vivo that is independent of b ad-
renoceptor activation (Straube and Frey, 2003). In
addition to facilitation by NE of enduring LTP in
the entorhinal cortex-to-granule cell perforant
pathway, there is LTP facilitation in the granule
cell-to-CA3 pathway when subthreshold LTP
stimuli are combined with b adrenoceptor activa-
tion (Hopkins and Johnston, 1984). Thus, weak
patterns for LTP recruit enduring LTP in both the
input and output components of DG circuitry
when NE is elevated.
Reward also promotes the conversion of early
phase perforant path-LTP to late phase perforant
path-LTP (Seidenbecher et al., 1997), and is known
to activate LC neurons (Sara and Segal, 1991).
Punishment has similar effects on activation of LC
(Sara and Segal, 1991) and the conversion of early
to late DG LTP. Conversions of early to late LTP
by reward and punishment depend b adrenoceptor
activation (Seidenbecher et al., 1997). These results
suggest reward- or punishment-related LC activa-
tion ‘reinforces’ plasticity in the DG. LC-NE may
also contribute to long-lasting circuit changes
through the promotion of the survival of new neu-
rons in DG (Rizk et al., 2006) and/or the promo-
tion of neurogenesis (Kulkarni et al., 2002).
311
Restoration of recently-decayed LTP in the DG
occurs following electrical stimulation of the LC
(Ezrokhi et al., 1999). This phenomenon may re-
late to other evidence for a role of NE in memory
retrieval (Sara and Devauges, 1989; Devauges and
Sara, 1991; Murchison et al., 2004). NE’s promo-
tion of theta EEG and the reported dependence of
engram spread on b adrenoceptor activation could
also contribute to retrieval processes (Flexner et
al., 1985; Givens, 1996). Aston-Jones and Cohen’s
proposal that phasic LC-NE optimizes decision-
driven behavior and decision-driven memory rep-
resentations provides further theoretical support
for an NE retrieval function (Aston-Jones and
Cohen, 2005).
Exploratory behavior is increased with DG NE
infusions (Flicker and Geyer, 1982) and this effect
is blocked by a b adrenoceptor antagonist. A tonic
LC-NE driven modulation of diversive explora-
tion (Usher et al., 1999; Mansour et al., 2003) is
consistent with lesion evidence suggesting the DG
has an important role in encoding spatial novelty
(Lee et al., 2005; Jerman et al., 2006).
Vasopressin in the DG facilitates memory for
passive avoidance and increases NE turnover in
DG (Kovacs et al., 1979b). Both consolidation and
retrieval are enhanced, and these effects require
intact LC-NE innervation (Kovacs et al., 1979a).
Vasopressin in dorsal hippocampus also facilitates
memory for spatial learning (Paban et al., 2003).
NE dependence of this effect remains to be as-
sessed.
Corticotrophin releasing factor infused in the
DG also enhances passive avoidance memory,
while loss of local NE innervation, or blockade of
local b adrenoceptors or NMDA receptors, pre-
vents this memory enhancement (Lee et al., 1993).
Lee et al. suggested that b adrenoceptors promote
NMDA-mediated synaptic plasticity to support
passive avoidance memory. A dependence of NE-
induced plasticity on NMDA receptors was also
described in some in vitro studies (Sarvey et al.,
1989).
As we come to better understand the behavioral
role of DG, new probes of NE’s function in the
DG should emerge. Aston-Jones and Cohen have
proposed, based on data from primates, that pha-
sic LC-NE activation facilitates decision-driven
312
responses or memories, while tonic LC-NE activity
facilitates exploration (Aston-Jones and Cohen,
2005). The latter hypothesis is supported by in-
creased exploration following NE infusion in DG
(Flicker and Geyer, 1982), but the relation of pha-
sic activation to decision-driven behavior and/or
memory is unclear with respect to DG activity.
Bouret and Sara’s more general description
(Bouret and Sara, 2005) of the function of the
LC-NE system as ‘network resetting’ is broadly
consistent with NE-associated changes in EEG
and with the promotion of long-term synaptic
plasticity in DG input and output.
The promotion of frequency-induced long-term
potentiation (Dragoi et al., 2003) by NE reviewed
above and the selective enhancement of the spatial
map matrix input from the medial entorhinal cor-
tex (Hafting et al., 2005) described earlier (see
Pathway selectivity) lead to the prediction that, for
the DG, strong activation of NE input would pro-
mote global remapping of spatial context. Weaker
activation of NE input and an associated transient
reduction in network inhibition (see Extracellular
unit recording, above) might enhance input gener-
alization, or pattern completion to assist retrieval,
but strong activation would alter the DG map.
Global remapping provides a new framework for
the encoding of associative learning, reducing in-
terference among associations and increasing
memory capacity. Global remapping is assumed
to underlie episodic memory. The prediction that
strong LC-NE activation recruits global remap-
ping in DG is consistent with the hypothesis that
activation of the LC and the related sympathetic
system is part of a response to strongly significant
events that promotes long-term memories for sig-
nificant events (Cahill et al., 1994). Global remap-
ping with novelty exposure has been shown for
CA3 cells (Leutgeb et al., 2006), which like DG
(Chawla et al., 2005; Rolls and Kesner, 2006), are
implicated in sparse encoding and orthogonal en-
vironmental representations. The occurrence of
global remapping in DG concomitant with strong
LC activation remains to be demonstrated, but is
suggested by the data reviewed here.
In summary, evoked potential and cellular data
support various effects of NE on information
processing in the DG. Transient effects of phasic
NE release include reduction of feedforward inhi-
bition and increases in granule cell excitability,
such that granule cell responses to entorhinal input
increase. NE’s modulation of feedback interneu-
rons enhances theta to facilitate throughput and
associative change in synaptic strength depending
on theta phase (Pavlides et al., 1988; Orr et al.,
2001). NE enhances and depresses inputs in a
pathway-dependent manner and recruits glial met-
abolic support for neuronal activation. Strong NE
input engages processes that recruit long-term
modifications of excitability and/or delayed in-
creases in synaptic strength in the DG pathway.
Together these effects support both attentional
and memory roles for DG NE.
Acknowledgment
The author wishes to express appreciation for the
support of Canada’s NSERC Granting Council
via Grant A9791.
References
Almaguer-Melian, W., Rojas-Reyes, Y., Alvare, A., Rosillo,
J.C., Frey, J.U. and Bergado, J.A. (2005) Long-term potent-
iation in the dentate gyrus in freely moving rats is reinforced
by intraventricular application of norepinephrine, but not
oxotremorine. Neurobiol. Learn. Mem., 83: 72–78.
Andres, M.E., Bustos, G. and Gysling, K. (1993) Regulation of
[3H]norepinephrine release by N-methyl-D-aspartate recep-
tors in minislices from the dentate gyrus and the CA1-CA3
area of the rat hippocampus. Biochem. Pharmacol., 46:
1983–1987.
Arango, V., Ernsberger, P., Reis, D.J. and Mann, J.J. (1990)
Demonstration of high- and low-affinity beta-adrenergic re-
ceptors in slide-mounted sections of rat and human brain.
Brain Res., 516: 113–121.
Aston-Jones, G. and Bloom, F.E. (1981a) Norepinephrine-con-
taining locus coeruleus neurons in behaving rats exhibit pro-
nounced responses to non-noxious environmental stimuli. J.
Neurosci., 1: 887–900.
Aston-Jones, G. and Bloom, F.E. (1981b) Activity of nor-
epinephrine-containing locus coeruleus neurons in behaving
rats anticipates fluctuations in the sleep- waking cycle. J.
Neurosci., 1: 876–886.
Aston-Jones, G. and Cohen, J.D. (2005) An integrative theory
of locus coeruleus-norepinephrine function: adaptive gain
and optimal performance. Annu. Rev. Neurosci., 28:
403–450.

Aston-Jones, G., Foote, S.L. and Segal, M. (1985) Impulse
conduction properties of noradrenergic locus coeruleus axons
projecting to monkey cerebrocortex. Neuroscience, 15:
765–777.
Babstock, D.M. and Harley, C.W. (1992) Paragigantocellularis
stimulation induces beta-adrenergic hippocampal potentiat-
ion. Brain Res. Bull., 28: 709–714.
Babstock, D.M. and Harley, C.W. (1993) Lateral olfactory
tract input to dentate gyrus is depressed by prior nor-
adrenergic activation using nucleus paragigantocellularis
stimulation. Brain Res., 629: 149–154.
Berridge, C.W., Bolen, S.J., Manley, M.S. and Foote, S.L.
(1996) Modulation of forebrain electroencephalographic
activity in halothane-anesthetized rat via actions of nor-
adrenergic beta-receptors within the medial septal region. J.
Neurosci., 16: 7010–7020.
Berridge, C.W. and Foote, S.L. (1991) Effects of locus co-
eruleus activation on electroencephalographic activity in
neocortex and hippocampus. J. Neurosci., 11: 3135–3145.
Berridge, C.W. and Waterhouse, B.D. (2003) The locus co-
eruleus-noradrenergic system: modulation of behavioral state
and state-dependent cognitive processes. Brain Res. Brain
Res. Rev., 42: 33–84.
Bijak, M. and Misgeld, U. (1995) Adrenergic modulation of
hilar neuron activity and granule cell inhibition in the
guinea-pig hippocampal slice. Neuroscience, 67: 541–550.
Bjorklund, A., Segal, M. and Stenevi, U. (1979) Functional
reinnervation of rat hippocampus by locus coeruleus im-
plants. Brain Res., 170: 409–426.
Blackstad, T.W., Fuxe, K. and Hokfelt, T. (1967) Noradren-
aline nerve terminals in the hippocampal region of the rat
and the guinea pig. Z. Zellforsch. Mikrosk. Anat., 78:
463–473.
Booze, R.M., Crisostomo, E.A. and Davis, J.N. (1993) Beta-
adrenergic receptors in the hippocampal and retrohippocam-
pal regions of rats and guinea pigs: autoradiographic and
immunohistochemical studies. Synapse, 13: 206–214.
Bouret, S. and Sara, S.J. (2005) Network reset: a simplified
overarching theory of locus coeruleus noradrenaline func-
tion. Trends Neurosci., 28: 574–582.
Bramham, C.R., Bacher-Svendsen, K. and Sarvey, J.M. (1997)
LTP in the lateral perforant path is beta-adrenergic receptor-
dependent. Neuroreport, 8: 719–724.
Bronzino, J.D., Kehoe, P., Hendriks, R., Vita, L., Golas, B.,
Vivona, C. and Morgane, P.J. (1999) Hippocampal neuro-
chemical and electrophysiological measures from freely
moving rats. Exp. Neurol., 155: 150–155.
Bronzino, J.D., Kehoe, P., Mallinson, K. and Fortin, D.A.
(2001) Increased extracellular release of hippocampal NE is
associated with tetanization of the medial perforant pathway
in the freely moving adult male rat. Hippocampus, 11:
423–429.
Brown, R.A.M., Walling, S.G., Milway, J.S. and Harley, C.W.
(2005) Locus ceruleus activation suppresses feedforward
interneurons and reduces beta-gamma electroencephalogram
frequencies while it enhances theta frequencies in rat dentate
gyrus. J. Neurosci., 25: 1985–1991.
313
Brunelli, M., Castellucci, V. and Kandel, E.R. (1976) Synaptic
facilitation and behavioral sensitization in Aplysia: possible
role of serotonin and cyclic AMP. Science, 194: 1178–1181.
Burgard, E.C., Decker, G. and Sarvey, J.M. (1989) NMDA
receptor antagonists block norepinephrine-induced long-last-
ing potentiation and long-term potentiation in rat dentate
gyrus. Brain Res., 482: 351–355.
Cahill, L., Prins, B., Weber, M. and McGaugh, J.L. (1994)
Beta-adrenergic activation and memory for emotional events.
Nature, 371: 702–704.
Chaulk, P.C. and Harley, C.W. (1998) Intracerebroventricular
norepinephrine potentiation of the perforant path-evoked
potential in dentate gyrus of anesthetized and awake rats: a
role for both alpha- and beta-adrenoceptor activation. Brain
Res., 787: 59–70.
Chawla, M.K., Guzowski, J.F., Ramirez-Amaya, V., Lipa,
P., Hoffman, K.L., Marriott, L.K., Worley, P.F., McNaugh-
ton, B.L. and Barnes, C.A. (2005) Sparse, environmentally
selective expression of Arc RNA in the upper blade of the
rodent fascia dentata by brief spatial experience. Hippocam-
pus, 15: 579–586.
Crutcher, K.A. and Davis, J.N. (1980) Hippocampal alpha- and
beta-adrenergic receptors: comparison of [3H]dihydroalpre-
nolol and [3H]WB 4101 binding with noradrenergic innerva-
tion in the rat. Brain Res., 182: 107–117.
Dahl, D. and Li, J. (1994) Induction of long-lasting potentiat-
ion by sequenced applications of isoproterenol. Neuroreport,
5: 657–660.
Dahl, D. and Sarvey, J.M. (1989) Norepinephrine induces
pathway-specific long-lasting potentiation and depression in
the hippocampal dentate gyrus. Proc. Natl. Acad. Sci.
U.S.A., 86: 4776–4780.
Dahl, D. and Sarvey, J.M. (1990) Beta-adrenergic agonist-in-
duced long-lasting synaptic modifications in hippocampal
dentate gyrus require activation of NMDA receptors, but
not electrical activation of afferents. Brain Res., 526:
347–350.
Dahl, D. and Winson, J. (1985) Action of norepinephrine in the
dentate gyrus. I. Stimulation of locus coeruleus. Exp. Brain
Res., 59: 491–496.
Dan, Y. and Poo, M.M. (2006) Spike timing-dependent plas-
ticity: from synapse to perception. Physiol. Rev., 86:
1033–1048.
Daoudal, G. and Debanne, D. (2003) Long-term plasticity of
intrinsic excitability: learning rules and mechanisms. Learn.
Mem., 10: 456–465.
Day, H.E., Campeau, S., Watson Jr., S.J. and Akil, H. (1997)
Distribution of alpha 1a-, alpha 1b- and alpha 1d-adrenergic
receptor mRNA in the rat brain and spinal cord. J. Chem.
Neuroanat., 13: 115–139.
Desai, M.A., Valli, M.J., Monn, J.A. and Schoepp, D.D. (1995)
1-BCP, a memory-enhancing agent, selectively potentiates
AMPA-induced [3H]norepinephrine release in rat hippocam-
pal slices. Neuropharmacology, 34: 141–147.
Devauges, V. and Sara, S.J. (1991) Memory retrieval enhance-
ment by locus coeruleus stimulation: evidence for mediation
by beta-receptors. Behav. Brain Res., 43: 93–97.
314
Dragoi, G., Harris, K.D. and Buzsaki, G. (2003) Place repre-
sentation within hippocampal networks is modified by long-
term potentiation. Neuron, 39: 843–853.
Duncan, G.E., Little, K.Y., Koplas, P.A., Kirkman, J.A.,
Breese, G.R. and Stumpf, W.E. (1991) Beta-adrenergic re-
ceptor distribution in human and rat hippocampal
formation: marked species differences. Brain Res., 561:
84–92.
Edwards, C., Nahorski, S.R. and Rogers, K.J. (1974) In vivo
changes of cerebral cyclic adenosine 30 ,50 -monophosphate
induced by biogenic amines: association with phosphorylase
activation. J. Neurochem., 22: 565–572.
Ezrokhi, V.L., Zosimovskii, V.A., Korshunov, V.A. and Mark-
evich, V.A. (1999) Restoration of decaying long-term po-
tentiation in the hippocampal formation by stimulation of
neuromodulatory nuclei in freely moving rats. Neuroscience,
88: 741–753.
Fara-On, M., Evans, J.H. and Harley, C.W. (2005) Idazoxan
activates rat forebrain glycogen phosphorylase in vivo: a his-
tochemical study. Brain Res., 1059: 83–92.
Flexner, J.B., Flexner, L.B., Church, A.C., Rainbow, T.C. and
Brunswick, D.J. (1985) Blockade of beta 1- but not of beta 2-
adrenergic receptors replicates propranolol’s suppression of
the cerebral spread of an engram in mice. Proc. Natl. Acad.
Sci. U.S.A., 82: 7458–7461.
Flicker, C. and Geyer, M.A. (1982) Behavior during hippo-
campal microinfusions. I. Norepinephrine and diversive ex-
ploration. Brain Res., 257: 79–103.
Florin-Lechner, S.M., Druhan, J.P., Aston-Jones, G. and
Valentino, R.J. (1996) Enhanced norepinephrine release in
prefrontal cortex with burst stimulation of the locus co-
eruleus. Brain Res., 742: 89–97.
Frick, A., Magee, J. and Johnston, D. (2004) LTP is accom-
panied by an enhanced local excitability of pyramidal neuron
dendrites. Nat. Neurosci., 7: 126–135.
Frizzell, L.M. and Harley, C.W. (1994) The N-methyl-D-as-
partate channel blocker ketamine does not attenuate, but
enhances, locus coeruleus-induced potentiation in rat dentate
gyrus. Brain Res., 663: 173–178.
Givens, B. (1996) Stimulus-evoked resetting of the dentate theta
rhythm: relation to working memory. Neuroreport, 8:
159–163.
Gray, J.A. and Ball, G.G. (1970) Frequency-specific relation
between hippocampal theta rhythm, behavior, and am-
obarbital action. Science, 168: 1246–1248.
Gray, J.A., McNaughton, N., James, D.T. and Kelly, P.H.
(1975) Effect of minor tranquillisers on hippocampal theta
rhythm mimicked by depletion of forebrain noradrenaline.
Nature, 258: 424–425.
Gray, R. and Johnston, D. (1987) Noradrenaline and beta-ad-
renoceptor agonists increase activity of voltage-dependent
calcium channels in hippocampal neurons. Nature, 327:
620–622.
Haas, H.L. and Rose, G.M. (1987) Noradrenaline blocks po-
tassium conductance in rat dentate granule cells in vitro.
Neurosci. Lett., 78: 171–174.
Hafting, T., Fyhn, M., Molden, S., Moser, M.B. and Moser,
E.I. (2005) Microstructure of a spatial map in the entorhinal
cortex. Nature, 436: 801–806.
Haring, J.H. and Davis, J.N. (1985) Differential distribution of
locus coeruleus projections to the hippocampal formation:
anatomical and biochemical evidence. Brain Res., 325:
366–369.
Harley, C., Milway, J.S. and Lacaille, J.C. (1989) Locus co-
eruleus potentiation of dentate gyrus responses: evidence for
two systems. Brain Res. Bull., 22: 643–650.
Harley, C. and Rusak, B. (1993) Daily variation in active gly-
cogen phosphorylase patches in the molecular layer of rat
dentate gyrus. Brain Res., 626: 310–317.
Harley, C.W. (1987) A role for norepinephrine in arousal,
emotion and learning: limbic modulation by norepinephrine
and the Kety hypothesis. Prog. Neuropsychopharmacol.
Biol. Psychiatry, 11: 419–458.
Harley, C.W. and Evans, S. (1988) Locus coeruleus-induced
enhancement of the perforant path-evoked potential: effects
of intradentate beta blockers. In: Woody, C.D., Alkon, D.L.
and McGaugh, J.L. (Eds.), Cellular Mechanisms of Condi-
tioning and Behavioral Plasticity. Plenum Publishing
Corporation, New York, New York, pp. 415–423.
Harley, C.W., Lalies, M.D. and Nutt, D.J. (1996) Estimating
the synaptic concentration of norepinephrine in dentate gyrus
which produces beta-receptor mediated long-lasting potent-
iation in vivo using microdialysis and intracerebroventricular
norepinephrine. Brain Res., 710: 293–298.
Harley, C.W. and Milway, J.S. (1986) Glutamate ejection in
the locus coeruleus enhances the perforant path-evoked
population spike in the dentate gyrus. Exp. Brain Res., 63:
143–150.
Harley, C.W. and Sara, S.J. (1992) Locus coeruleus bursts in-
duced by glutamate trigger delayed perforant path spike am-
plitude potentiation in the dentate gyrus. Exp. Brain Res., 89:
581–587.
Hopkins, W.F. and Johnston, D. (1984) Frequency-dependent
noradrenergic modulation of long-term potentiation in the
hippocampus. Science, 226: 350–352.
Hortnagl, H., Berger, M.L., Sperk, G. and Pifl, C. (1991)
Regional heterogeneity in the distribution of neurotransmit-
ter markers in the rat hippocampus. Neuroscience, 45:
261–272.
Jerman, T., Kesner, R.P. and Hunsaker, M.R. (2006) Discon-
nection analysis of CA3 and DG in mediating encoding but
not retrieval in a spatial maze learning task. Learn. Mem., 13:
458–464.
Jones, L.S., Gauger, L.L. and Davis, J.N. (1985) Anatomy of
brain alpha 1-adrenergic receptors: in vitro autoradiography
with [125I]-heat. J. Comp. Neurol., 231: 190–208.
Jung, M.W. and McNaughton, B.L. (1993) Spatial selectivity of
unit activity in the hippocampal granular layer. Hippocam-
pus, 3: 165–182.
Kitchigina, V., Vankov, A., Harley, C. and Sara, S.J. (1997)
Novelty-elicited, noradrenaline-dependent enhancement of
excitability in the dentate gyrus. Eur. J. Neurosci., 9: 41–47.

Klukowski, G. and Harley, C.W. (1994) Locus coeruleus ac-
tivation induces perforant path-evoked population spike po-
tentiation in the dentate gyrus of awake rat. Exp. Brain Res.,
102: 165–170.
Koda, L.Y. and Bloom, F.E. (1977) A light and electron mi-
croscopic study of noradrenergic terminals in the rat dentate
gyrus. Brain Res., 120: 327–335.
Kopell, N., Ermentrout, G.B., Whittington, M.A. and Traub,
R.D. (2000) Gamma rhythms and beta rhythms have differ-
ent synchronization properties. Proc. Natl. Acad. Sci.
U.S.A., 97: 1867–1872.
Kosaka, T., Katsumaru, H., Hama, K., Wu, J.Y. and Heiz-
mann, C.W. (1987) GABAergic neurons containing the
Ca2+-binding protein parvalbumin in the rat hippocampus
and dentate gyrus. Brain Res., 419: 119–130.
Kovacs, G.L., Bohus, B. and Versteeg, D.H. (1979a) The
effects of vasopressin on memory processes: the role of
noradrenergic neurotransmission. Neuroscience, 4:
1529–1537.
Kovacs, G.L., Bohus, B., Versteeg, D.H., de Kloet, E.R. and
de, W.D. (1979b) Effect of oxytocin and vasopressin on
memory consolidation: sites of action and catecholaminergic
correlates after local microinjection into limbic-midbrain
structures. Brain Res., 175: 303–314.
Kulkarni, V.A., Jha, S. and Vaidya, V.A. (2002) Depletion of
norepinephrine decreases the proliferation, but does not in-
fluence the survival and differentiation, of granule cell pro-
genitors in the adult rat hippocampus. Eur. J. Neurosci., 16:
2008–2012.
Kusaka, K., Morinobu, S., Kawano, K. and Yamawaki, S.
(2004) Effect of neonatal isolation on the noradrenergic
transduction system in the rat hippocampal slice. Synapse,
54: 223–232.
Lacaille, J.C. and Harley, C.W. (1985) The action of
norepinephrine in the dentate gyrus: beta-mediated
facilitation of evoked potentials in vitro. Brain Res., 358:
210–220.
Lacaille, J.C. and Schwartzkroin, P.A. (1988) Intracellular
responses of rat hippocampal granule cells in vitro to
discrete applications of norepinephrine. Neurosci. Lett., 89:
176–181.
Lancaster, B., Hu, H., Ramakers, G.M. and Storm, J.F. (2001)
Interaction between synaptic excitation and slow afterhyper-
polarization current in rat hippocampal pyramidal cells. J.
Physiol., 536: 809–823.
Lapierre, Y., Beaudet, A., Demianczuk, N. and Descarries, L.
(1973) Noradrenergic axon terminals in the cerebral cortex of
rat. II. Quantitative data revealed by light and electron mi-
croscope radioautography of the frontal cortex. Brain Res.,
63: 175–182.
Lee, E.H., Lee, C.P., Wang, H.I. and Lin, W.R. (1993)
Hippocampal CRF, NE, and NMDA system interactions in
memory processing in the rat. Synapse, 14: 144–153.
Lee, I., Hunsaker, M.R. and Kesner, R.P. (2005) The role of
hippocampal subregions in detecting spatial novelty. Behav.
Neurosci., 119: 145–153.
315
Leutgeb, S., Leutgeb, J.K., Moser, E.I. and Moser, M.B. (2006)
Fast rate coding in hippocampal CA3 cell ensembles.
Hippocampus, 16: 765–774.
Loy, R., Koziell, D.A., Lindsey, J.D. and Moore, R.Y. (1980)
Noradrenergic innervation of the adult rat hippocampal for-
mation. J. Comp. Neurol., 189: 699–710.
Lynch, M.A. and Bliss, T.V. (1986) Noradrenaline modulates
the release of [14C]glutamate from dentate but not from
CA1/CA3 slices of rat hippocampus. Neuropharmacology,
25: 493–498.
Mansour, A.A., Babstock, D.M., Penney, J.H., Martin, G.M.,
McLean, J.H. and Harley, C.W. (2003) Novel objects in a
holeboard probe the role of the locus coeruleus in curiosity:
support for two modes of attention in the rat. Behav.
Neurosci., 117: 621–631.
Maura, G., Pittaluga, A., Ulivi, M. and Raiteri, M. (1988) En-
hancement of endogenous GABA release from rat synapto-
somal preparations is mediated by alpha 2-adrenoceptors
pharmacologically different from alpha 2-autoreceptors. Eur.
J. Pharmacol., 157: 23–29.
McIntyre, C.K., Hatfield, T. and McGaugh, J.L. (2002) Am-
ygdala norepinephrine levels after training predict inhibitory
avoidance retention performance in rats. Eur. J. Neurosci.,
16: 1223–1226.
McNaughton, B.L., Battaglia, F.P., Jensen, O., Moser, E.I. and
Moser, M.B. (2006) Path integration and the neural basis of
the ‘cognitive map’. Nat. Rev. Neurosci., 7: 663–678.
Milner, T.A. and Bacon, C.E. (1989a) GABAergic neurons in
the rat hippocampal formation: ultrastructure and synaptic
relationships with catecholaminergic terminals. J. Neurosci.,
9: 3410–3427.
Milner, T.A. and Bacon, C.E. (1989b) Ultrastructural local-
ization of tyrosine hydroxylase-like immunoreactivity in the
rat hippocampal formation. J. Comp. Neurol., 281:
479–495.
Milner, T.A., Lee, A., Aicher, S.A. and Rosin, D.L. (1998)
Hippocampal alpha2a-adrenergic receptors are located pre-
dominantly presynaptically but are also found postsynaptic-
ally and in selective astrocytes. J. Comp. Neurol., 395:
310–327.
Milner, T.A., Shah, P. and Pierce, J.P. (2000) Beta-adrenergic
receptors primarily are located on the dendrites of granule
cells and interneurons but also are found on astrocytes and a
few presynaptic profiles in the rat dentate gyrus. Synapse, 36:
178–193.
Minneman, K.P., Hegstrand, L.R. and Molinoff, P.B. (1979)
Simultaneous determination of beta-1 and beta-2-adrenergic
receptors in tissues containing both receptor subtypes. Mol.
Pharmacol., 16: 34–46.
Moudy, A.M., Kunkel, D.D. and Schwartzkroin, P.A.
(1993) Development of dopamine-beta-hydroxylase-posi-
tive fiber innervation of the rat hippocampus. Synapse, 15:
307–318.
Murchison, C.F., Zhang, X.Y., Zhang, W.P., Ouyang, M., Lee,
A. and Thomas, S.A. (2004) A distinct role for nor-
epinephrine in memory retrieval. Cell, 117: 131–143.
316
Nahorski, S.R., Rogers, K.J. and Edwards, C. (1975) Cerebral
glycogenolysis and stimulation of beta-adrenoreceptors and
histamine H2 receptors. Brain Res., 92: 529–533.
Neuman, R.S. and Harley, C.W. (1983) Long-lasting potent-
iation of the dentate gyrus population spike by nor-
epinephrine. Brain Res., 273: 162–165.
Nicholas, A.P., Pieribone, V. and Hokfelt, T. (1993a) Distri-
butions of mRNAs for alpha-2 adrenergic receptor subtypes
in rat brain: an in situ hybridization study. J. Comp. Neurol.,
328: 575–594.
Nicholas, A.P., Pieribone, V.A. and Hokfelt, T. (1993b) Cel-
lular localization of messenger RNA for beta-1 and beta-2
adrenergic receptors in rat brain: an in situ hybridization
study. Neuroscience, 56: 1023–1039.
Nitz, D.A. and McNaughton, B.L. (1999) Hippocampal EEG
and unit activity responses to modulation of serotonergic
median raphe neurons in the freely behaving rat. Learn.
Mem., 6: 153–167.
Oleskevich, S., Descarries, L. and Lacaille, J.C. (1989) Quan-
tified distribution of the noradrenaline innervation in the
hippocampus of adult rat. J. Neurosci., 9: 3803–3815.
Orr, G., Rao, G., Houston, F.P., McNaughton, B.L. and
Barnes, C.A. (2001) Hippocampal synaptic plasticity is mod-
ulated by theta rhythm in the fascia dentata of adult and aged
freely behaving rats. Hippocampus, 11: 647–654.
Paban, V., Soumireu-Mourat, B. and escio-Lautier, B. (2003)
Behavioral effects of arginine8-vasopressin in the Hebb-Will-
iams maze. Behav. Brain Res., 141: 1–9.
Palamarchouk, V.S., Swiergiel, A.H. and Dunn, A.J. (2002)
Hippocampal noradrenergic responses to CRF injected into
the locus coeruleus of unanesthetized rats. Brain Res., 950:
31–38.
Palamarchouk, V.S., Zhang, J., Zhou, G., Swiergiel, A.H. and
Dunn, A.J. (2000) Hippocampal norepinephrine-like voltam-
metric responses following infusion of corticotropin-releasing
factor into the locus coeruleus. Brain Res. Bull., 51: 319–326.
Pang, K. and Rose, G.M. (1987) Differential effects of nor-
epinephrine on hippocampal complex-spike and theta-neu-
rons. Brain Res., 425: 146–158.
Parfitt, K.D., Doze, V.A., Madison, D.V. and Browning, M.D.
(1992) Isoproterenol increases the phosphorylation of the
synapsins and increases synaptic transmission in dentate
gyrus, but not in area CA1, of the hippocampus. Hippocam-
pus, 2: 59–64.
Parfitt, K.D., Hoffer, B.J. and Browning, M.D. (1991) Nor-
epinephrine and isoproterenol increase the phosphorylation
of synapsin I and synapsin II in dentate slices of young but
not aged Fisher 344 rats. Proc. Natl. Acad. Sci. U.S.A., 88:
2361–2365.
Pavlides, C., Greenstein, Y.J., Grudman, M. and Winson, J.
(1988) Long-term potentiation in the dentate gyrus is induced
preferentially on the positive phase of theta-rhythm. Brain
Res., 439: 383–387.
Phelps, C.H. (1972) Barbiturate-induced glycogen accumula-
tion in brain. An electron microscopic study. Brain Res., 39:
225–234.
Pieribone, V.A., Nicholas, A.P., Dagerlind, A. and Hokfelt, T.
(1994) Distribution of alpha 1 adrenoceptors in rat brain
revealed by in situ hybridization experiments utilizing sub-
type-specific probes. J. Neurosci., 14: 4252–4268.
Pittaluga, A., Bonfanti, A. and Raiteri, M. (2000) Somatostatin
potentiates NMDA receptor function via activation of
InsP(3) receptors and PKC leading to removal of the Mg2+
block without depolarization. Br. J. Pharmacol., 130:
557–566.
Pittaluga, A., Feligioni, M., Ghersi, C., Gemignani, A. and
Raiteri, M. (2001) Potentiation of NMDA receptor function
through somatostatin release: a possible mechanism for the
cognition-enhancing activity of GABA(B) receptor antago-
nists. Neuropharmacology, 41: 301–310.
Pittaluga, A., Feligioni, M., Longordo, F., Arvigo, M. and
Raiteri, M. (2005) Somatostatin-induced activation and up-
regulation of N-methyl-D-aspartate receptor function: medi-
ation through calmodulin-dependent protein kinase II,
phospholipase C, protein kinase C, and tyrosine kinase in
hippocampal noradrenergic nerve endings. J. Pharmacol.
Exp. Ther., 313: 242–249.
Pittaluga, A. and Raiteri, M. (1987) GABAergic nerve termi-
nals in rat hippocampus possess alpha 2-adrenoceptors reg-
ulating GABA release. Neurosci. Lett., 76: 363–367.
Pittaluga, A. and Raiteri, M. (1992) N-methyl-D-aspartic acid
(NMDA) and non-NMDA receptors regulating hippocampal
norepinephrine release. I. Location on axon terminals and
pharmacological characterization. J. Pharmacol. Exp. Ther.,
260: 232–237.
Pittaluga, A., Vaccari, D. and Raiteri, M. (1997) The ‘‘kynuren-
ate test,’’ a biochemical assay for putative cognition enhanc-
ers. J. Pharmacol. Exp. Ther., 283: 82–90.
Quach, T.T., Rose, C. and Schwartz, J.C. (1978) [3H]Glycogen
hydrolysis in brain slices: responses to neurotransmitters and
modulation of noradrenaline receptors. J. Neurochem., 30:
1335–1341.
Rainbow, T.C., Parsons, B. and Wolfe, B.B. (1984) Quantita-
tive autoradiography of beta 1- and beta 2-adrenergic recep-
tors in rat brain. Proc. Natl. Acad. Sci. U.S.A., 81:
1585–1589.
Raiteri, M., Garrone, B. and Pittaluga, A. (1992) N-methyl-D-
aspartic acid (NMDA) and non-NMDA receptors regulating
hippocampal norepinephrine release. II. Evidence for func-
tional cooperation and for coexistence on the same axon ter-
minal. J. Pharmacol. Exp. Ther., 260: 238–242.
Reid, A.T. and Harley, C.W. (2005) Unpaired perforant
path and locus ceruleus activations only transiently potenti-
ate the evoked dentate gyrus population spike. Online
Abstracts Viewer/Itinerary Planner and CD-ROM. Ref
Type: Abstract.
Ribak, C.E. (1992) Local circuitry of GABAergic basket cells in
the dentate gyrus. Epilepsy Res. Suppl., 7: 29–47.
Ribak, C.E., Nitsch, R. and Seress, L. (1990) Proportion of
parvalbumin-positive basket cells in the GABAergic inner-
vation of pyramidal and granule cells of the rat hippocampal
formation. J. Comp. Neurol., 300: 449–461.

Risso, F., Grilli, M., Parodi, M., Bado, M., Raiteri, M. and
Marchi, M. (2004) Nicotine exerts a permissive role on
NMDA receptor function in hippocampal noradrenergic ter-
minals. Neuropharmacology, 47: 65–71.
Rizk, P., Salazar, J., Raisman-Vozari, R., Marien, M., Ruberg,
M., Colpaert, F. and Debeir, T. (2006) The alpha2-ad-
renoceptor antagonist dexefaroxan enhances hippocampal
neurogenesis by increasing the survival and differentiation of
new granule cells. Neuropsychopharmacology, 31: 1146–1157.
Rolls, E.T. and Kesner, R.P. (2006) A computational theory of
hippocampal function, and empirical tests of the theory.
Prog. Neurobiol., 79: 1–48.
Rose, G.M. and Pang, K.C. (1989) Differential effect of nor-
epinephrine upon granule cells and interneurons in the dent-
ate gyrus. Brain Res., 488: 353–356.
Sara, S.J. and Devauges, V. (1989) Idazoxan, an alpha-2 an-
tagonist, facilitates memory retrieval in the rat. Behav. Neu-
ral Biol., 51: 401–411.
Sara, S.J., Dyon-Laurent, C. and Herve, A. (1995) Novelty
seeking behavior in the rat is dependent upon the integrity of
the noradrenergic system. Brain Res. Cogn. Brain Res., 2:
181–187.
Sara, S.J. and Segal, M. (1991) Plasticity of sensory responses
of locus coeruleus neurons in the behaving rat: implications
for cognition. Prog. Brain Res., 88: 571–585.
Sarvey, J.M., Burgard, E.C. and Decker, G. (1989) Long-term
potentiation: studies in the hippocampal slice. J. Neurosci.
Methods, 28: 109–124.
Schacher, S., Castellucci, V.F. and Kandel, E.R. (1988) cAMP
evokes long-term facilitation in Aplysia sensory neurons that
requires new protein synthesis. Science, 240: 1667–1669.
Scheinin, M., Lomasney, J.W., Hayden-Hixson, D.M., Scham-
bra, U.B., Caron, M.G., Lefkowitz, R.J. and Fremeau Jr.,
R.T. (1994) Distribution of alpha 2-adrenergic receptor sub-
type gene expression in rat brain. Brain Res. Mol. Brain Res.,
21: 133–149.
Segal, M. (1976) Brain stem afferents to the rat medial septum.
J. Physiol., 261: 617–631.
Segal, M. and Bloom, F.E. (1976a) The action of nor-
epinephrine in the rat hippocampus. III. Hippocampal cel-
lular responses to locus coeruleus stimulation in the awake
rat. Brain Res., 107: 499–511.
Segal, M. and Bloom, F.E. (1976b) The action of nor-
epinephrine in the rat hippocampus. IV. The effects of locus
coeruleus stimulation on evoked hippocampal unit activity.
Brain Res., 107: 513–525.
Seidenbecher, T., Reymann, K.G. and Balschun, D. (1997) A
post-tetanic time window for the reinforcement of long-term
potentiation by appetitive and aversive stimuli. Proc. Natl.
Acad. Sci. U.S.A., 94: 1494–1499.
Siggins, G.R., Battenberg, E.F., Hoffer, B.J., Bloom, F.E. and
Steiner, A.L. (1973) Noradrenergic stimulation of cyclic
adenosine monophosphate in rat Purkinje neurons: an
immunocytochemical study. Science, 179: 585–588.
Stanton, P.K. and Sarvey, J.M. (1985a) Depletion of nor-
epinephrine, but not serotonin, reduces long-term
317
potentiation in the dentate gyrus of rat hippocampal slices.
J. Neurosci., 5: 2169–2176.
Stanton, P.K. and Sarvey, J.M. (1985b) The effect of high-
frequency electrical stimulation and norepinephrine on cyclic
AMP levels in normal versus norepinephrine-depleted rat
hippocampal slices. Brain Res., 358: 343–348.
Stanton, P.K. and Sarvey, J.M. (1985c) Blockade of nor-
epinephrine-induced long-lasting potentiation in the hippo-
campal dentate gyrus by an inhibitor of protein synthesis.
Brain Res., 361: 276–283.
Stanton, P.K. and Sarvey, J.M. (1987) Norepinephrine regu-
lates long-term potentiation of both the population spike and
dendritic EPSP in hippocampal dentate gyrus. Brain Res.
Bull., 18: 115–119.
Straube, T. and Frey, J.U. (2003) Involvement of beta-ad-
renergic receptors in protein synthesis-dependent late
long-term potentiation (LTP) in the dentate gyrus of
freely moving rats: The critical role of the LTP induction
strength. Neuroscience, 119: 473–479.
Straube, T., Korz, V., Balschun, D. and Frey, J.U. (2003) Re-
quirement of beta-adrenergic receptor activation and protein
synthesis for LTP-reinforcement by novelty in rat dentate
gyrus. J. Physiol., 552: 953–960.
Subbarao, K.V. and Hertz, L. (1991) Stimulation of energy
metabolism by alpha-adrenergic agonists in primary cultures
of astrocytes. J. Neurosci. Res., 28: 399–405.
Uecker, A., Barnes, C.A., McNaughton, B.L. and Reiman,
E.M. (1997) Hippocampal glycogen metabolism, EEG, and
behavior. Behav. Neurosci., 111: 283–291.
U’Prichard, D.C., Reisine, T.D., Mason, S.T., Fibiger, H.C.
and Yamamura, H.I. (1980) Modulation of rat brain
alpha- and beta-adrenergic receptor populations by
lesion of the dorsal noradrenergic bundle. Brain Res., 187:
143–154.
Usher, M., Cohen, J.D., Servan-Schreiber, D., Rajkowski, J.
and Aston-Jones, G. (1999) The role of locus coeruleus
in the regulation of cognitive performance. Science, 283:
549–554.
Vankov, A., Herve-Minvielle, A. and Sara, S.J. (1995) Re-
sponse to novelty and its rapid habituation in locus coeruleus
neurons of the freely exploring rat. Eur. J. Neurosci., 7:
1180–1187.
Walling, S.G. and Harley, C.W. (2004) Locus ceruleus activa-
tion initiates delayed synaptic potentiation of perforant path
input to the dentate gyrus in awake rats: a novel beta-ad-
renergic- and protein synthesis-dependent mammalian plas-
ticity mechanism. J. Neurosci., 24: 598–604.
Walling, S.G., Nutt, D.J., Lalies, M.D. and Harley, C.W.
(2004) Orexin-A infusion in the locus ceruleus triggers nor-
epinephrine (NE) release and NE-induced long-term potent-
iation in the dentate gyrus. J. Neurosci., 24: 7421–7426.
Wang, G.S., Chang, N.C., Wu, S.C. and Chang, A.C. (2002)
Regulated expression of alpha2B adrenoceptor during devel-
opment. Dev. Dyn., 225: 142–152.
Washburn, M. and Moises, H.C. (1989) Electrophysiological
correlates of presynaptic alpha 2-receptor-mediated
318
inhibition of norepinephrine release at locus coeruleus
synapses in dentate gyrus. J. Neurosci., 9: 2131–2140.
Winson, J. and Dahl, D. (1985) Action of norepinephrine in the
dentate gyrus. II. Iontophoretic studies. Exp. Brain Res., 59:
497–506.
Winzer-Serhan, U.H., Raymon, H.K., Broide, R.S., Chen, Y.
and Leslie, F.M. (1997a) Expression of alpha 2 adrenocep-
tors during rat brain development. I. Alpha 2A messenger
RNA expression. Neuroscience, 76: 241–260.
Winzer-Serhan, U.H., Raymon, H.K., Broide, R.S., Chen, Y. and
Leslie, F.M. (1997b) Expression of alpha 2 adrenoceptors during
rat brain development. II. Alpha 2C messenger RNA expression
and [3H]rauwolscine binding. Neuroscience, 76: 261–272.
Yavich, L., Jakala, P. and Tanila, H. (2005) Noradrenaline
overflow in mouse dentate gyrus following locus coeruleus
and natural stimulation: real-time monitoring by in vivo
voltammetry. J. Neurochem., 95: 641–650.
Zilles, K., Gross, G., Schleicher, A., Schildgen, S., Bauer, A.,
Bahro, M., Schwendemann, G., Zech, K. and Kolassa, N.
(1991) Regional and laminar distributions of alpha 1-ad-
renoceptors and their subtypes in human and rat hippocam-
pus. Neuroscience, 40: 307–320.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 19

Endocannabinoids in the dentate gyrus

Charles J. Frazier1,2,

1
Department of Pharmacodynamics, University of Florida, College of Pharmacy, Gainesville, FL 32610, USA
2
Department of Neuroscience, University of Florida, College of Medicine, Gainesville, FL 32610, USA

Abstract: Recent years have produced rapid and enormous growth in our understanding of end-
ocannabinoid-mediated signaling in the CNS. While much of the recent progress has focused on other areas
of the brain, a significant body of evidence has developed that indicates the presence of a robust system for
endocannabinoid-mediated signaling in the dentate gyrus. This chapter will provide an overview of our
current understanding of that system based on available anatomical and physiological data.

Keywords: dentate gyrus; cannabinoids; CB1; retrograde transmission; short term plasticity; presynaptic
modulation.

Introduction depolarization-induced suppression of inhibition

Both chronic and acute use of cannabis (mari-
juana) produces a wide array of physiological,
cognitive, and analgesic effects in humans that
have been widely recognized for centuries. The
active ingredient in marijuana, delta-9-tetra-
hydrocannabinol (D9-THC), was first identified in
1964 (Gaoni and Mechoulam, 1964); the first pri-
mary cannabinoid receptor in the CNS (CB1) was
identified and cloned in 1990 (Howlett et al., 1990;
Matsuda et al., 1990), and the principal endog-
enous ligands, arachidonoylethanolamide (AEA)
and 2-arachidonoylglycerol (2-AG), were identi-
fied soon thereafter (Devane et al., 1992; Sugiura
et al., 1995).
Efforts to understand the neurophysiological
consequences of CB1 receptor activation received
an enormous boost in 2001 with the discovery that
endogenous cannabinoids (ECs) act as retrograde
signaling molecules in a phenomenon known as
Corresponding author. Tel.: +1 352 392 3447;
Fax: +1 352 392 9187; E-mail: cjfraz@ufl.edu
(DSI) (Ohno-Shosaku et al., 2001; Wilson and Ni-
coll, 2001). As originally described by Pitler and
Alger in hippocampal pyramidal cells (1992), and
by Marty in cerebellar Purkinje cells (Llano et al.,
1991), DSI is a form of short term synaptic plas-
ticity whereby depolarization of a single neuron
results in transient inhibition of GABAergic affer-
ents to that same neuron. A principal role for ECs
in DSI is indicated by the fact DSI is blocked by
CB1 antagonists, occluded by CB1 receptor ago-
nists, and absent in CB1 / animals (Kreitzer and
Regehr, 2001a; Ohno-Shosaku et al., 2001; Wilson
and Nicoll, 2001; Wilson et al., 2001; Varma et al.,
2001; Yoshida et al., 2002). The retrograde nature
of the EC-mediated signaling was originally (and
most simply) indicated by failure of depolariza-
tion-induced EC release (or bath application syn-
thetic agonists) to inhibit responses to local
application of exogenous GABA, despite effec-
tively reducing evoked responses. This and other
physiological data is strongly supported by an ar-
ray of immunohistochemical studies that indicate
prominent expression of CB1 receptors both

DOI: 10.1016/S0079-6123(07)63019-2 319

320
presynaptically and on GABAergic terminals (for
general review see Freund et al., 2003, and below).
Although it has only been a short time since these
landmark studies, EC-dependent DSI governed by
fundamentally similar mechanisms has now been
identified in numerous other brain areas including
the amygdala, substantia nigra, basal ganglia, neo-
cortex, brainstem, and recently, in both granule
cells and mossy cells of the dentate gyrus (Trettel
and Levine, 2003; Yanovsky et al., 2003; Bodor et
al., 2005; Isokawa and Alger, 2005; Mukhtarov et
al., 2005; Zhu and Lovinger, 2005; Engler et al.,
2006; Hofmann et al., 2006). Cumulatively, these
findings highlight what is now the obviously broad
significance of EC-mediated retrograde signaling in
modulating inhibitory transmission in the CNS, but
nevertheless, represent only a small percentage of
recent advances. For example, it is now clear that
cannabinoid receptors are expressed on many glut-
amatergic terminals as well. Accordingly, depolari-
zation-induced EC-mediated inhibition of
excitation (DSE) has now been described in the
cerebellum (Kreitzer and Regehr, 2001b), ventral
tegmental area (Melis et al., 2004a, b), hypo-
thalamus (Di et al., 2005), amygdala (Domenici et
al., 2006), and in the hippocampus (Ohno-Shosaku
et al., 2002b), where the role of ECs in modulating
Schaffer collateral inputs to CA1 pyramidal cells
remains a particularly active area of investigation
(Hajos and Freund, 2002; Hoffman et al., 2005;
Katona et al., 2006; Takahashi and Castillo, 2006).
Further, efforts to understand the role of postsy-
naptic metabotropic glutamate receptors in EC re-
lease have proven instrumental in identifying EC-
mediated long-term depression (LTD) in a number
of areas. At present, both heterosynaptic and ho-
mosynaptic forms of EC-dependent LTD have been
described, affecting both GABAergic and glut-
amatergic synapses respectively (Marsicano et al.,
2002; Robbe et al., 2002; Chevaleyre and Castillo,
2003; Hoffman et al., 2003). While presynaptic ex-
pression still seems to be the norm in this form of
EC-dependent synaptic plasticity, there has even
been a recent example of EC-mediated, mGluR
dependent LTD in the cerebellum that is expressed
postsynaptically (Safo and Regehr, 2005). Given
the extensive nature of EC-mediated effects on
synaptic transmission, it should perhaps not come
as a surprise that fascinating metaplastic roles for
ECs are also now coming to light. For example,
Castillo and colleagues have recently and elegantly
demonstrated that EC-mediated LTD can lower the
threshold for traditional NMDA-dependent LTP in
spatially confined parts of the CA1 pyramidal cell
dendritic tree (Chevaleyre and Castillo, 2004).
These types of findings are mentioned here in
brief simply to underscore our growing apprecia-
tion for the enormous breadth of EC-dependent
signaling in CNS physiology, and to highlight the
extremely rapid nature of recent progress. In my
view, this trend shows no sign of slowing. In fact,
in this chapter, I will attempt to make the case that
the dentate gyrus is an area of significant interest,
and strong potential, for this field; albeit one that
has received comparatively little attention to date.
In making that argument I will highlight a signifi-
cant body of existing evidence clearly indicating
the presence of a robust system for EC-dependent
signaling in the dentate gyrus, I will review the
growing list of studies that reveal specific neuro-
physiological consequences of EC-dependent
signaling at identified synapses in the dentate,
and I will cover recent work suggesting several
novel roles for EC signaling (from neuroprotection
to neurogenesis) in which the dentate may be an
area of particular interest. Potential future direc-
tions for some of this work will also be discussed.
For a broader overview of specific roles for ECs in
CNS physiology the interested reader is referred to
any number of recent and outstanding reviews
(Schlicker and Kathmann, 2001; Alger, 2002; Wil-
son and Nicoll, 2002; Freund et al., 2003; Cheval-
eyre et al., 2006; Marsicano and Lutz, 2006).
A brief overview of the endocannabinoid system
There are currently two known cannabinoid re-
ceptors: the CB1 receptor, cloned in 1990 (Mats-
uda et al., 1990), and the CB2 receptor, identified 3
years later (Munro et al., 1993). Both receptors are
classic metabotropic receptors coupled to Gi/o. Of
these, the CB1 receptor appears to be intimately
involved in mediating most central effects of ECs.
It is among the most prominently expressed me-
tabotropic receptors in the mammalian CNS, with

overall expression levels approaching that of iono-
tropic receptors for glutamate and GABA. Partic-
ularly high levels of CB1 expression are found in
hippocampus (including the dentate gyrus), cortex,
cerebellum, and basal ganglia (Herkenham et al.,
1991a, c; Tsou et al., 1998a). This distribution is
consistent with well-recognized effects of can-
nabinoids on movement and memory. The CB2
receptor by contrast, has been reported in the CNS
only in striatum and cerebellum (Skaper et al.,
1996; Van Sickle et al., 2005), but is far more
abundant in the immune system where it is
thought to be involved in regulation of both im-
mune function and inflammatory responses (for
review see Berdyshev, 2000; De et al., 2000). Since
the CB2 receptor is believed to be absent in the
dentate, it will not be reviewed further here.
There are likely several endogenous ligands for
the CB1 receptor, of which two have been partic-
ularly well characterized to date: arachidonoyl-
ethanolamide (anandamide, abbreviated AEA)
and 2-AG. Both of these ligands are fatty acid
derivatives, although they differ significantly in
terms of their synthetic pathway. AEA is thought
to be synthesized by phospholipase D induced hy-
drolysis of a lipid precursor, N-arachidonoyl
phosphatidylethanolamine (N-arachidonoyl-PE)
(Di Marzo et al., 1994, 1999; Cadas et al., 1997).
One of the primary immediate precursors for
synthesis of 2-AG, by contrast, appears to be di-
acylglycerol, which is produced from phosphatidy-
linositol by phospholipase C, and subsequently
converted to 2-AG by diacylglycerol lipase
(DAGL) (Stella et al., 1997; Piomelli, 2003).
Because both AEA and 2-AG are membrane per-
meant, conventional vesicular release mechanisms
are implausible. Thus, regulation of EC-mediated
signaling is likely to depend principally on regula-
tion of the synthetic pathways described above, on
uptake from the extracellular space, and on subse-
quent breakdown. These processes also differ some-
what between AEA and 2-AG. Synthesis of AEA is
increased significantly by rapid increases in intra-
cellular calcium concentration (that reach low mi-
cromolar levels) such as occur with membrane
depolarization. The calcium dependence of AEA
production is tied to the fact that availability of N-
arachidonoyl-PE depends heavily on the activity of
321
a highly calcium-dependent enzyme, N-acyltransf-
erase, which is capable of producing N-arachido-
noyl-PE from phosphatidylethanolamine (Di Marzo
et al., 1994; Cadas et al., 1996; Piomelli, 2003).
While it is clear that increases in intracellular cal-
cium can also increase production of 2-AG through
a PLC and DAG dependent pathway (Stella et al.,
1997), the overall story here is somewhat more
complicated. For example, in some systems calcium
dependent production of 2-AG is likely to be sig-
nificantly enhanced if the calcium influx occurs co-
incident with synaptic activation of certain
metabotropic receptors (notably metabotropic
glutamate receptors and/or muscarinic acetylcholine
receptors) capable of activating a calcium-dependent
isoform of phospholipase C (Hashimotodani et al.,
2005; Maejima et al., 2005). On the other hand,
there also appear to be metabolic routes through
which metabotropic receptors can promote largely
calcium-independent production of 2-AG (Maejima
et al., 2001; Varma et al., 2001; Ohno-Shosaku et al.,
2002a, 2005; Fukudome et al., 2004).
Once AEA has reached the extracellular space,
uptake is mediated by a fast, selective, saturable,
temperature-dependent, and thus apparently car-
rier-mediated process (Beltramo et al., 1997; Hillard
et al., 1997; Piomelli et al., 1999). Subsequently,
AEA is rapidly broken down to arachidonic acid
and ethanolamine by fatty acid amide hydrolase
(FAAH), an enzyme that has an expression pattern
in the CNS that closely corresponds to that of the
CB1 receptor (Hillard et al., 1995; Cravatt et al.,
1996; Egertova et al., 1998; Piomelli et al., 1999;
Deutsch et al., 2002). The corresponding story is
again more complicated for 2-AG. While it is clear
that the anandamide transporter can also facilitate
uptake of 2-AG, and that FAAH can also break it
down, it is too early to rule out the possibility of an
additional transport mechanism for 2-AG, and it is
clear that at least one other major route for break-
down, in this case mediated by monoglyceride
lipase (MGL), is available (Dinh et al., 2002).
Activation of metabotropic CB1 receptors lo-
cated primarily on presynaptic terminals may be
coupled to adenylyl cyclase, or directly to either
Ca2+ or K+ conductances, and is typically (but
perhaps not exclusively) observed to inhibit cal-
cium dependent release of neurotransmitter. The
322
general consensus is that such activation of en-
dogenous CB1 receptors typically occurs through
retrograde transmission of endocannabinoids syn-
thesized postsynaptically, and further that 2-AG
and/or AEA are likely to be the retrograde mes-
sengers in most if not all forms of EC-mediated
retrograde signaling identified to date. It is inter-
esting to note that 2-AG is present in brain at
concentrations approximately 200-fold greater
than AEA, although this may simply reflect a
more central role in lipid metabolism and not nec-
essarily a similarly dominant role in retrograde
signaling (Sugiura et al., 1995; Stella et al., 1997).
For interesting and recent discussion of the rela-
tive roles of 2-AG and AEA in retrograde signa-
ling see Chevaleyre et al. (2006).
While this represents a brief and general summary
of the EC system as currently understood in the
brain (including the dentate) it is important to re-
iterate that this field is advancing rapidly. Significant
advances in our conceptions of this system are cer-
tainly possible, and perhaps even probable. One
area where our knowledge is currently developing at
a rapid pace concerns CB1 receptor-mediated mod-
ulation of glutamatergic transmission (see below).
Further, it is likely that additional sites of EC action
will be identified, which may include currently un-
cloned CB-like receptors and/or conventional non-
CB receptors not previously associated with EC ac-
tion. Thus it seems likely that our understanding of
the possible modulatory effects of ECs on synaptic
transmission in the CNS will only broaden with
time. A more extensive discussion of the molecular
aspects of EC signaling is available in a number of
other reviews (Di Marzo and Deutsch, 1998; Lutz,
2002; Piomelli, 2003; Sugiura et al., 2006).
Markers of the EC system in the dentate gyrus
CB1 receptor expression in the dentate gyrus: initial
findings indicate prominent role in GABAergic
neurons
Early efforts to map the distribution of putative CB
receptors in brain were hindered by the fact that
they relied largely on radiolabeled derivatives of D9-
THC that had generally poor affinity for the
receptors and a strong tendency to partition into
biological membranes (Harris et al., 1978; Roth and
Williams, 1979; Thomas et al., 1992). However, in
1991 two landmark studies capitalized on the fa-
vorable properties of a synthetic CB1 agonist,
CP55,940, to provide one of the first complete sur-
veys of putative cannabinoid receptor binding sites
in the CNS (Herkenham et al., 1991a, b). These
studies demonstrated that the molecular layer of the
dentate gyrus was among the most intensely labeled
areas in the rat brain. Comparatively low to mod-
erate levels of 3H-CP55,940 binding were also re-
ported in the hilus, with only sparse labeling over
the granule cell layer. Interestingly, these studies
also reported differential binding of 3H-CP55,940
along the septo-temporal axis of the hippocampus,
with denser binding on the dorsal (septal) end.
These findings, which were reported from rat brain,
were largely consistent with later studies using the
same radioligand in human (Glass et al., 1997).
While the original studies with 3H-CP55,940
represented a significant advance in their day, it is
noteworthy that they were completed at about the
same time that the CB1 receptor was isolated and
cloned (Matsuda et al., 1990). That feat enabled
two additional approaches to the study of CB1
receptor expression and distribution. The first to
be employed was in situ hybridization for CB1
mRNA (Mailleux and Vanderhaeghen, 1992;
Mailleux et al., 1992; Matsuda et al., 1993; Marsi-
cano and Lutz, 1999). In general, these studies
noted high levels of CB1 mRNA in a subpopula-
tion of cells in the subgranular region of the hilus,
where 3H-CP55,940 was low to moderate, with
comparatively less concentrated expression of CB1
mRNA in the molecular layer, where 3H-CP55,940
binding had been intense. This differential distri-
bution between CB1 mRNA expression and 3H-
CP55,940 binding was the basis for one of the first
arguments that functional CB1 expression may be
relegated largely to processes distal from the soma.
The second major advance enabled by cloning
of the CB1 receptor was the immunohistochemical
characterization of receptor distribution in the
CNS. This technique offered several significant
advantages over earlier efforts. First, because the
initial antibodies employed were targeted to spe-
cific amino acid sequences on the N-terminal of

the cloned CB1 receptor, they offered significantly
greater specificity than earlier radioligand binding
studies. Second, the higher resolution of immuno-
histochemical techniques allowed the first detailed
examination of CB1 expression at the subcellular
level. Third, the eventual application of elegant
dual-labeling immunohistochemical techniques al-
lowed simultaneous analysis of CB1 expression
and cell phenotype. As such, our current under-
standing of CB1 expression and distribution in the
CNS depends heavily on these techniques, and re-
sults relevant to the dentate will be reviewed in
some detail below.
The first immunohistochemical studies of CB1
distribution in the CNS clearly demonstrated in-
tense immunoreactivity in the molecular layer of the
dentate gyrus (Pettit et al., 1998; Tsou et al., 1998a;
Katona et al., 1999). Consistent with early binding
studies, the intensity of this signal was reported
to be highest in the inner third of the molecular
layer, and on the septal end of the hippocampus
(Fig. 1AA, Tsou et al., 1998a). However, based on
the greater resolution of these techniques, it now
became apparent that this area specifically con-
tained many CB1-immunoreactive nerve fibers, but
very few labeled cell bodies. By contrast, intensely
immunoreactive neuronal cell bodies were identified
in the hilus, just below the granule cell layer. These
cells were reported to represent as much as 62% of
all CB1-immunoreactive cell bodies in the dentate,
and had characteristic pyramidal morphology, with
a single primary dendrite that often extended with-
out branching through the granule cell layer toward
the molecular layer (Fig. 1AB C, Tsou et al., 1998a;
Katona et al., 1999). The remaining CB1-positive
cell bodies were located largely in the hilus and ex-
hibited a variety of morphologies.
The subsequent application of dual-labeling
immunohistochemical techniques represented
a major advance in this area, as it allowed the
determination that the vast majority of CB1-
immunoreactive cell bodies were also positive for
CCK. Specifically, between 80 and 95% of all
CCK-positive cell bodies in the dentate were re-
ported to also be CB1-positive, with a comparably
high percentage of CB1-positive cells being CCK-
positive. In stark contrast, virtually none of the
parvalbumin-positive basket cells expressed CB1
323
(Katona et al., 1999; Tsou et al., 1999). Further
examination of these results at the electron micro-
scopic level indicated cellular expression of CB1
was largely or exclusively confined to cytoplasmic
organelles, while surface expression of CB1 was
very high on CCK-positive axon terminals that
were often clustered around principal cell somas
and proximal dendrites (Katona et al., 1999). Al-
though these particular observations were origi-
nally made primarily in hippocampal cells in CA1
and CA3, it now seems reasonable to predict that
surface expression of CB1 on GABAergic neurons
in the dentate may also be restricted to nerve ter-
minals. This hypothesis is consistent with the
differential distribution of 3H-CP55,940 binding
and in situ hybridization described earlier, and is
also strongly supported by a study from the
Freund laboratory the following year that specifi-
cally identified CB1-positive axon terminals in the
hilus (Acsady et al., 2000). Although this study did
not directly perform dual-labeling experiments for
both CB1 and CCK, it did make the striking ob-
servation that both CB1-positive and (in separate
experiments) CCK-positive terminals displayed
very similar and yet highly unusual target selec-
tivity. Specifically they found these terminals made
abundant synaptic contacts on the soma and prox-
imal dendrites of the GluR2/3 and calcitonin gene-
related peptide (CGRP)-positive hilar mossy cells,
but largely avoided both parvalbumin and sub-
stance P (and thus also CCK)-immunoreactive in-
terneurons in the hilus.
Cumulatively, the data described so far clearly
indicates prominent expression of CB1 receptors in
the dentate, particularly on presynaptic terminals of
CCK-positive interneurons, many of which have
cell bodies located in the subgranular zone. The
specific distribution of CB1-positive terminals fur-
ther suggests a likely role for CB1 activation in
regulating GABAergic transmission to both gran-
ule cells and hilar mossy cells. Nevertheless the ap-
parent absence of CCK and CB1positive terminals
making synaptic contact with other basket cells
clearly distinguishes hilar circuitry from that in
other hippocampal subfields, and may hold clues to
understanding the unique physiology of this area.
While the vast majority of the work described
above was completed in either Sprague-Dawley or
324

Fig. 1. CB1 receptor expression in the dentate gyrus. (AA) An early study demonstrates intense immunoreactivity for CB1 receptors in
the dentate gyrus, especially in the inner molecular layer (mo). (AB C) Although dentate granule cells appear to be CB1 negative, there
are many intensely stained neurons found at the base of the granule cell layer, with characteristic pyramidal morphology. Other
intensely stained neurons were found in the hilus (not shown). Note that labeled cell bodies were generally absent in the molecular
layer. Scale bars: 200 mm, 50 mm, 20 mm in AA–C respectively. Adapted with permission from Tsou et al. (1998a). (B) A recent
immunohistochemical analysis of CB1 receptor expression in wild type (WT), CaMK-CB1 / , GABA-CB1 / , and complete CB1 /
mice provided clear evidence of CB1 expression on glutamatergic terminals (as well as GABAergic terminals) in the inner molecular
layer of the dentate gyrus. Areas indicated by black boxes in the top row are shown in higher magnification in the bottom row. Scale
bar in H is 100 mm. Adapted with permission from Monory et al. (2006).

Wistar rats, the basic features of CB1 expression in
the dentate described so far appear to be largely
conserved in gray mouse lemur, primate, and hu-
man brain (Ong and Mackie, 1999; Katona et al.,
2000; Harkany et al., 2005). Further, although
much of this work has relied on the N-terminal
antibodies to the CB1 receptor that are identical or
similar to the one originally developed by Tsou et
al. (1998a), key features of this expression pattern
have also been observed using antibodies that target
intracellular sites on the C-terminal end of the CB1
receptor (Egertova and Elphick, 2000; Hajos et al.,
2000). Nevertheless, at this juncture it is important
to emphasize that the data as summarized above do
not represent the full picture with respect to likely
sites of EC action in the dentate. Our current un-
derstanding of this topic with respect to glut-
amatergic systems is reviewed below.
CB1 receptor expression in the dentate gyrus:
emphasis on glutamatergic neurons
Compared to the results summarized above, the
question of CB1 receptor expression by glut-
amatergic neurons has so far been a contentious

one. This is particularly true in areas CA1 and
CA3 where apparent expression of low levels of
CB1 mRNA in pyramidal neurons as indicated by
in situ hybridization generally has not been easy to
reconcile with the apparent lack of CB1 immuno-
reactivity on glutamatergic cell bodies or axon
terminals as indicated by immunohistochemical
techniques. This discrepancy has been made all the
more acute by divergent reports from a number of
outstanding laboratories regarding the physiolog-
ical effects of ECs on glutamatergic transmission
in the Schaffer collateral pathway (Hajos and
Freund, 2002; Hoffman et al., 2005; Katona et al.,
2006; Takahashi and Castillo, 2006). However, in
general, the current details of this debate are out-
side the scope of this review. The dentate gyrus has
to a large extent been spared from a similar debate
in part because, unlike in area CA1 and CA3, in
situ hybridization tends to agree with both
immunohistochemical analysis and early radiolig-
and binding studies in suggesting that dentate
granule cells are CB1-negative. Nevertheless, it
would be inaccurate to say there have been no di-
vergent results in this area at all. For example, a
unique N-terminal antibody targeting amino acids
83–91 of the CB1 receptor employed by Pettit et al.
(1998) does label dentate granule cells, while the
Tsou et al. (1998a) antibody that fails to label
dentate granule cells in rats and humans has been
reported to do so in non-human primates. While
the precise reasons for these discrepancies in the
dentate gyrus have still not been completely re-
solved, very recent work using novel C-terminal
antibodies to CB1 has simultaneously reinforced
the conclusion that dentate granule cells are CB1-
negative, while elegantly demonstrating a poten-
tially major role for EC signaling via the other
glutamatergic cell type in the dentate, the hilar
mossy cells.
The first of two significant reports in this area
was from the laboratory of Tamas Freund
(Katona et al., 2006). This study used a guinea
pig anti-CB1 antibody originally characterized by
Fukudome et al. (2004) that targets amino acids
443–473 on the C-terminal end of the protein. The
pattern of CB1 immunoreactivity observed with
this antibody had similar distribution to that pre-
viously reported, however the intensity of labeling
325
was notably greater than had been observed pre-
viously, particularly in the inner third of the mo-
lecular layer. Further, upon examination of the
results at the electron microscopic level it became
clear that, in contrast to results with earlier anti-
bodies, as many as 80% of asymmetrical (i.e. ex-
citatory) synapses in the inner molecular layer, and
30–50% of asymmetrical synapses in other layers
were unambiguously positive for CB1. No differ-
ences in this pattern were observed between
C57BL/6 and CD1 mice, and the authors noted
that preliminary data suggests similar results will
be obtained using this antibody in both rat and
human hippocampus.
A second extremely recent study has reproduced
and extended these findings using yet another
unique antibody, in this case a goat anti-CB1 an-
tibody targeting the C-terminal 77 amino acids of
the protein (Monory et al., 2006). While providing
other significant insights to be reviewed later in
this chapter, this study confirms the presence of
CB1-positive glutamatergic terminals in the inner
third of the molecular layer by elegantly demon-
strating clear immunohistochemical labeling for
CB1 that is resistant to selective CB1 knockouts
targeted specifically at GABAergic neurons, but
reduced by selective CB1 knockouts targeting
glutamatergic cells (Fig. 1B, for more on this tech-
nology, also see Marsicano et al., 2003). The ob-
servation that apparently glutamate-specific CB1
labeling was concentrated in the inner third of the
molecular layer, combined with the observation
that dentate granule cells continued to lack
immunoreactivity for CB1 in all models, suggested
the hypothesis that glutamate-specific CB1 labe-
ling could be related to presynaptic expression on
the axon terminals of hilar mossy cells. Consistent
with that hypothesis, these authors further use
dual-labeling in situ hybridization techniques to
show strong co-localization of mRNA for CB1
and for the vesicular glutamate transporter type 1
(VGlutT1) both in the soma of hilar mossy cells
and in presynaptic terminals forming asymmetric
synaptic contacts in the inner molecular layer.
These findings do not conflict with earlier work
using other antibodies that indicated robust ex-
pression of CB1 on CCK-positive axon terminals
in this area. Instead, they indicate additional
326
expression of CB1 on glutamatergic terminals
likely belonging to hilar mossy cells that was not
previously detected. This is a striking conclusion
by any measure, and I suspect it will have a sig-
nificant impact on future work in this area. Nev-
ertheless, the precise reasons that CB1 expression
by glutamatergic terminals in the molecular layer
was not detected earlier and continues to be some-
what unclear. One likely possibility is that the N-
terminal antibodies generally employed in earlier
work simply lack the sensitivity to detect lower
levels of CB1 expression present on glutamatergic
terminals. Another compatible possibility that has
been raised, but to my knowledge not yet formally
reported, is that there exists a ‘CB1 receptor-in-
teracting protein’ that is exclusively expressed by
excitatory cells and is capable of interfering with
sites targeted by some of the earlier C-terminal
antibodies (Niehaus et al., 2004).
CB1 receptor expression in the dentate gyrus: co-
expression with other receptor subtypes
While a concerted effort has been made to charac-
terize CB1 expression with respect to glutamatergic
vs. GABAergic neurons and presynaptic vs. somatic
expression as described above, some otherwise iso-
lated reports have specifically examined co-expres-
sion of CB1 with other receptor subtypes.
Noteworthy examples in the dentate include a re-
port by Cristino et al. (2006) who used dual-labeling
immunohistochemical techniques to suggest co-ex-
pression of both CB1 and the vanilloid receptor
TRPV1 in many non-pyramidal bipolar neurons of
the dentate, particularly in the molecular layer. Fur-
ther, Morales and Backman (2002) used double
labeling in situ hybridization techniques to demon-
strate a strong co-localization of mRNA for both
CB1 receptors and 5-HT3A receptors in the sub-
granular interneurons of the dentate hilus. This re-
sult was generally consistent with Hermann et al.
(2002) who also demonstrated co-localization of
CB1 with 5-HT1B, and 5-HT3 receptors, and further
noted the extent of co-expression was higher in
neurons with high levels of CB1 mRNA. This study
also was noteworthy in identifying co-expression of
CB1 with D2 receptors throughout the dentate.
These findings suggest the possibility of significant
cross talk between several transmitter systems in
regulating the activity of CCK-positive interneurons
in the dentate.
Other markers of the EC system in the dentate
gyrus
Although a few additional approaches have been
employed, most efforts to demonstrate the pres-
ence of the EC system in the hippocampus and
dentate gyrus without relying on detection of CB1
have focused on the two primary enzymes for
degradation of AEA and 2-AG reviewed earlier:
FAAH and MGL, respectively.
At this point, immunohistochemical studies tar-
geting various areas on the C-terminal end of
FAAH have indicated that expression is predom-
inantly postsynaptic and largely confined to hippo-
campal principal cells and dentate granule cells
(Tsou et al., 1998b; Romero et al., 2002; Egertova
et al., 2003; Gulyas et al., 2004). This distribution is
consistent with that reported for mRNA for FAAH
(Thomas et al., 1997), and indicates that FAAH
expression is largely restricted to neurons known to
receive CB1-positive afferent inputs. The diffuse
labeling for FAAH outside the cell layers noted by
Egertova et al. (2003) may be an artifact of tissue
preparation, or as noted by the authors, could in-
dicate an ability of principal cells to excrete FAAH
into the extracellular space. Interestingly, Gulyas et
al. (2004) has also noted that intracellular expres-
sion of FAAH was most often localized to the sur-
face of organelles associated with calcium storage.
In contrast to the predominantly postsynaptic
localization of FAAH, MGL appears to be ex-
pressed primarily presynaptically in both the hip-
pocampus and dentate gyrus. Specifically,
immunoreactivity for MGL has been identified
on the mossy fiber axon terminals of dentate gran-
ule cells, as well as on the terminals of CA3 py-
ramidal cells and some interneurons (Gulyas et al.,
2004). In contrast, absence of MGL immunoreac-
tivity in certain areas of CA3 and the molecular
layer noted in the same study suggested that CA1
and perforant path inputs are likely to lack this
enzyme.

Another noteworthy technique in this area has
been to examine the expression of DAGL, a bio-
synthetic enzyme central to the production of 2-
AG. Two very recent studies that were published
almost simultaneously have both demonstrated
that DAGL, as detected by immunohistochemical
analysis using several different antibodies, is highly
expressed by principal neurons in the hippocam-
pus and granule cells (and likely mossy cells) of the
dentate gyrus, and has further indicated that such
expression is almost completely restricted to the
head and neck of dendritic spines (Katona et al.,
2006; Yoshida et al., 2006). Katona et al. (2006)
also showed specifically that presynaptic terminals
opposite DAGL positive spines are CB1-positive,
indicating that DAGL is perfectly positioned to
promote retrograde transmission, presumably me-
diated by 2-AG at numerous sites throughout the
hippocampus and dentate gyrus.
Physiological role for ECs in the dentate
Early efforts to examine the neurophysiological
effects of cannabinoids began well before isolation
and cloning of the CB1 receptor. Both in vivo and
in vitro approaches were employed; experimental
measures ranged from auditory and sensory
evoked potentials to extracellular field potentials
and unit recordings. Using such techniques, a
number of studies specifically demonstrated effects
of cannabinoids on identified cells or circuits in the
dentate gyrus (see Kujtan et al., 1983; Campbell et
al., 1986a, b; Wilkison and Pontzer, 1987; Hamp-
son et al., 1989). However, just as cloning of the
CB1 receptor and development of powerful
immunohistochemical techniques dramatically ex-
panded our understanding of CB1 expression in
the brain, major advances in electrophysiological
techniques have now provided the technology nec-
essary to dissect the effects of both endogenous
and exogenous cannabinoids at the synaptic level.
Admittedly, applications of such technology to the
study of cannabinoid-dependent systems in the
dentate gyrus are only recently beginning to ap-
pear in the literature. Nevertheless, the emphasis
below will be on such recent findings, with a par-
ticular effort to note when and where physiological
327
findings to date are consistent with predictions
based on the existing anatomical information re-
viewed above.
DSI in the dentate gyrus
Anatomical studies reviewed earlier strongly sug-
gest that CB1-positive GABAergic terminals of
CCK-positive basket cells make numerous synap-
tic contacts with both granule cells and hilar mossy
cells. These findings suggest such inputs may be
modulated by CB1 agonists. Further, the presence
of FAAH and/or DAGL in both granule cells and
mossy cells suggests they may represent sources of
ECs. Detailed physiological studies have now
strongly reinforced these conclusions.
The first evidence that GABAergic inputs to
dentate granule cells were sensitive to exogenous
cannabinoids was published in 2000 when Hajos et
al. used whole-cell recording techniques to dem-
onstrate that evoked inhibitory postsynaptic cur-
rents (eIPSCs) recorded from dentate granule cells
were reduced to 3676% of baseline by bath ap-
plication of WIN55,212-2 in wild type but not in
CB1 / mice (Hajos et al., 2000). Several years
later, Nakatsuka et al. (2003) similarly reported a
WIN55,212-2-mediated, AM-251-sensitive reduc-
tion in both frequency and amplitude of sponta-
neous IPSCs (sIPSCs) recorded from dentate
granule cells in the human dentate gyrus. These
findings indicated that GABAergic inputs to dent-
ate granule cells are indeed inhibited by CB1 re-
ceptor activation, however they did not directly
address the question of whether granule cells were
capable of activating these receptors by signaling
via ECs.
That question was answered in the affirmative
with the first complete characterization of DSI in
dentate granule cells in 2005 (Isokawa and Alger,
2005). This study specifically demonstrated that
direct depolarization of a single dentate granule
cell reduces the amplitude of both sIPSCs and
IPSCs evoked by stimulation of the molecular
layer (Fig. 2A). Endocannabinoids were impli-
cated in this process because both CP55,942 and
WIN55,212-2 mimicked and occluded DSI and
because DSI was blocked by the CB1 antagonist
328

Fig. 2. DSI in dentate granule cells and hilar mossy cells. (AA) Dentate gyrus of the hippocampal slice (scale, 50 mm). (AB) Differential
interference contrast image of dentate granule cells (scale, 10 mm). (AC) Single dentate granule cell filled, and visualized with flu-
orescence microscopy during whole cell recording (scale, 10 mm). (AD) DSI of eIPSCs (left, top) and sIPSCs (right, top) is observed in
dentate granule cells. In both cases, magnitude of DSI is reduced by bath application of thapsigargin (TG, 2 mM, bottom). (AE)
Thapsigargin also reduced the calcium transients produced by DSI-inducing depolarizing voltage steps. Adapted with permission from
Isokawa and Alger (2005). (B) Depolarization induced release of endogenous cannabinoids from hilar mossy cells preferentially
inhibits calcium-dependent exocytosis. DSI in of eIPSCs is apparent in an individual mossy cell. DSI of miniature IPSCs is absent in
the same cell. Following bath application of KCl and CaCl2, DSI of mIPSCs becomes apparent. Summary data are presented in BD,
where numbers on the bars are n values. Adapted with permission from Hofmann et al. (2006). (See Color Plate 19.2 in color plate
section.)

SR141716A. These authors further demonstrated
that the magnitude of DSI correlated with the ex-
tent of depolarization-induced increases in [Ca2+]i
and that DSI could be blocked by chelating
postsynaptic calcium with internal BAPTA. Fi-
nally, this study indicated a particular role for
calcium release from ryanodine sensitive internal
calcium stores in the mechanism of DSI, as deple-
tion of those stores reduced both [Ca2+]i and the
magnitude of DSI.
Interestingly, an earlier study had demonstrated
that although DSI of sIPSCs was absent in dentate
granule cells in control conditions, it could be de-
tected 1 week or more following induction of

febrile seizures (Chen et al., 2003). Induction of
febrile seizures also enhanced EC-mediated signa-
ling in area CA1, apparently due to increases in
CB1 expression that were restricted to CCK-pos-
itive GABAergic terminals. While the finding that
febrile seizures upregulate EC-mediated signaling
throughout the hippocampus and dentate gyrus
remains quite intriguing, the apparent discrepancy
with respect to the ability to detect DSI in dentate
granule cells in control conditions remains unex-
plained.
Also consistent with predictions based on ana-
tomical data is a recent report from my laboratory
which provides the first published characterization
of DSI as observed in hilar mossy cells (Hofmann
et al., 2006). Specifically, we reported DSI of
eIPSCs in hilar mossy cells that depends on both
postsynaptic calcium influx and presynaptic CB1
receptors. The magnitude of DSI in hilar mossy
cells was directly dependent on depolarization du-
ration, and enhanced by bath application of car-
bachol (CCh). Further, the presynaptic mechanism
of inhibition was found to be largely or exclusively
selective for calcium-dependent release events (Fig.
2B). We also used dual whole cell recording tech-
niques to demonstrate that there are tight spatial
constraints on diffusion of ECs released from hilar
mossy cells. Finally, we were intrigued to observe
two different forms of expression of DSI of sIPSCs
that depended on whether the sIPSCs had high
spectral power at theta frequencies. Cumulatively,
these results suggested a prominent role for EC-
mediated signaling between hilar mossy cells and
GABAergic afferents in normal hilar function, and
raise interesting questions about the effects of both
ECs and CCh on coordinated activity in GAB-
Aergic networks. An earlier report from another
laboratory indicating DSI of sIPSCs could be de-
tected in hilar mossy cells appeared in abstract
form (Howard et al., 2003).
Evidence that ECs modulate glutamatergic
transmission in the dentate gyrus
The axon terminals of mossy fibers, perforant path
inputs, mossy cells, and back-projecting CA3 py-
ramidal cells all form glutamatergic synapses
329
within the dentate gyrus. Nevertheless, I think it
is fair to say that despite a number of studies
clearly indicating antiepileptic (Consroe et al.,
1975; Consroe and Wolkin, 1977; Wallace et al.,
2001, 2002, 2003; Blair et al., 2006), and neuro-
protective (Nagayama et al., 1999; Braida et al.,
2000; Van der Stelt et al., 2001a, b, 2002;
Khaspekov et al., 2004) effects of cannabinoids,
there is a notable lack of physiological studies that
directly test the hypothesis that ECs can modulate
excitatory transmission at these synapses. For ex-
ample, to my knowledge, there are currently no
published reports directly testing effects of ECs on
mossy fiber transmission. Preliminary data from
my laboratory is consistent with the weight of ex-
isting anatomical evidence in suggesting that iso-
lated mossy fiber inputs to CA3 pyramidal cells are
indeed CB1-negative; however it must be noted
that neither our preliminary work nor existing an-
atomical data can at present explicitly rule out the
possibility of additional non-CB1 receptors for
endocannabinoids at these terminals.
There are somewhat more available data with
respect to perforant path inputs to dentate granule
cells. Although these inputs were initially reported
to be insensitive to D9-THC using in vivo extra-
cellular recording techniques (Wilkison and Po-
ntzer, 1987), two later reports using different
agonists and activation strategies have been con-
sistent with the hypothesis that these terminals
may express presynaptic CB1 receptors. One such
study explicitly demonstrated that bath applica-
tion of WIN55,212-2 produced a rightward shift in
the input/output curve, and a reduction in paired
pulse facilitation, of perforant path synaptic po-
tentials recorded from the outer third of the mo-
lecular layer (Kirby et al., 1995). Further, another
study examining field potential responses pro-
duced by stimulation of the medial perforant path
strikingly found that reductions in fEPSPs pro-
duced by an acetylcholinesterase inhibitor
(physostigmine) were blocked by the CB1 recep-
tor antagonist AM-251, but not by methoctra-
mine, an antagonist presumably selective for
presynaptic (m2) muscarinic acetylcholine recep-
tors. These results suggested that cholinergic fa-
cilitation of EC release can lead to tonic EC-
mediated inhibition of perforant path inputs to
330
dentate granule cells, and as such may be consist-
ent with anatomical studies which indicate low
levels of CB1 expression in the outer two thirds of
the molecular layer.
There is also very limited available data that
directly addresses whether ECs play a role in mod-
ulating release of glutamate from mossy cell axon
terminals. Likely sensitivity to inhibition by CB1
activation has been suggested, of course, by recent
and striking immunohistochemical findings re-
viewed earlier, but it is also implicated by the
demonstration that WIN55,212-2 sensitive eIPSCs
can be produced in dentate granule cells by stim-
ulation in the inner molecular layer, and perhaps
most notably by the observation that CB1 receptor
knockouts that are selective for glutamatergic neu-
rons reduce the threshold for kainic acid-induced
seizures (Marsicano et al., 2003; Monory et al.,
2006). Nevertheless, it is likely that much remains
to be learned from a detailed electrophysiological
examination of EC-mediated signaling at this
synapse. Finally, there is also a complete lack of
information in the literature about the EC sensi-
tivity of back projecting associational/commis-
sural inputs to hilar neurons or dentate granule
cells. As the functional consequences of CB1 re-
ceptor activation on Schaffer collateral inputs to
CA1 pyramidal cells is currently a matter of much
debate and investigation (see ‘Introduction’) it will
not be reviewed further here.
A role for endocannabinoids in neurogenesis?
In contrast to original conclusions, groundbreak-
ing work over the last 10–15 years has clearly
demonstrated that neurogenesis does occur in se-
lective regions of the adult mammalian CNS, most
notably in the olfactory bulb and dentate gyrus
(for recent review, see Lledo et al., 2006). Indeed,
the subgranular zone of the dentate gyrus, previ-
ously recognized for intense expression of CB1 re-
ceptors in a subset of inhibitory interneurons, has
now also been recognized as a primary neuropro-
liferative zone. Further, Jiang et al. (2005) has used
immunohistochemical techniques coupled with in-
jections of 5-bromo-2-deoxyuridine (BrdU, a base
analog of thymidine) to show that dividing cells in
the subgranular region of the adult rat dentate
gyrus are immunoreactive for CB1. These striking
findings, coupled with work completed in culture,
have clearly suggested the hypothesis that EC-me-
diated signaling might modulate neurogenesis in
vivo. Consistent with that hypothesis, Jiang et al.
(2005) also demonstrated that chronic administra-
tion of a synthetic cannabinoid (HU210) increased
neurogenesis in the adult dentate gyrus, coincident
with anxiolytic and antidepressant like effects. At
present, two other studies have made a compatible
observation that BrdU-labeling of newborn cells in
the adult dentate gyrus is dramatically reduced in
CB1 knockout animals compared to wild type
mice (Jin et al., 2004; Kim et al., 2006). However,
the full story here is likely to be quite complicated
because apparently paradoxical increases in ne-
urogenesis produced by administration of CB1
antagonists, including SR141716A and AM-251,
prior to BrdU-labeling have been noted by several
laboratories (Rueda et al., 2002; Jin et al., 2004).
Further, in at least one case, an AEA-mediated
reduction in neurogenesis in the dentate gyrus has
been reported (Rueda et al., 2002). One possible
explanation is that apparent inhibitory effects of
cannabinoids on neurogenesis, when they occur,
may be CB1 receptor-independent, and perhaps
mediated by direct activation of VR1. This is con-
sistent with the observation by Jin et al. (2004) that
SR141716A-mediated increases in neurogenesis
were preserved in CB1 knockouts, but absent in
VR1 knockouts, and also compatible with a much
earlier report indicating that both AEA and
SR141716A may interact with VR1 receptors
(Zygmunt et al., 1999). In contrast, the most re-
cent mechanism suggested for CB1-dependent in-
creases in neurogenesis implicate down regulation
of nitric oxide synthase (Di Marzo et al., 2002;
Kim et al., 2006).
Non-CB receptor targets of ECs relevant to the
dentate
A final topic worthy of particular consideration here
is the possible abundance of non-CB1 targets
for EC-mediated signaling in the dentate gyrus. Al-
though 2-AG has been postulated to be the

retrograde messenger in many forms of EC-medi-
ated signaling, a role for AEA is far from ruled out.
In fact, a role for AEA is implied by the presence of
FAAH in hippocampal principal cells and dentate
granule cells, and has occasionally been implicated
in depolarization-induced as opposed to synaptical-
ly driven release of ECs (for review, see Chevaleyre
et al., 2006). The extent to which EC-mediated
signaling involves AEA is a particularly important
question in large part because AEA and its met-
abolites, far more than 2-AG, have been associated
with action at non-CB receptor sites (for review, see
Di Marzo et al., 2002).
One such site that may have an important role
to play in the dentate gyrus is the VR1 receptor. Of
course, VR1 receptor expression is typically asso-
ciated with sensory fibers, and thus much of the
data implicating an interaction of AEA and VR1 is
based on apparent AEA-mediated increases in
VR1-dependent release of CGRP. However, AEA
and its analogs have also been shown to directly
activate recombinant VR1 receptors in several
types of assays and expression systems (Bisogno et
al., 2001; Di Marzo et al., 2001; Ralevic et al.,
2001). Further, recent immunohistochemical data
has indicated a prominent expression of VR1 in
the brain, with particularly high levels in several
areas, including the dentate gyrus (Toth et al.,
2005). Collectively these findings suggest an inter-
action between AEA and VR1, which may be rel-
evant in understanding EC-mediated signaling in
the dentate. This argument is further supported,
albeit indirectly, by reports in other areas of the
CNS where VR1 activation by AEA and/or caps-
aicin has been shown to directly modulate synaptic
transmission (Marinelli et al., 2002, 2003). Other
potential targets for AEA that could be significant
in the dentate include a7-containing nicotinic ace-
tylcholine receptors and 5-HT3 receptors. Inter-
estingly, in both these cases, action of AEA has
been shown to have inhibitory effects on recom-
binant receptors expressed in Xenopus oocytes (Oz
et al., 2002, 2003), while additional lines of evi-
dence implicating cannabinoids in serotonergic
function have been previously reviewed (Morales,
2006). Further, a potential non-competitive inter-
action between AEA and muscarinic acetylcholine
receptors has been suggested in both binding and
331
functional assays (Christopoulos and Wilson,
2001; Lanzafame et al., 2004).
Conclusions and future directions
The data reviewed here indicate the presence of a
robust system for EC-mediated signaling in the
dentate gyrus that is clearly composed of both
synthetic and degradative enzymes for end-
ocannabinoids as well as cannabinoid receptors.
A significant body of work reviewed here has
largely coalesced to indicate that there is promi-
nent expression of CB1 receptors on the axon ter-
minals of CCK-positive basket cells in the dentate,
and that these CB1-positive terminals likely form
numerous synaptic contacts with both granule
cells and hilar mossy cells. Consistent with that
view, recent work has demonstrated robust mod-
ulation of GABAergic synapses to both granule
cells and mossy cells that does indeed depend on
retrograde activation of CB1 receptors by ECs
(Isokawa and Alger, 2005; Hofmann et al., 2006).
Until very recently, available data suggested that
we would be unlikely to observe major effects of
ECs at other types of synapses in the dentate.
However, dramatic improvements in immuno-
histochemical techniques have now clearly chal-
lenged that conclusion by indicating robust
expression of CB1 receptors on excitatory termi-
nals in the inner third of the molecular layer that
are likely to arise from hilar mossy cells, and that
appear to be key players in determining threshold
to kainic acid-induced seizures (Marsicano et al.,
2003; Katona et al., 2006; Monory et al., 2006).
These striking findings have strongly reinforced
the notion of the dentate gyrus as a key player in
EC-mediated signaling, and may highlight it as an
attractive system for future research on EC-medi-
ated modulation of glutamatergic transmission.
That being said, it seems clear that we still have
much to learn from a continued characterization
of the EC system at the electron microscopic level,
particularly when coupled with detailed physio-
logical studies at the synaptic level. In fact, a
number of issues beyond transmitter phenotype
are likely to be of particular importance. These
include developing a better understanding of
332
physiological stimuli that trigger EC-mediated
signaling and more accurate measures of the scope
of the effects. This will likely involve further ex-
amination of the role of postsynaptic metabotrop-
ic receptors in the release process and careful
evaluation of potential heterosynaptic and meta-
plastic signaling mechanisms. At a larger level, it
will also be important to understand broader
effects of ECs on network activity, where both in
vitro and in vivo studies focused on the dentate
gyrus will likely be central to elucidating the ne-
uroprotective and antiepileptic effects of can-
nabinoids. Finally, the potential role of ECs in
neurogenesis is an extremely important topic that
clearly has broad implications for neurobiology.
Although the dentate gyrus has arguably been
partially overlooked in the rapid pace of the last 5
years, it now seems clear that those motivated to
understand EC-mediated signaling in the CNS
have cause to look toward the dentate, while con-
versely, those who seek a more thorough under-
standing of the dentate may find significant value
in further studies of cannabinoids.
References
Acsady, L., Katona, I., Martinez-Guijarro, F.J., Buzsaki, G.
and Freund, T.F. (2000) Unusual target selectivity of peri-
somatic inhibitory cells in the hilar region of the rat hippo-
campus. J. Neurosci., 20: 6907–6919.
Alger, B.E. (2002) Retrograde signaling in the regulation of
synaptic transmission: focus on endocannabinoids. Prog.
Neurobiol., 68: 247–286.
Beltramo, M., Stella, N., Calignano, A., Lin, S.Y., Ma-
kriyannis, A. and Piomelli, D. (1997) Functional role of
high-affinity anandamide transport, as revealed by selective
inhibition. Science, 277: 1094–1097.
Berdyshev, E.V. (2000) Cannabinoid receptors and the regula-
tion of immune response. Chem. Phys. Lipids, 108: 169–190.
Bisogno, T., Hanus, L., De, P.L., Tchilibon, S., Ponde, D.E.,
Brandi, I., Moriello, A.S., Davis, J.B., Mechoulam, R. and
Di Marzo, V. (2001) Molecular targets for cannabidiol and
its synthetic analogues: effect on vanilloid VR1 receptors and
on the cellular uptake and enzymatic hydrolysis of anand-
amide. Br. J. Pharmacol., 134: 845–852.
Blair, R.E., Deshpande, L.S., Sombati, S., Falenski, K.W.,
Martin, B.R. and Delorenzo, R.J. (2006) Activation of the
cannabinoid type-1 receptor mediates the anticonvulsant
properties of cannabinoids in the hippocampal neuronal cul-
ture models of acquired epilepsy and status epilepticus. J.
Pharmacol. Exp. Ther., 317: 1072–1078.
Bodor, A.L., Katona, I., Nyiri, G., Mackie, K., Ledent, C.,
Hajos, N. and Freund, T.F. (2005) Endocannabinoid signa-
ling in rat somatosensory cortex: laminar differences and in-
volvement of specific interneuron types. J. Neurosci., 25:
6845–6856.
Braida, D., Pozzi, M. and Sala, M. (2000) CP 55,940 protects
against ischemia-induced electroencephalographic flattening
and hyperlocomotion in Mongolian gerbils. Neurosci. Lett.,
296: 69–72.
Cadas, H., di, T.E. and Piomelli, D. (1997) Occurrence and
biosynthesis of endogenous cannabinoid precursor, N-ara-
chidonoyl phosphatidylethanolamine, in rat brain. J. Neuro-
sci., 17: 1226–1242.
Cadas, H., Gaillet, S., Beltramo, M., Venance, L. and Piomelli,
D. (1996) Biosynthesis of an endogenous cannabinoid pre-
cursor in neurons and its control by calcium and cAMP. J.
Neurosci., 16: 3934–3942.
Campbell, K.A., Foster, T.C., Hampson, R.E. and Deadwyler,
S.A. (1986a) Delta 9-tetrahydrocannabinol differentially
affects sensory-evoked potentials in the rat dentate gyrus. J.
Pharmacol. Exp. Ther., 239: 936–940.
Campbell, K.A., Foster, T.C., Hampson, R.E. and Deadwyler,
S.A. (1986b) Effects of delta 9-tetrahydrocannabinol on sen-
sory-evoked discharges of granule cells in the dentate gyrus
of behaving rats. J. Pharmacol. Exp. Ther., 239: 941–945.
Chen, K., Ratzliff, A., Hilgenberg, L., Gulyas, A., Freund,
T.F., Smith, M., Dinh, T.P., Piomelli, D., Mackie, K. and
Soltesz, I. (2003) Long-term plasticity of endocannabinoid
signaling induced by developmental febrile seizures. Neuron,
39: 599–611.
Chevaleyre, V. and Castillo, P.E. (2003) Heterosynaptic LTD of
hippocampal GABAergic synapses: a novel role of end-
ocannabinoids in regulating excitability. Neuron, 38:
461–472.
Chevaleyre, V. and Castillo, P.E. (2004) Endocannabinoid-medi-
ated metaplasticity in the hippocampus. Neuron, 43: 871–881.
Chevaleyre, V., Takahashi, K.A. and Castillo, P.E. (2006)
Endocannabinoid-mediated synaptic plasticity in the CNS.
Annu. Rev. Neurosci., 29: 37–76.
Christopoulos, A. and Wilson, K. (2001) Interaction of anand-
amide with the M(1) and M(4) muscarinic acetylcholine re-
ceptors. Brain Res., 915: 70–78.
Consroe, P. and Wolkin, A. (1977) Cannabidiol: antiepileptic
drug comparisons and interactions in experimentally induced
seizures in rats. J. Pharmacol. Exp. Ther., 201: 26–32.
Consroe, P.F., Wood, G.C. and Buchsbaum, H. (1975) Antic-
onvulsant nature of marihuana smoking. JAMA, 234:
306–307.
Cravatt, B.F., Giang, D.K., Mayfield, S.P., Boger, D.L.,
Lerner, R.A. and Gilula, N.B. (1996) Molecular character-
ization of an enzyme that degrades neuromodulatory fatty-
acid amides. Nature, 384: 83–87.
Cristino, L., De, P.L., Pryce, G., Baker, D., Guglielmotti, V.
and Di Marzo, V. (2006) Immunohistochemical localization
of cannabinoid type 1 and vanilloid transient receptor po-
tential vanilloid type 1 receptors in the mouse brain. Neuro-
science, 139: 1405–1415.

De, P.L., Melck, D., Bisogno, T. and Di Marzo, V. (2000)
Endocannabinoids and fatty acid amides in cancer, inflam-
mation and related disorders. Chem. Phys. Lipids, 108:
191–209.
Deutsch, D.G., Ueda, N. and Yamamoto, S. (2002) The fatty
acid amide hydrolase (FAAH). Prostaglandins Leukot. Es-
sent. Fatty Acids, 66: 201–210.
Devane, W.A., Hanus, L., Breuer, A., Pertwee, R.G., Steven-
son, L.A., Griffin, G., Gibson, D., Mandelbaum, A., Etinger,
A. and Mechoulam, R. (1992) Isolation and structure of a
brain constituent that binds to the cannabinoid receptor.
Science, 258: 1946–1949.
Di Marzo, V., Bisogno, T., De, P.L., Brandi, I., Jefferson, R.G.,
Winckler, R.L., Davis, J.B., Dasse, O., Mahadevan, A., Ra-
zdan, R.K. and Martin, B.R. (2001) Highly selective CB(1)
cannabinoid receptor ligands and novel CB(1)/VR(1) van-
illoid receptor ‘‘hybrid’’ ligands. Biochem. Biophys. Res. Co-
mmun., 281: 444–451.
Di Marzo, V., Bisogno, T., De, P.L., Melck, D. and Martin,
B.R. (1999) Cannabimimetic fatty acid derivatives: the
anandamide family and other endocannabinoids. Curr.
Med. Chem., 6: 721–744.
Di Marzo, V., De, P.L., Fezza, F., Ligresti, A. and Bisogno, T.
(2002) Anandamide receptors. Prostaglandins Leukot. Es-
sent. Fatty Acids, 66: 377–391.
Di Marzo, V. and Deutsch, D.G. (1998) Biochemistry of the
endogenous ligands of cannabinoid receptors. Neurobiol.
Dis., 5: 386–404.
Di Marzo, V., Fontana, A., Cadas, H., Schinelli, S., Cimino,
G., Schwartz, J.C. and Piomelli, D. (1994) Formation and
inactivation of endogenous cannabinoid anandamide in cen-
tral neurons. Nature, 372: 686–691.
Di, S., Boudaba, C., Popescu, I.R., Weng, F.J., Harris, C.,
Marcheselli, V.L., Bazan, N.G. and Tasker, J.G. (2005) Ac-
tivity-dependent release and actions of endocannabinoids in
the rat hypothalamic supraoptic nucleus. J. Physiol., 569:
751–760.
Dinh, T.P., Carpenter, D., Leslie, F.M., Freund, T.F., Katona,
I., Sensi, S.L., Kathuria, S. and Piomelli, D. (2002) Brain
monoglyceride lipase participating in endocannabinoid inac-
tivation. Proc. Natl. Acad. Sci. U.S.A., 99: 10819–10824.
Domenici, M.R., Azad, S.C., Marsicano, G., Schierloh, A.,
Wotjak, C.T., Dodt, H.U., Zieglgansberger, W., Lutz, B. and
Rammes, G. (2006) Cannabinoid receptor type 1 located on
presynaptic terminals of principal neurons in the forebrain
controls glutamatergic synaptic transmission. J. Neurosci.,
26: 5794–5799.
Egertova, M., Cravatt, B.F. and Elphick, M.R. (2003) Com-
parative analysis of fatty acid amide hydrolase and cb(1)
cannabinoid receptor expression in the mouse brain: evidence
of a widespread role for fatty acid amide hydrolase in reg-
ulation of endocannabinoid signaling. Neuroscience, 119:
481–496.
Egertova, M. and Elphick, M.R. (2000) Localisation of can-
nabinoid receptors in the rat brain using antibodies to the
intracellular C-terminal tail of CB. J. Comp. Neurol., 422:
159–171.
333
Egertova, M., Giang, D.K., Cravatt, B.F. and Elphick, M.R.
(1998) A new perspective on cannabinoid signalling: com-
plementary localization of fatty acid amide hydrolase and the
CB1 receptor in rat brain. Proc. Biol. Sci., 265: 2081–2085.
Engler, B., Freiman, I., Urbanski, M. and Szabo, B. (2006)
Effects of Exogenous and Endogenous Cannabinoids on
GABAergic Neurotransmission between the Caudate-Put-
amen and the Globus Pallidus in the Mouse. J. Pharmacol.
Exp. Ther., 316: 608–617.
Freund, T.F., Katona, I. and Piomelli, D. (2003) Role of en-
dogenous cannabinoids in synaptic signaling. Physiol. Rev.,
83: 1017–1066.
Fukudome, Y., Ohno-Shosaku, T., Matsui, M., Omori, Y.,
Fukaya, M., Tsubokawa, H., Taketo, M.M., Watanabe, M.,
Manabe, T. and Kano, M. (2004) Two distinct classes of
muscarinic action on hippocampal inhibitory synapses: M2-
mediated direct suppression and M1/M3-mediated indirect
suppression through endocannabinoid signalling. Eur. J. Ne-
urosci., 19: 2682–2692.
Gaoni, Y. and Mechoulam, R. (1964) Isolation, structure and
partial synthesis of an active constituent of hashish. J. Am.
Chem. Soc., 86: 1646–1647.
Glass, M., Dragunow, M. and Faull, R.L. (1997) Cannabinoid
receptors in the human brain: a detailed anatomical and
quantitative autoradiographic study in the fetal, neonatal
and adult human brain. Neuroscience, 77: 299–318.
Gulyas, A.I., Cravatt, B.F., Bracey, M.H., Dinh, T.P., Piomelli,
D., Boscia, F. and Freund, T.F. (2004) Segregation of two
endocannabinoid-hydrolyzing enzymes into pre- and postsy-
naptic compartments in the rat hippocampus, cerebellum and
amygdala. Eur. J. Neurosci., 20: 441–458.
Hajos, N. and Freund, T.F. (2002) Pharmacological separation
of cannabinoid sensitive receptors on hippocampal excitatory
and inhibitory fibers. Neuropharmacology, 43: 503–510.
Hajos, N., Katona, I., Naiem, S.S., Mackie, K., Ledent, C.,
Mody, I. and Freund, T.F. (2000) Cannabinoids inhibit hip-
pocampal GABAergic transmission and network oscillations.
Eur. J. Neurosci., 12: 3239–3249.
Hampson, R.E., Foster, T.C. and Deadwyler, S.A. (1989)
Effects of delta-9-tetrahydrocannabinol on sensory evoked
hippocampal activity in the rat: principal components anal-
ysis and sequential dependency. J. Pharmacol. Exp. Ther.,
251: 870–877.
Harkany, T., Dobszay, M.B., Cayetanot, F., Hartig, W., Siege-
mund, T., Aujard, F. and Mackie, K. (2005) Redistribution
of CB1 cannabinoid receptors during evolution of cholinergic
basal forebrain territories and their cortical projection areas:
a comparison between the gray mouse lemur (Microcebus
murinus, primates) and rat. Neuroscience, 135: 595–609.
Harris, L.S., Carchman, R.A. and Martin, B.R. (1978) Evi-
dence for the existence of specific cannabinoid binding sites.
Life Sci., 22: 1131–1137.
Hashimotodani, Y., Ohno-Shosaku, T., Tsubokawa, H., Ogata,
H., Emoto, K., Maejima, T., Araishi, K., Shin, H.S. and
Kano, M. (2005) Phospholipase Cbeta serves as a coincidence
detector through its Ca2+ dependency for triggering retro-
grade endocannabinoid signal. Neuron, 45: 257–268.
334
Herkenham, M., Lynn, A.B., de Costa, B.R. and Richfield,
E.K. (1991a) Neuronal localization of cannabinoid receptors
in the basal ganglia of the rat. Brain Res., 547: 267–274.
Herkenham, M., Lynn, A.B., Johnson, M.R., Melvin, L.S., de
Costa, B.R. and Rice, K.C. (1991b) Characterization and
localization of cannabinoid receptors in rat brain: a quanti-
tative in vitro autoradiographic study. J. Neurosci., 11:
563–583.
Herkenham, M., Lynn, A.B., Johnson, M.R., Melvin, L.S., de
Costa, B.R. and Rice, K.C. (1991c) Characterization and
localization of cannabinoid receptors in rat brain: a quanti-
tative in vitro autoradiographic study. J. Neurosci., 11:
563–583.
Hermann, H., Marsicano, G. and Lutz, B. (2002) Coexpression
of the cannabinoid receptor type 1 with dopamine and se-
rotonin receptors in distinct neuronal subpopulations of the
adult mouse forebrain. Neuroscience, 109: 451–460.
Hillard, C.J., Edgemond, W.S., Jarrahian, A. and Campbell,
W.B. (1997) Accumulation of N-arachidonoylethanolamine
(anandamide) into cerebellar granule cells occurs via facili-
tated diffusion. J. Neurochem., 69: 631–638.
Hillard, C.J., Wilkison, D.M., Edgemond, W.S. and Campbell,
W.B. (1995) Characterization of the kinetics and distribution
of N-arachidonylethanolamine (anandamide) hydrolysis by
rat brain. Biochim. Biophys. Acta, 1257: 249–256.
Hoffman, A.F., Macgill, A.M., Smith, D., Oz, M. and Lupica,
C.R. (2005) Species and strain differences in the expression of
a novel glutamate-modulating cannabinoid receptor in the
rodent hippocampus. Eur. J. Neurosci., 22: 2387–2391.
Hoffman, A.F., Oz, M., Caulder, T. and Lupica, C.R. (2003)
Functional tolerance and blockade of long-term depression
at synapses in the nucleus accumbens after chronic can-
nabinoid exposure. J. Neurosci., 23: 4815–4820.
Hofmann, M.E., Nahir, B. and Frazier, C.J. (2006) End-
ocannabinoid mediated depolarization-induced suppression
of inhibition in hilar mossy cells of the rat dentate gyrus. J.
Neurophysiol., 96: 2501–2512.
Howard, A.L., Ratzliff, A.D.H. and Soltesz, I. (2003) Explor-
ing Endocannabinoid signaling in the dentate gyrus. Soc.
Neurosci. Abstr., 808.11.
Howlett, A.C., Bidaut-Russell, M., Devane, W.A., Melvin,
L.S., Johnson, M.R. and Herkenham, M. (1990) The can-
nabinoid receptor: biochemical, anatomical and behavioral
characterization. Trends Neurosci., 13: 420–423.
Isokawa, M. and Alger, B.E. (2005) Retrograde end-
ocannabinoid regulation of GABAergic inhibition in the rat
dentate gyrus granule cell. J. Physiol., 567: 1001–1010.
Jiang, W., Zhang, Y., Xiao, L., Van, C.J., Ji, S.P., Bai, G. and
Zhang, X. (2005) Cannabinoids promote embryonic and
adult hippocampus neurogenesis and produce anxiolytic- and
antidepressant-like effects. J. Clin. Invest., 115: 3104–3116.
Jin, K., Xie, L., Kim, S.H., Parmentier-Batteur, S., Sun, Y.,
Mao, X.O., Childs, J. and Greenberg, D.A. (2004) Defective
adult neurogenesis in CB1 cannabinoid receptor knockout
mice. Mol. Pharmacol., 66: 204–208.
Katona, I., Sperlagh, B., Magloczky, Z., Santha, E., Kofalvi,
A., Czirjak, S., Mackie, K., Vizi, E.S. and Freund, T.F.
(2000) GABAergic interneurons are the targets of can-
nabinoid actions in the human hippocampus. Neuroscience,
100: 797–804.
Katona, I., Sperlagh, B., Sik, A., Kafalvi, A., Vizi, E.S., Mac-
kie, K. and Freund, T.F. (1999) Presynaptically located CB1
cannabinoid receptors regulate GABA release from axon
terminals of specific hippocampal interneurons. J. Neurosci.,
19: 4544–4558.
Katona, I., Urban, G.M., Wallace, M., Ledent, C., Jung, K.M.,
Piomelli, D., Mackie, K. and Freund, T.F. (2006) Molecular
composition of the endocannabinoid system at glutamatergic
synapses. J. Neurosci., 26: 5628–5637.
Khaspekov, L.G., Brenz Verca, M.S., Frumkina, L.E., Her-
mann, H., Marsicano, G. and Lutz, B. (2004) Involvement of
brain-derived neurotrophic factor in cannabinoid receptor-
dependent protection against excitotoxicity. Eur. J. Neuro-
sci., 19: 1691–1698.
Kim, S.H., Won, S.J., Mao, X.O., Ledent, C., Jin, K. and
Greenberg, D.A. (2006) Role for neuronal nitric oxide synt-
hase in cannabinoid-induced neurogenesis. J. Pharmacol.
Exp. Ther., 319: 150–154.
Kirby, M.T., Hampson, R.E. and Deadwyler, S.A. (1995) Can-
nabinoids selectively decrease paired-pulse facilitation of per-
forant path synaptic potentials in the dentate gyrus in vitro.
Brain Res., 688: 114–120.
Kreitzer, A.C. and Regehr, W.G. (2001a) Cerebellar depolari-
zation-induced suppression of inhibition is mediated by en-
dogenous cannabinoids. J. Neurosci., 21: RC174.
Kreitzer, A.C. and Regehr, W.G. (2001b) Retrograde inhibition
of presynaptic calcium influx by endogenous cannabinoids
at excitatory synapses onto Purkinje cells. Neuron, 29:
717–727.
Kujtan, P.W., Carlen, P.L. and Kapur, B.M. (1983) Delta 9-
tetrahydrocannabinol and cannabidiol: dose-dependent
effects on evoked potentials in the hippocampal slice. Can.
J. Physiol. Pharmacol., 61: 420–426.
Lanzafame, A.A., Guida, E. and Christopoulos, A. (2004)
Effects of anandamide on the binding and signaling proper-
ties of M1 muscarinic acetylcholine receptors. Biochem.
Pharmacol., 68: 2207–2219.
Llano, I., Leresche, N. and Marty, A. (1991) Calcium entry
increases the sensitivity of cerebellar Purkinje cells to applied
GABA and decreases inhibitory synaptic currents. Neuron,
6: 565–574.
Lledo, P.M., Alonso, M. and Grubb, M.S. (2006) Adult ne-
urogenesis and functional plasticity in neuronal circuits. Nat.
Rev. Neurosci., 7: 179–193.
Lutz, B. (2002) Molecular biology of cannabinoid receptors.
Prostaglandins Leukot. Essent. Fatty Acids, 66: 123–142.
Maejima, T., Hashimoto, K., Yoshida, T., Aiba, A. and Kano,
M. (2001) Presynaptic inhibition caused by retrograde signal
from metabotropic glutamate to cannabinoid receptors. Neu-
ron, 31: 463–475.
Maejima, T., Oka, S., Hashimotodani, Y., Ohno-Shosaku, T.,
Aiba, A., Wu, D., Waku, K., Sugiura, T. and Kano, M.
(2005) Synaptically driven endocannabinoid release requires
Ca2+-assisted metabotropic glutamate receptor subtype 1 to

phospholipase C {beta}4 signaling cascade in the cerebellum.
J. Neurosci., 25: 6826–6835.
Mailleux, P., Parmentier, M. and Vanderhaeghen, J.J. (1992)
Distribution of cannabinoid receptor messenger RNA in the
human brain: an in situ hybridization histochemistry with
oligonucleotides. Neurosci. Lett., 143: 200–204.
Mailleux, P. and Vanderhaeghen, J.J. (1992) Distribution of
neuronal cannabinoid receptor in the adult rat brain: a com-
parative receptor binding radioautography and in situ hy-
bridization histochemistry. Neuroscience, 48: 655–668.
Marinelli, S., Di Marzo, V., Berretta, N., Matias, I., Maccarr-
one, M., Bernardi, G. and Mercuri, N.B. (2003) Presynaptic
facilitation of glutamatergic synapses to dopaminergic neu-
rons of the rat substantia nigra by endogenous stimulation of
vanilloid receptors. J. Neurosci., 23: 3136–3144.
Marinelli, S., Vaughan, C.W., Christie, M.J. and Connor, M.
(2002) Capsaicin activation of glutamatergic synaptic trans-
mission in the rat locus coeruleus in vitro. J. Physiol., 543:
531–540.
Marsicano, G., Goodenough, S., Monory, K., Hermann, H.,
Eder, M., Cannich, A., Azad, S.C., Cascio, M.G., Gutierrez,
S.O., Van der Stelt, M., Lopez-Rodriguez, M.L., Casanova,
E., Schutz, G., Zieglgansberger, W., Di Marzo, V., Behl, C.
and Lutz, B. (2003) CB1 cannabinoid receptors and on-de-
mand defense against excitotoxicity. Science, 302: 84–88.
Marsicano, G. and Lutz, B. (1999) Expression of the can-
nabinoid receptor CB1 in distinct neuronal subpopulations in
the adult mouse forebrain. Eur. J. Neurosci., 11: 4213–4225.
Marsicano, G. and Lutz, B. (2006) Neuromodulatory functions
of the endocannabinoid system. J. Endocrinol. Invest., 29:
27–46.
Marsicano, G., Wotjak, C.T., Azad, S.C., Bisogno, T., Ram-
mes, G., Cascio, M.G., Hermann, H., Tang, J., Hofmann, C.,
Zieglgansberger, W., Di Marzo, V. and Lutz, B. (2002) The
endogenous cannabinoid system controls extinction of avers-
ive memories. Nature, 418: 530–534.
Matsuda, L.A., Bonner, T.I. and Lolait, S.J. (1993) Localiza-
tion of cannabinoid receptor mRNA in rat brain. J. Comp.
Neurol., 327: 535–550.
Matsuda, L.A., Lolait, S.J., Brownstein, M.J., Young, A.C. and
Bonner, T.I. (1990) Structure of a cannabinoid receptor and
functional expression of the cloned cDNA. Nature, 346:
561–564.
Melis, M., Perra, S., Muntoni, A.L., Pillolla, G., Lutz, B.,
Marsicano, G., Di Marzo, V., Gessa, G.L. and Pistis, M.
(2004a) Prefrontal cortex stimulation induces 2-arachido-
noyl-glycerol-mediated suppression of excitation in dopa-
mine neurons. J. Neurosci., 24: 10707–10715.
Melis, M., Pistis, M., Perra, S., Muntoni, A.L., Pillolla, G. and
Gessa, G.L. (2004b) Endocannabinoids mediate presynaptic
inhibition of glutamatergic transmission in rat ventral teg-
mental area dopamine neurons through activation of CB1
receptors. J. Neurosci., 24: 53–62.
Monory, K., Massa, F., Egertova, M., Eder, M., Blaudzun, H.,
Westenbroek, R., Kelsch, W., Jacob, W., Marsch, R., Ekker,
M., Long, J., Rubenstein, J.L., Goebbels, S., Nave, K.A.,
During, M., Klugmann, M., Wolfel, B., Dodt, H.U.,
335
Zieglgansberger, W., Wotjak, C.T., Mackie, K., Elphick,
M.R., Marsicano, G. and Lutz, B. (2006) The end-
ocannabinoid system controls key epileptogenic circuits in
the hippocampus. Neuron, 51: 455–466.
Morales, M. (2006) Cannabinoids and the central serotonergic
system. In: Onaivi E.S., Sugiura T. and Di Marzo V. (Eds.),
Endocannabinoids. CRC Press, Boca Raton, FL, pp.
249–260.
Morales, M. and Backman, C. (2002) Coexistence of serotonin
3 (5-HT3) and CB1 cannabinoid receptors in interneurons of
hippocampus and dentate gyrus. Hippocampus, 12: 756–764.
Mukhtarov, M., Ragozzino, D. and Bregestovski, P. (2005)
Dual Ca2+ modulation of glycinergic synaptic currents in
rodent hypoglossal motoneurones. J. Physiol., 569: 817–831.
Munro, S., Thomas, K.L. and bu-Shaar, M. (1993) Molecular
characterization of a peripheral receptor for cannabinoids.
Nature, 365: 61–65.
Nagayama, T., Sinor, A.D., Simon, R.P., Chen, J., Graham,
S.H., Jin, K. and Greenberg, D.A. (1999) Cannabinoids and
neuroprotection in global and focal cerebral ischemia and in
neuronal cultures. J. Neurosci., 19: 2987–2995.
Nakatsuka, T., Chen, H.X., Roper, S.N. and Gu, J.G. (2003)
Cannabinoid receptor-1 activation suppresses inhibitory
synaptic activity in human dentate gyrus. Neuropharmacol-
ogy, 45: 116–121.
Niehaus, J.L., Wallis, K.T., Elphick, M.R. and Lewis, D.L.,
(2004) CRIP1B, a novel CB1 cannabinoid receptor interact-
ing protein, interacts with CB1 at the distal C-terminal tail.
Soc. Neurosci. Abstr., 623.19.
Ohno-Shosaku, T., Hashimotodani, Y., Maejima, T. and Kano,
M. (2005) Calcium signaling and synaptic modulation: reg-
ulation of endocannabinoid-mediated synaptic modulation
by calcium. Cell Calcium, 38: 369–374.
Ohno-Shosaku, T., Maejima, T. and Kano, M. (2001) Endog-
enous cannabinoids mediate retrograde signals from depo-
larized postsynaptic neurons to presynaptic terminals.
Neuron, 29: 729–738.
Ohno-Shosaku, T., Shosaku, J., Tsubokawa, H. and Kano, M.
(2002a) Cooperative endocannabinoid production by neu-
ronal depolarization and group I metabotropic glutamate
receptor activation. Eur. J. Neurosci., 15: 953–961.
Ohno-Shosaku, T., Tsubokawa, H., Mizushima, I., Yoneda,
N., Zimmer, A. and Kano, M. (2002b) Presynaptic can-
nabinoid sensitivity is a major determinant of depolarization-
induced retrograde suppression at hippocampal synapses. J.
Neurosci., 22: 3864–3872.
Ong, W.Y. and Mackie, K. (1999) A light and electron micro-
scopic study of the CB1 cannabinoid receptor in primate
brain. Neuroscience, 92: 1177–1191.
Oz, M., Ravindran, A., az-Ruiz, O., Zhang, L. and Morales, M.
(2003) The endogenous cannabinoid anandamide inhibits al-
pha7 nicotinic acetylcholine receptor-mediated responses in
Xenopus oocytes. J. Pharmacol. Exp. Ther., 306: 1003–1010.
Oz, M., Zhang, L. and Morales, M. (2002) Endogenous can-
nabinoid, anandamide, acts as a noncompetitive inhibitor on
5-HT3 receptor-mediated responses in Xenopus oocytes.
Synapse, 46: 150–156.
336
Pettit, D.A., Harrison, M.P., Olson, J.M., Spencer, R.F. and
Cabral, G.A. (1998) Immunohistochemical localization of the
neural cannabinoid receptor in rat brain. J. Neurosci. Res.,
51: 391–402.
Piomelli, D. (2003) The molecular logic of endocannabinoid
signalling. Nat. Rev. Neurosci., 4: 873–884.
Piomelli, D., Beltramo, M., Glasnapp, S., Lin, S.Y., Gout-
opoulos, A., Xie, X.Q. and Makriyannis, A. (1999) Structural
determinants for recognition and translocation by the anand-
amide transporter. Proc. Natl. Acad. Sci. U.S.A., 96:
5802–5807.
Pitler, T.A. and Alger, B.E. (1992) Postsynaptic spike firing
reduces synaptic GABAA responses in hippocampal pyram-
idal cells. J. Neurosci., 12: 4122–4132.
Ralevic, V., Kendall, D.A., Jerman, J.C., Middlemiss, D.N. and
Smart, D. (2001) Cannabinoid activation of recombinant and
endogenous vanilloid receptors. Eur. J. Pharmacol., 424:
211–219.
Robbe, D., Kopf, M., Remaury, A., Bockaert, J. and Manzoni,
O.J. (2002) Endogenous cannabinoids mediate long-term
synaptic depression in the nucleus accumbens. Proc. Natl.
Acad. Sci. U.S.A., 99: 8384–8388.
Romero, J., Hillard, C.J., Calero, M. and Rabano, A. (2002)
Fatty acid amide hydrolase localization in the human central
nervous system: an immunohistochemical study. Brain Res.
Mol. Brain Res., 100: 85–93.
Roth, S.H. and Williams, P.J. (1979) The non-specific mem-
brane binding properties of delta9-tetrahydrocannabinol and
the effects of various solubilizers. J. Pharm. Pharmacol., 31:
224–230.
Rueda, D., Navarro, B., Martinez-Serrano, A., Guzman, M.
and Galve-Roperh, I. (2002) The endocannabinoid anand-
amide inhibits neuronal progenitor cell differentiation
through attenuation of the Rap1/B-Raf/ERK pathway. J.
Biol. Chem., 277: 46645–46650.
Safo, P.K. and Regehr, W.G. (2005) Endocannabinoids control
the induction of cerebellar LTD. Neuron, 48: 647–659.
Schlicker, E. and Kathmann, M. (2001) Modulation of trans-
mitter release via presynaptic cannabinoid receptors. Trends
Pharmacol. Sci., 22: 565–572.
Skaper, S.D., Buriani, A., Dal, T.R., Petrelli, L., Romanello, S.,
Facci, L. and Leon, A. (1996) The ALIAmide palm-
itoylethanolamide and cannabinoids, but not anandamide,
are protective in a delayed postglutamate paradigm of ex-
citotoxic death in cerebellar granule neurons. Proc. Natl.
Acad. Sci. U.S.A., 93: 3984–3989.
Stella, N., Schweitzer, P. and Piomelli, D. (1997) A second en-
dogenous cannabinoid that modulates long-term potentiat-
ion. Nature, 388: 773–778.
Sugiura, T., Kondo, S., Sukagawa, A., Nakane, S., Shinoda, A.,
Itoh, K., Yamashita, A. and Waku, K. (1995) 2-Arachidonoyl-
glycerol: a possible endogenous cannabinoid receptor ligand in
brain. Biochem. Biophys. Res. Commun., 215: 89–97.
Sugiura, T., Oka, S., Ikeda, S. and Waku, K. (2006) Occur-
rence, biosynthesis, and metabolism of endocannabinoids. In:
Onaivi E.S., Sugiura T. and Di Marzo V. (Eds.), End-
ocannabinoids. CRC Press, Boca Raton, FL, pp. 177–214.
Takahashi, K.A. and Castillo, P.E. (2006) The CB1 can-
nabinoid receptor mediates glutamatergic synaptic suppres-
sion in the hippocampus. Neuroscience, 139: 795–802.
Thomas, B.F., Wei, X. and Martin, B.R. (1992) Characteriza-
tion and autoradiographic localization of the cannabinoid
binding site in rat brain using [3H]11-OH-delta 9-THC-
DMH. J. Pharmacol. Exp. Ther., 263: 1383–1390.
Thomas, E.A., Cravatt, B.F., Danielson, P.E., Gilula, N.B. and
Sutcliffe, J.G. (1997) Fatty acid amide hydrolase, the degra-
dative enzyme for anandamide and oleamide, has selective
distribution in neurons within the rat central nervous system.
J. Neurosci. Res., 50: 1047–1052.
Toth, A., Boczan, J., Kedei, N., Lizanecz, E., Bagi, Z., Papp,
Z., Edes, I., Csiba, L. and Blumberg, P.M. (2005) Expression
and distribution of vanilloid receptor 1 (TRPV1) in the adult
rat brain. Brain Res. Mol. Brain Res., 135: 162–168.
Trettel, J. and Levine, E.S. (2003) Endocannabinoids mediate
rapid retrograde signaling at interneuron right-arrow pyram-
idal neuron synapses of the neocortex. J. Neurophysiol., 89:
2334–2338.
Tsou, K., Brown, S., Sanudo-Pena, M.C., Mackie, K. and
Walker, J.M. (1998a) Immunohistochemical distribution of
cannabinoid CB1 receptors in the rat central nervous system.
Neuroscience, 83: 393–411.
Tsou, K., Mackie, K., Sanudo-Pena, M.C. and Walker, J.M.
(1999) Cannabinoid CB1 receptors are localized primarily on
cholecystokinin-containing GABAergic interneurons in the
rat hippocampal formation. Neuroscience, 93: 969–975.
Tsou, K., Nogueron, M.I., Muthian, S., Sanudo-Pena, M.C.,
Hillard, C.J., Deutsch, D.G. and Walker, J.M. (1998b) Fatty
acid amide hydrolase is located preferentially in large neu-
rons in the rat central nervous system as revealed by
immunohistochemistry. Neurosci. Lett., 254: 137–140.
Van der Stelt, M., Veldhuis, W.B., Bar, P.R., Veldink, G.A.,
Vliegenthart, J.F. and Nicolay, K. (2001a) Neuroprotection
by delta9-tetrahydrocannabinol, the main active compound
in marijuana, against ouabain-induced in vivo excitotoxicity.
J. Neurosci., 21: 6475–6479.
Van der Stelt, M., Veldhuis, W.B., van Haaften, G.W., Fezza,
F., Bisogno, T., Bar, P.R., Veldink, G.A., Vliegenthart, J.F.,
Di, M.V. and Nicolay, K. (2001b) Exogenous anandamide
protects rat brain against acute neuronal injury in vivo. J.
Neurosci., 21: 8765–8771.
Van der Stelt, M., Veldhuis, W.B., Maccarrone, M., Bar, P.R.,
Nicolay, K., Veldink, G.A., Di, M.V. and Vliegenthart, J.F.
(2002) Acute neuronal injury, excitotoxicity, and the end-
ocannabinoid system. Mol. Neurobiol., 26: 317–346.
Van Sickle, M.D., Duncan, M., Kingsley, P.J., Mouihate, A.,
Urbani, P., Mackie, K., Stella, N., Makriyannis, A., Piomelli,
D., Davison, J.S., Marnett, L.J., Di Marzo, V., Pittman, Q.J.,
Patel, K.D. and Sharkey, K.A. (2005) Identification and
functional characterization of brainstem cannabinoid CB2
receptors. Science, 310: 329–332.
Varma, N., Carlson, G.C., Ledent, C. and Alger, B.E. (2001)
Metabotropic glutamate receptors drive the end-
ocannabinoid system in hippocampus. J. Neurosci., 21:
RC188.

Wallace, M.J., Blair, R.E., Falenski, K.W., Martin, B.R. and
Delorenzo, R.J. (2003) The endogenous cannabinoid system
regulates seizure frequency and duration in a model of tem-
poral lobe epilepsy. J. Pharmacol. Exp. Ther., 307: 129–137.
Wallace, M.J., Martin, B.R. and Delorenzo, R.J. (2002) Evi-
dence for a physiological role of endocannabinoids in the
modulation of seizure threshold and severity. Eur. J. Pharma-
col., 452: 295–301.
Wallace, M.J., Wiley, J.L., Martin, B.R. and Delorenzo, R.J.
(2001) Assessment of the role of CB1 receptors in cannabinoid
anticonvulsant effects. Eur. J. Pharmacol., 428: 51–57.
Wilkison, D.M. and Pontzer, N.J. (1987) The actions of THC
on the intact hippocampus: a comparison of dentate and
CA1 responses. Brain Res. Bull., 19: 63–67.
Wilson, R.I., Kunos, G. and Nicoll, R.A. (2001) Presynaptic
specificity of endocannabinoid signaling in the hippocampus.
Neuron, 31: 453–462.
Wilson, R.I. and Nicoll, R.A. (2001) Endogenous cannabinoids
mediate retrograde signalling at hippocampal synapses. Na-
ture, 410: 588–592.
Wilson, R.I. and Nicoll, R.A. (2002) Endocannabinoid signa-
ling in the brain. Science, 296: 678–682.
337
Yanovsky, Y., Mades, S. and Misgeld, U. (2003) Retrograde
signaling changes the venue of postsynaptic inhibition in rat
substantia nigra. Neuroscience, 122: 317–328.
Yoshida, T., Fukaya, M., Uchigashima, M., Miura, E.,
Kamiya, H., Kano, M. and Watanabe, M. (2006) Localiza-
tion of diacylglycerol lipase-{alpha} around postsynaptic
spine suggests close proximity between production site of an
endocannabinoid, 2-arachidonoyl-glycerol, and presynaptic
cannabinoid CB1 receptor. J. Neurosci., 26: 4740–4751.
Yoshida, T., Hashimoto, K., Zimmer, A., Maejima, T., Araishi,
K. and Kano, M. (2002) The cannabinoid CB1 receptor me-
diates retrograde signals for depolarization-induced suppres-
sion of inhibition in cerebellar Purkinje cells. J. Neurosci., 22:
1690–1697.
Zhu, P.J. and Lovinger, D.M. (2005) Retrograde end-
ocannabinoid signaling in a postsynaptic neuron/synaptic
bouton preparation from basolateral amygdala. J. Neurosci.,
25: 6199–6207.
Zygmunt, P.M., Petersson, J., Andersson, D.A., Chuang, H.,
Sorgard, M., Di Marzo, V., Julius, D. and Hogestatt, E.D.
(1999) Vanilloid receptors on sensory nerves mediate the
vasodilator action of anandamide. Nature, 400: 452–457.
Plate 19.2. DSI in dentate granule cells and hilar mossy cells. (AA) Dentate gyrus of the hippocampal slice (scale, 50 mm). (AB)
Differential interference contrast image of dentate granule cells (scale, 10 mm). (AC) Single dentate granule cell filled, and visualized
with fluorescence microscopy during whole cell recording (scale, 10 mm). (AD) DSI of eIPSCs (left, top) and sIPSCs (right, top) is
observed in dentate granule cells. In both cases, magnitude of DSI is reduced by bath application of thapsigargin (TG, 2 mM, bottom).
(AE) Thapsigargin also reduced the calcium transients produced by DSI-inducing depolarizing voltage steps. Adapted with permission
from Isokawa and Alger (2005). (B) Depolarization induced release of endogenous cannabinoids from hilar mossy cells preferentially
inhibits calcium-dependent exocytosis. DSI in of eIPSCs is apparent in an individual mossy cell. DSI of miniature IPSCs is absent in
the same cell. Following bath application of KCl and CaCl2, DSI of mIPSCs becomes apparent. Summary data are presented in BD,
where numbers on the bars are n values. Adapted with permission from Hofmann et al. (2006). (For B/W version, see page 328 in the
volume.)
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 20

Pro-inflammatory cytokines and their effects

in the dentate gyrus

Mark Pickering and John J. O’Connor

UCD School of Biomolecular and Biomedical Science, Conway Institute of Biomolecular and Biomedical Research,
University College Dublin, Belfield, Dublin 4, Ireland

Abstract: The older notion of a central nervous system existing in essential isolation from the immune
system has changed dramatically in recent years as the body of evidence relating to the interactions between
these two systems has grown. Here we address the role of a particular subset of immune modulatory
molecules, the pro-inflammatory cytokines, in regulating neuronal function and viability in the dentate
gyrus of the hippocampus. These inflammatory mediators are known to be elevated in many neuropath-
ological conditions, such as Alzheimer’s disease, Parkinson’s disease and ischaemic injury that follows
stroke. Pro-inflammatory cytokines, such as tumour necrosis factor-a (TNF-a), interleukin 1-b (IL-1b) and
interleukin 18 (IL-18), have been shown to regulate neurotoxicity; although, due to the complexity of the
cytokine action in neurons and glia, the effect may be either facilitatory or protective, depending on the
circumstances. As well as their role in neurotoxicity and neuroprotection, the pro-inflammatory cytokines
have also been shown to be potent regulators of synaptic function. In particular, TNF-a, IL-1b and IL-18
have all been shown to inhibit long-term potentiation, a form of neuronal plasticity widely believed to
underlie learning and memory, both in the early p38 mitogen activated protein kinase-dependant phase and
the later protein synthesis-dependant phase. In this article we address the mechanisms underlying these
cytokine effects in the dentate gyrus of the hippocampus.

Keywords: TNF-a; IL-1b; IL-18; long-term potentiation; dentate gyrus; mGluR5

Introduction Chun, 2001). Chief amongst these are the cyto-

The older notion of two ‘‘super-systems’’, nervous
and immune, existing in relative isolation from
each other due largely to the blood-brain barrier,
has given way in recent years to a new consensus,
as it became apparent that many immune mole-
cules may be used by the nervous system in inter-
cellular communication (Boulanger et al., 2001;
Corresponding author. Tel.: +353 1 716 6765;
Fax: +353 1 716 7417; E-mail: john.oconnor@ucd.ie
kines — multifunctional proteins that play crucial
roles in cellular communication and activation.
Cytokines have been classified as pro-inflamma-
tory or anti-inflammatory depending on the bal-
ance of their effects on the immune system
(Mosmann et al., 1986). Cytokines may have an
indirect modulatory effect on the nervous system
via their effects on the hypothalamic-pituitary-
adrenal axis (Besedovsky et al., 1991). However,
here we will address the direct effect roles of the
pro-inflammatory cytokines, in particular tumour
necrosis factor-a (TNF-a), interleukin-1b (IL-1b)

DOI: 10.1016/S0079-6123(07)63020-9 339

340

and interleukin-18 (IL-18) on the CNS in general, brain by astrocytes, microglial and some neurons
and the dentate gyrus in particular. (Lieberman et al., 1989; Morganti-Kossman et al.,
1997; Chung and Benveniste, 1990). Under various
pathological conditions, such as trauma, ischemia
Distribution of pro-inflammatory cytokines
and inflammatory diseases, the expression and
in the CNS
release of TNF-a is rapidly increased, in some
cases as early as 1 h after the brain insult and well
Direct action of the pro-inflammatory cytokines
before neuronal death (Liu et al., 1994; Wang et al.,
TNF-a and IL-1b on the CNS has been known for
1994; Allan and Rothwell, 2001)
some time (Plata-Salaman et al., 1988). Elevated
Two different TNFR (p55 and p75) have been
CNS expression of various pro-inflammatory cyto-
identified (Beutler and Van Huffel, 1994a, b;
kines have been noted in many neuropathological
Wajant and Scheurich, 2001) and shown to medi-
situations, both chronic, such as Alzheimer’s disease
ate differential cellular responses using distinct
(Cacquevel et al., 2004) and multiple sclerosis
pathways (Kinouchi et al., 1991; Tartaglia et al.,
(Merrill, 1992), and acute, such as ischemia and
1991). For instance, the signal transduction path-
stroke (Liu et al., 1994; Klein et al., 2000; Yu and
way used by the p55 TNFR results in the activa-
Lau, 2000), and infection (Waage et al., 1989).
tion of the transcription factor nuclear factor
Additionally, pro-inflammatory cytokines are con-
kappaB (NF-kB) (Kolesnick and Golde, 1994;
stitutively expressed in unperturbed cells of the
Goodman and Mattson, 1996; Mattson et al.,
CNS (Benveniste and Benos, 1995; Yu and Lau,
1997a, b) (see Fig. 1). TNF-a activates the NF-kB
2000).
family of transcription factors, which are ubiqui-
While pro-inflammatory cytokine binding sites
tously expressed and are pivotal in controlling
are found throughout the hippocampus, in addi-
diverse cellular processes, including immune re-
tion to other brain regions, there is some evidence
sponses, cell proliferation and differentiation
that some of the functions of these cytokines are
(Israel, 2000; Silverman and Maniatis, 2001;
distinct in different hippocampal regions. For ex-
Ghosh and Karin, 2002; Li and Verma, 2002).
ample, it has been demonstrated that local injec-
Lack of the p50 subunit of NF-kB increases the
tion of IL-1 receptor antagonist (IL-1ra) into the
vulnerability of hippocampal neurons to excito-
dentate gyrus and CA3 regions of the hippocam-
toxicity (Yu et al., 1999). Increasingly, it has
pus leads to a reduction in the febrile response of
become evident that NF-kB also plays important
rats to peripheral injection of lipopolysaccharide
roles in the CNS (O’Neill and Kaltschmidt, 1997)
(Cartmell et al., 1999), whereas IL-1ra injection
and recently Sheridan et al. (2006) have shown
has no effect in the CA1 region of the hippocam-
that TNF-a can selectively upregulate NF-kB
pus. Here, we will address some of the many roles
expression in hippocampal sub-neuronal popula-
played by TNF-a and the pro-inflammatory IL1-
tions. For example in the dentate gyrus and CA3
type cytokines in the dentate gyrus, highlighting
regions there were clear subpopulations with
differences between the dentate gyrus and the
respect to activation of NF-kB. In the dentate
pyramidal cell regions of the hippocampus where
gyrus there is a clear distinction between the outer
appropriate.
4/5th of the stratum granulosum, the cells of which
expressed high nuclear NF-kB activity and the
Action of TNF-a in the dentate gyrus inner 1–2 cell layers that expressed much lower lev-
els of the active form of this transcription factor.
The pro-inflammatory cytokine TNF-a is a One of the most active areas of investigation
17-kDa peptide and forms multimers which are ac- regarding TNF-a is in relation to its role in
tive in binding TNF receptors (TNFR) that are neurotoxicity. Although TNF-a was originally
constitutively expressed on both neurons and glia in named for its degeneration-inducing action in some
the central nervous system (Benveniste and Benos, types of tumor cells, our current understanding of
1995). TNF-a can be synthesized and released in the TNF-a is such that it cannot be simply considered
 341

Fig. 1. Putative TNF-a signalling pathway in the dentate gyrus. Downstream signalling by TNF-a involves recruitment by the TNF-a
receptor of the death domain containing adaptor protein, TNF receptor-associated death domain (TRADD) and the TNF receptor-
associated factor 2 (TRAF-2) which then recruits the receptor interacting protein (RIP) eventually leading to activation and trans-
location of NF-kB to the nucleus.

as an inducer or facilitator of cell death. On the one
hand, TNF-a has been shown to induce cell death
in septo-hippocampal cultures. Zhao et al. (2001)
detected alpha-spectrin fragments in these septo-
hippocampal cultures treated with TNF-a and
found elevated 120 kDa fragments, indicative of
caspase-3 activity, but not 145-kDa fragments, in-
dicative of calpain activity. Unlike calpain, which is
associated with both necrotic and apoptotic cell
death, caspase-3 is exclusively characteristic of
apoptosis-like cell death (Armstrong et al., 1996;
Wang et al., 1996; Nath et al., 1998).
On the other hand, a number of investigators
have presented evidence that, under different con-
ditions, TNF-a may play a protective role against
neuronal cell death. For example TNF-a treat-
ment can protect against focal cerebral
ischemia (Nawashiro et al., 1997). Also, while it
has been shown that TNF-a mediates damage to
myelin and oligodendrocytes (Selmaj and Raine,
1998), Garcia et al. (1992) found that it was not
toxic to rat-cultured CNS neurons, including those
from the spinal cord, ventral mesencephalon, cer-
ebellum, septum and striatum, in vitro. TNF-a
under in vitro conditions may protect neurons
against metabolic, excitotoxic or oxidative insults
by promoting maintenance of intracellular Ca2+
homeostasis, suppression of reactive oxygen spe-
cies (Cheng et al., 1994), and by activation of
transcription factor NF-kB (Barger et al., 1995).
342
Mice genetically deficient in TNF-receptor p55
(R1) or both R1 and p75 (R2), also show exacer-
bated neuronal damage compared to wild type
controls following middle cerebral artery occlusion
(Bruce et al., 1996; Gary et al., 1998).
Some of the reasons for the manifest complexity
of the relationship between TNF-a and neuronal
cell death may be due to differences between the
role of TNF-a in the early, acute post-injury
phase, where it appears to be largely detrimental,
and late post-injury phase, where it may be pro-
tective because it activates cellular repair mecha-
nisms. It must also be remembered that there is
some evidence that the different TNF receptors
may have differing roles in TNF-a toxicity.
It has been known for some time that the p55
TNF-R1 plays a major role in the anti-bacterial
response and sensitivity to septic shock (Pfeffer
et al., 1993), but the role of TNF-R2 was unclear
until the membrane form of TNF receptor was
recognized as the physiological activator of this
TNF receptor (Grell et al., 1995). The more recent
development of TNF-R1 and TNF-R2 knockout
mice has allowed opportunities to tease apart the
roles of the separate receptor pathways. While
looking at cell death in retinal ischemia-reperfu-
sion experiments in these knockout mice, Fontaine
et al. (2002) demonstrated opposing actions of the
two TNF receptors, with TNF-R1 aggravating
neuronal damage and TNF-R2 promoting neuro-
protection via an Akt/PKB signal pathway. The
work of Fontaine et al. (2002) is in agreement with
Kassiotis and Kollias (2001), who also showed a
dual role for TNF-a in experimental autoimmune
encephalomyelitis (EAE), a mouse model for
multiple sclerosis. Indeed, this differential role
for the two TNF receptors may be of key impor-
tance in understanding TNF-a systems. By ‘‘fine
tuning’’ the balance between the two receptors, the
exact role of TNF-a in cell damage can be con-
trolled from species to species and tissue to tissue.
This makes the TNF-a system very evolutionary
versatile and advantageous.
Another aspect of the complexity may relate to
the role of TNF-a in the interactions between the
neurons and glia. Increasingly, glial cells are no
longer seen as passive supporters of neurons, but
as active participants in information processing
(for review, see Haydon, 2001; Volterra and
Steinhauser, 2004); neuronal–glial interactions
are seen as important to many neural processes
and phenomena. Of particular interest in the con-
text of this review is the relationship between
neurons, glial cells and TNF-a in glutamate ex-
citotoxicity.
Role of TNF-a in excitotoxicity
Excitotoxicity in general is linked to excessive
glutamate activation of receptors, particularly the
N-methyl-D-aspartate (NMDA) receptor. Cell
death resulting from excessive levels of glutamate
and overstimulation of glutamate receptors is
known to be caused by impaired uptake of gluta-
mate by glial cells (Choi, 1988). In vivo, it has been
shown that mice lacking expression of the excita-
tory amino-acid transporter, EAAT2/GLT-1,
develop epilepsy and increased susceptibility to
acute injury as a result of excessive extracellular
glutamate (Tanaka et al., 1997). The expression of
this transporter has been shown recently to be
both positively and negatively regulated by NF-kB
(Sitcheran et al., 2005). This study showed that the
increased binding of NF-kB to the EAAT2 pro-
moter in H4 astroglioma cells was regulated by
epidermal growth factor (EGF), but decreased
expression was caused by TNF-a inducing the
classical IkappaB degradation pathway to trigger
NF-kB nuclear translocation and DNA binding to
repress EAAT2 expression. In this situation, the
presence of elevated TNF-a concentrations leads
to elevated extracellular glutamate concentration,
thereby increasing the risk of glutamate excitotox-
icity. Indeed, Zou and Crews (2005) also showed
that in rat organotypic hippocampal slice cultures,
which possess a cytoarchitecture comparable to
that in vivo, TNF-a increased glutamate neuro-
toxicity. They also demonstrated that the effect was
mediated by NMDA and not AMPA receptors.
In addition to this, TNF-a and glutamate
have also been implicated together recently in
b-amyloid-induced microglia-related cell death.
Abundant activated microglia are prominent in
the brains of patients with Alzheimer’s disease
(Griffin et al., 1989), and are associated with beta

amyloid plaques (Griffin et al., 1995; Frautschy
et al., 1998). It has been proposed that inefficient
phagocytosis of peptide by microglia could lead to
hyperactivation of cells and release of inflamma-
tory mediators and neurotoxic factors, thereby
contributing to neurodegenerative processes
(Akiyama et al., 2000). It is widely believed that
the microglia play a direct role in the neuronal
death in Alzheimer’s disease. Floden et al. (2005)
showed that applying media from b-amyloid-
stimulated microglial cultures to neurons led to
neuronal cell death that was dependant on the
synergistic co-activation of TNF-a and NMDA
receptors; the NMDA receptor antagonists meman-
tine and 2-amino-5-phosphopetanoic acid as well as
soluble TNF-R1 applied to the neurons protected
them from cell death. Interestingly, blockade of
either the TNF-R1 or NMDA receptors alone was
insufficient to induce neuronal cell death.
There is, however, evidence that other glutamate
receptors are involved in the relationship between
neuronal cell death, glial cells and TNF-a. Taylor
et al. (2005) examined the effect of metabotropic
glutamate (mGlu) receptor stimulation on TNF-a
release. They found that stimulating rat primary-
cultured microglial mGlus for 24 h induced micro-
glial activation. This in turn induced caspase-3
activation in cerebellar granule neurons in culture,
both those treated with microglial-conditioned
media as well as in neuronal/microglial co-cul-
tures. This microglial neurotoxicity was mediated
by TNF-a released by the microglia via neuronal
TNF-R1 and caspase-3 activation. Importantly, it
was the specific group II mGluR agonist 2S,20 R,30 -
2-(20 ,30 -dicarboxy-cyclopropyl)glycine, which led
to the release of TNF-a and consequent toxicity;
N-acetyl-L-aspartyl-L-glutamate, a specific mGlu3
agonist, did not induce microglial activation or
neurotoxicity. TNF-a was only neurotoxic in the
presence of microglia or microglial-conditioned
medium; possibly this was due to the presence of
microglial-derived Fas ligand. However, TNF-a
neurotoxicity was prevented when the neurons
were exposed to conditioned medium from micro-
glia stimulated by the specific group III agonist L-
2-amino-4-phosphono-butyric acid.
Glial–glial interactions also influence the effects
of TNF-a. Bezzi et al. (2001) showed that astrocyte
343
glutamate release induced by activation of the
chemokine receptor CXCR4 is accompanied by
release of TNF-a, and that the TNF-a release is
dramatically enhanced by microglia. In light of
this evidence of glial–glial and glial–neuronal inter-
actions, it is possible to see how a cascading
neuronal–glial interaction could occur, leading to
cell death. Activation of TNF receptors on as-
trocytes leads to increased extracellular glutamate,
which may lead to neurotoxicity in itself, while also
activating mGlu2 on microglia, which releases more
TNF-a, in addition to other pro-inflammatory
cytokines. Also, it has been known for some time
that the neurotoxic effects of TNF-a may
be mediated by activation of glutamate AMPA
receptor subtypes (Gelbard et al., 1993).
TNF-a and synaptic transmission
The subject of TNF-a in glial–neuronal interac-
tions also emerges when looking at synaptic trans-
mission. Beattie et al. (2002) showed that glial
TNF-a causes an increase in surface expression of
neuronal AMPA receptors, which would increase
synaptic efficacy, and that the removal of endog-
enous TNF-a induced a decrease in AMPA recep-
tor expression at the cell surface. Further evidence
for this relationship between TNF-a and AMPA
receptor density came from Stellwagen and Male-
nka (2006), who found the same elevated AMPA
receptor density in the presence of elevated
TNF-a. This TNF-a-induced AMPA receptor ex-
ocytosis has recently been shown to be mediated
by activation of TNF-R1 receptors through a PI3
kinase-dependent pathway. (Stellwagen et al.,
2005). The newly expressed AMPA receptors were
shown to have lower stoichiometric amounts
of GluR2, making the receptors permeable to
Ca2+ ions. Additionally, this study also showed a
surprising concurrent endocytosis of inhibitory
GABA receptors.
TNF-a and synaptic plasticity
Changes in neuronal excitability brought about by
TNF-a have important implications for synaptic
344
plasticity (Carroll et al., 2001). Indeed, TNF-a is
known to act as a regulator of synaptic plasticity
in the dentate gyrus, in addition to playing a role
in apoptotic events. As previously mentioned,
elevated levels of TNF-a have been observed in
several neuropathological states that are associ-
ated with learning and memory deficits, such as
Alzheimer’s disease, leading to the search for a
possible role in plasticity. To this end, much work
has been done concerning the role of TNF-a in the
hippocampus. TNF-a has been shown to regulate
the development of the hippocampus; TNF-a
alpha knockout mice demonstrate decreased
arbourization of the apical dendrites of the CA1
and CA3 regions while at the same time causing
accelerated development in the dentate gyrus
(Golan et al., 2004), probably via activation of
the TNF-R2 receptor, which, as mentioned earlier,
does not lead to caspase-3 activation, but is known
to transduce the trophic effect of TNF-a by lead-
ing to the activation of several transcription fac-
tors (Yang et al., 2002).
Two forms of synaptic plasticity in the hippo-
campus, long-term potentiation (LTP) and long-
term depression (LTD) involve glutamate receptor
activation and increased intracellular Ca2+ levels,
with LTP induction dependant on the activation
of Ca2+/calmodulin kinase II, PKC and PKA, and
LTD on activation of serine/threonine phospha-
tases 1, 2A and 2B (Mayford et al., 1995; Coussens
and Teyler, 1996; Silva et al., 1997). LTP is a long-
lasting increase in synaptic efficacy, which is
thought to be an important underlying mecha-
nism of learning and memory formation (Bliss and
Collindridge, 1993).
TNF-a has in fact been shown to inhibit LTP in
the CA1 region, as well as the dentate gyrus of
the rat hippocampus, when pathophysiological
levels of TNF-a are used (Tancredi et al., 1992;
Cunningham et al., 1996; Butler et al., 2004;
Cumiskey et al., 2004, 2007). There are, however,
differences in the mechanisms by which TNF-a
inhibits early phase (i.e., o120 min post-tetanic
stimulation) and late phase (>120 min) LTP. But-
ler et al. (2002) investigated these differences in the
dentate gyrus region of rat hippocampal slices.
They found, using immunohistochemical tech-
niques, that TNF-a activated the p38 MAP
kinase. Inhibition of p38 MAP kinase with SB
203580 blocked the impairment of the early phase
of LTP by TNF-a, but the late phase was un-
affected, suggesting that the inhibition of LTP by
TNF-a is biphasic in character. If the early stage
inhibition of LTP is mediated by a p38 MAP kin-
ase pathway, it may be that the late phase inhibi-
tion of LTP by TNF-a is regulated by altered
protein synthesis; not an unlikely possibility given
the aforementioned regulation of NF-kB by TNF-
a. As well as a role for the p38 MAPK, (Cumiskey
et al., 2004; Cumiskey and O’ Connor, 2006) also
provided evidence for a role of metabotropic gluta-
mate receptors in the TNF-a-mediated inhibition of
LTP in the dentate gyrus (see Fig. 2).
In addition to the effect of TNF-a on LTP, it
has been shown that TNF receptor knockout mice
demonstrate an impairment of LTD in the CA1
region of the hippocampus (Albensi and Mattson,
2000). The effect is mimicked by NF-kB decoy
DNA, which implicates the TNF-a/NF-kB signal-
ling pathway in LTD. Additionally, Wang et al.
(2006) showed that metabotropic glutamate recep-
tor mediated LTD in the dentate, induced with
(RS)-3,5-dihydroxyphenylglycine (DHPG) could
not be induced in TNF-R1 knockout mice.
Interestingly, the findings relating to the effect of
TNF-a on synaptic plasticity seem to have some
behavioural correlates in vivo. TNF-a knockout
mice showed increased performance in spatial mem-
ory and learning as measured in the Morris water
maze task when compared to wild type animals
(Golan et al., 2004). Conversely, Aloe et al. (1999)
demonstrated a significant impairment in spatial
learning in two lines of mice that over-expressed
human recombinant TNF-a, also using the same
water maze task. However, these results must be
interpreted in the context of the study showing de-
velopmental differences mentioned previously; the
‘‘substrate’’ of learning and memory is not neces-
sarily the same in the knockout and wild type mice.
Recent work has also shown that TNF-a plays a
crucial role in the development of synaptic scaling;
a form of homeostatic synaptic plasticity that
scales synaptic strengths to compensate for pro-
longed changes in activity. This form of plasticity
is essential to stabilize activity across networks
(Turrigiano and Nelson, 2004), as it regulates
 345

Fig. 2. Schematic diagram of the role of the mGluR5 and NMDA receptor in mediating the TNF-a effects on LTP. TNF-a inhibits the
early phase of LTP by activation of TNF-R1 (p55) and is dependent on p38 activation. Activation of the G-protein-linked mGluRs
leads to inositol-1,4,5-trisphosphate (IP3) receptor-mediated Ca2+ release via phospholipase C (PLC). TNF-a may also alter Ca2+
concentrations. Combined activation of TNF-R and mGluR5 may lead to the impairment of LTP. TNF-a has negative effects on the
NMDA receptor EPSPs.

synaptic ‘‘gain’’, thus preventing excessive quiet-
ening or activity across the network. Stellwagen
and Malenka (2006) showed, using TNF receptor
knockout mice, that signalling from the constitu-
tive expression of TNF-a was not necessary for the
induction of LTP or LTD, but was necessary for
the development of synaptic scaling in the longer
term. Additionally, they showed that the source of
this TNF-a was not neuronal but was glial. This
finding is not surprising when one considers that
the glia are in a unique position to sense the overall
level of local network activity (presumably using
the index of glutamate overflow from synapses)
and adjust the activity accordingly.
While Stellwagen and Malenka (2006) did not
directly implicate TNF-a in synaptic scaling in the
dentate gyrus, the hippocampal slice component
of their study instead focusing on the CA1 region
in intact hippocampal slices, the evidence from
dissociated hippocampal cultures (presumably
containing a cohort of granule neurons) would
seem to suggest that TNF-a may also play a role in
the development of synaptic scaling in the dentate
gyrus.
Action of IL-1b in the dentate gyrus
IL-1b receptors
As described for TNF-a, IL-1 receptors have also
been shown to be present in many brain regions,
with high levels in the hippocampus and hypo-
thalamus (Ban et al., 1991). There are a number of
receptors to which IL-1 can bind (O’Neill, 1997).
The principal mediator of IL-1 signalling in the
CNS is believed to be the type I IL-1 receptor
(IL-1R1), which initiates intracellular events upon
IL-1b binding (O’Neill, 1996), although the type II
IL-1 receptor (IL-1R2) is also present, which
although having a similar affinity for IL-1b
to the type I receptor, has no ability to initiate
signal transduction events, and so may serve as a
‘‘decoy’’ receptor, attenuating IL-1b-induced sig-
nalling. Other receptors are the IL-1 receptor
accessory protein (IL-RAcp) that partakes in
IL-1R1 signalling and the IL-1 receptor-related
protein (IL-1Rcp), which is the receptor for IL-18,
a member of the IL-1 family (Okamura et al.,
1998), and will be discussed in detail later.
346
The presence of IL-1 receptors in brain was
shown in 1990 by Takao et al. using 125I-labelled
IL-1a (both IL-1a and IL-1b can activate IL-1 re-
ceptors) binding to crude membrane preparations
of mouse hippocampus. Their autoradiographic
localization studies revealed very low densities of
[1251]-IL-1a binding sites throughout the brain,
with highest densities present in the molecular and
granular layers of the dentate gyrus of the hippo-
campus. This finding was later confirmed by fur-
ther binding studies (Haour et al., 1990; Ban et al.,
1991) and then by affinity cross-linking of IL-1a
(Parnet et al., 1994). Further characterization of
the distribution of IL-1 receptors in brain utilized
antisense cRNA against the type I IL-1 receptor
in mouse brain (Cunningham et al., 1992), and
RT-PCR (Parnet et al., 1994). These studies
demonstrate the presence of both type I and type
II IL-1 receptors in mouse hippocampus, with
particularly high expression in the cell-dense gran-
ule cell layer of the dentate gyrus.
Much like TNF-R2 signalling, IL-1 signalling
leads to a change in gene expression — up to 90
genes have been shown to be affected by IL-1b
(O’Neill and Green, 1998). These changes include
upregulation of mRNA transcription for a number
of cytokines, growth factors, adhesion molecules
and acute-phase proteins. Again, similar to TNF-
a, IL-1b regulates gene transcription is by activa-
tion of the transcription factor NF-kB. IL-1 binds
to IL-1R1, which is complexed with the IL-1 Racp.
This in turn leads to recruitment of the IL-1
receptor-associated kinase (IRAK; O’Neill and
Green, 1998). IRAK then associates with TRAF6
(tumour necrosis factor (TNF)-receptor-associated
factor), causing activation of IcB kinase (IKK).
IKK activation leads to phosphorylation of IKB,
which marks it for ubiquitination and subsequent
degradation (O’Neill, 1997; O’Neill and Green,
1998). The p50-p65 heterodimer, now free of 1cB,
can translocate to the nucleus, bind to its nuclear
KB site and activate gene transcription, leading to
expression of proteins such as IL-6, cycloxygenase-
2, inducible-nitric oxide synthase, IL-1b, TNF-a
and interferons (O’Neill and Kaltschmidt, 1997;
see Fig. 3).
One family of protein kinases that have been
very strongly implicated in IL-1 signalling are the
MAP kinases. The three MAP kinase cascades: the
p42/44 MAP kinases (also known as extracellular
regulated kinases — ERKs); the p38 MAP kinases
(also known as reactivating kinase — RK) and jun
kinases (JNK; also known as p54 or stress-
activated protein kinases — SAPKs) have all been
implicated in IL-1 signalling.
IL-1b and synaptic plasticity
IL-1b has been found to exert a number of effects
in the CNS, and has been implicated in a number
of processes such as centrally-mediated fever, slow
wave sleep patterns, appetite suppression and
neurodegeneration (for reviews, see Rothwell and
Hopkins, 1995; Rothwell, 1999). Oitzl et al. (1993)
showed that IL-1b can disrupt the acquisition of
spatial learning in the Morris water maze via
a mechanism which seemed to be independent of
IL-1b’s pyrogenic properties. More recently, it was
shown that IL-1R knockout mice exhibited deficits
in a number of learning paradigms, as well as in
LTP, suggesting that there is a physiological role for
IL-1b in learning and memory, and possibly the in-
duction of synaptic plasticity (Avital et al., 2003);
although, as with the TNFR knockout studies,
the evidence is confounded by possible effects of
cytokine receptor deficits on development. The first
reported action of IL-1b on synaptic plasticity was
inhibition of NMDA receptor-independent LTP
in the mouse mossy-fibre-CA3 pathway in vitro
(Katuski et al., 1990). Concentrations as low as
50 pg/ml were found to produce a significant
inhibition of LTP, and this inhibitory effect
was blocked by Lys-D-Pro-Thr, a tripeptide ana-
logue of IL-1b. No effects on baseline transmission
were observed. In the Schaeffer-collateral-CA1
pathway, Bellinger et al. (1993) reported an inhi-
bition by IL-1b of the potentiation of both the
field EPSP and the population spike induced by
high-frequency stimulation. IL-1b treatment was
also reported to cause a transient decrease in pop-
ulation spike amplitude and a persistent decrease
in EPSP slope. Bellinger et al. (1993) reported
preliminary results that suggested that treatment
of slices with the IL-1ra led to the induction of a
larger LTP than control conditions. A later study
 347

Fig. 3. Putative IL-1b signalling pathway in the dentate gyrus. In IL-1 signalling, the signal-mediating IL-1 receptor type I (IL-1R1)
forms a heterodimer with a second molecule the IL-1 receptor accessory protein (IL-1RAcP). Binding of IL-1b to this receptor complex
leads to activation of the transcription factor NFB through different signalling molecules. DNA-binding motifs for NFB are found in
the promotors and enhancers of many genes that are known to be activated upon inflammation.

of the MPP-dentate granule cell pathway showed
an inhibition of the induction of LTP by treatment
with IL-1b, and this effect was antagonized by co-
application of IL-1ra, and again, as in the CA3
region, IL-1b was found to have no effect on
baseline transmission in the dentate gyrus
(Cunningham et al., 1996). This IL-1b-induced
inhibition of LTP was accompanied by a decrease
in Ca2+ influx into slices compared with tetanized
slices, suggesting that IL-1b inhibits LTP in the
dentate gyrus by inhibiting the Ca2+ influx neces-
sary for LTP induction.
Schneider et al. (1998) showed that in the CA1
region, expression of IL-1b transcripts was signifi-
cantly upregulated 1 h following tetanization both in
vitro and in vivo, which may explain why
Bellinger et al. (1993) found that treatment of slices
solely with IL-1ra produces a larger potentiation
of EPSPs than in untreated slices. Along with the
increased gene expression of IL-1b during LTP,
Schneider et al. (1998) and Coogan et al. (1999) have
shown that the maintenance phase of LTP can be
blocked by administration of IL-1ra. This suggests
a requirement for IL-1b in the sustained expression
of LTP, i.e., the endogenous IL-1b production by
tetanic stimulation is required for sustained expres-
sion of LTP, while, somewhat paradoxically, the
presence of elevated IL-1b can inhibit the initial
induction of LTP. It seems that it is the sequence
and timing of IL-1b changes that is important in
regulating its effect on LTP.
In the dentate gyrus, the IL-1b-induced inhibition
of LTP in slices was found to be concomi-
tant with a decrease in Ca2+ influx following tetanic
stimulation, either pre- or postsynaptically, as
judged from synaptosomal preparations (Cunning-
ham et al., 1996). Further investigations into the
actions of IL-1b in the dentate gyrus showed that
treatment of slices with IL-1b caused a large de-
pression of the pharmacologically isolated NMDA
348
receptor EPSP (NMDA-EPSP; Coogan and O’Con-
nor, 1997). These effects were thought to be specific
for IL-1R1, because the depression was fully antag-
onized by IL-1ra. This might explain the decrease in
Ca2+ influx, possibly through the NMDA receptor
channel, and so may serve to inhibit LTP induction
by preventing sufficient Ca2+ influx via the NMDA
receptor to trigger the intracellular processes that
underlie the expression of LTP (Bliss and Collind-
ridge, 1993). The fact that IL-1b seems to have an
effect on NMDA receptor-mediated EPSPs suggests
a postsynaptic locus of action for IL-1b in the dent-
ate gyrus. The inhibitory effects of IL-1b have not
been restricted to tetanically induced LTP since it
has been shown that IL-1b can inhibit a tetraethyl-
ammonium-induced synaptic potentiation in the rat
dentate gyrus (Coogan and O’Connor, 1999).
IL-1b has also been shown to inhibit the release of
glutamate from synaptosomes prepared from hippo-
campi of young rats, via a mechanism that seemed to
be coupled to phospholipase A2(Murray et al., 1997).
However, IL-1b was not observed to have any effect
on glutamate release from synaptosomes and hippo-
campi from aged animals (Murray et al., 1997;
Lynch, 1998b) The inhibition of LTP by IL-1b in
vivo has been proposed to be mediated by a decrease
in glutamate release and also via an induced increase
in the formation of reactive oxygen radicals, leading
to an increase in lipid peroxidation and a decrease in
arachidonic acid levels (Murray and Lynch, 1998).
As arachidonic acid has been shown to act as a
retrograde (postsynaptic-to-presynaptic) messenger
during LTP induction (Medina and Izquierdo,
1995), this has been suggested as a mechanism of
IL-1b-induced inhibition of LTP in the dentate gyrus.
As discussed above, the mitogen-activated protein
(MAP) kinases have been implicated in the induc-
tion of LTP. PD98059, a specific inhibitor of MAP
kinase kinase (MEK) and thus, p42/44 MAP kinase
activation (Pang et al., 1995), has been shown to
block induction of LTP in the area CA1 as well as
the dentate gyrus of the hippocampus (English and
Sweatt, 1996; Coogan et al., 1999). When these
effects were further investigated to try to elucidate
the intracellular pathway through which IL-1b
exerts its action on LTP, it was found that pre-
incubation of slices with the p38 MAP kinase in-
hibitor SB203580 inhibited the effects of IL-1b on
both LTP induction and NMDA-EPSPs in the
dentate gyrus (Coogan et al., 1999). SB203580 did
cause a small but significant increase in baseline
transmission. It is possible that there may be a high
constitutive expression of p38 MAP kinase activity
in the hippocampus, which may exert an inhibitory
effect on synaptic transmission, and its inhibition
may lead to the observed increase in synaptic trans-
mission. When the effects of a lower dose of
PD98059 (10 mM, a dose that does not inhibit
LTP) were examined on the IL-1b inhibition of
LTP, it was found that it did not antagonize both
the inhibition of LTP or the inhibition of the iso-
lated NMDA receptor-mediated EPSP (Coogan et
al., 1999). In addition to p38 MAPK, a role for c-
jun n-terminal kinase in the inhibition of LTP by
IL-1b and long-term depression has also been dem-
onstrated (Curran et al., 2003). A role for MAP
kinases in LTD has also been shown in the
CA3–CA1 synapse (Bolshakov et al., 2000).
IL-1b and Ca2+ channels
While, as previously mentioned, IL-1 receptor acti-
vation leads to altered gene expression, since most
of the inhibitory effects of IL-1b in the hippocampus
have been reported following acute treatment of
slices, it seems unlikely that gene transcription plays
a role in the observed effects. Therefore it seems
more likely that IL-1b mediates its effects by acti-
vation of cytoplasmic cascades with cytoplasmic/
membrane bound substrates. The mechanism by
which IL-1b exerts its effects on hippocampal LTP
remains controversial, although a number of possi-
bilities have been raised. In dissociated guinea-pig
CA1 pyramidal neurones it has been shown that IL-
1b rapidly induces a depression of inward, voltage-
dependent Ca2+ currents, (both peak and late cur-
rent) in a dose-dependent fashion from 0.197 to
3.1 ng/ml. This could be reversed by pretreatment
with IL-1ra (Plata-Salaman and Ffrench-Mullen,
1992), indicating a specific effect on IL-1R1.
Further investigation of this effect showed it to be
inhibited by a non-hydrolysable GTP analogue and
pertussis toxin, and also by protein kinase C inhib-
itors H-7 and staurosporine (Plata-Salaman and
Ffrench-Mullen, 1994), suggesting that the effect on
 349

Ca2+ channels is mediated by G-proteins and pro-
tein kinase C, although H-7 and staurosporine can
also inhibit other kinases, such as protein kinase A
and Ca2+-calmodulin kinase (Hikada et al., 1991).
IL-1ra was found to completely block (excess of 25)
the PKC-dependent IL-1b induced decrease in Ca2+
currents. It is therefore likely that IL-1b interacts
with the IL-1R1 membrane site. It must be noted,
however, that low concentrations of IL-1b (3.5 ng/
ml) have been shown to decrease intracellular Ca2+
concentrations while higher levels (100 ng/ml) mark-
edly increase it (Campbell and Lynch, 1998).
IL-18
IL-18 is a member of the IL-1 pro-inflammatory
cytokine family, closely related to IL-1b (Bazan
et al., 1996), and, like other cytokines, can be
detected in many brain regions, including the hip-
pocampus (Culhane et al., 1998). It is expressed
mainly in glia (both astrocytes and microglia) and
is secreted in response to LPS stimulation (Conti
et al., 1999). Again, like TNF-a and IL-1b, acute
application of IL-18 has been shown to inhibit the
induction of LTP in the dentate gyrus, without
affecting baseline neurotransmission (Curran and
O’Connor, 2001, 2003; Cumiskey et al., 2007).
Another similarity between IL-1b and IL-18 is the
observation that IL-18 also activates p38 MAP
kinase (Thomassen et al., 1998), which, as men-
tioned above, is implicated in the IL-1b inhibition
of LTP. Recent work has also shown interesting
similarities between the mechanism of IL-18 inhi-
bition of LTP and the mechanism by which TNF-a
inhibits LTP. Cumiskey et al. (2007) showed that,
like TNF-a, IL-18 inhibition of LTP involves
mGluRs. The study suggested that IL-18 inhibits
LTP in the dentate gyrus by an LTD-like mech-
anism. In these experiments, the application of the
group II mGluR antagonist MTPG prevented the
inhibition of LTP by IL-18. Group II and III
mGluRs are negatively coupled to adenylyl cyclase
(Tanabe et al., 1992), and their activation can induce
synaptic depression (Manahan-Vaughan and Rey-
mann, 1997), thus suggesting that IL-18 might be
acting through an LTD-like mechanism.

Fig. 4. Schematic diagram of the role of the mGluRs and IL-18R in LTP in the dentate gyrus. IL-18 inhibits the early phase of LTP by
activation of IL-18R. Activation of the G-protein-linked mGluR5 leads to inositol-1,4,5-trisphosphate (IP3) receptor-mediated Ca2+
release via phospholipase C (PLC). Activation of the group II mGluRs leads to inhibition of cAMP. p38 has also been implicated in
IL-18 mediated inhibition of LTP.
350
An interesting aspect of the involvement of
mGluRs in the effect of IL-18 on synaptic trans-
mission and plasticity comes from an apparently
contradictory study. Kanno et al. (2004) found that
IL-18 facilitates basal synaptic transmission as a
result of stimulating presynaptic glutamate release
and enhancing AMPA receptor responses in the
CA1, but had no significant effect on LTP in that
region. The key difference would appear to be the
region of the hippocampus under investigation. The
expression of mGluRs is known to be very different
in the CA1 and dentate regions (Fotuhi et al.,
1994), with high levels of mGluR2, mGluR5 and
mGluR7 and low levels of mGluR1, mGluR3 and
mGluR4 in the dentate, while the CA1 exhibits low
mGluR5 and high mGluR1 distribution. This
would seem to suggest that the effect of IL-18 on
synaptic transmission and plasticity is dependant in
some way in the balance between mGluR1 and
mGluR5 signalling. The exact nature of this rela-
tionship will remain unknown until further studies
can shed light on the signalling paths from IL-18 to
the mGluRs, which would also provide further in-
sight into the complex nature of IL-18 signalling in
the hippocampus (Fig. 4).
Concluding remarks
The overriding theme among studies of pro-inflam-
matory cytokines in the dentate gyrus, and other
brain regions, is that of complexity. A primary ex-
ample is the potential for opposing actions. In ad-
dition, the action of these cytokines on neurons is
often subtle and multi-faceted, with very different
short- and long-term consequences, making their
elucidation all the more challenging. A more com-
prehensive understanding cannot be achieved until
the interactions between the cytokine receptors and
other neuronal mechanisms are better understood.
Acknowledgements
We would like to thank the Health Education
Authority, Ireland, PRTLI Cycle 3 for funding
and Dr Derval Cumiskey for assistance on figure
production.
References
Akiyama, H., Barger, S., Barnum, S., Bradt, B., Bauer, J., Cole,
G.M., Cooper, N.R., Eikelenboom, P., Emmerling, M., Fie-
bich, B.L., Finch, C.E., Frautschy, S., Griffin, W.S., Hampel,
H., Hull, M., Landreth, G., Lue, L., Mrak, R., MacKenzie,
I.R., McGeer, P.L., O’Banion, M.K., Pachter, J., Pasinetti,
G., Plata-Salaman, C., Rogers, J., Rydel, R., Shen, Y., Streit,
W., Strohmeyer, R., Tooyoma, I., Van Muiswinkel, F.L.,
Veerhuis, R., Walker, D., Webster, S., Wegrzyniak, B.,
Wenk, G. and Wyss-Coray, T. (2000) Inflammation and
Alzheimer’s disease. Neurobiol. Aging, 21: 383–421.
Albensi, B.C. and Mattson, M.P. (2000) Evidence for the
involvement of TNF-a and NF-kappaB in hippocampal
synaptic plasticity. Synapse, 35: 151–159.
Allan, S.M. and Rothwell, N.J. (2001) Cytokines and acute
neurodegeneration. Nat. Rev. Neurosci., 2: 734–744.
Aloe, L., Properzi, F., Probert, L., Akassoglou, K., Kassiotis,
G., Micera, A. and Fiore, L. (1999) Learning abilities, NGF
and BDNF brain levels in two lines of TNFalpha transgenic
mice, one characterized by neurological disorders, the other
phenotypically normal. Brain Res., 840: 125–137.
Armstrong, R.C., Aja, T., Xiang, J., Gaur, S., Krebs, J.F.,
Hoang, K., Bai, X., Korsmeyer, S.J., Karanewsky, D.S.,
Fritz, L.C. and Tomaselli, K.J. (1996) Fas-induced activation
of the cell death-related protease CPP32 is inhibited by Bcl-2
and by ICE family protease inhibitors. J. Biol. Chem., 271:
16850–16855.
Avital, A., Goshen, I., Kamsler, A., Segal, M., Iverfeldt, K.,
Richter-Levin, G. and Yirmiya, R. (2003) Impaired inter-
leukin-1 signaling is associated with deficits in hippocampal
memory processes and neural plasticity. Hippocampus, 13:
826–834.
Ban, E., Milon, G., Prunhom, N., Fillion, G. and Haour, F.
(1991) Receptors for interleukin-1 (a and b) in mouse brain:
mapping and neuronal localization in hippocampus. Neuro-
science, 43: 21–30.
Barger, S.W., Horster, D., Furukawa, K., Goodman, Y.,
Krieglstein, J. and Mattson, M.P. (1995) Tumor necrosis
factors alpha and beta protect neurons against amyloid beta-
peptide toxicity: evidence for involvement of a kappa
B-binding factor and attenuation of peroxide and Ca2+ ac-
cumulation. Proc. Natl. Acad. Sci. U.S.A., 92: 9328–9332.
Bazan, J.F., Timans, J.C. and Kastelein, R.A. (1996) A newly
defined interleukin-1b. Nature, 379: 591.
Beattie, E.C., Stellwagen, D., Morishita, W., Bresnahan, J.C.,
Ha, B.K., Von Zastrow, M., Beattie, M.S. and Malenka,
R.C. (2002) Control of synaptic strength by glial TNFalpha.
Science, 295: 2282–2285.
Bellinger, F.P., Madamba, S. and Siggins, G.R. (1993)
Interleukin-1b inhibits synaptic strength and long-term
potentiation in the rat CAI hippocampus. Brain Res., 628:
227–234.
Benveniste, E.N. and Benos, D.J. (1995) TNF-alpha- and IFN-
gamma-mediated signal transduction pathways: effects on
glial cell gene expression and function. FASEB J., 9:
1577–1584.

Besedovsky, H.O., del Rey, A., Klusman, I., Furukawa, H.,
Monge Arditi, G. and Kabiersch, A. (1991) Cytokines as
modulators of the hypothalamus-pituitary-adrenal axis.
J. Steroid Biochem. Mol. Biol., 40: 613–618.
Beutler, B. and Van Huffel, C. (1994a) An evolutionary and
functional approach to the TNF receptor/ligand family. Ann.
N.Y. Acad. Sci., 730: 118–133.
Beutler, B. and Van Huffel, C. (1994b) Unraveling function in
the TNF ligand and receptor families. Science, 264: 667–668.
Bezzi, P., Domercq, M., Brambilla, L., Galli, R., Schols, D., De
Clercq, E., Vescovi, A., Bagetta, G., Kollias, G., Meldolesi, J.
and Volterra, A. (2001) CXCR4-activated astrocyte gluta-
mate release via TNF: amplification by microglia triggers
neurotoxicity. Nat. Neurosci., 4: 702–707.
Bliss, T.V.P. and Collindridge, G.L. (1993) A synaptic model of
memory: long-term potentiation in the hippocampus. Nature,
361: 31–38.
Bolshakov, V.Y., Carboni, L., Cobb, M.H., Siegelbaum, S.A.
and Belardetti, F. (2000) Dual MAP kinase pathways medi-
ate opposing forms of long-term plasticity at CA3–CA1
synapses. Nat. Neurosci., 3: 1107–1112.
Boulanger, L.M., Huh, G.S. and Shatz, C.J. (2001) Neuronal
plasticity and cellular immunity: shared molecular mecha-
nisms. Curr. Opin. Neurobiol., 11: 568–578.
Bruce, A.J., Boling, W., Kindy, M.S., Peschon, J., Kraemer,
P.J., Carpenter, M.K., Holtsberg, F.W. and Mattson, M.P.
(1996) Altered neuronal and microglial responses to excito-
toxic and ischemic brain injury in mice lacking TNF recep-
tors. Nat. Med., 2: 788–794.
Butler, M., O’Connor, J.J. and Moynagh, P. (2004) Dissection
of TNF-a inhibition of LTP reveals a p38 MAPK-dependent
mechanism which maps to early but not late-phase LTP.
Neuroscience, 124: 319–326.
Butler, M.P., Moynagh, P.N. and O’Connor, J.J. (2002)
Methods of detection of the transcription factor NF-kappa
B in rat hippocampal slices. J. Neurosci. Methods, 119:
185–190.
Cacquevel, M., Lebeurrier, N., Cheenne, S. and Vivien, D.
(2004) Cytokines in neuroinflammation and Alzheimer’s dis-
ease. Curr. Drug Targets, 5(6): 529–534.
Campbell, V. and Lynch, M.A. (1998) Biphasic modulation
of [Ca2] by interleukin-1b in rat cortical synatosomes.
Involvement of a PTX-sensitive G-protein and p42 MAP
kinase. Neuroreport, 9: 9–12.
Carroll, R.C., Beattie, E.C., von Zastrow, M. and Malenka,
R.C. (2001) Role of AMPA receptor endocytosis in synaptic
plasticity. Nat. Rev. Neurosci., 2(5): 315–324.
Cartmell, T., Luheshi, G.N. and Rothwell, N.J. (1999) Brain
sites of action of endogenous interleukin-1 in the febrile
response to localized inflammation in the rat. J. Physiol.,
518(2): 585–594.
Cheng, B., Christakos, S. and Mattson, M.P. (1994) Tumor
necrosis factors protect neurons against excitotoxic/
metabolic insults and promote maintenance of Ca2+ home-
ostasis. Neuron, 12: 139–153.
Choi, D.W. (1988) Glutamate neurotoxicity and diseases of the
nervous system. Neuron, 1: 623–634.
351
Chun, J. (2001) Selected comparison of immune and nervous
system development. Adv. Immunol., 77: 297–322.
Chung, I.Y. and Benveniste, E.N. (1990) Tumor necrosis
factor-alpha production by astrocytes. Induction by lip-
opolysaccharide, IFN-gamma, and IL-1 beta. J. Immunol.,
144: 2999–3007.
Conti, B., Park, L.C., Calingasan, N.Y., Kim, Y., Kim, H.,
Bae, Y., Gibson, G.E. and Joh, T.H. (1999) Cultures of
astrocytes and microglia express interleukin 18. Brain Res.
Mol. Brain Res., 67: 46–52.
Coogan, A. and O’Connor, J.J. (1997) Inhibition of NMDA
receptor-mediated synaptic transmission in the rat dentate
gyrus in vitro by IL-1b. Neuroreport, 8(9): 2107–2110.
Coogan, A. and O’Connor, J.J. (1999) Interleukin-1b inhibits
a tetraethylammonium-induced synaptic potentiation in the
rat dentate gyrus in vitro. Eur. J. Pharmacol., 374(2):
197–206.
Coogan, A., O’Neill, L.A.J. and O’Connor, J.J. (1999) The p38
MAP kinase inhibitor SB203580 antagonises the inhibitory
effect of interleukin-1b on long-term potentiation in the rat
dentate gyrus in vitro. Neuroscience, 93(1): 57–69.
Coussens, C.M. and Teyler, T.J. (1996) Protein kinase and
phosphatase activity regulate the form of synaptic plasticity
expressed. Synapse, 24: 97–103.
Culhane, A.C., Hall, M.D., Rothwell, N.J. and Luheshi, G.N.
(1998) Cloning of rat brain interleukin-18 cDNA. Mol.
Psychiatry, 3: 362–366.
Cumiskey, D., Butler, M.P., Moynagh, P.N. and O’Connor, JJ.
(2004) The inhibitory effect of tumour necrosis factor-a on
long-term potentiation is attenuated by type 1 metabotropic
glutamate receptor blockade. J. Physiol., 560P: C23.
Cumiskey, D. and O’ Connor, J.J. (2006) A role for mGluRs
and Ca2+ in the inhibitory effect of tumour necrosis factor-a
on long-term potentiation. Proc. Br. Pharmacol. Soc., 3(4):
055P.
Cumiskey, D., Pickering, M. and O’Connor, J.J. (2007)
Interleukin-18 mediated inhibition of LTP in the rat dentate
gyrus is attenuated in the presence of mGluR antagonists.
Neurosci. Lett., 412: 206–210.
Cunningham, A.J., Murray, C.A., O’Neill, L.A., Lynch, M.A.
and O’Connor, J.J. (1996) Interleukin-1 beta (IL-1 beta)
and tumour necrosis factor (TNF) inhibit long-term po-
tentiation in the rat dentate gyrus in vitro. Neurosci. Lett.,
203: 17–20.
Cunningham, E.T., Wada, E., Carter, D.B., Tracey, D.E.,
Battery, J.F. and De Souza, E.B. (1992) In situ histochemical
localization of type I interleukin-1 messenger RNA in the
central nervous system, pituitary and adrenal gland of the
mouse. J. Neurosci., 12: 1101–1114.
Curran, B. and O’Connor, J.J. (2001) The novel pro-inflam-
matory cytokine interleukin-18 (IL-18) inhibits long-term
potentiation in the rat hippocampus in vitro. Neuroscience,
108(1): 83–90.
Curran, B. and O’Connor, J.J. (2003) The inhibition of long-
term potentiation by pro-inflammatory cytokines is attenu-
ated in the presence of nicotine. Neurosci. Lett., 344:
103–106.
352
Curran, B.P., Murray, H.J. and O’Connor, J.J. (2003) A role
for c-jun n-terminal kinase in the inhibition of long-term
potentiation by interleukin-1beta and long-term depression
in the rat dentate gyrus in vitro. Neuroscience, 118:
347–357.
English, J.D. and Sweatt, J.D. (1996) Activation of p42
MAPkinase in hippocampal LTP. J. Biol. Chem., 271:
24329–24332.
Floden, A.M., Li, S. and Combs, C.K. (2005) Beta-amyloid-
stimulated microglia induce neuron death via synergistic
stimulation of tumor necrosis factor alpha and NMDA
receptors. J. Neurosci., 25: 2566–2575.
Fontaine, V., Mohand-Said, S., Hanoteau, N., Fuchs, C.,
Pfizenmaier, K. and Eisel, U. (2002) Neurodegenerative and
neuroprotective effects of tumor necrosis factor (TNF) in
retinal ischemia: opposite roles of TNF receptor 1 and TNF
receptor 2. J. Neurosci., 22: 1–7.
Fotuhi, M., Standaert, D.G., Testa, C.M., Penney Jr., J.B.
and Young, A.B. (1994) Differential expression of metabo-
tropic glutamate receptors in the hippocampusand ent-
orhinal cortex of the rat. Brain Res. Mol. Brain Res.,
21(3–4): 283–292.
Frautschy, SA., Yng, F., Irrizarry, M., Hyman, B., Saido, T.C.,
Hsiao, K. and Cole, G.M. (1998) Microglial response to
amyloid plaques in APPsw transgenic mice. Am. J. Pathol.,
152: 307–317.
Garcia, J.E., Nonner, D., Ross, D. and Barrett, J.N. (1992)
Neurotoxic components in normal serum. Exp. Neurol., 118:
309–316.
Gary, D.S., Bruce-Keller, A.J., Kindy, M.S. and Mattson, M.P.
(1998) Ischemic and excitotoxic brain injury is enhanced in
mice lacking the p55 tumor necrosis factor receptor. J. Cereb.
Blood Flow Metab., 18: 1283–1287.
Gelbard, H.A., Dzenko, K.A., DiLoreto, D., del Cerro, C.,
del Cerro, M. and Epstein, L.G. (1993) Neurotoxic effects of
tumor necrosis factor alpha in primary human neuronal
cultures are mediated by activation of the glutamate AMPA
receptor subtype: implications for AIDS neuropathogenesis.
Dev. Neurosci., 15: 417–422.
Ghosh, S. and Karin, M. (2002) Missing pieces in the
NF-kappaB puzzle. Cell, 109(Suppl.): S81–S96.
Golan, H., Levav1, T., Mendelsohn, A. and Huleihel, M. (2004)
Involvement of tumor necrosis factor alpha in hippocampal
development and function. Cereb. Cortex, 14: 97–105.
Goodman, Y. and Mattson, M.P. (1996) Ceramide protects
hippocampal neurons against excitotoxic and oxidative
insults, and amyloid beta-peptide toxicity. J. Neurochem.,
66: 869–872.
Grell, M., Wajant, H., Löhden, M., Maxeiner, B., Georgopou-
los, S., Kollias, G., Lesslauer, W., Pfizenmaier, K. and
Scheurich, P. (1995) The transmembrane form of tumor
necrosis factor (TNF) is the prime activating ligand of the 80
kDa TNF receptor. Cell, 83: 793–802.
Griffin, W.S.T., Sheng, J.G., Roberts, G.W. and Mrak, R.E.
(1995) Interleukin-1 expression in different plaque types
in Alzheimer’s disease, significance in plaque evolution.
J. Neuropathol. Exp. Neurol., 54.
Griffin, W.S.T., Stanley, L.C., Ling, C., MacLeod, V., Perrot,
L.J., White III, C.L., et al. (1989) Brain interleukin-1 and S100
immunoreactivity are elevated in Alzheimer disease and in
Down syndrome. Proc. Natl. Acad. Sci. U.S.A., 86: 7611–7615.
Haour, F.G., Ban, E.M., Milon, G.M., Baran, D. and Fillion,
G.M. (1990) Brain interleukin-1 receptors: characterization
and modulation after lipopolysaccharide injection. Prog.
Neuroendrochronol., 3: 196–204.
Haydon, P.G. (2001) Glia: listening and talking to the synapse.
Nat. Rev. Neurosci., 2: 185–193.
Hikada, H., Watanabe, M. and Kobayashi, R. (1991) Proper-
ties and use of H-series compounds as protein kinase inhib-
itors. Metab. Enzymol., 201: 328–336.
Israel, A. (2000) The IKK complex: an integrator of all signals
that activate NF-kappaB? Trends Cell Biol., 10: 129–133.
Kanno, T., Nagata, T., Yamamoto, S., Okamura, H. and
Nishizaki, T. (2004) Interleukin-265 18 stimulates synaptic-
ally released glutamate and enhances postsynaptic AMPA
receptor responses in the CA1 region of mouse hippocampal
slices. Brain Res., 1012: 190–193.
Kassiotis, G. and Kollias, G. (2001) Uncoupling the proin-
flammatory from the immunosuppressive properties of
tumor necrosis factor (TNF) at the p55 TNF receptor level:
implications for pathogenesis and therapy of autoimmune
demyelination. J. Exp. Med., 193: 427–434.
Katuski, H., Nakai, S., Hirai, Y., Akaji, K.I., Kiso, Y. and
Satoh, M. (1990) Interleukin-l, inhibits long-term potentiat-
ion in the CA3 region of mouse hippocampal slices. Eur. J.
Pharmacol., 181: 323–326.
Kinouchi, K., Brown, G., Pasternak, G. and Donner, D.B.
(1991) Identification and characterization of receptors for
tumor necrosis factor-alpha in the brain. Biochem. Biophys.
Res. Commun., 181: 1532–1538.
Klein, B.D., White, H.S. and Callahan, K.S. (2000) Cytokine
and intracellular signaling regulation of tissue factor expres-
sion in astrocytes. Neurochem. Int., 36: 441–449.
Kolesnick, R. and Golde, D.W. (1994) The sphingomyelin
pathway in tumor necrosis factor and interleukin-1 signaling.
Cell, 77: 325–328.
Li, Q. and Verma, I.M. (2002) NF-kappaB regulation in the
immune system. Nat. Rev. Immunol., 2: 725–734.
Lieberman, A.P., Pitha, P.M., Shin, H.S. and Shin, M.L.
(1989) Production of tumor necrosis factor and other
cytokines by astrocytes stimulated with lipopolysaccharide
or a neurotropic virus. Proc. Natl. Acad. Sci. U.S.A., 86:
6348–6352.
Liu, T., Clark, R.K., McDonnell, P.C., Young, P.R., White,
R.F., Barone, F.C. and Feuerstein, G.Z. (1994) Tumor
necrosis factor-alpha expression in ischemic neurons. Stroke,
25: 1481–1488.
Lynch, M.A. (1998) Age-related impairment in long-term
potentiation in hippocampus: a role for the cytokine, inter-
leukin-1 beta? Prog. Neurobiol., 56: 571–589.
Manahan-Vaughan, D. and Reymann, K.G. (1997) Group 1
metabotropic glutamate receptors contribute to slow-onset
potentiation in the rat CA1 region in vivo. Neuropharma-
cology, 36: 1533–1538.

Mattson, M.P., Barger, S.W., Furukawa, K., Bruce, A.J.,
Wyss-Coray, T., Mark, R.J. and Mucke, L. (1997a) Cellular
signaling roles of TGF beta, TNF alpha and beta APP in
brain injury responses and Alzheimer’s disease. Brain Res.
Brain Res. Rev., 23: 47–61.
Mattson, M.P., Goodman, Y., Luo, H., Fu, W. and Furukawa,
K. (1997b) Activation of NF-kappaB protects hippocampal
neurons against oxidative stress-induced apoptosis: evidence
for induction of manganese superoxide dismutase and sup-
pression of peroxynitrite production and protein tyrosine
nitration. J. Neurosci. Res., 49: 681–697.
Mayford, M., Wang, J., Kandel, E.R. and O’Dell, T.J. (1995)
CaMKII regulates the frequency response function of hip-
pocampal synapses for the production of both LTD and
LTP. Cell, 81: 891–904.
Medina, J.H. and Izquierdo, I. (1995) Retrograde messengers,
long-term potentiation and memory. Behav. Brain Res., 58:
1–8.
Merrill, J.E. (1992) Pro-inflammatory and antiinflammatory
cytokines in multiple sclerosis and central nervous system
acquired immunodeficiency syndrome. J. Immunother., 12:
167–170.
Morganti-Kossman, M.C., Lenzlinger, P.M., Hans, V., Stahel,
P., Csuka, E., Ammann, E., Stocker, R., Trentz, O. and
Kossmann, T. (1997) Production of cytokines following brain
injury: beneficial and deleterious for the damaged tissue.
Mol. Psychiatry, 2: 133–136.
Mosmann, T.R., Cherwinski, H., Bond, M.W., Giedlin, M.A.
and Coffman, R.L. (1986) Two types of murine helper T
cell clone. I. Definition according to profiles of lymphokine
activities and secreted proteins. J. Immunol., 136:
2348–2357.
Murray, C.A. and Lynch, M.A. (1998) Evidence that increased
hippocampal expression of the cytokine interleukin-1b is a
common trigger for age- and stress-induced impairments in
long-term potentiation. J. Neurosci., 18: 2974–2981.
Murray, C.A., McGahon, B., McBennet, S. and Lynch, M.A.
(1997) Interleukin-l/1 inhibits glutamate release in hippo-
campus of young, but not aged, rats. Neurobiol. Ageing, 18:
343–348.
Nath, R., Robert, A., McGinnis, K.M. and Wang, K.K.W.
(1998) Evidence for activation of caspase-3-like protease in
excitotoxins and hypoxia/hypoglycemia-injured cerebrocor-
tical neurons. J. Neurochem., 71: 186–195.
Nawashiro, H., Tasaki, K., Ruetzler, C.A. and Hallenbeck,
J.M. (1997) TNF-alpha pretreatment induces protective
effects against focal cerebral ischemia in mice. J. Cereb.
Blood Flow Metab., 17: 483–490.
O’Neill, L.A. and Kaltschmidt, C. (1997) NF-kappa B: a crucial
transcription factor for glial and neuronal cell function.
Trends Neurosci., 20: 252–258.
O’Neill, L.A.J. (1996) Pharmacological targets in the immune
response. Interleukin-l receptors and signal transduction.
Biochem. Soc. Trans., 24: 207–211.
O’Neill, L.A.J. (1997) Towards an understanding of the signal
transduction pathway for interleukin-1. Biochim. Biophys.
Acta, 1266: 31–44.
353
O’Neill, L.A.J. and Green, C. (1998) Signal transduction path-
ways activated by the IL-1 receptor family: ancient signalling
machinery in mammals, insects and plants. J. Leukoc. Biol.,
63: 650–657.
Oitzl, M.S., van Oers, H., Schobitz, B. and de Kloet, E.R.
(1993) Interleukin-1 beta, but not interleukin-6, impairs spa-
tial navigation learning. Brain Res., 613: 160–163.
Okamura, H., Tsutsui, H., Kashiwamura, S., Yoshimoto, T.
and Nakanishi, K. (1998) Interleukin-18: a novel cytokine
that augments both innate and acquired immunity. Adv.
Immunol., 70: 281–312.
Pang, L., Sawada, T., Deckers, S.J. and Saltiel, A.R. (1995)
Inhibition of MAP kinase kinase blocks the differentiation of
PC-12 cells induced by nerve growth factor. J. Biol. Chem.,
270: 13585–13588.
Parnet, P., Amindari, S., Wu, C., Brunke-Reese, D., Goujon,
E., Weyhenmeyer, J.A., Dantzer, R. and Kelley, K.W. (1994)
Expression of type I and type II interleukin-I receptors in
mouse brain. Mol. Brain Res., 27: 63–70.
Pfeffer, K., Matsuyama, T., Kunidg, T.M., Wakeham, A.,
Kishihara, K., Shahinian, A., Wiegmann, K., Ohashi, P.S.,
Krönke, M. and Mak, T.W. (1993) Mice deficient for the 55 kd
tumor necrosis factor receptor are resistant to endotoxic shock,
yet succumb to L. monocytogenes infection. Cell, 73: 457–467.
Plata-Salaman, C.R. and Ffrench-Mullen, J.M.H. (1992)
Interleukin-lb depresses calcium currents in CA 1 hippocam-
pal neurons at pathophysiological concentrations. Brain Res.
Bull., 29: 221–223.
Plata-Salaman, C.R. and Ffrench-Mullen, J.M.H. (1994) Inter-
leukin-lb inhibits Ca2, channel currents in hippocampal neu-
rons through protein kinase C. Eur. J. Pharmacol., 266: 1–10.
Plata-Salaman, C.R., Oomura, Y. and Kai, Y. (1988) Tumor
necrosis factor and interleukin-1 beta: suppression of food
intake by direct action in the central nervous system. Brain
Res., 448: 106–114.
Rothwell, N. (1999) Cytokines — killers in the brain. J.
Physiol., 514: 3–17.
Rothwell, N.J. and Hopkins, S.J. (1995) Cytokines and the
nervous system II: actions and mechanisms of action. Trends
Neurosci., 18: 130–136.
Schneider, H., Prossi, F., Balschun, D., Wagner, A., Del Rey,
A. and Besedovsky, H.O. (1998) A neuromodulatory role of
interleukin-1b in the hippocampus. Proc. Natl. Acad. Sci.
U.S.A., 95: 7778–7783.
Selmaj, K.W. and Raine, C.S. (1998) Tumor necrosis factor
mediates myelin and oligodendrocyte damage in vitro. Ann.
Neurol., 23: 339–346.
Sheridan, G., Pickering, M., Moynagh, P.N., O’Connor, J.J. and
Murphy, K.J. (2006) Tumour necrosis factor-a selectively up-
regulates nuclear factor kb expression in hippocampal sub-
neuronal populations. Proc. Br. Pharmacol. Soc., 3(4): 113P.
Silva, A.J., Smith, A.M. and Giese, K.P. (1997) Gene targeting
and the biology of learning and memory. Annu. Rev. Genet.,
31: 527–546.
Silverman, N. and Maniatis, T. (2001) NF-kappaB signaling
pathways in mammalian and insect innate immunity. Genes
Dev., 15: 2321–2342.
354
Sitcheran, R., Gupta, P., Fisher, P.B. and Baldwin, A.S. (2005)
Positive and negative regulation of EAAT2 by NF-kappaB: a
role for N-myc in TNFalpha-controlled repression. EMBO
J., 24(3): 510–520.
Stellwagen, D., Beattie, E.C., Seo, J.Y. and Malenka, R.C.
(2005) Differential regulation of AMPA receptor and
GABA receptor trafficking by tumor necrosis factor-alpha.
J. Neurosci., 25: 3219–3228.
Stellwagen, D. and Malenka, R.C. (2006) Synaptic scaling
mediated by glial TNF-alpha. Nature, 440: 1054–1059.
Tanabe, Y., Masu, M., Ishii, T., Shigemoto, R. and Nakanishi,
S. (1992) A family of metabotropic glutamate receptors.
Neuron, 8: 169–179.
Tanaka, K., Watase, K., Manabe, T., Yamada, K., Watanabe,
M., Takahashi, K., Iwama, H., Nishikawa, T., Ichihara, N.,
Kikuchi, T., Okuyama, S., Kawashima, N., Hori, S., Taki-
moto, M. and Wada, K. (1997) Epilepsy and exacerbation
of brain injury in mice lacking the glutamate transporter
GLT-1. Science, 276: 1699–1702.
Tancredi, V., D’Arcangelo, G., Grassi, F., Tarroni, P.,
Palmieri, G., Santoni, A. and Eusebi, F. (1992) Tumor ne-
crosis factor alters synaptic transmission in rat hippocampal
slices. Neurosci. Lett., 146: 176–178.
Tartaglia, L.A., Weber, R.F., Figari, I.S., Reynolds, C.,
Palladino Jr., M.A. and Goeddel, D.V. (1991) The two
different receptors for tumor necrosis factor mediate distinct
cellular responses. Proc. Natl. Acad. Sci. U.S.A., 88:
9292–9296.
Taylor, D.L., Jones, F., Kubota, E.S. and Pocock, J.M. (2005)
Stimulation of microglial metabotropic glutamate receptor
mGlu2 triggers tumor necrosis factor alpha-induced neuro-
toxicity in concert with microglial-derived Fas ligand. J.
Neurosci., 25(11): 2952–2964.
Thomassen, E., Bird, T.A., Renshaw, B.R., Kennedy, M.K.
and Sims, J.E. (1998) Binding of interleukin-18 to the
interleukin-1 receptor homologous receptor IL-1Rrp1
leads to activation of signaling pathways similar to those
used by interleukin-1. J. Interferon Cytokine Res., 18:
1077–1088.
Turrigiano, G.G. and Nelson, S.B. (2004) Homeostatic plastic-
ity in the developing nervous system. Nat. Rev. Neurosci., 5:
97–107.
Volterra, A. and Steinhauser, C. (2004) Glial modulation of
synaptic transmission in the hippocampus. Glia, 47(3):
249–257.
Waage, A., Halstensen, A., Shalaby, R., Brandtzaeg, P.,
Kierulf, P. and Espevik, T. (1989) Local production of
tumor necrosis factor alpha, interleukin 1, and interleukin 6
in meningococcal meningitis. Relation to the inflammatory
response. J. Exp. Med., 170: 1859–1867.
Wajant, H. and Scheurich, P. (2001) Tumor necrosis factor
receptor associated factor (TRAF) 2 and its role in TNF
signaling. Int. J. Biochem. Cell Biol., 33: 19–32.
Wang, K.K.W., Nath, R., Raser, K.J. and
Hajimohammadreza, I. (1996) Maitotoxin induces calpain
activation in SH-SY5Y neuroblastoma cells and cerebrocor-
tical cultures. Arch. Biochem. Biophys., 331: 208–214.
Wang, Q., Chang, L., Rowan, M.J. and Anwyl, R. (2006)
Developmental dependence, the role of the kinases p38
MAPK and PKC, and the involvement of tumor necrosis
factor-R1 in the induction of mGlu-5 LTD in the dentate
gyrus. Neuroscience, 144: 110–118.
Wang, X., Yue, T.L., Barone, F.C., White, R.F., Gagnon, R.C.
and Feuerstein, G.Z. (1994) Concomitant cortical expression
of TNF-alpha and IL-1 beta mRNAs follows early response
gene expression in transient focal ischemia. Mol. Chem.
Neuropathol., 23: 103–114.
Yang, L., Lindholm, K., Konishi, Y., Li, R. and Shen, Y.
(2002) Target depletion of distinct tumor necrosis factor
receptor subtype reveals hippocampal neuron death and
survival through different signal transduction pathways.
J. Neurosci., 22: 3025–3032.
Yu, A.C. and Lau, L.T. (2000) Expression of interleukin-1
alpha, tumor necrosis factor alpha and interleukin-6 genes in
astrocytes under ischemic injury. Neurochem. Int., 36:
369–377.
Yu, Z., Zhou, D., Bruce-Keller, A.J., Kindy, M.S. and
Mattson, M.P. (1999) Lack of the p50 subunit of nuclear
factor-kappaB increases the vulnerability of hippocampal
neurons to excitotoxic injury. J. Neurosci., 19: 8856–8865.
Zhao, X., Bausano, B., Pike, B.R., Newcomb-Fernandez, J.K.,
Wang, K.K., Shohami, E., Ringger, N.C., DeFord, S.M.,
Anderson, D.K. and Hayes, R.L. (2001) TNF-alpha stimu-
lates caspase-3 activation and apoptotic cell death in
primary septo-hippocampal cultures. J. Neurosci. Res., 64:
121–131.
Zou, J.Y. and Crews, F.T. (2005) TNF alpha potentiates gluta-
mate neurotoxicity by inhibiting glutamate uptake in organ-
otypic brain slice cultures: neuroprotection by NF kappa B
inhibition. Brain Res., 1034(1–2): 11–24.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 21

Role of corticosteroid hormones in the dentate gyrus

Marian Joëls

Swammerdam Institute of Life Sciences, Center for NeuroScience (SILS-CNS), University of Amsterdam,
Kruislaan 320, 1098 SM Amsterdam, The Netherlands

Abstract: Dentate granule cells are enriched with receptors for the stress hormone corticosterone, i.e., the
high-affinity mineralocorticoid receptor (MR), which is already extensively occupied with low levels of the
hormone, and the glucocorticoid receptor (GR), which is particularly activated after stress. More than any
other cell type in the brain studied so far, dentate granule cells require hormone levels to be within the
physiological range. In the absence of corticosteroids, proliferation and apoptotic cell death are dramat-
ically enhanced. Dendritic morphology and synaptic transmission are compromised. Conversely, prolonged
exposure of animals to a high level of corticosterone suppresses neurogenesis and presumably makes
dentate granule cells more vulnerable to delayed cell death. These corticosteroid effects on dentate cell and
network function are translated into behavioral consequences, in health and disease.

Keywords: chronic stress; adrenalectomy; mineralocorticoid receptor; glucocorticoid receptor;

neurogenesis; apoptosis; electrophysiology

Systems activated by stress latter causes an array of behavioral changes in-

When organisms are exposed to a stressful situa-
tion, sensory information about the situation is
processed in the brain, and output regarding the
event is forwarded to the hypothalamus (de Kloet
et al., 1998, 2005; see Fig. 1). From there two
hormonal systems are activated, aimed to face the
challenge and eventually normalize the disturbed
balance: the sympatho-adrenomedullar system,
which leads to the release of adrenaline from the
adrenal medulla; and the hypothalamo-pituitary-
adrenal system, which results in elevated levels of
the adrenocortical hormone (corticosterone in ro-
dents, cortisol in humans). Adrenaline and corti-
costerone are not only active in peripheral organs
but can also strongly affect brain function. The
Corresponding author. Tel.: +31-20-5257626;
Fax: +31-20-5257709; E-mail: joels@science.uva.nl
cluding increased vigilance, alertness, focused
attention, and selection of appropriate strategies,
which together result in behavioral adaptation.
Moreover, the hormones help to store relevant
information in brain, for future use.
The initial phases of behavioral adaptation pri-
marily involve fast-acting catecholamines, i.e.,
mainly noradrenaline, neuropeptides such as
corticotrophin releasing hormone and vasopressin
(de Kloet et al., 2005), and possibly neurosteroids
(Stell et al., 2003). Moreover, it has become clear
in recent years that corticosterone may also play a
role in these early phases, through non-genomic
pathways (Di et al., 2003; Karst et al., 2005). The
main role of corticosteroids, though, is to change
brain function through slow and long-lasting
actions. These long-lasting actions are mediated
by corticosteroid hormone receptors, of which
two main types have been recognized in the brain

DOI: 10.1016/S0079-6123(07)63021-0 355

356

Fig. 1. Exposure of a rat to stress may activate many brain regions (depending on the type of stressor), including the amygdala (Amy),
hippocampus (Hipp) and prefrontal cortex (PFC). These areas project to the hypothalamus (HYP). Stimulation of cells in the
hypothalamus leads to the activation of the fast-acting sympatho-adrenomedullar system (lower right) and the slower-acting hypo-
thalamo-pituitary-adrenal axis (lower left). Both systems not only affect the function of peripheral organs but also feed back to the
brain, via adrenaline and corticosterone respectively. Adrenaline can, via intermediate steps involving the nucleus tractus solitarius,
give rise to central release of noradrenaline (NA) from the locus coeruleus (LC), which then exerts widespread influence on other areas
such as the amygdala, prefrontal cortex and hippocampus. Corticosterone is distributed throughout the brain but acts only at those
sites where receptors are enriched. Two receptor types are known in the brain, i.e., the mineralocorticoid (MR) and the glucocorticoid
receptor (GR). Principal cells in the hippocampal CA1 area and the dentate gyrus (DG; see inset) are among the few cells in the brain
that express MR (gray dots) as well as GR (black dots). CA3 neurons primarily express MR. In most other parts of the brain, GRs
prevail. ANS ¼ autonomic nervous system; ACTH ¼ adrenocorticotropin hormone; CRH ¼ corticotropin releasing hormone.
Adapted with permission from Elsevier (Joëls et al., 2006).

(Reul and de Kloet, 1985). The first type, i.e., the
mineralocorticoid receptor (MR), is identical
to the protein found in peripheral organs like the
kidney where it is involved in mineral balance.
This receptor displays a very high affinity for
the mineralocorticoid aldosterone as well as for
corticosterone (Kd
0.5 nM). In kidney tissue, al-
dosterone preferentially binds to the MR, since
corticosterone is converted into a biologically
inactive 11-keto isoform, by the enzyme 11-b-
hydroxysteroid dehydrogenase type 2 (Seckl and
Walker, 2003). In most brain regions this enzyme
is not present during adulthood, so corticosterone
(which is present in a 100-fold excess over aldos-
terone) under normal conditions is the most likely
occupant of MRs in brain. MRs show a restricted

distribution within the brain, with high abundance
in some limbic areas, like the hippocampal subre-
gions CA1, CA3 and dentate gyrus, the lateral
septum and central amygdala, and in motor nuclei
of the brain stem (de Kloet et al., 1998). The sec-
ond receptor type in brain is the glucocorticoid
receptor (GR), which has an approximately 10-
fold lower affinity for corticosterone than the MR.
GRs have a relatively low affinity for aldosterone,
but effectively bind synthetic steroids like dexa-
methasone. This receptor is much more ubiquitous
than the MR and can be found in nearly all
brain regions — both in neurons and glial cells —
although some areas show a high enrichment,
such as the paraventricular nucleus of the hypo-
thalamus, the hippocampal CA1 region, and the
dentate gyrus. Recent studies have shown that
both the MR and GR gene give rise to several
splice variants, and the proteins are known to have
multiple isoforms due to alternative translation
initiation sites as well as abundant posttransla-
tional modifications caused by acetylation, phos-
phorylation, ubiquitination and sumoylation
(Pascual-Le Tallec and Lombes, 2005; Zhou and
Cidlowski, 2005). The functional relevance of
these variants in brain, though, is presently not
understood.
As is evident from the above overview, some
cells in brain express both MR and GR. This is for
instance the case for granule cells in the dentate
gyrus. The degree of activation of the two receptor
types, however, depends on the available levels of
the hormone. When the organism is at rest and
circadian release of corticosteroids is at its nadir,
MRs are already activated to a substantial degree
by corticosterone, but occupation of GRs by corti-
costerone is below 20% (de Kloet et al., 1998).
After exposure to stress or at the circadian
peak, GRs will become increasingly occupied by
corticosterone. Under physiological conditions,
corticosteroid receptor occupancy in dentate
granule cells will therefore range from a situation
of predominant MR activation to a situation
where both receptor types are fully occupied.
Electrophysiological studies over the past decades
have investigated both conditions. In addition,
studies have examined what happens to dentate
cells in the complete absence of corticosteroid
357
hormones, or when animals have been exposed to
excess corticosteroid levels (e.g., due to chronic
stress). Each of these conditions will be addressed
in the following sections.
Dentate function in the absence of corticosteroid
hormones
Cell turnover and morphology
The dentate gyrus is quite unique in that it requires
corticosteroid receptor activation to prevent apop-
tosis and to preserve neuronal integrity. Nearly
two decades ago it was reported for the first time
that complete removal of corticosterone by adre-
nalectomy (ADX) can lead to substantial loss of
dentate granule cells (Sloviter et al., 1989; Gould
et al., 1990), due to a process showing all the hall-
marks of apoptosis (Sloviter et al., 1993a). The
degree of apoptosis varies greatly among individ-
ual animals, but rarely involves more than 25% of
the neurons in the first week after ADX. Interest-
ingly, about one out of five animals do not show
any apoptosis upon ADX (Sloviter et al., 1993b;
Stienstra et al., 1998). It is still not fully under-
stood why these animals escape from apoptosis,
but most likely these animals still have some re-
sidual corticosterone producing cells (despite the
fact that circulating corticosterone levels are below
the detection limit). Approximately 3 days are re-
quired for cells to fully undergo the apoptosis pro-
gram upon removal of corticosterone (Hu et al.,
1997). Substitution with a low dose of corticoster-
one, which preferentially occupies the MR, or with
aldosterone can fully prevent apoptosis (Woolley
et al., 1991; Stienstra et al., 1998), pointing to a
crucial role of the MR in preventing apoptosis.
Why some cells undergo apoptosis but neigh-
boring cells that are seemingly comparable stay
healthy is not easy to understand. Presumably,
relatively ‘young’ granule cells are more resistant
to apoptosis (Cameron and Gould, 1996). We
tested the hypothesis that surviving cells differ in
their gene expression profile triggered by ADX,
resulting in resistance against apoptosis (Nair
et al., 2004; Fig. 2). Multiple genes expressed in
surviving individual granule cells were assessed by
358

Fig. 2. (A) Slices from animals that were adrenalectomized (ADX) 1, 2 or 3 days previously, or sham-operated rats, were stained in
vitro with Hoechst 33258, a stain for nuclear chromatin. Apoptotic granule cells with condensed nuclear chromatin are brightly colored
(arrows) and could not be recorded with whole cell patch clamp electrodes. Neighboring cells, which were not apoptotic, were recorded
(cross), and RNA was subsequently isolated from the cell. This procedure allows the collection of RNA from cells that have not yet
entered the apoptotic pathways. (B) After RNA collection, multiple gene expression in surviving individual granule cells was assessed
by linear antisense RNA amplification and hybridization to slot blots containing various neuronal cDNAs. (C) Compared to cells from
sham-operated control rats (average of all data from the 3 days, normalized to 10%, black squares), the ratio between Bcl-2 and Bax
RNA expression was significantly enhanced in 2- and 3-day ADX cells (open squares), which at that time had still resisted entering the
apoptotic pathway. A similar enhanced expression ratio was seen 3 days after ADX for calbindin relative to calcium-calmodulin kinase
II (CamKII). (D) To confirm that Bcl-2 and calbindin conferred resistance to entering the apoptotic route, these two genes were
overexpressed unilaterally in the dentate gyrus shortly before ADX; a control virus was injected into the contralateral hemisphere. A
significant reduction in the number of apoptotic cells 3 days after ADX was seen in the ipsilateral versus contralateral (open bars)
dentate gyrus of animals receiving the Bcl-2-expressing virus (gray bars), but not the calbindin-expressing virus (black bar). Adapted
with permission from Blackwell (Nair et al., 2004).

linear antisense RNA amplification and hybridi-
zation to slot blots containing various neuronal
cDNAs. Hierarchical clustering and principal
component analysis was performed on two phys-
iological variables and 14 mRNA ratios from cells
collected 1, 2 or 3 days after ADX or sham op-
eration. The results indicated that surviving 3-day
ADX granule cells display lower membrane
capacitance, lower relative N-methyl-D-aspartate
(NMDA) R1 mRNA expression and higher
relative MR, alpha1A voltage-gated calcium chan-
nel, Bcl-2 and NMDA R2C mRNA expression; it
is important to realize that overall dentate expres-
sion of these products may show changes in the
opposite direction (e.g., Cardenas et al., 2002),
most likely caused by the cells destined to die.
Some 1- and 2-day ADX cells clustered with these
3-day survivors; therefore, one or more compo-
nents of their mRNA expression profile may rep-
resent predictive markers for apoptosis resistance.

The functional relevance of two candidate genes
was tested by in vivo local over-expression in the
same model system; of these, Bcl-2 conferred par-
tial protection when induced shortly before ADX.
Collectively, the data indicate that cells with a low
membrane capacitance and high input resistance
— i.e., presumably ‘younger’ dentate granule cells
(Liu et al., 1996) — can trigger a gene expression
profile upon ADX which prevents initiation of the
apoptotic program, whereas ‘older’ cells may no
longer be able to do this.
Probably not all degradation in the dentate fol-
lowing ADX is due to apoptosis. Interestingly, it
was shown that loss of markers of mature neurons
can occur, rather than cell death, and this can be
reversed by a serotonin-1A receptor agonist
(Huang et al., 1997). This could reflect that the
latter restored antigenicity; more likely, though, it
could indicate that ADX leads to neurite retrac-
tion, which is restored by a serotonin-1A receptor
agonist. In agreement with the latter, 3-D recon-
struction of surviving granule cells after ADX
revealed that the dendritic length of higher order
segments is significantly reduced in animals with
undetectable corticosterone levels (Wossink et al.,
2001). Yet, in ADX animals with residual circu-
lating corticosterone no such reduced dendritic
length was observed, indicating that minute
amounts of corticosterone suffice to maintain den-
dritic integrity of granule cells. Growth factors,
which are sensitive to the presence of corticosteroids
in the dentate gyrus, may play a role in maintaining
the dendritic structure (Chao and McEwen, 1994).
ADX not only promotes apoptosis and neuro-
degeneration but also neurogenesis (Cameron and
Gould, 1994). Already within a day of corticos-
terone removal the degree of cell proliferation in
the dentate gyrus was enhanced (Cameron et al.,
1995). This process may depend on intact ent-
orhinal input and NMDA receptor-dependent
transmission (Cameron et al., 1995). Data support
that facilitation of neurogenesis after ADX not
only requires MR activation but also GR activa-
tion (Wong and Herbert, 2005). The importance
of corticosterone and the MR for neurogenesis
and dentate granule cell number also follows
from observations in mutant mice. Thus, in a
pro-opiomelanocortin null mutant mouse (where
359
activation of the adrenal cortex by adreno-
corticotropin is absent) granule cell density was
decreased, due to diminished cell proliferation
(Ostwald et al., 2006). Mice with a genetic disrup-
tion of the MR also exhibited decreased granule
cell density and a significant reduction in neuro-
genesis (Gass et al., 2000).
Physiology
If removal of corticosterone leads to loss of gran-
ule cells within 3 days, this would be expected to
impair functional network properties in the dent-
ate, particularly so if the loss involves mature
neurons that are fully integrated into the network.
The fact that proliferation is enhanced shortly
after ADX is less relevant at early time points,
because at that time the newborn cells are not
yet incorporated into the network. Later on, of
course, this burst in neurogenesis may have func-
tional consequences too.
Initial field potential recordings indeed corrob-
orated the notion that apoptosis impairs synaptic
transmission in the dentate gyrus. Thus, stimula-
tion of the perforant path in vitro leads to a field
excitatory postsynaptic potential (fEPSP) in the
molecular layer of the dentate gyrus. Three days
after ADX, both the slope and the amplitude of
this fEPSP were reduced by approximately 30%
(Stienstra et al., 1998). No reduction in fEPSP was
observed in those animals that did not display
apoptosis. Substitution of animals with a low dose
of corticosterone, sufficient to activate the brain
MRs and prevent apoptosis, fully normalized the
fEPSPS properties.
Yet, three observations indicate that the reduced
field response is not a consequence of apoptotic
cell loss. First, the reduced fEPSP slope and am-
plitude were already present 1 day after ADX, i.e.,
several days before the first apoptotic cells can be
discerned (Stienstra et al., 1998). Second, admin-
istration of corticosterone in vitro to slices from
3-day ADX rats fully restored the amplitude and
slope of the fEPSP, despite the presence of apo-
ptotic cells in the slices (Stienstra and Joëls, 2000).
Finally, intracellular recording of dentate granule
cells — which due to the method is confined to
360

non-apoptotic cells with an intact plasma mem-
brane — also revealed a 40% reduction of the
EPSP amplitude (Joëls et al., 2001; Fig. 3A). This
was observed both for the NMDA- and AMPA
receptor-mediated components of the synaptic
response. The latter suggests that postsynaptic
glutamate receptors are unlikely to be directly
changed after ADX, but rather that dendritic
segments carrying both types of receptors, or pre-
synaptic afferents, are lost.
Most passive (resting membrane potential, input
resistance) and active membrane properties (action
potential height and width) were on average not
significantly changed 3 days after ADX (Joëls
et al., 2001). As these properties depend on the age
of the cells (Liu et al., 1996), these observations
also indicate that the functional consequences of
apoptotic cell death of mature dentate cells are
probably limited. Rather, in parallel with (and not
as a consequence of) the change in cell turnover,
an extensive synaptic reorganization seems to
occur which largely alters synaptic properties of
the surviving cells.
In addition to these altered responses to low-
frequency stimulation of perforant path afferents,
it has also been found that long-term potentiation
(LTP) is reduced after ADX (Smriga et al., 1996;
Krugers et al., 2007), although one report found
no change (Pavlides et al., 1994). Again, the
reduced LTP occurs independent of altered cell
turnover, as LTP (but not cell turnover) can be
fully normalized by adding corticosterone to slices
from 3 day ADX rats (Krugers et al., 2007).
Ionic conductances of dentate granule cells
have also been studied after ADX. More specifi-
cally, the influx of calcium (Ca) through high

Fig. 3. (A) Typical excitatory postsynaptic potentials (EPSPs) induced in dentate granule cells by stimlation of the perforant path. The
EPSP amplitude and area under the signal (voltage  time) was reduced after ADX (A2) compared to sham operated controls (A1) or
ADX rats treated in vivo with a low dose of corticosterone sufficient to occupy the MR but not GR (A3). When the AMPA and
NMDA receptor-mediated components of the EPSPs were pharmacologically isolated, it became apparent that both were reduced by
40–50% 3 days after ADX (A4). Adapted with permission from the American Physiological Society (Joëls et al., 2001). (B) The
distribution of the dentate Ca-current amplitudes in sham-operated control animals can be fitted well with two Gaussian curves (dark
gray line), suggesting that under control conditions most cells have relatively small Ca-current amplitudes, but a subpopulation of
dentate cells displays a large Ca-current amplitude. At short intervals after ADX (1–2 days, light gray line), both populations can still
be discerned. The curve is shifted to the right, reflecting an ADX-dependent increase in Ca-current amplitude. Some days later (stippled
line; 3–7 days after ADX, i.e., at a time point when some of the cells have died by apoptosis), the peak of the amplitude distribution is
more or less at the same position, suggesting that in most neurons the Ca-current amplitude is not further enhanced. Importantly,
though, the curve became unimodal: the subpopulation with large Ca-current amplitudes were no longer present, suggesting that those
cells that had a large Ca-current amplitude to start with died by apoptosis. Adapted with permission from Blackwell (Karst and Joëls,
2001).

voltage-activated channels was investigated (Karst
and Joëls, 2001). Shortly after ADX (i.e., 1–2 days
later) Ca-current amplitude and density were in-
creased, whereas ADX rats treated with a low dose
of corticosterone (thus preferentially occupying
the MR) exhibited small Ca currents. Interestingly,
reduced Ca influx was observed at later times after
ADX. Further investigation of the amplitude dis-
tribution (Fig. 3B) revealed that while the current
amplitude distribution in general was shifted to
the right after ADX, a particular subpopulation of
cells with large Ca-current amplitudes had com-
pletely disappeared. It was reasoned that ADX
enhances Ca influx of all dentate granule cells,
similar to what had been observed earlier in CA1
neurons (Karst et al., 1994). In cells with a large
Ca influx — i.e., presumably the relatively ‘old’
cells (Thibault and Landfield, 1996) — the ADX-
induced increase in Ca influx may be just sufficient
to make them vulnerable to cell death. Combined
with a reduced potency to trigger genes that would
confer resistance to apoptotic cell death, this
enhanced Ca influx may cause the disappearance
of these ‘older’ cells 3–7 days after ADX.
In summary, low levels of corticosterone suffi-
cient to activate MRs in dentate granule cells seem
to be necessary to preserve neuronal integrity in
this area. In the absence of corticosteroids and
MR activation, apoptosis and neurogenesis are
increased. Even prior to apoptosis, synaptic reor-
ganization seems to occur, resulting in reduced
synaptic efficacy and plasticity. At this early stage
too, Ca influx is enhanced. Relatively ‘young’ gran-
ule cells may withstand these adverse conditions, as
they probably have a small Ca influx to start with,
and seem to be able to increase the expression of
survival genes such as Bcl-2. Yet, ‘older’ cells are
particularly vulnerable and may be destined to
degenerate and/or enter the apoptotic pathway.
Physiological variations in corticosteroid levels
Cell turnover
In contrast to what is seen after ADX (or chronic
stress, see below), neurogenesis and apoptosis in
the dentate gyrus are only mildly affected by
361
exposure to a single stressor. It was found that
proliferation is indeed suppressed shortly after
stress exposure (Heine et al., 2004b). Apoptosis is
increased, not only after stress (Heine et al.,
2004b), but also after a single injection with a
high dose of dexamethasone (Hassan et al., 1996),
involving shifts in the expression of pro- and anti-
apoptotic genes (Almeida et al., 2000). The effects
of acute stress, though, are largely normalized
within 24 h (Heine et al., 2004b), indicating that
the impact of a single stressor is probably limited.
Physiology
In cells which carry both MRs and GRs, such as
pyramidal neurons in the CA1 region or granule
cells in the dentate gryus, physiological fluctua-
tions in corticosterone levels will result in a recep-
tor occupation ranging from predominant MR
occupation when the organism is at rest and at the
circadian nadir, to concomitant MR and GR
occupation after stress or at the circadian peak. In
CA1 neurons, these two conditions are associated
with distinctly different physiological responses
(Joëls, 1997). Thus, compared to the situation
where corticosteroid receptors are unoccupied,
conditions of predominant MR occupation are
associated with restricted Ca influx in the soma
and primary dendrites, limited cell-firing frequency
accommodation, stable glutamate/GABAergic
transmission, small hyperpolarizing responses to
serotonin (5-HT) and efficient LTP. All of these
properties help to promote ongoing transfer of in-
formation through this area. Concomitant activa-
tion of GRs leads in a slow gene-mediated fashion
to enhanced Ca influx, prominent cell-firing
frequency accommodation, large hyperpolarizing
responses to serotonin (5-HT), reduction of exci-
tatory responses to noradrenaline, and reduction
of LTP; there seems to be an activity (energy)-
dependent attenuation of glutamate and gamma
amino butyric acid (GABA) mediated responses,
although postsynaptic responses via AMPA
receptors are enhanced within a time-window of
2–4 h after GR activation (Karst and Joëls, 2005).
All in all, GR activation generally may result in
normalization of activity that was temporary
362
raised by transmitters that reach the CA1 region at
the time of the stress exposure. As is also evident
from the above discussion, steroid-dependency
for many cell properties in the CA1 area follows
a U-shaped curve (Joëls, 2006). For instance, Ca
currents are large in the absence of corticosterone,
small with predominant MR activation, and
large again when both receptor types are activated
(Fig. 4).
Although the consequences of physiological
fluctuations in corticosteroid levels are far less in-
vestigated in the dentate gyrus than in the CA1
region, a U-shaped dose dependency may not exist
for all properties in the dentate (Joëls, 2006; see
Fig. 4). In the dentate, as opposed to the CA1
region (Kerr et al., 1992, 1994), no significant
difference in Ca-current amplitude was observed in
slices with predominant MR versus slices with
concomitant MR and GR activation (Van Gemert
and Joëls, 2006). It should be noted, though, that
these data were obtained in animals that had been
handled daily for 3 weeks, which in itself may
affect cellular function; still, preliminary data
in naı̈ve animals confirm this observation (Van
Gemert et al., unpublished observation). Similarly,
AMPA receptor-mediated synaptic responses were
comparable in slices (from handled rats) with pre-
dominant MR activation and slices where both
receptor types were activated (Karst and Joëls,
2003). This was also seen for the slope of the fE-
PSP (Stienstra and Joëls, 2000). Hyperpolarizing
responses to 5-HT do show some differences be-
tween conditions of low and high corticosterone
concentrations (Karten et al., 2001), but these are
not as distinct as in the CA1 area. A gene expres-
sion survey of single dentate cells from handled
rats — involving, e.g., transcripts for the subunits
of glutamate and GABAA receptors and of Ca
channels — also revealed very little differences due
to activation of GRs (Qin et al., 2004). Appar-
ently, a rise in corticosterone level does often not
lead to efficient activation of GR or GR-depend-
ent pathways in the dentate gyrus. This could be
due to local differences in GR variants, in post-
translational modification of the GR or in the
presence of proteins that interact with GR.
The steroid sensitivity of LTP in the dentate
gyrus seems to be somewhat more complex to
interpret. Initial experiments reported that selec-
tive activation of MRs in ADX rats prolongs LTP
(Pavlides et al., 1994), whereas very high doses of
corticosterone reduce LTP in vivo (Pavlides et al.,
1993) and show a strong tendency toward a de-
crease in vitro (Alfarez et al., 2003). Later studies
showed that forced and uncontrollable swim stress
prolongs LTP (Korz and Frey, 2003; Kavushansky
et al., 2006), via a rapid MR-dependent mecha-
nism (Korz and Frey, 2003) involving mitogen-
activated protein kinases (Ahmed et al., 2006) and
depending on intact input from the basolateral
amygdala (Korz and Frey, 2005). This suggests
that LTP in the dentate can be reduced by corti-
costerone (as in the CA1 area), but that stress
in the intact brain causes a different effect, due to
modulatory synaptic inputs and the effect of
stress-released factors other than corticosteroid
hormones (e.g., peptides or neurosteroids).
In conclusion, activation of MRs in dentate
granule cells is very important for the physiology
of the area. In general it serves the same role as in
the CA1 area, i.e., it promotes the transfer of sig-
nals through the dentate gyrus. To what extent
additional GR activation plays a role in the dent-
ate is at this moment less clear. The available data
seems to indicate that the strong effects of GR
activation (on top of MR activation) as seen in the
CA1 area do not take place in the dentate gyrus.
Dentate function after chronic stress
Cell turnover
Cell proliferation and apoptosis have been exten-
sively studied in the dentate in relation to chronic
stress, a condition that is known to be a risk factor
in the etiology of psychiatric diseases like major
depression (McEwen, 2003; de Kloet et al., 2005).
This research has recently been boosted by the
observation that anti-depressants promote neuro-
genesis and that, in fact, this proliferative action is
necessary for these compounds to exert their
behavioral effects in an animal model (Santarelli
et al., 2003; see further Chapter 38 in this book).
In general, chronic stress — be it from repeated
restraining, from a psychosocial source, or from a
 363

100
90
80
70
60
50
response (% of maximum)

40
30 CA1
CA1
20

100
90
80
70

60
50
40
30 DG
DG
20

MR
MR

GR
GR

stress
Fig. 4. Dose-response relationships for cellular effects of corticosterone in the CA1 hippocampal area and the dentate gyrus (DG). All
responses are expressed as a percentage of the maximal response. The concentration of corticosterone is a rough estimate of the local
concentration, based on the solutions perfused on in vitro preparations or derived from the plasma concentration when fluctuations in
hormone levels were measured in vivo. In the CA1 area, both the amplitude of depolarization-induced Ca currents (open squares) and
the hyperpolarization caused by serotonin-1A receptor activation (filled circles) display a clear U-shaped dose dependency. The
descending limb is linked to activation of MRs, while the ascending limb is associated with gradual GR activation, as occurs after
stress; relative occupation of the two receptors with increasing corticosterone levels is depicted at the bottom. DG granule neurons
show a clear effect on the field potential (filled squares) and single cell response (closed triangles) caused by activation of glutamatergic
AMPA receptors; this effect is linked to MR activation. Although these cells also abundantly express GRs, high doses of corti-
costerone do not lead to additional changes, except when animals had been exposed to 3 weeks of unpredictable stress prior to
recording (open triangles). Adapted with permission from Elsevier (Joëls, 2006).
364
combination of physical and psychological stress-
ors — has been reported to reduce cell prolifera-
tion in the dentate gyrus, at least in males (Czeh
et al., 2001; Pham et al., 2003; Westenbroek et al.,
2004; Heine et al., 2004b). This seems to be largely
due to the elevation in corticosterone level asso-
ciated with chronic stress, as repeated exogenous
administration of corticosterone (which enhances
corticosteroid levels but suppresses circulating lev-
els of CRH and ACTH) also decreased the number
of adult newborn cells, due to a reduction in cell
survival (Wong and Herbert, 2004). Importantly,
administration of a GR antagonist can fully nor-
malize the reduction in cell proliferation and/or
survival after repeated exogenous corticosterone
administration (Mayer et al., 2006). It is presently
still unclear, however, whether exogenous corti-
costerone and chronic stress indeed have the exact
same effects on all aspects of neurogenesis, i.e.,
proliferation, migration, differentiation and sur-
vival of newborn cells. Whether or not chronic
stress results in a reduction of the dentate volume
may depend on the duration or severity of the
stress protocol (Pham et al., 2003; Heine et al.,
2004b).
Surprisingly, apoptosis was reduced after
chronic stress when analyzed over the entire dent-
ate gyrus and hilus (Heine et al., 2004b). This may
signify that chronic stress slows down the cell
cycle, causing fewer neurons to be born, but also
fewer to die. This is corroborated by the observa-
tion that chronic stress increases the level of
p27Kip, an inhibitor of the cell cycle (Heine
et al., 2004a). Both the effect on neurogenesis and
on apoptosis observed after 3 weeks of unpredict-
able stress or restraint stress were largely reversed
when animals were allowed to recover from stress
for 3 weeks (Pham et al., 2003; Heine et al., 2004b).
As progenitor cells do not express GRs (Garcia
et al., 2004), corticosteroid-dependent effects on
their proliferation requires other cells in the neigh-
borhood which do contain GRs and can produce
factors that indirectly affect the function of pro-
genitor cells. Several potential intermediate factors
have been investigated, but the emerging picture
is somewhat confusing. It was found that, while
acute stressors lead to a transient increase in brain-
derived neurotrophic factor (BDNF) mRNA levels
in the dentate gyrus (Marmigere et al., 2003), re-
peated restraint stress reduces BDNF mRNA lev-
els (Smith et al., 1995). In contrast, NT-3 mRNA
levels were increased. Stress did not affect the ex-
pression of neurotrophin-4, or tyrosine receptor
kinases (trkB or C). More recently, though, it was
found that chronic stress increases the levels of the
catalytic but not of the truncated form of trkB
(Nibuya et al., 1999). As many newborn cells are
found in close association with the vasculature,
mediators in this system were also investigated.
A significant reduction in the level of vascular
endothelial growth factor and its receptor Flk-1
was observed after 3 weeks of unpredictable stress
(Heine et al., 2005). These levels were restored
when the period of stress was followed by 3 weeks
of rest. In addition to these growth factors, several
cell adhesion factors have been studied. Effects of
chronic over-exposure to corticosterone varied,
however, because it was found to enhance the
expression of neural cell adhesion molecule in the
dentate gyrus (Pham et al., 2003), decrease it
(Venero et al., 2002; Nacher et al., 2004) or leave it
unaffected (Van Gemert et al., 2006). Transcripts
for the adhesion molecule L1 were increased
(Venero et al., 2002). Clearly, the crucial factors
mediating the effect of chronic stress on progenitor
cell function still need to be identified.
Physiology
While temporary elevations in corticosterone level,
leading to brief activation of GRs, seem to be rel-
atively ineffective in the dentate gyrus of handled
or naı̈ve rats (as opposed to the CA1 area), such
elevations do become very effective when occur-
ring against a background of chronic stress. This
was found to be true for GR-evoked changes in
gene expression as well as functional endpoints.
For instance, high doses of corticosterone in
vitro do not change the AMPA receptor-mediated
responses in handled animals, but a large increase
in the amplitude of excitatory postsynaptic cur-
rents was observed with the same dose of corti-
costerone when slices were prepared from
chronically-stressed rats (Karst and Joëls, 2003).
A significant enhancement was also seen with

respect to the amplitude of the sustained Ca cur-
rent, when comparing corticosterone-treated cells
from control with those of chronically-stressed
rats (Van Gemert and Joëls, 2006). This effect
of chronic stress was also seen for the relative
expression of the subunit, which forms the pore
of the L-type (sustained) Ca channel (Qin et al.,
2004).
It has been found, both in vivo (Pavlides et al.,
2002) and in vitro (Alfarez et al., 2003), that in-
duction of LTP is very much impaired in the dent-
ate gyrus of chronically stressed rats. Additional
administration of a high dose of corticosterone is
ineffective (Alfarez et al., 2003), as LTP cannot be
reduced more than it is already in tissue exposed to
low concentrations of the hormone.
Taken together, it is evident that repeated
exposure of the dentate gyrus to high levels of co-
rticosterone reduces the number of newborn cells,
through mechanisms that are presently only partly
understood. Over a period of weeks this may not
lead to a large reduction in the overall number of
cells or volume of the adult dentate gyrus. Periods
of stress relief may even partially normalize the
changes in cell turnover. However, prolonged
ongoing stress to which the organism cannot
adapt may substantially change the composition
of dentate gyrus granule cells. This could result in
cognitive disturbances, and contribute to the pre-
cipitation of clinical symptoms. Interestingly, with
respect to physiological properties, dentate granule
cells seem to respond stronger to activation of
GRs (and thus presumably to acute stressors)
when the organism has experienced some weeks of
unpredictable stress. The ensuing increased excit-
ability in combination with enhanced Ca influx
could make cells more vulnerable to excitotoxic
damage.
Functional relevance of corticosteroid effects in the
dentate gyrus
Corticosteroid dose dependency in the dentate gyrus
While many brain regions are affected by stress-
related hormones, the dentate gyrus seems to be
exquisitely sensitive to concentrations of such
365
hormones that go beyond the borders of a re-
stricted, physiological range. When corticosteroid
hormone concentrations drop below a critical
level, or when they are repeatedly elevated, neu-
ronal integrity in the dentate gyrus is at stake. In
the CA1 area of the hippocampus, conditions that
lead to predominant activation of the MR have
been proposed to facilitate ongoing activity as well
as promote neuronal viability. In the dentate
gyrus, however, activation of MRs is an absolute
necessity to preserve neurites and to restrain ex-
cessive cell turnover. On the other end of the spec-
trum, repeated stress seems to slow the cycle of
proliferating cells and put granule cells at risk,
by increasing their response to glutamate and
enhancing voltage-dependent Ca influx. Clearly,
much of the unique sensitivity of the dentate gyrus
to fluctuating corticosteroid levels is caused by the
presence of progenitor cells in this area, which do
not carry corticosteroid receptors themselves but
nevertheless are very much influenced in their ac-
tivity through neighboring cells that do express
MR and GR.
In CA1 pyramidal cells, which like dentate
granule cells are among the few types of neurons
that co-express MR and GR, physiological shifts
in corticosteroid level (and hence in receptor oc-
cupation ratios) lead to distinct changes in func-
tional properties. As discussed above, a U-shaped
dose dependency has been observed for cortico-
steroid actions in this region. By contrast, dentate
granule cells seem to react in a more linear fashion,
with a relatively limited response to temporary
activation of GRs. For instance, (1) absence of
corticosterone leads to increased neurogenesis and
(2) exposure to an acute stressor leads to a tem-
porary reduction in neurogenesis; (3) repetitive
stress exposure results in a more prominent and
sustained suppression of neurogenesis (Fig. 5).
Another example pertains to glutamate receptor-
mediated transmission: (1) in the absence of corti-
costerone, responses to glutamate are small; (2)
activation of MRs increases the responses; (3)
while additional GR activation does not seem to
further enhance the response in control rats, such
activation does lead to a marked enhancement of
AMPA receptor-mediated responses in animals
with a history of chronic stress exposure.
366

A C handled

6000 * Acute stress

Mol. layer 24 hrs after

# BrdUpositive cells
5000 acute stress

4000
SGZ
h
hilus 3000

2000

GCL 1000

B D
9 * SHAM 3500
* handled
# H3 Thymidine-labeled cells

8
3000 24 hrs after

# BrdUpositive cells
7 ADX
chronic
2500 stress
6

5 2000

4
1500
3
1000
2
1 500

0 0

Fig. 5. (A) Immunohistochemical image of the dentate gyrus, showing proliferating cells mostly in the subgranular zone (SGZ) and far
less in the granular cell layer (GCL), molecular layer and hilus. Proliferating cells were marked with Ki-67. (B) One week after ADX or
sham operation, animals received an injection of 3H thymidine to label dividing cells. Three weeks later, dentate cells with a neuronal
phenotype were counted and expressed in number per 106 mm2. The number of newborn neurons was significantly enhanced after
ADX. Adapted with permission from Elsevier (Cameron and Gould, 1994). (C) Injection of BrdU just before exposure to acute stress
revealed a significant reduction in the number of newborn cells in the dentate gyrus (excluding the hilus). When BrdU was injected one
day after acute stress exposure, the difference in the number of newborn cells no longer was evident. Apparently, acute stress only
results in a transient suppression of proliferation in the dentate gyrus. Adapted from Heine et al., 2004b. (D). If BrdU was injected one
day after chronic stress exposure, the number of newborn cells was significantly reduced, indicating that chronic stress induces a much
longer-lasting suppression of cell proliferation. Adapted from Heine et al., 2004b and unpublished observations.

Importantly, the altered glutamate responses do not
depend on the changes in neurogenesis, but never-
theless seem to have the same steroid dependency.
Why two cell types that apparently share the same
steroid receptor expression pattern — i.e., CA1
pyramidal neurons and dentate granule cells — dis-
play such a different steroid dose dependency is not
understood, but most likely local factors like the
intracellular protein composition and network prop-
erties play an important role.
Functional relevance in health and disease
The effects of corticosteroid hormones on physi-
ological properties of hippocampal neurons will
have consequences for behavioral functions in
which these neurons participate. Indeed, numerous
studies have described how stress and/or selective
manipulation of MR or GR occupancy alter hip-
pocampal-dependent behavior. With respect to
tests requiring spatial memory, it was found that
MR occupation is important for reaction to novel
information as well as determination of behavioral
strategy (Oitzl and de Kloet, 1992; Oitzl et al.,
1994). Additional GR activation within the learn-
ing context is required for consolidation of spatial
information (Oitzl and de Kloet, 1992; Oitzl et al.,
2001). Importantly, GR activation that is not
related to the learning context may impair rather
than improve memory consolidation (De Kloet
et al., 1999). It has been proposed that convergence

in time and space (i.e., brain area) is necessary for
the facilitating effects of stress on memory to take
place (Joëls et al., 2006). It should also be realized
that due to stress, many transmitter and hormone
systems are activated, so that not only corticoster-
one levels but also those of, e.g., CRH and nor-
adrenaline will be elevated. This is of great relevance
since dentate gyrus function, and in particular LTP,
is known to be enhanced both by CRH (Wang et al.,
2000) and noradrenaline (e.g., Bramham et al.,
1997). Next to these short-lasting stressors, many
studies have addressed the effects of chronic stress
on memory function. It was found that chronic
stress exposure impairs spatial memory in males
(Luine et al., 1994; Wright and Conrad, 2005); yet,
memory improvement was observed in females
(Luine, 2002; McLaughlin et al., 2005). It is un-
clear, however, to what extent the spatial orientation
studies reflect functions of the dentate gyrus.
In this respect, it may be also of interest to con-
sider effects of stress on anxiety-related behavior,
which was shown to involve amygdala nuclei as well
as hippocampal subregions (LeDoux, 2000; McG-
augh, 2004). An extensive line of research has re-
vealed that inhibitory avoidance memory is
facilitated by stress (McGaugh and Roozendaal,
2002). Contextual fear conditioning is also pro-
moted by additional stress exposure (Cordero
et al., 2003; Donley et al., 2005). The facilitation
of avoidance behavior critically depends on GR ac-
tivation in the basolateral amygdala and requires
interactions between noradrenaline and corticoster-
one to occur within a restricted time-frame (McG-
augh and Roozendaal, 2002). Although intra-
hippocampal injections of a GR agonist could also
improve avoidance memory, these effects only
occurred when the basolateral amygdala was intact
(Roozendaal et al., 1999). A clear interaction be-
tween basolateral amygdala and the hippocampus,
particularly the dentate gyrus, was also demon-
strated with electrophysiology. Thus, perforant
path-induced LTP in the dentate was shown to be
facilitated by amygdala stimulation applied within
1 h before tetanic stimulation of the perforant path
(Akirav and Richter-Levin, 2002). With longer time
delays, amygdala stimulation reduced perforant
path LTP. It was shown that both effects involve
noradrenaline as well as corticosterone.
367
More distinct functional changes may occur in
the dentate gyrus when the corticosteroid levels are
at extreme levels for a considerable period of time.
As discussed above, this will occur when cortico-
steroid levels are extremely low, such as after ADX
in experimental animals. In humans though,
extreme adrenal insufficiency will rarely pass un-
noticed for any length of time, so the consequences
for brain function will be usually curtailed by re-
placement therapy. By contrast, prolonged periods
of exposure to uncontrollable and/or unpredicta-
ble stressors are not uncommon. Animal studies
indicate that renewed exposure to stress of an
organism that already has a history of repeated
stress experiences induces particularly strong
changes in gene expression and cell physiology of
dentate cells, such as enhanced response to exci-
tatory input and increased voltage-dependent Ca
influx. The consequences of these changes in phys-
iology will be particularly prominent when they
coincide with extensive depolarization of the dent-
ate area. This could occur when the dentate is
challenged by demanding behavioral events but
also in association with pathological conditions
such as ischemic insults (Krugers et al., 2000). This
fits with the general theory that (prolonged) expo-
sure to glucocorticoids may exacerbate damage
inflicted to hippocampal cells by pathological con-
ditions (Sapolsky, 1996; McEwen, 2004).
A better understanding of the functional conse-
quence of corticosteroid actions in the dentate,
however, awaits studies that more specifically ad-
dress this issue. It will be very informative to study
the influences of variations in corticosteroid level
using behavioral tests that more precisely reflect
dentate gyrus function as well as pathological con-
ditions that critically depend on the contribution
of this area.
References
Ahmed, T., Frey, J.U. and Korz, V. (2006) Long-term effects of
brief acute stress on cellular signaling and hippocampal LTP.
J. Neurosci., 26: 3951–3958.
Akirav, I. and Richter-Levin, G. (2002) Mechanisms of am-
ygdala modulation of hippocampal plasticity. J. Neurosci.,
22: 9912–9921.
Alfarez, D.N., Joëls, M. and Krugers, H.J. (2003) Chronic
unpredictable stress impairs long-term potentiation in rat
368
hippocampal CA1 area and dentate gyrus in vitro. Eur. J.
Neurosci., 17: 1928–1934.
Almeida, O.F., Conde, G.L., Crochemore, C., Demeneix, B.A.,
Fischer, D., Hassan, A.H., Meyer, M., Holsboer, F. and
Michaelidis, T.M. (2000) Subtle shifts in the ratio between
pro- and antiapoptotic molecules after activation of cortico-
steroid receptors decide neuronal fate. FASEB J., 14:
779–790.
Bramham, C.R., Bacher-Svendsen, K. and Sarvey, J.M. (1997)
LTP in the lateral perforant path is beta-adrenergic receptor-
dependent. Neuroreport, 8: 719–724.
Cameron, H.A. and Gould, E. (1994) Adult neurogenesis is
regulated by adrenal steroids in the dentate gyrus. Neurosci-
ence, 61: 203–209.
Cameron, H.A. and Gould, E. (1996) Distinct populations
of cells in the adult dentate gyrus undergo mitosis or
apoptosis in response to adrenalectomy. J. Comp. Neurol.,
369: 56–63.
Cameron, H.A., McEwen, B.S. and Gould, E. (1995) Regula-
tion of adult neurogenesis by excitatory input and NMDA
receptor activation in the dentate gyrus. J. Neurosci., 15:
4687–4692.
Cardenas, S.P., Parra, C., Bravo, J., Morales, P., Lara, H.E.,
Herrera-Marschitz, M. and Fiedler, J.L. (2002) Corticoster-
one differentially regulates bax, bcl-2 and bcl-x mRNA levels
in the rat hippocampus. Neurosci. Lett., 331: 9–12.
Chao, H.M. and McEwen, B.S. (1994) Glucocorticoids and the
expression of mRNAs for neurotrophins, their receptors and
GAP-43 in the rat hippocampus. Brain Res. Mol. Brain Res.,
26: 271–276.
Cordero, M.I., Venero, C., Kruyt, N.D. and Sandi, C. (2003)
Prior exposure to a single stress session facilitates subsequent
contextual fear conditioning in rats. Evidence for a role of
corticosterone. Horm. Behav., 44: 338–345.
Czeh, B., Michaelis, T., Watanabe, T., Frahm, J., de Biurrun,
G., van Kampen, M., Bartolomucci, A. and Fuchs, E. (2001)
Stress-induced changes in cerebral metabolites, hippocampal
volume, and cell proliferation are prevented by antidepres-
sant treatment with tianeptine. Proc. Natl. Acad. Sci. U.S.A.,
98: 12796–12801.
De Kloet, E.R., Joëls, M. and Holsboer, F. (2005) Stress and
the brain: from adaptation to disease. Nat. Rev. Neurosci., 6:
463–475.
De Kloet, E.R., Oitzl, M.S. and Joëls, M. (1999) Stress and
cognition: are corticosteroids good or bad guys? Trends
Neurosci., 22: 422–426.
De Kloet, E.R., Vreugdenhil, E., Oitzl, M.S. and Joëls, M.
(1998) Brain corticosteroid receptor balance in health and
disease. Endocr. Rev., 19: 269–301.
Di, S., Malcher-Lopes, R., Halmos, K.C. and Tasker, J.G.
(2003) Nongenomic glucocorticoid inhibition via end-
ocannabinoid release in the hypothalamus: a fast feedback
mechanism. J. Neurosci., 23: 4850–4857.
Donley, M.P., Schulkin, J. and Rosen, J.B. (2005) Glucocorti-
coid receptor antagonism in the basolateral amygdala and
ventral hippocampus interferes with long-term memory of
contextual fear. Behav. Brain Res., 164: 197–205.
Garcia, A., Steiner, B., Kronenberg, G., Bick-Sander, A. and
Kempermann, G. (2004) Age-dependent expression of
glucocorticoid — and mineralocorticoid receptors on neural
precursor cell populations in the adult murine hippocampus.
Aging Cell, 3: 363–371.
Gass, P., Kretz, O., Wolfer, D.P., Berger, S., Tronche, F.,
Reichardt, H.M., Kellendonk, C., Lipp, H.P., Schmid, W.
and Schütz, G. (2000) Genetic disruption of mineral-
ocorticoid receptor leads to impaired neurogenesis and
granule cell degeneration in the hippocampus of adult mice.
EMBO Rep., 1: 447–451.
Gould, E., Woolley, C.S. and McEwen, B.S. (1990) Short-term
glucocorticoid manipulations affect neuronal morphology
and survival in the adult dentate gyrus. Neuroscience, 37:
367–375.
Hassan, A.H., von Rosenstiel, P., Patchev, V.K., Holsboer, F.
and Almeida, O.F. (1996) Exacerbation of apoptosis in
the dentate gyrus of the aged rat by dexamethasone and
the protective role of corticosterone. Exp. Neurol., 140:
43–52.
Heine, V.M., Maslam, S., Joëls, M. and Lucassen, P.J. (2004a)
Increased P27KIP1 protein expression in the dentate gyrus
of chronically stressed rats indicates G1 arrest involvement.
Neuroscience, 129: 593–601.
Heine, V.M., Maslam, S., Zareno, J., Joëls, M. and Lucassen,
P.J. (2004b) Suppressed proliferation and apoptotic changes
in the rat dentate gyrus after acute and chronic stress are
reversible. Eur. J. Neurosci., 19: 131–144.
Heine, V.M., Zareno, J., Maslam, S., Joëls, M. and Lucassen,
P.J. (2005) Chronic stress in the adult dentate gyrus re-
duces cell proliferation near the vasculature and VEGF
and Flk-1 protein expression. Eur. J. Neurosci., 21:
1304–1314.
Hu, Z., Yuri, K., Ozawa, H., Lu, H. and Kawata, M. (1997)
The in vivo time course for elimination of adrenalectomy-
induced apoptotic profiles from the granule cell layer of the
rat hippocampus. J. Neurosci., 17: 3981–3989.
Huang, J., Strafaci, J.A. and Azmitia, E.C. (1997) 5-HT1A
receptor agonist reverses adrenalectomy-induced loss of
granule neuronal morphology in the rat dentate gyrus.
Neurochem. Res., 22: 1329–1337.
Joëls, M. (1997) Steroid hormones and excitability in the mam-
malian brain. Front. Neuroendocrinol., 18: 2–48.
Joëls, M. (2006) Corticosteroid effects in the brain: U-shape it.
Trends Pharmacol. Sci., 27: 244–250.
Joëls, M., Pu, Z., Wiegert, O., Oitzl, M.S. and Krugers, H.J.
(2006) Learning under stress: how does it work? Trends
Cogn. Sci., 10: 152–158.
Joëls, M., Stienstra, C. and Karten, Y. (2001) Effect of
adrenalectomy on membrane properties and synaptic
potentials in rat dentate granule cells. J. Neurophysiol., 85:
699–707.
Karst, H., Berger, S., Turiault, M., Tronche, F., Schütz, G. and
Joëls, M. (2005) Mineralocorticoid receptors are indispensa-
ble for nongenomic modulation of hippocampal glutamate
transmission by corticosterone. Proc. Natl. Acad. Sci.
U.S.A., 102: 19204–19207.

Karst, H. and Joëls, M. (2001) Calcium currents in rat dentate
granule cells are altered after adrenalectomy. Eur. J. Neuro-
sci., 14: 503–512.
Karst, H. and Joëls, M. (2003) Effect of chronic stress on
synaptic currents in rat hippocampal dentate gyrus neurons.
J. Neurophysiol., 89: 625–633.
Karst, H. and Joëls, M. (2005) Corticosterone slowly
enhances miniature excitatory postsynaptic current ampli-
tude in mice CA1 hippocampal cells. J. Neurophysiol., 94:
3479–3486.
Karst, H., Wadman, W.J. and Joëls, M. (1994) Corticosteroid
receptor-dependent modulation of calcium currents in rat
hippocampal CA1 neurons. Brain Res., 649: 234–242.
Karten, Y.J., Stienstra, C.M. and Joëls, M. (2001) Corticoster-
oid effects on serotonin responses in granule cells of the rat
dentate gyrus. J. Neuroendocrinol., 13: 233–238.
Kavushansky, A., Vouimba, R.M., Cohen, H. and Richter-
Levin, G. (2006) Activity and plasticity in the CA1, the
dentate gyrus, and the amygdala following controllable vs.
uncontrollable water stress. Hippocampus, 16: 35–42.
Kerr, D.S., Campbell, L.W., Thibault, O. and Landfield, P.W.
(1992) Hippocampal glucocorticoid receptor activation
enhances voltage-dependent Ca2+ conductances: relevance
to brain aging. Proc. Natl. Acad. Sci. U.S.A., 89: 8527–8531.
Korz, V. and Frey, J.U. (2003) Stress-related modulation of
hippocampal long-term potentiation in rats: involvement of
adrenal steroid receptors. J. Neurosci., 23: 7281–7287.
Korz, V. and Frey, J.U. (2005) Bidirectional modulation of
hippocampal long-term potentiation under stress and no-
stress conditions in basolateral amygdala-lesioned and intact
rats. J. Neurosci., 25: 7393–7400.
Krugers, H.J., van der Linden, S., van Olst, E., Alfarez, D.N.,
Maslam, S., Lucassen, P.J. and Joëls, M. (2007) Dissociation
between apoptosis, neurogenesis and synaptic potentiation in
the dentate gyrus of adrenalectomized rats. Synapse, 61:
221–230.
Krugers, H.J., Maslam, S., Korf, J. and Joëls, M. (2000) The
corticosterone synthesis inhibitor metyrapone prevents
hypoxia/ischemia-induced loss of synaptic function in the
rat hippocampus. Stroke, 31: 1162–1172.
LeDoux, J.E. (2000) Emotion circuits in the brain. Annu. Rev.
Neurosci., 23: 155–184.
Liu, Y.B., Lio, P.A., Pasternak, J.F. and Trommer, B.L. (1996)
Developmental changes in membrane properties and postsy-
naptic currents of granule cells in rat dentate gyrus. J.
Neurophysiol., 76: 1074–1088.
Luine, V. (2002) Sex differences in chronic stress effects on
memory in rats. Stress, 5: 205–216.
Luine, V., Villegas, M., Martinez, C. and McEwen, B.S. (1994)
Repeated stress causes reversible impairments of spatial
memory performance. Brain Res., 639: 167–170.
Marmigere, F., Givalois, L., Rage, F., Arancibia, S. and Tapia-
Arancibia, L. (2003) Rapid induction of BDNF expression in
the hippocampus during immobilization stress challenge in
adult rats. Hippocampus, 13: 646–655.
Mayer, J.L., Klumpers, L., Maslam, S., de Kloet, E.R., Joëls,
M. and Lucassen, P.J. (2006) Brief treatment with the
369
glucocorticoid receptor antagonist mifepristone normalises
the corticosterone-induced reduction of adult hippocampal
neurogenesis. J. Neuroendocrinol., 18: 629–631.
McEwen, B.S. (2003) Mood disorders and allostatic load. Biol.
Psychiatry, 54: 200–207.
McEwen, B.S. (2004) Protection and damage from acute and
chronic stress: allostasis and allostatic overload and relevance
to the pathophysiology of psychiatric disorders. Ann. N.Y.
Acad. Sci., 1032: 1–7.
McGaugh, J.L. (2004) The amygdala modulates the consolida-
tion of memories of emotionally arousing experiences. Annu.
Rev. Neurosci., 27: 1–28.
McGaugh, J.L. and Roozendaal, B. (2002) Role of adrenal
stress hormones in forming lasting memories in the brain.
Curr. Opin. Neurobiol., 12: 205–210.
McLaughlin, K.J., Baran, S.E., Wright, R.L. and Conrad,
C.D. (2005) Chronic stress enhances spatial memory in
ovariectomized female rats despite CA3 dendritic retrac-
tion: possible involvement of CA1 neurons. Neuroscience,
135: 1045–1054.
Nacher, J., Gomez-Climent, M.A. and McEwen, B. (2004)
Chronic non-invasive glucocorticoid administration
decreases polysialylated neural cell adhesion molecule
expression in the adult rat dentate gyrus. Neurosci. Lett.,
370: 40–44.
Nair, S.M., Karst, H., Dumas, T., Phillips, R., Sapolsky, R.M.,
Rumpff-van Essen, L., Maslam, S., Lucassen, P.J. and Joëls,
M. (2004) Gene expression profiles associated with survival
of individual rat dentate cells after endogenous corticosteroid
deprivation. Eur. J. Neurosci., 20: 3233–3243.
Nibuya, M., Takahashi, M., Russell, D.S. and Duman, R.S.
(1999) Repeated stress increases catalytic TrkB mRNA in rat
hippocampus. Neurosci. Lett., 67: 81–84.
Oitzl, M.S., Fluttert, M. and de Kloet, E.R. (1994) The effect of
corticosterone on reactivity to spatial novelty is mediated by
central mineralocorticosteroid receptors. Eur. J. Neurosci., 6:
1072–1079.
Oitzl, M.S. and de Kloet, E.R. (1992) Selective corticosteroid
antagonists modulate specific aspects of spatial orientation
learning. Behav. Neurosci., 106: 62–71.
Oitzl, M.S., Reichardt, H.M., Joëls, M. and de Kloet, E.R.
(2001) Point mutation in the mouse glucocorticoid receptor
preventing DNA binding impairs spatial memory. Proc. Natl.
Acad. Sci. U.S.A., 98: 12790–12795.
Ostwald, D., Karpac, J. and Hochgeschwender, U. (2006)
Effects on hippocampus of lifelong absence of glucocorti-
coids in the pro-opiomelanocortin null mutant mouse
reveal complex relationship between glucocorticoids and
hippocampal structure and function. J. Mol. Neurosci., 28:
291–302.
Pascual-Le Tallec, L. and Lombes, M. (2005) The mineral-
ocorticoid receptor: a journey exploring its diversity and
specificity of action. Mol. Endocrinol., 19: 2211–2221.
Pavlides, C., Kimura, A., Magarinos, A.M. and McEwen,
B.S. (1994) Type I adrenal steroid receptors prolong hip-
pocampal long-term potentiation. Neuroreport, 5:
2673–2677.
370
Pavlides, C., Nivon, L.G. and McEwen, B.S. (2002) Effects
of chronic stress on hippocampal long-term potentiation.
Hippocampus, 12: 245–257.
Pavlides, C., Watanabe, Y. and McEwen, B.S. (1993) Effects
of glucocorticoids on hippocampal long-term potentiation.
Hippocampus, 3: 183–192.
Pham, K., Nachler, J., Hof, P.R. and McEwen, B.S. (2003)
Repeated restraint stress suppresses neurogenesis and induces
biphasic PSA-NCAM expression in the adult rat dentate
gyrus. Eur. J. Neurosci., 17: 879–886.
Qin, Y., Karst, H. and Joëls, M. (2004) Chronic unpredictable
stress alters gene expression in rat single dentate granule cells.
J. Neurochem., 89: 364–374.
Reul, J.M. and de Kloet, E.R. (1985) Two receptor systems for
corticosterone in rat brain: microdistribution and differential
occupation. Endocrinology, 117: 2505–2511.
Roozendaal, B., Nguyen, B.T., Power, A.E. and McGaugh,
J.L. (1999) Basolateral amygdala noradrenergic influence
enables enhancement of memory consolidation induced by
hippocampal glucocorticoid receptor activation. Proc. Natl.
Acad. Sci. U.S.A., 96: 11642–11647.
Santarelli, L., Saxe, M., Gross, C., Surget, A., Battaglia, F.,
Dulawa, S., Weisstaub, N., Lee, J., Duman, R., Arancio, O.,
Belzung, C. and Hen, R. (2003) Requirement of hippocampal
neurogenesis for the behavioral effects of antidepressants.
Science, 301: 805–809.
Sapolsky, R.M. (1996) Stress, glucocorticoids, and damage to the
nervous system: the current state of confusion. Stress, 1: 1–19.
Seckl, J.R. and Walker, B.R. (2003) Minireview: 11beta-
hydroxysteroid dehydrogenase type 1 — a tissue-specific
amplifier of glucocorticoid action. Endocrinology, 142:
1371–1376.
Sloviter, R.S., Dean, E. and Neubort, S. (1993a) Electron mi-
croscopic analysis of adrenalectomy-induced hippocampal
granule cell degeneration in the rat: apoptosis in the adult
central nervous system. J. Comp. Neurol., 330: 337–351.
Sloviter, R.S., Sollas, A.L., Dean, E. and Neurbort, S. (1993b)
Adrenalectomy-induced granule cell degeneration in the rat
hippocampal dentate gyrus: characterization of an in vivo
model of controlled neuronal death. J. Comp. Neurol., 330:
324–336.
Sloviter, R.S., Valiquette, G., Abrams, G.M., Ronk, E.C.,
Sollas, A.L., Paul, L.A. and Neubort, S. (1989) Selective loss
of hippocampal granule cells in the mature rat brain after
adrenalectomy. Science, 243: 535–538.
Smith, M.A., Makino, S., Kvetnansky, R. and Post, R.M.
(1995) Stress and glucocorticoids affect the expression of
brain-derived neurotrophic factor and neurotrophin-3
mRNAs in the hippocampus. J. Neurosci., 15: 1768–1777.
Smriga, M., Saito, H. and Nishiyama, N. (1996) Hippocampal
long- and short-term potentiation is modulated by adrenal-
ectomy and corticosterone. Neuroendocrinology, 64: 35–41.
Stell, B.M., Brickley, S.G., Tang, C.Y., Farrant, M. and Mody,
I. (2003) Neuroactive steroids reduce neuronal excitability by
selectively enhancing tonic inhibition mediated by delta
subunit-containing GABAA receptors. Proc. Natl. Acad.
Sci. U.S.A., 100: 14439–14444.
Stienstra, C.M., van der Graaf, F., Bosma, A., Karten, Y.J.,
Hesen, W. and Joëls, M. (1998) Synaptic transmission in the
rat dentate gyrus after adrenalectomy. Neuroscience, 85:
1061–1071.
Stienstra, C.M. and Joëls, M. (2000) Effect of corticosteroid
treatment in vitro on adrenalectomy-induced impairment
of synaptic transmission in the rat dentate gyrus. J.
Neuroendocrinol., 12: 199–205.
Thibault, O. and Landfield, P.W. (1996) Increase in single
L-type calcium channels in hippocampal neurons during
aging. Science, 272: 1017–1020.
Van Gemert, N.G. and Joëls, M. (2006) Effect of chronic stress
and RU38486 treatment on voltage-dependent calcium cur-
rents in rat hippocampal dentate gyrus. J. Neuroendocrinol.,
18: 732–741.
Van Gemert, N.G., van Riel, E., Meijer, O.C., Fehr, S.,
Schachner, M. and Joëls, M. (2006) No effect of prolonged
corticosterone over-exposure on NCAM, SGK1, and RGS4
mRNA expression in rat hippocampus. Brain Res., 1093:
161–166.
Venero, C., Tilling, T., Hermans-Borgmeyer, I., Schmidt, R.,
Schachner, M. and Sandi, C. (2002) Chronic stress induces
opposite changes in the mRNA expression of the cell
adhesion molecules NCAM and L1. Neuroscience, 115:
1211–1219.
Wang, H.L., Tsai, L.Y. and Lee, E.H. (2000) Corticotropin-
releasing factor produces a protein synthesis — dependent
long-lasting potentiation in dentate gyrus neurons. J.
Neurophysiol., 83: 343–349.
Westenbroek, C., den Boer, J.A., Veenhuis, M. and ter Horst,
G.J. (2004) Chronic stress and social housing differentially
affect neurogenesis in male and female rats. Brain Res. Bull.,
64: 303–308.
Wong, E.Y. and Herbert, J. (2004) The corticoid environment:
a determining factor for neural progenitors’ survival in the
adult hippocampus. Eur. J. Neurosci., 20: 2491–2498.
Wong, E.Y. and Herbert, J. (2005) Roles of mineralocorticoid
and glucocorticoid receptors in the regulation of progenitor
proliferation in the adult hippocampus. Eur. J. Neurosci., 22:
785–792.
Woolley, C.S., Gould, E., Sakai, R.R., Spencer, R.L. and
McEwen, B.S. (1991) Effects of aldosterone or RU28362
treatment on adrenalectomy-induced cell death in the dentate
gyrus of the adult rat. Brain Res., 554: 312–315.
Wossink, J., Karst, H., Mayboroda, O. and Joëls, M. (2001)
Morphological and functional properties of rat dentate gran-
ule cells after adrenalectomy. Neuroscience, 108: 263–272.
Wright, R.L. and Conrad, C.D. (2005) Chronic stress leaves
novelty-seeking behavior intact while impairing spatial rec-
ognition memory in the Y-maze. Stress, 8: 151–154.
Zhou, J. and Cidlowski, J.A. (2005) The human glucocorticoid
receptor: one gene, multiple proteins and diverse responses.
Steroids, 70: 407–417.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 22

Neurotrophins in the dentate gyrus

Devin K. Binder

Department of Neurological Surgery, University of California, Irvine, CA 92868, USA

Abstract: Since the discovery of nerve growth factor (NGF) in the 1950s and brain-derived neurotrophic
factor (BDNF) in the 1980s, a great deal of evidence has mounted for the roles of neurotrophins (NGF;
BDNF; neurotrophin-3, NT-3; and neurotrophin-4/5, NT-4/5) in development, physiology, and pathology.
BDNF in particular has important roles in neural development and cell survival, as well as appearing
essential to molecular mechanisms of synaptic plasticity and larger scale structural rearrangements of axons
and dendrites. Basic activity-related changes in the central nervous system (CNS) are thought to depend on
BDNF modulation of synaptic transmission. Pathologic levels of BDNF-dependent synaptic plasticity may
contribute to conditions such as epilepsy and chronic pain sensitization, whereas application of the trophic
properties of BDNF may lead to novel therapeutic options in neurodegenerative diseases and perhaps even
in neuropsychiatric disorders. In this chapter, I review neurotrophin structure, signal transduction mech-
anisms, localization and regulation within the nervous system, and various potential roles in disease.
Modulation of neurotrophin action holds significant potential for novel therapies for a variety of
neurological and psychiatric disorders.

Keywords: brain-derived neurotrophic factor; neurotrophin-3; neurotrophin-4/5; nerve growth factor;

epilepsy

Introduction to neurotrophins antiparallel b-strands and cysteine residues in

Neurotrophin structure
Each neurotrophin (including nerve growth factor,
NGF; brain-derived neurotrophic factor, BDNF;
neurotrophin-3, NT-3; neurotrophin-4/5, NT-4/5;
see below) consists of a noncovalently linked ho-
modimer and contains (1) a signal peptide following
the initiation codon; (2) a pro-region containing an
N-linked glycosylation site and a proteolytic cleav-
age site for furin-like pro-protein convertases, fol-
lowed by the mature sequence; (3) a distinctive
three-dimensional structure containing two pairs of
Corresponding author. Tel.: +1 (714) 456-6966;
Fax: +1 (714) 456-8212; E-mail: dbinder@uci.edu
a cystine knot motif. Initially produced as prone-
urotrophins, prohormone convertases such as furin
cleave the proneurotrophins (molecular weight,
MW
30 kDa) to the mature neurotrophin (MW

14 kDa) (Chao and Bothwell, 2002). Proneuro-
trophins have altered binding characteristics and
distinct biologic activity in comparison with mature
neurotrophins (Lee et al., 2001b). Mature ne-
urotrophins are noncovalently linked homodimers
with MW approximately 28 kDa. Dimerization
appears essential for neurotrophin (NT) receptor
activation. BDNF shares approximately 50% amino
acid identity with NGF, NT-3, and NT-4/5.
The BDNF gene (in humans mapped to chro-
mosome 11p) has four 50 exons (exons I–IV) that
are associated with distinct promoters, and one 30

DOI: 10.1016/S0079-6123(07)63022-2 371

372
exon (exon V) that encodes the mature BDNF
protein (Metsis et al., 1993; Timmusk et al.,
1993b). Eight distinct mRNAs are transcribed,
with transcripts containing exons I–III expressed
predominantly in brain and exon IV found in lung
and heart (Timmusk et al., 1993b).
Neurotrophin signal transduction
Each NT binds one or more of the tropomyosin-
related kinase (trk) receptors, members of the
family of receptor tyrosine kinases (RTKs)
(Patapoutian and Reichardt, 2001). Trk proteins
are transmembrane RTKs homologous to other
RTKs such as the epidermal growth factor (EGF)
receptor and insulin receptor family. Ligand-
induced receptor dimerization results in kinase ac-
tivation; subsequent receptor autophosphorylation
on multiple tyrosine residues creates specific binding
sites for intracellular target proteins, which bind to
the activated receptor via SH2 domains (Barbacid,
1994; Patapoutian and Reichardt, 2001). These in-
clude PLCg1 (phospholipase C), p85 (the noncata-
lytic subunit of PI-3 kinase), and Shc (SH2-
containing sequence); activation of these target pro-
teins can then lead to a variety of intracellular
signaling cascades such as the Ras-MAP (mitogen-
activated protein) kinase cascade and phosphorylat-
ion of cyclic AMP-response element binding protein
(CREB) (Patapoutian and Reichardt, 2001; Segal,
2003).
Binding specificity is conferred via the juxtamem-
brane Ig-like domain of the extracellular portion of
the receptor in the following pattern (Urfer et al.,
1995). TrkA binds NGF (with low-affinity binding
by NT-3 in some systems); trkB binds BDNF and
NT-4/5 with lower-affinity binding by NT-3; and
trkC binds NT-3 (Barbacid, 1994). Trk receptors
exist in both a full-length (trkB.FL) form as well as
truncated (trkB.T1, trkB.T2) forms lacking the kin-
ase domain (Eide et al., 1996; Fryer et al., 1997).
Although most functions attributed to BDNF are
associated with full-length trkB, several roles have
been suggested for truncated receptors, including
growth and development (Fryer et al., 1997;
Yacoubian and Lo, 2000; Luikart et al., 2003) and
negative modulation of trkB receptor expression
and function (Eide et al., 1996; Haapasalo et al.,
2001, 2002). Expression of truncated trk receptors
on astrocytes is upregulated following injury (Frisen
et al., 1993) and may modulate neuronal vulnera-
bility (Saarelainen et al., 2000a) and sequestration
of BDNF in astrocytes (Biffo et al., 1995; Roback
et al., 1995; Alderson et al., 2000). Recent studies
have shown that BDNF activates glial calcium
signaling by truncated trk receptors (Climent et al.,
2000; Rose et al., 2003).
In addition, all of the NTs bind to the p75
receptor, designated p75NTR. p75NTR, related to
proteins of the tumor necrosis factor (TNFR) su-
perfamily, has a glycosylated extracellular region
involved in ligand binding, a transmembrane re-
gion, and a short cytoplasmic sequence lacking
intrinsic catalytic activity (Chao and Hempstead,
1995; Dechant and Barde, 2002). NT binding to
p75NTR is linked to several intracellular signal
transduction pathways, including nuclear factor-
kB (NF-kB), Jun kinase, and sphingomyelin
hydrolysis (Dechant and Barde, 2002). p75NTR
signaling mediates biologic actions distinct from
those of the trk receptors, notably the initiation
of programed cell death (apoptosis) (Casaccia-
Bonnefil et al., 1996; Frade et al., 1996; Roux
et al., 1999; Dechant and Barde, 2002). It has also
been suggested that p75 may serve to determine
NT binding specificity (Esposito et al., 2001; Lee
et al., 2001a; Zaccaro et al., 2001).
Nerve growth factor (NGF)
NGF was discovered in the early 1950s by Rita
Levi-Montalcini and Viktor Hamburger due to its
trophic (survival and growth-promoting) effects on
sensory and sympathetic neurons (Levi-Montalcini
and Hamburger, 1951). In addition, NGF supports
the survival and neurotransmitter synthesis of
cholinergic neurons in the central nervous system
(CNS). In the brain, it is synthesized primarily in
cholinergic target tissues such as the cortex, hippo-
campal pyramidal layer, and striatum (Gall and
Isackson, 1989; Rylett and Williams, 1994). The
trkA NT receptor is expressed primarily on the
axons of NGF-dependent cholinergic neurons
(Sobreviela et al., 1994).
 373

Table 1. Seizure regulation of NGF expression

Reference Methods Results

NGF mRNA
Gall and Isackson (1989)
Gall and Lauterborn (1992)
Ernfors et al. (1991)
Bengzon et al. (1993)
Schmidt-Kastner et al. (1996)
Mudo et al. (1996)
Sato et al. (1996)
Morimoto et al. (1998)
NGF protein
Bengzon et al. (1992)
Hilar electrolytic lesion
Hilar electrolytic lesion
Rapid kindling (ventral
hippocampal stimulation)
Traditional kindling (ventral
hippocampal stimulation
CA1–CA2)
Pilocarpine
Pilocarpine
Traditional amygdala kindling
Traditional amygdala kindling
Rapid hippocampal kindling
(ventral hippocampal
stimulation) NGF ELISA
Increases in DG (4 h) and cortex (17 h)
Increase in DG, max at 6 h (10  ) and 24 h (6  )
(biphasic)
Increase max at 1 h (DG), 4 h (PC)
Increase — did not do time course but studied
relationship between development of kindling
and neurotrophin induction — found similar
NGF induction (approximately 2  at 2 h time
point) regardless of kindling stage
Increase in DG max at 3–6 h
Increase in DG max at 3 h (approximately 2.5  )
Increase in CA1, CA3, perirhinal cortex 1 h after
stage 5 kindled seizure
Increase in DG, max at 2 h (2  )
After 40 stimulations, NGF protein levels
(measured by 2-site ELISA) increased to 150% in
DG at 7 days, 260% in PC at 12 h, and 170% in
parietal cortex at 24 h

NGF gene regulation Basal levels of trkA mRNA in the hippocampus

NGF expression levels are regulated by activity. This
has been most clearly demonstrated following the
intense activity associated with seizures (Table 1).
Hilar electrolytic lesion-induced (Gall and Isackson,
1989; Gall and Lauterborn, 1992; Lauterborn et al.,
1994) or kindled seizures (Ernfors et al., 1991;
Bengzon et al., 1993; Sato et al., 1996; Morimoto
et al., 1998) induce a rapid and transient expression
of NGF mRNA in dentate gyrus granule cells as
well as piriform cortex. Similarly, pilocarpine-
induced status epilepticus increases NGF mRNA
expression in dentate gyrus, maximum at approxi-
mately 3 h (Mudo et al., 1996; Schmidt-Kastner
et al., 1996).
Basal levels of NGF protein in the hippocampus
are very low (Narisawa-Saito and Nawa, 1996).
Studies with NGF ELISAs indicate that NGF
protein levels do increase following kindling, es-
pecially in dentate gyrus, piriform cortex, and pa-
rietal cortex (Bengzon et al., 1992). In contrast to
the mRNA increases, NGF protein levels remain
elevated for at least 7 days (Bengzon et al., 1992).
are very low (Cellerino, 1996; Mudo et al., 1996),
and do not appear to increase following kindling
(Bengzon et al., 1993; Merlio et al., 1993) or pi-
locarpine status epilepticus (Mudo et al., 1996)
(Table 2).
Brain-derived neurotrophic factor (BDNF)
In 1982, BDNF, the second member of the ‘‘NT’’
family of neurotrophic factors, was shown to pro-
mote survival of a subpopulation of dorsal root
ganglion neurons, and subsequently purified from
pig brain (Barde et al., 1982). The amino acid
sequence of BDNF was found to have a strong
homology with NGF. Since then, other members
of the NT family such as NT-3 (Maisonpierre
et al., 1990b) and NT-4/5 (Hallbook et al., 1991; Ip
et al., 1992) have been described, each with a dis-
tinct profile of trophic effects on subpopulations of
neurons in the peripheral and CNS.
BDNF mRNA has a widespread distribution in
the CNS (Merlio et al., 1993; Conner et al., 1997),
374

Table 2. Seizure regulation of trk receptor expression

Reference Methods Results

TrkA mRNA
Cellerino (1996) Basal hippocampal Basal not detectable by 35S; can only
detect with 33P
Mudo et al. (1996) Pilocarpine Basal not detectable and pilocarpine
status did not increase levels
Bengzon et al. (1993) Hippocampal kindling No change
Merlio et al. (1993) Rapid hippocampal kindling No change
TrkB mRNA
Bengzon et al. (1993) Rapid hippocampal kindling Increased in DG 2 h (2  ) after focal or
(ventral hippocampal stimulation) generalized seizures
Merlio et al. (1993) Rapid hippocampal kindling Increased threefold in DG at 30 min after
(ventral hippocampal stimulation) 40 stimulations
Elmer et al. (1996a, b) Rapid hippocampal kindling Increased in DG max at 2 h
(ventral hippocampal stimulation)
Humpel et al. (1993) PTZ Increased twofold in DG at 3 h
Nibuya et al. (1995) ECS Increased fivefold (DG) at 2 h
Schmidt-Kastner et al. (1996) Pilocarpine Increase in DG, CA1–3 at 3–6 h
Mudo et al. (1996) Pilocarpine Increase in DG, amygdala, PC at 3 h
Hughes et al. (1998) Hippocampal afterdischarge Increase in DG at 4 h
TrkC mRNA
Bengzon et al. (1993) Rapid hippocampal kindling Increased in DG 2 h (1.5  ) after focal or
(ventral hippocampal stimulation) generalized seizures
Merlio et al. (1993) Rapid hippocampal kindling No change after 40 rapid K stimulations
(ventral hippocampal stimulation)
Mudo et al. (1995) ICV KA or ICV bicuculline Increase max at 3 h (bicuculline) or 12 h
(KA) confined to DG

including limbic forebrain, neocortex, and is abun-
dant in all principal neurons of the hippocampus
(Ernfors et al., 1990; Hofer et al., 1990; Wetmore
et al., 1990). Like BDNF mRNA, constitutive
BDNF protein expression is widespread (Conner
et al., 1997; Yan et al., 1997b), localized on neu-
ronal cell bodies, axons, and dendrites. The mossy-
fiber axons of hippocampal dentate granule cells
display especially intense BDNF immunoreactivity
(Conner et al., 1997). The principal receptor for
BDNF, trkB, is a receptor tyrosine kinase, which
is found in both catalytic and truncated forms in
the adult forebrain (Fryer et al., 1996; Drake et al.,
1999). TrkB mRNA and protein are found in hip-
pocampus (Merlio et al., 1992; Altar et al., 1994;
Yan et al., 1997a). Truncated trkB is also found in
the ependymal cells lining the ventricular cavities,
effectively limiting diffusion of intraventricularly
administered BDNF (Yan et al., 1994; Anderson
et al., 1995).
Localization, transport and release
Unlike the classical target-derived trophic factor
model in which NTs — such as NGF — are
retrogradely transported, there is now abundant
evidence that BDNF is also anterogradely trans-
ported in brain. First, BDNF protein is localized
to nerve terminals (Conner et al., 1997), and path-
way transection or axonal transport inhibition
abrogates this terminal expression (Altar et al.,
1997; Conner et al., 1997; Altar and DiStefano,
1998). Second, higher resolution studies have
shown that BDNF is associated with dense-core
vesicles (Fawcett et al., 1997; Altar and DiStefano,
1998), which are the primary site for neuropeptide
storage and release from nerve terminals. Third,
further functional studies have supported the
anterograde transport hypothesis (Fawcett et al.,
1998, 2000). Fourth, pro-BDNF is shuttled from
the trans-Golgi network into secretory granules,

where it is cleaved by prohormone convertase 1
(PC1) (Farhadi et al., 2000).
In addition, emerging evidence suggests that
both BDNF and trk receptors may undergo reg-
ulated intracellular transport. For example, sei-
zures lead to redistribution of BDNF mRNA from
hippocampal CA3 cell bodies to their apical dend-
rites (Bregola et al., 2000; Simonato et al., 2002).
Trk signaling is now thought to include retrograde
transport of intact NT-trk complexes to the neu-
ronal cell body (Miller and Kaplan, 2001; Ginty
and Segal, 2002).
Recent evidence indicates that NTs are released
acutely following neuronal depolarization (Gries-
beck et al., 1999; Mowla et al., 1999; Hartmann et
al., 2001; Egan et al., 2003; Goggi et al., 2003;
Brigadski et al., 2005). In fact, direct activity-
dependent pre- to postsynaptic transneuronal
transfer of BDNF has been demonstrated using
fluorescently labeled BDNF (Kohara et al., 2001).
The released form of BDNF is thought to be pro-
BDNF (Mowla et al., 2001), raising the possibility
of postsecretory proteolytic processing by mem-
brane-associated or extracellular proteases in the
modulation of BDNF action (Lee et al., 2001b).
BDNF gene regulation
A multitude of stimuli have been described that
alter BDNF gene expression in both physiologic
and pathologic states (Lindholm et al., 1994).
Physiologic stimuli are known to increase BDNF
mRNA content. For example, light stimulation
increases BDNF mRNA in visual cortex (Castrén
et al., 1992), osmotic stimulation increases BDNF
mRNA in the hypothalamus (Castrén et al., 1995;
Dias et al., 2003), and whisker stimulation in-
creases BDNF mRNA expression in somatosen-
sory barrel cortex (Rocamora et al., 1996).
Electrical stimuli that induce long-term potentiat-
ion (LTP) in the hippocampus, a cellular model of
learning and memory, increase BDNF and NGF
expression (Patterson et al., 1992; Castrén et al.,
1993; Bramham et al., 1996). Even physical exer-
cise has been shown to increase NGF and BDNF
expression in hippocampus (Neeper et al., 1995).
Interestingly, BDNF levels vary across the estrous
375
cycle, which correlate with its effects on neural
excitability (Scharfman et al., 2003).
Distinct BDNF 50 exons are differentially reg-
ulated by stimuli such as neural activity. For
example, exons I–III, but not exon IV, increase
after kainic acid-induced seizures (Timmusk et al.,
1993b) or other stimuli that increase activity
(Lauterborn et al., 1996; Tao et al., 2002). Pro-
tein synthesis is required for the effects of activity
on exon I and II, but not III and IV, raising the
possibility that the latter act as immediate early
genes (Lauterborn et al., 1996; Castrén et al.,
1998). The transcription factor CaRF (calcium
response factor) activates transcription of exon III
under the control of a calcium response element,
CaRE1 (Tao et al., 2002). CREB, which can be
stimulated by diverse stimuli ranging from activity
to chronic antidepressant treatment (Nibuya et al.,
1995, 1996; Shieh et al., 1998; Tao et al., 1998;
Shieh and Ghosh, 1999), also modulates exon III
transcription. Recent evidence also indicates that
neural activity triggers calcium-dependent phos-
phorylation and release of MeCP2 (methyl-CpG
binding protein 2) from BDNF promoter III to
derepress transcription (Chen et al., 2003).
Pathologic states are also associated with altera-
tion in BDNF gene expression. For example,
seizures dramatically upregulate BDNF mRNA
(Table 3). A wide variety of seizure paradigms
(kindling; kainic acid; pilocarpine; pentylenetetrazol,
PTZ; electroconvulsive shock, ECS) rapidly and
dramatically increase expression of BDNF mRNA
in dentate gyrus as well as in other areas of the hip-
pocampus and cortex (Ernfors et al., 1991; Isackson
et al., 1991; Gall and Lauterborn, 1992; Dugich-
Djordjevic et al., 1992a, b; Bengzon et al., 1993;
Humpel et al., 1993; Nibuya et al., 1995; Mudo
et al., 1996; Sato et al., 1996; Schmidt-Kastner et al.,
1996). This is associated with a transient upregula-
tion of BDNF protein (Nawa et al., 1995; Elmer
et al., 1996b; Hughes et al., 1998).
TrkB mRNA and protein in the dentate gyrus
are also upregulated following various seizure
protocols (Bengzon et al., 1993; Merlio et al.,
1993; Elmer et al., 1996a) (Table 2). TrkB mRNA
expression is increased in dentate granule cells
2–6 h after rapid electrical kindling, hippocampal
after discharge, PTZ kindling, ECS, or pilocarpine
376

Table 3. Seizure regulation of BDNF expression

Reference Methods Results

BDNF mRNA
Isackson et al. (1991)
Gall and Lauterborn (1992)
Ernfors et al. (1991)
Dugich-Djordjevic et al. (1992b)
Dugich-Djordjevic et al. (1992a)
Bengzon et al. (1993)
Humpel et al. (1993)
Nibuya et al. (1995)
Schmidt-Kastner et al. (1996)
Mudo et al. (1996)
Sato et al. (1996)
BDNF protein
Nawa et al. (1995)
Elmer et al. (1996a, b)
Hughes et al. (1998)
Hilar electrolytic lesion
Perforant path stimulation (1 AD)
Rapid kindling (ventral
hippocampal stimulation)
KA
KA
Traditional kindling (ventral
hippocampal stimulation CA1–2)
PTZ kindling (30 mg/kg i.p.
followed by convulsive dose
(50 mg/kg))
ECS
Pilocarpine
Pilocarpine
Traditional amygdala kindling
Hilar electrolytic lesion
BDNF ELISA
Rapid kindling (ventral
hippocampal stimulation)
BDNF ELISA
Hippocampal after discharge
Increase, onset o1.5 h, max at 6 h (12  )
Increase, onset 20 min, max at 4 h (12  )
Increase max at 30 min (DG/PC), 1 h (CA1)
Increase in all hippocampus (P21, P40), no
change despite seizures (P8)
Increase max at 30 min (10  in DG, 2–6  in
CA1, CA3, CA4); later in cortex
Increase — did not do time course but studied
relationship between development of kindling
and neurotrophin induction — found similar
BDNF induction (approximately 9  at 2 h time
point) regardless of kindling stage
Increase max at 3 h (DG, PC, amygdala) after
acute convulsive PTZ
Increase 30-fold 2 h after ECS (DG), fivefold
(PC)
Increase max at 3–6 h (DG, other hippocampal,
neocortex, PC, striatum, thalamus)
Increase max at 3–6 h (DG, amygdala, PC)
4–5-fold increase in DG 1 h after stage 5 kindled
seizure
Highest levels of basal BDNF in hippocampus
(followed by hypothalamus, neocortex,
cerebellum, thalamus and striatum)
Fourfold induction of BDNF protein levels in
hippocampus, maximum at 24 h after HL, down
by 1 week
Basal levels of BDNF highest in
DGCA3>CA1>PC
After 1 AD: increase to 150% in DG at 6 h, 200%
in CA3 at 12 h, 50% in PC at 6 h
After 40 ADs: increase to 200% in DG at 6 and
24 h, 150% in CA3 at 6 h, 300% in PC at 2 and
6h
After 7 ADs: increase to 200% in CA3 at 24 h
Increase in DG at 4 h

status epilepticus (Bengzon et al., 1993; Humpel et
al., 1993; Nibuya et al., 1995; Elmer et al., 1996a;
Mudo et al., 1996; Schmidt-Kastner et al., 1996;
Hughes et al., 1998). Subcellular studies have dem-
onstrated targeting of BDNF and trkB mRNAs
to dendrites in CA3 neurons following kindled
seizures (Simonato et al., 2002).
Role(s) of BDNF during development
In vitro and in vivo studies have demonstrated
that BDNF has survival- and growth-promoting
actions on a variety of CNS neurons, including
dorsal root ganglion cells, dopaminergic and
cholinergic neurons, retinal ganglion cells, and

hippocampal and cortical neurons (Johnson et al.,
1986; Alderson et al., 1990; Hyman et al., 1991;
Knusel et al., 1991; Acheson et al., 1995; Patel and
McNamara, 1995; Huang and Reichardt, 2001).
Certain peripheral sensory neurons, especially
those in vestibular and nodose-petrosal ganglia,
depend on the presence of BDNF because BDNF
homozygous knockout (BDNF/) mice lack
these neurons (Huang and Reichardt, 2001). Un-
like NGF, sympathetic neurons are not affected,
nor are motor neurons. BDNF/ mice fail to
thrive, demonstrate lack of proper coordination of
movement and balance, and ultimately die by 3
weeks of age. However, heterozygous BDNF
knockout (BDNF+/) mice are viable, and ex-
hibit a variety of phenotypes, including obesity
(Lyons et al., 1999; Kernie et al., 2000), decreased
seizure susceptibility (Kokaia et al., 1995), and
impaired spatial learning (Linnarsson et al., 1997).
Interestingly, conditional postnatal BDNF gene
deletion (Rios et al., 2001) and reduction in trkB
expression (Xu et al., 2003) also cause obesity.
Physiologic regulation of BDNF gene expres-
sion may be very important in the development
of the brain. For example, BDNF contributes to
activity-dependent development of the visual cor-
tex. Provision of excess BDNF (Cabelli et al.,
1995) or blockade of BDNF signaling (Cabelli
et al., 1997) leads to abnormal patterning of ocular
dominance columns during a critical period of
visual cortex development. This suggests a role for
BDNF in axonal pathfinding during development.
BDNF also has powerful effects on dendritic mor-
phology (McAllister et al., 1997; Murphy et al.,
1998; Horch and Katz, 2002; Tolwani et al., 2002).
Effects on synaptic transmission
BDNF has an enormous range of physiologic ac-
tions at both developing and mature synapses, over-
all enhancing synaptic transmission by both pre- and
postsynaptic mechanisms. The first studies of BDNF
effects on synaptic transmission showed that BDNF
increased the frequency of miniature excitatory
postsynaptic currents (EPSCs) at Xenopus neuro-
muscular synapses (Lohof et al., 1993). Since then,
numerous studies have examined the actions of
377
BDNF. Overall, BDNF appears to strengthen
excitatory (glutamatergic) synapses and weaken in-
hibitory (GABAergic) synapses. Schuman and
colleagues demonstrated that exposure of adult rat
hippocampal slices to BDNF led to a long-lasting
potentiation of synaptic strength at Schaffer collat-
eral-CA1 synapses (Kang and Schuman, 1995). Sub-
sequent studies have supported a role of BDNF in
LTP (Korte et al., 1995; Korte et al., 1996; Patterson
et al., 1996; Kang et al., 1997; Xu et al., 2000). For
example, incubation of hippocampal or visual cor-
tical slices with trkB inhibitors inhibits LTP (Figurov
et al., 1996), and hippocampal slices from BDNF/
mice exhibit impaired LTP induction (Korte et al.,
1995), which is restored by reintroduction of BDNF
(Korte et al., 1996; Patterson et al., 1996).
Whether BDNF-induced synaptic potentiation
occurs primarily by a presynaptic action (e.g.
through enhancement of glutamate release)
or postsynaptically (e.g. via phosphorylation of
neurotransmitter receptors) is intensely debated
(Schinder and Poo, 2000). A number of studies
have provided evidence for a presynaptic locus
(Xu et al., 2000; Tyler et al., 2002; see also Kafitz
et al., 1999), yet evidence for postsynaptic actions
has also been obtained (Black, 1999; Thakker-
Varia et al., 2001; reviewed in Poo, 2001). Both
pre- and postsynaptic trkB receptors in the hip-
pocampus may be important (Drake et al., 1999).
A role for BDNF in GABAergic synapses was
first raised by studies showing that BDNF influ-
ences GABAergic neuronal phenotype (Marty et al.,
1996; McLean Bolton et al., 2000). Subsequently,
BDNF was shown to decrease inhibitory (GAB-
Aergic) synaptic transmission (Tanaka et al., 1997;
Frerking et al., 1998; Wardle and Poo, 2003). Re-
cent evidence shows that BDNF can modulate the
function of GABAA receptors via modulation
of phosphorylation state (Jovanovic et al., 2004).
Interestingly, BDNF may also regulate the efficacy
of GABAergic synapses by direct downregulation of
the neuronal K+-Cl cotransporter, which would
impair neuronal Cl extrusion and weaken GAB-
Aergic inhibition (Rivera et al., 2002). Similarly,
a recent paper found differential effects of BDNF
on GABA-mediated currents in excitatory and
inhibitory neuron subpopulations, selectively de-
creasing the efficacy of inhibitory neurotransmission
378
by downregulation of Cl transport (Wardle and
Poo, 2003).
Effect on neurogenesis
An important feature of the dentate gyrus is the
lifelong production of new granule cells from pro-
genitor cells located in the subgranular zone
(Gould and McEwen, 1993; Scharfman, 2004).
BDNF has also been found to enhance neurogen-
esis. Intraventricular infusion of BDNF or adeno-
viral-induced BDNF activity increases the number
of neurons in the adult olfactory bulb, striatum,
septum, and thalamus (Zigova et al., 1998;
Benraiss et al., 2001; Pencea et al., 2001), which
can be potentiated by concurrent inhibition of glial
differentiation of subependymal progenitor cells
(Chmielnicki et al., 2004). Intrahippocampal infu-
sion of BDNF into adult rats leads to increased
dentate gyrus neurogenesis, accompanied by in-
creased numbers of ectopic granule cells (Scharf-
man et al., 2005). BDNF+/ mice show decreased
numbers of BrdU-labeled cells in the dentate gyrus
(Lee et al., 2002). Studies of cultured progenitor
cells have elucidated some of the signaling mech-
anisms, which appear to involve trkB activation,
followed by activation of the MAP kinase and PI3-
kinase pathways (Barnabe-Heider and Miller,
2003) and downstream modification of basic he-
lix-loop-helix transcription factors (Ito et al.,
2003). Although some studies have concluded that
the primary effect of BDNF is on proliferation
(Katoh-Semba et al., 2002), other experiments
suggest an important effect on survival (Lee et al.,
2002). The effects of BDNF may depend on a
previous history of ischemic damage (Larsson
et al., 2002; Gustafsson et al., 2003).
Effects on learning and memory
Learning and memory depend on persistent selective
modification of synapses between CNS neurons.
Since BDNF appears to be involved in activity-
dependent synaptic plasticity, there is great interest
in its role in learning and memory (Yamada and
Nabeshima, 2003). The hippocampus, which is re-
quired for many forms of long-term memory in
humans and animals, appears to be an important site
of BDNF action. Rapid and selective induction
of BDNF expression in the hippocampus during
contextual learning has been demonstrated (Hall
et al., 2000), and function-blocking antibodies to
BDNF (Alonso et al., 2002), BDNF knockout
(Linnarsson et al., 1997), knockout of forebrain
trkB signaling (Minichiello et al., 1999), or overex-
pression of truncated trkB (Saarelainen et al., 2000b)
in mice impairs spatial learning. Another study dem-
onstrated upregulation of BDNF in monkey parietal
cortex associated with tool-use learning (Ishibashi
et al., 2002). In humans, a valine to methionine
polymorphism at the 50 pro-region of the human
BDNF protein was found to be associated with
poorer episodic memory; in vitro, neurons trans-
fected with met-BDNF-green fluorescence protein
(GFP) exhibited reduced depolarization-induced
BDNF secretion (Egan et al., 2003).
Neurotrophin-3 (NT-3)
NT-3, first described in 1990 (Maisonpierre et al.,
1990b), is similar to BDNF in several ways. Like
BDNF, NT-3 mRNA and protein are widely dis-
tributed in the adult CNS (Maisonpierre et al.,
1990a, b, Zhou and Rush, 1994; Katoh-Semba
et al., 1996). While the preferred receptor for NT-3
is trkC, NT-3 can also bind to trkA and trkB
(Barbacid, 1994; Ryden and Ibanez, 1996; Huang
and Reichardt, 2003). Like BDNF, NT-3 is in-
volved in synaptic transmission and neuronal
excitability (Thoenen, 1995). Addition of NT-3
to hippocampal slices enhances synaptic strength
at Schaffer collateral-CA1 synapses (Kang and
Schuman, 1995). NT-3 enhances paired-pulse fa-
cilitation in the perforant path-dentate gyrus path-
way (Kokaia et al., 1998; Asztely et al., 2000). Like
BDNF, NT-3 reduces GABAergic inhibition (Kim
et al., 1994). Also like BDNF, NT-3 enhances the
survival and differentiation of neural progenitor
cells (Barnabe-Heider and Miller, 2003).
NT-3 gene regulation
However, whereas NGF and BDNF levels increase
after seizures, NT-3 levels are reduced in dentate

Table 4. Seizure regulation of NT-3 and NT-4 expression
Reference Methods
NT-3 mRNA
Gall and Lauterborn (1992) Hilar electrolytic lesion
Bengzon et al. (1993) Traditional kindling (ventral
hippocampal stimulation CA1-CA2)
Schmidt-Kastner and Olson (1995) Pilocarpine
Mudo et al. (1996) Pilocarpine
Kim et al. (1998) Amygdala kindling
NT-4 mRNA
Timmusk et al. (1993a, b) RNAse protection analysis from
different tissues
Mudo et al. (1996) Pilocarpine
gyrus granule neurons (Gall and Lauterborn,
1992; Bengzon et al., 1993; Schmidt-Kastner and
Olson, 1995; Mudo et al., 1996; Kim et al., 1998)
(Table 4). This suggests that the potential role of
NT-3 in seizure progression is different. Whether
trkC is elevated appears to depend on the model
used; rapid kindling induces no change (Merlio et
al., 1993) or a transient increase (Bengzon et al.,
1993) in trkC mRNA levels, and ICV KA or ICV
bicuculline transiently increase trkC mRNA in
dentate granule cells (Mudo et al., 1995).
Neurotrophin-4/5
The fourth member of the NT family, NT-4/5, was
discovered after NGF, BDNF, and NT-3 (Hall-
book et al., 1991; Ip et al., 1992). Levels of NT-4/5
in the brain are very low at baseline (Timmusk
et al., 1993a; Katoh-Semba et al., 2003) and are
not increased by seizures (Timmusk et al., 1993a;
Mudo et al., 1996) (Table 4). NT-4/5/ mice, un-
like BDNF/ mice, are normal and long-lived
with no obvious neurological deficits (Conover
et al., 1995; Liu et al., 1995). The only loss of
neurons in NT-4/5/ mice appears to be a reduc-
tion in the number of sensory neurons in the no-
dose-petrosal and geniculate ganglia (Conover
et al., 1995; Liu et al., 1995). Provision of NT-4/
379
Results
Decrease, onset 12 h max 12 h (20%) in DG
Decrease — did not do time course but studied
relationship between development of kindling
and neurotrophin induction — found similar NT-
3 decrease (
50%) regardless of kindling stage
Decrease max at 3 h
Decrease max at 12–24 h (40–50%)
Decrease in DG
Could not detect NT-4 by in situ but did find very
low levels in adult brain using RNAse protection
No increase in NT-4 mRNA following either
systemic KA or hippocampal stimulation
Basal not detectable and pilocarpine status did
not increase levels
5 protects hippocampal and cortical neurons
against energy deprivation-induced injury (Cheng
et al., 1994) and adrenalectomy-induced apoptosis
of hippocampal granule cells (Qiao et al., 1996).
Application of NT-4/5 enhances excitatory synap-
tic transmission in cultured hippocampal neurons
(Lessmann et al., 1994).
Roles of neurotrophins in epilepsy models
The discovery that limbic seizures increase NGF
mRNA levels (Gall and Isackson, 1989) led to the
idea that seizure-induced expression of neurotrophic
factors may contribute to the lasting structural and
functional changes underlying epileptogenesis (Gall
et al., 1991, 1997; Jankowsky and Patterson, 2001).
NGF and epilepsy
Is there a functional role for NGF gene upregula-
tion in epileptogenesis? Indeed, intraventricular
administration of NGF antibodies retards amygdala
kindling (Funabashi et al., 1988) and blocks kin-
dling-induced mossy-fiber sprouting (Van der Zee
et al., 1995). Similarly, an NGF inhibitory peptide
inhibits amygdala kindling and mossy-fiber sprout-
ing (Rashid et al., 1995). Conversely, intraventricu-
lar NGF infusion was found to facilitate amygdala
380
and hippocampal kindling and increase mossy-fiber
sprouting (Adams et al., 1997).
Whether NGF exerts its effects on kindling and
kindling-induced morphological changes via trkA
or p75NTR has been investigated. Inhibition of
Ras, a downstream effector of trkA, inhibits kin-
dling and kindling-associated mossy-fiber sprout-
ing (Li et al., 2003). Peptide inhibitors of NGF
binding to trkA but not to p75NTR can inhibit
kindling, whereas both trkA and p75NTR inhibi-
tion can inhibit mossy-fiber sprouting (Li et al.,
2005).
What is the locus of effects of NGF on hippo-
campal kindling? As described above, there is little
evidence for trkA expression in hippocampus
(Sobreviela et al., 1994; Cellerino, 1996). Simi-
larly, there is little expression of p75 in hippocam-
pus at baseline (Pioro and Cuello, 1990; Sobreviela
et al., 1994). It is likely that NGF-dependent
effects are due to modulation of the cholinergic
system, as both trkA and p75NTR (Hofer et al.,
1990) receptors are most strongly expressed in the
basal forebrain cholinergic neurons which project
to hippocampus (Sobreviela et al., 1994). Consist-
ent with this hypothesis is that cholinergic agonists
and antagonists produce effects on kindling and
sprouting parallel to those of NGF (Adams et al.,
2002).
BDNF and epilepsy
Abundant in vitro and in vivo evidence implicates
BDNF in the cascade of electrophysiologic and
behavioral changes underlying the epileptic state
(Binder et al., 2001). BDNF mRNA and protein
are markedly upregulated in the hippocampus by
seizure activity in animal models (Ernfors et al.,
1991; Isackson et al., 1991; Lindvall et al., 1994;
Nibuya et al., 1995). Infusion of trkB receptor
body (a chimera of human IgG-Fc domain and the
extracellular domain of the trkB receptor) (Binder
et al., 1999b) or use of BDNF+/ (Kokaia et al.,
1995) or truncated trkB-overexpressing (Lahteinen
et al., 2002) mice inhibits epileptogenesis in animal
models. Conversely, direct application of BDNF
induces hyperexcitability in vitro (Scharfman,
1997; Scharfman et al., 1999), overexpression of
BDNF in transgenic mice leads to spontaneous
seizures (Croll et al., 1999), and intrahippocampal
infusion of BDNF is sufficient to induce seizure
activity in vivo (Scharfman et al., 2002).
A separate group of experiments has demon-
strated that chronic BDNF infusion can inhibit
kindling (Larmet et al., 1995; Osehobo et al., 1996;
Reibel et al., 2000b). These inhibitory effects ap-
pear to be due to trkB receptor downregulation
following chronic BDNF administration, and
hence are still consistent with the ‘‘pro-
epileptogenic BDNF’’ hypothesis. This interpreta-
tion is supported by the observation that chronic
exposure to BDNF in vitro leads to downregula-
tion of trkB mRNA and protein (Knusel et al.,
1997). Similarly, continuous in vivo intrahippo-
campal BDNF infusion results in downregulation
of trkB protein by as much as 80% (Frank et al.,
1996). Thus, whereas chronic BDNF infusion in-
hibits kindling progression, acute microinjections
of BDNF enhance epileptogenesis in the absence
of effect on trkB expression (Xu et al., 2004). Fur-
thermore, chronic infusions of BDNF may upreg-
ulate the inhibitory neuropeptide Y (NPY) (Reibel
et al., 2000a).
Whether BDNF has a significant effect on sei-
zure-associated mossy-fiber sprouting is not clear.
While mossy-fiber sprouting has been reported in
BDNF+/ mice and following BDNF infusion
(Kokaia et al., 1995; Scharfman et al., 2002), there
is no effect on mossy-fiber sprouting in BDNF-
overexpressing mice or following chronic infusion
or bolus injection of BDNF in other studies (Qiao
et al., 2001; Xu et al., 2004). However, BDNF
overexpression does increase dendritic length and
complexity in the hippocampus (Tolwani et al.,
2002). The relative role of BDNF on effect synap-
tic changes vs. larger scale morphological changes
during epileptogenesis remains to be clarified.
The anatomic locus of action of NTs during
epileptogenesis has been clarified with the study of
trk receptor activation (see below).
NT-3 and epilepsy
In comparison with BDNF, what is the evidence
for a role of NT-3 in epileptogenesis? In NT-3+/

mice, which have
30% reduction in basal NT-3
mRNA levels, amygdala kindling was markedly
retarded (Elmer et al., 1997). However, compen-
satory changes in BDNF and trkB mRNA levels in
these mice made these data difficult to interpret
(Elmer et al., 1997). Chronic intraventricular in-
fusion of NT-3 retards the development of behavi-
oral seizures (Xu et al., 2002), probably in part via
downregulation of trk phosphorylation (Xu et al.,
2002).
What about the effects of NT-3 on kindling-
induced mossy-fiber sprouting in the dentate
gyrus? Chronic infusion of NT-3 inhibits kin-
dling-associated mossy-fiber sprouting (Xu et al.,
2002). However, this effect is unclear as infusion of
NT-3 in the absence of kindling actually enhances
sprouting of mossy fibers in the inner molecular
layer of the dentate gyrus and CA3 stratum oriens
(Xu et al., 2002).
NT-4 and epilepsy
Unlike NGF, BDNF, and NT-3 levels, levels of
NT-4/5 do not appear to be regulated by seizure
activity (Timmusk et al., 1993a; Mudo et al.,
1996). The amygdala kindling phenotype of NT-4/
5/ mice was studied (He et al., 2006). No aspect
of the development or persistence of amygdala
kindling was different between NT-4/5/ and
wild-type mice (He et al., 2006).
Trk receptor activation following seizure activity
The ability to monitor trk receptor activation fol-
lowing seizures using phospho-specific trk antibodies
enabled identification of the anatomy, time course,
and threshold characteristics of trk receptor activa-
tion in the hippocampus following seizure activity
(Binder et al., 1999a). Kainate-induced status epi-
lepticus or hippocampal electrographic seizures in-
crease phospho-trk immunoreactivity selectively in
the hippocampus, primarily confined to the dentate
hilus and CA3 stratum lucidum. This seizure-
induced phospho-trk immunoreactivity is marked
but transient, maximal at 24–48 h but back to base-
line by 1 week. The seizure duration threshold for
increase in phospho-trk immunoreactivity appears to
381
correspond to the previously reported threshold for
increase in BDNF gene expression. These observa-
tions are examined in greater detail in the next few
sections.
Anatomy of seizure-induced phospho-trk
immunoreactivity
Following seizure activity, phospho-trk immuno-
reactivity is selectively increased in dentate hilus
and CA3 stratum lucidum of hippocampus (Bin-
der et al., 1999a). This distribution precisely coin-
cides with the ‘‘mossy fiber’’ pathway of dentate
granule cell axon terminals. In addition, this an-
atomic pattern coincides with the distribution of
both basal and seizure-induced BDNF protein.
Basal BDNF protein is also localized in hilus and
CA3 stratum lucidum (Conner et al., 1997), and
seizures increase levels of BDNF protein in dentate
gyrus and CA3 (Elmer et al., 1998) and BDNF
immunoreactivity in hilus and CA3 stratum lucid-
um (Smith et al., 1997; Yan et al., 1997b; Rudge et
al., 1998; Vezzani et al., 1999). This precise an-
atomic colocalization of increased phospho-trk
immunoreactivity and increases in BDNF protein
suggests that the phospho-trk immunoreactivity is
caused by seizure-induced increases in BDNF.
BDNF, but not NGF, is known to increase levels
of NPY (Croll et al., 1994), and kindling and
kainate-induced seizures increase NPY immuno-
reactivity in hilus and CA3 stratum lucidum
(Marksteiner et al., 1990; Tønder et al., 1994),
further implicating seizure-induced BDNF acting
in the mossy-fiber pathway. While NGF mRNA
content is upregulated by seizures, the anatomic
distribution of increased NGF protein is not
known. Thus, these anatomic considerations are
most consistent with a role for BDNF.
Time course of seizure-induced phospho-trk
immunoreactivity
The time course of known BDNF upregulation
following seizures coincides temporally with in-
creased phospho-trk immunoreactivity. Using hip-
pocampal microdissection and a two-site ELISA
for BDNF, Elmer et al. showed that after seven
382
ventral hippocampal electrographic seizures, the
maximum increase in BDNF protein occurs at 12 h
in dentate gyrus and 24 h in CA3 (Elmer et al.,
1998). Similarly, maximum increases in BDNF
protein following hilus lesion-induced (Nawa et
al., 1995) or kainate-induced (Rudge et al., 1998)
seizures occur at approximately 24 h in hippocam-
pus. Importantly, BDNF protein levels in both of
these studies returned to baseline after 1 week,
similar to phospho-trk immunoreactivity. In con-
trast, Bengzon et al. found maximal NGF protein
content (measured by two-site immunoassay) 7
days after a similar rapid kindling protocol (Ben-
gzon et al., 1992) and did not see NGF protein
increases at earlier time points. Similarly, Lowen-
stein et al. found maximal NGF-like neurotrophic
activity of hippocampal extracts from animals 1
week after KA treatment (Lowenstein et al., 1993).
Thus, the time-course data favor a role for BDNF
rather than NGF in seizure-induced phospho-trk
immunoreactivity.
Seizure duration threshold for increased phospho-
trk immunoreactivity
The seizure duration threshold for increase in
phospho-trk immunoreactivity further supports a
role for BDNF. Consistently, increased phospho-
trk immunoreactivity was observed only in hippo-
campal kindled animals with ESD
70 s (Binder
et al., unpublished data). In a similar ventral hip-
pocampal stimulation protocol, Bengzon et al. ob-
served increases in BDNF mRNA content in
dentate granule cells in an all-or-none manner
above an electrographic seizure duration of ap-
proximately 70 s (Bengzon et al., 1993). Like the
increases in mRNA content, increases in phospho-
trk immunoreactivity appeared to be ‘‘all-
or-none’’ as no differences were noted in intensity
of immunoreactivity between kainate-treated and
7 hippocampal ES-treated animals despite marked
differences in seizure duration (hours for kainate
vs. seconds for 7 hippocampal ESs) (Binder et al.,
unpublished data). This strong similarity between
thresholds as well as all-or-none characteristics
suggests that such prior increases in BDNF
mRNA content may not only be necessary for
any increase in phospho-trk immunoreactivity but
also sufficient for maximal increase in phospho-trk
immunoreactivity following seizures.
Evidence that the trk receptor activated by seizures
is trkB
Indirect evidence suggests that BDNF-induced
trkB activation is responsible for the increased
phospho-trk immunoreactivity following seizures.
First, the mRNA content of NGF and BDNF is
increased following seizures (Ernfors et al., 1991;
Isackson et al., 1991; Lindvall et al., 1994; Nibuya
et al., 1995) whereas dentate granule NT-3 mRNA
content is decreased (Gall et al., 1991; Gall and
Lauterborn, 1992; Bengzon et al., 1993; Schmidt-
Kastner and Olson, 1995; Mudo et al., 1996). Sec-
ond, protein levels of NGF and BDNF increase
after seizure activity (Bengzon et al., 1992; Elmer
et al., 1998). Third, the time-course data described
above implicate BDNF rather than NGF. Fourth,
mRNA levels of the other NT known to activate
trkB, NT-4, are very low in adult brain (Timmusk
et al., 1993a) and do not increase after seizures
(Mudo et al., 1996). Fifth, unlike trkB and trkC,
levels of expression of trkA in hippocampus are
barely detectable (Barbacid, 1994; Cellerino,
1996), suggesting that trkA is unlikely to mediate
seizure-induced increases in phospho-trk immuno-
reactivity.
In order to more directly analyze the role of the
trkB receptor in seizure-induced trk receptor acti-
vation, He et al. studied trk receptor phosphor-
ylation in a mouse mutant with a single point
mutation at the shc site (Y490 in humans, Y515 in
mice) of the trkB receptor (He et al., 2002). Ho-
mozygous trkBshc/shc (Y515F) mice were generated
by Minichiello et al. and interestingly display loss
of NT-4-dependent neurons but have no major
effects on BDNF responses (Minichiello et al.,
1999). He et al. found that following amygdala
kindling stimulation, phospho-trk immunoreactiv-
ity is increased in wild-type mice in a similar pat-
tern (hilus and CA3 stratum lucidum) to that seen
in the rat experiments (described above). The

trkBshc/shc homozygous mice displayed absence of
seizure-induced phospho-trk immunoreactivity,
and the heterozygotes displayed intermediate
immunoreactivity (He et al., 2002). These experi-
ments suggest that the trk receptor activated
during kindling stimulation is indeed trkB.
Interestingly, the Y515F point mutation had
no effect on kindling development in the same
study (He et al., 2002). This is remarkably con-
sistent with the lack of effect of this mutation on
synaptic LTP (Korte et al., 2000). More recently,
this group has generated a distinct mouse with a
point mutation at the PLC site. Unlike trkBshc/shc
mice, trkBPLC/PLC mice exhibit impaired LTP
(Minichiello et al., 2002). This direct comparison
of distinct trkB tyrosine mutants implicates the
PLC signaling pathway as opposed to the MAPK
pathway in trkB activation-induced synaptic
plasticity.
Similarly, other studies have shown that spe-
cific stimuli may cause tyrosine-specific phos-
phorylation of the trkB receptor (i.e. at other
tyrosines but not at the shc site). For example,
Saarelainen et al., in studying the role of en-
dogenous BDNF and trkB signaling in the
mechanism of action of antidepressant drugs,
found that acute and chronic antidepressant
treatment caused trkB receptor phosphorylat-
ion and activation, but the pY674/5 site was
selectively phosphorylated compared to the
pY490 (shc) site (Saarelainen et al., 2003). The
further development of phosphorylation state-
specific antibodies to distinct tyrosines (pY674/
5, pY785) may prove to be of use in dissecting
tyrosine site-specific trkB signal transduction in
vivo in a variety of paradigms. Furthermore,
these results can be compared with antibodies
that recognize activated intracellular signaling
pathways (e.g. phosphoCREB) (Finkbeiner et
al., 1997).
Cellular site of seizure-induced phospho-trk
immunoreactivity
What is the likely cellular site of seizure-induced
phospho-trk immunoreactivity? The light micro-
scopic distribution of phospho-trk immunoreactivity
383
after seizure (dentate hilus and CA3 stratum lucidum
of hippocampus) corresponds to the mossy-fiber
pathway of dentate granule cell axon terminals
(Binder et al., 1999a). This suggests that the cellular
site of phospho-trk immunoreactivity is either on
mossy-fiber axons and/or targets. Localization on
mossy-fiber axons represents a parsimonious expla-
nation for both hilar and CA3 stratum lucidum
immunoreactivity. In contrast, localization on targets
requires immunoreactivity on both targets in hilus
(hilar interneurons) and in CA3 stratum lucidum
(pyramidal cell dendrites and/or stratum lucidum in-
terneurons).
Anatomic consideration of trkB-like immuno-
reactivity may lend insight into the likely cellular
site of phospho-trk immunoreactivity. In some
published experiments, an affinity-purified anti-
body directed against an extracellular trkB peptide
sequence was used, which does not distinguish be-
tween full-length and truncated (Barbacid, 1994)
trkB receptors. The earlier studies (using light mi-
croscopy) demonstrated that trkB-like immunore-
activity is preferentially distributed on cell bodies
and dendrites of both cortical and hippocampal
neurons (Fryer et al., 1996; Yan et al., 1997a).
Pyramidal neurons in hippocampus in particular
demonstrate marked trkB immunoreactivity on
cell bodies and dendrites in comparison with axons
(Fryer et al., 1996; Yan et al., 1997a). These stud-
ies utilized an antibody raised against the extra-
cellular portion of trkB (trkB2336) common to
both full-length and truncated forms. A more
recent and comprehensive study of cellular and
subcellular localization of trkB immunoreactivity
was carried out by Drake et al. (1999). These
investigators used acytoplasmic-domain antibody
(trkB-in) to selectively label the full-length form of
trkB and carried out both light and electron
microscopic analysis. Their conclusion was that
full-length trkB immunoreactivity exists in glut-
amatergic granule and pyramidal cells and was
most intense in axons, axon terminals, and den-
dritic spines and to a lesser extent in somata and
dendritic shafts. Occasionally, interneurons were
also labeled. Thus, phospho-trkB immunoreactiv-
ity could represent pre- and/or postsynaptic acti-
vation of trkB receptors in the mossy-fiber
pathway.
384
Potential models for induction of phospho-trk
immunoreactivity by seizure activity
Throughout the brain, BDNF immunoreactivity
appears to be preferentially localized in cell bodies
and axons compared to dendrites (Conner et al.,
1997). In addition, unlike the classical target-de-
rived trophic factor model in which NTs are ret-
rogradely transported, abundant recent evidence
suggests that CNS BDNF appears to be ante-
rogradely transported (Von Bartheld et al., 1996a;
Zhou and Rush, 1996; Altar et al., 1997; Conner et
al., 1997; Fawcett et al., 1998; Tonra et al., 1998).
This evidence, together with the anatomic distri-
bution of BDNF immunoreactivity in hippocam-
pus in a mossy fiber-like pattern, suggests that
BDNF protein in hilus and CA3 stratum lucidum
was synthesized in granule cell bodies and ante-
rogradely transported to mossy-fiber terminals.
Furthermore, following seizures there may be
increased anterograde transport of BDNF. First,
using hippocampal microdissections of dentate
gyrus (which contained hilus) and CA3 (which
contained stratum lucidum), Elmer et al. showed
that maximal BDNF protein levels after seizures
were at 12 h in dentate gyrus but 24 h in CA3
(Elmer et al., 1998). This suggests anterograde
transport of seizure-induced BDNF protein. More
recent evidence regarding the time course of
BDNF immunoreactivity following seizures dem-
onstrates that there is increased BDNF immuno-
reactivity in dentate granule cells at 4 h followed
by subsequent increases in hilus and finally in-
creases in CA3 stratum lucidum at about 24 h
(Vezzani et al., 1999) (C. Gall, personal commu-
nication). Furthermore, this anterograde ‘‘move-
ment’’ of BDNF immunoreactivity was abrogated
by the axonal transport inhibitor colchicine (C.
Gall, personal communication).
These considerations lead to a model in which
CA3 stratum lucidum phospho-trk immunoreactiv-
ity is a consequence of seizure-induced BDNF
release from mossy-fiber axons activating trkB
receptors on dendrites of CA3 pyramidal cells and
hilar interneurons. Supporting a postsynaptic site
for trk receptor activation is the evidence that full-
length trkB receptors are localized to the postsy-
naptic density (Wu et al., 1996). Alternatively,
dendritic BDNF mRNA targeting may underlie
another potential cellular mechanism for BDNF
translation, release, and trk receptor activation
(Simonato et al., 2002). Determining the ultrastruc-
tural distribution of phospho-trk immunoreactivity
would be necessary to distinguish these possibilities.
Since the other primary target of mossy-fiber
axons in CA3 is dendrites of stratum lucidum
interneurons (Spruston et al., 1997), it is possible
that phospho-trk immunoreactivity in stratum
lucidum could reflect activation of trk receptors on
interneurons as well as CA3 pyramidal cell dend-
rites. Indeed, quantitative analysis of mossy-fiber
targets in CA3 suggests that the number of synaptic
contacts onto GABAergic interneurons vastly out-
numbers those onto CA3 dendrites (Acsady et al.,
1998). Indeed, any interneuron with dendrites tra-
versing stratum lucidum could be a target of mossy-
fiber axons. However, it is unclear whether func-
tional trkB receptors exist on stratum lucidum in-
terneurons, as in situ hybridization studies show
trkB mRNA localization predominantly in granule
and pyramidal cells of hippocampus (Bengzon et al.,
1993) and only occasional interneurons were found
to be trkB-immunoreactive in the EM study (Drake
et al., 1999).
Furthermore, recent evidence indicates that
activated trk receptors may be endocytosed and
retrogradely transported while still tyrosine phos-
phorylated (Grimes et al., 1996; Von Bartheld et al.,
1996b; Bhattacharyya et al., 1997; Riccio et al.,
1997; Senger and Campenot, 1997). Therefore,
mossy fiber-like phospho-trk immunoreactivity
could in part reflect not only distal synaptic sites
of trk activation but also in-progress retrograde
transport of activated trk from CA3 within the
mossy fibers. Thus, the increase in phospho-trk
immunoreactivity observed in the dentate hilus may
represent activated trk from mossy-fiber terminals in
hilus or CA3.
Role of BDNF in other pathologic conditions
Pain
BDNF also may play an important neuromodu-
latory role in pain transduction (Malcangio and

Lessmann, 2003). BDNF is synthesized by dorsal
horn neurons and markedly upregulated in
inflammatory injury to peripheral nerves (along
with NGF) (Fukuoka et al., 2001). BDNF acutely
sensitizes nociceptive afferents and elicits hype-
ralgesia which is abrogated by BDNF inhibitors
(Kerr et al., 1999; Thompson et al., 1999; Pezet
et al., 2002). Central pain sensitization is an activ-
ity-dependent increase in excitability of dorsal
horn neurons leading to a clinically intractable
condition termed ‘‘neuropathic pain’’ in which
normally nonpainful somatosensory stimuli (touch
and pressure) become exquisitely painful (all-
odynia). Electrophysiological and behavioral data
demonstrate that inhibition of BDNF signal trans-
duction inhibits central pain sensitization (Kerr
et al., 1999; Pezet et al., 2002).
Neurodegenerative diseases
The idea that degenerative diseases of the nervous
system may result from insufficient supply of
neurotrophic factors has generated great interest in
BDNF as a potential therapeutic agent. Many
reports have documented evidence of decreased
expression of BDNF in neurological disease (Murer
et al., 2001). Selective reduction of BDNF mRNA in
the hippocampus has been reported in Alzheimer’s
disease specimens (Phillips et al., 1991; Ferrer et al.,
1999), although selective upregulation appears to
occur in plaque-related glial cells in an animal model
(Burbach et al., 2004). Decreased BDNF protein has
been demonstrated in the substantia nigra in Par-
kinson’s disease (Howells et al., 2000). BDNF pro-
motes survival of all major neuronal types affected
in Alzheimer’s and Parkinson’s disease, such as hip-
pocampal and neocortical neurons, cholinergic se-
ptal and basal forebrain neurons, and nigral
dopaminergic neurons.
Interestingly, recent work has implicated BDNF
in Huntington’s disease as well. Huntingtin, the
protein mutated in Huntington’s disease, upregulates
BDNF transcription, and loss of huntingtin-
mediated BDNF transcription leads to loss of tro-
phic support to striatal neurons which subsequently
degenerate in the hallmark pathology of the dis-
order (Zuccato et al., 2001). A recent study has
385
demonstrated that huntingtin normally inhibits the
neuron restrictive silencer element (NRSE) involved
in tonic repression of transcription from BDNF
promoter II (Zuccato et al., 2003). In all of these
disorders, provision of BDNF or increasing endog-
enous BDNF production may conceivably be ther-
apeutic if applied in the appropriate spatiotemporal
context (Spires et al., 2004).
Rett syndrome
Rett syndrome is an X-linked postnatal neurode-
velopmental disorder that strikes approximately 1
in 10,000 girls. It is characterized by regression of
normal development after about the age of 1 year
and eventually leads to several mental and physical
impairment, including cognitive and movement defi-
cits and breathing abnormalities. In 1999, Rett syn-
drome was linked to mutations in the MECP2 gene
on the X chromosome (Amir et al., 1999). MeCP2,
the protein product of the MECP2 gene, is a me-
thyl-CpG binding protein, known to bind DNA
regulatory regions to silence gene expression. In
2003, it was discovered that one of the genes nor-
mally turned off by MeCP2 is BDNF (Chen et al.,
2003; Martinowich et al., 2003). Recently, a strain of
mice missing the mouse version of MECP2 (Mecp2)
has been found to have abnormally low levels of
BDNF (Chang et al., 2006). These mice exhibit sev-
eral features of human Rett syndrome. Increasing
BDNF production in mice lacking Mecp2 restored
mobility and extended life span (Chang et al., 2006).
Neural activity triggers phosphorylation of MeCP2
that detaches it from the regulatory region of the
BDNF gene and allows BDNF transcription (Zhou
et al., 2006). Further study of MeCP2–BDNF in-
teractions may lead to novel insights and treatment
strategies for Rett syndrome. Interestingly, MECP2
abnormalities are starting to be found in other
neurodevelopmental disorders such as autism, sug-
gesting that BDNF dysregulation may also have
a more widespread role in the pathophysiology of
these conditions.
Neuropsychiatric disease
BDNF signaling may also be involved in affective
behaviors (Altar, 1999). Environmental stresses
386
such as immobilization that induce depression also
decrease BDNF mRNA (Smith et al., 1995). Con-
versely, physical exercise is associated with
decreased depression and increased BDNF mRNA
(Russo-Neustadt et al., 1999; Cotman and Berch-
told, 2002). Existing treatments for depression are
thought to act primarily by increasing endogenous
monoaminergic (i.e. serotonergic and nor-
adrenergic) synaptic transmission, and recent stud-
ies have shown that effective antidepressants
increase BDNF mRNA (Dias et al., 2003) and
protein (Chen et al., 2001; Altar et al., 2003). Ex-
ogenous delivery of BDNF promotes the function
and sprouting of serotonergic neurons in adult rat
brains (Mamounas et al., 1995), and BDNF-defi-
cient mice are also deficient in serotonergic inner-
vation (Lyons et al., 1999). Acute local BDNF
infusion has antidepressant-like effects in rats
(Shirayama et al., 2002). Thus, new pharmacolog-
ic strategies are focused on the potential antide-
pressant role of BDNF.
It has also been hypothesized that BDNF may
be involved in bipolar disorder (Tsai, 2004). In-
terestingly, lithium, a major drug for the treatment
of bipolar disorder, increases BDNF and trkB
activation in cerebral cortical neurons (Hashimoto
et al., 2002). BDNF is an attractive candidate gene
for susceptibility to bipolar disorder, and some
(Neves-Pereira et al., 2002; Sklar et al., 2002) but
not other (Hong et al., 2003; Nakata et al., 2003)
studies suggest linkage between BDNF poly-
morphisms and disease susceptibility (Green and
Craddock, 2003). How alterations in BDNF ac-
tivity may relate to fluctuating bouts of mania and
depression in bipolar disorder is still a matter of
speculation.
Perspective
Since the discovery of NGF in the 1950s and
BDNF in the 1980s, a great deal of evidence has
mounted for the roles of NGF, BDNF, NT-3, and
NT-4/5 in development, physiology, and pathol-
ogy. BDNF in particular has important roles in
neural development and cell survival, as well as
appearing essential to molecular mechanisms of
synaptic plasticity and larger scale structural
rearrangements of axons and dendrites. Basic ac-
tivity-related changes in the CNS are thought to
depend on BDNF modulation of synaptic trans-
mission. Pathologic levels of BDNF-dependent
synaptic plasticity may contribute to conditions
such as epilepsy and chronic pain sensitization,
whereas application of the trophic properties of
BDNF may lead to novel therapeutic options in
neurodegenerative diseases and perhaps even in
neuropsychiatric disorders.
The role of BDNF in epilepsy provides a partic-
ularly good example of the pleiotropic effects of
BDNF on excitability. The hippocampus and
closely associated limbic structures are thought to
be particularly important in the pro-epileptogenic
effects of BDNF (Binder et al., 1999a), and in-
creased BDNF expression in the hippocampus is
found in specimens from patients with temporal
lobe epilepsy (Mathern et al., 1997; Takahashi et al.,
1999). It is hoped that understanding of the hyper-
excitability associated with BDNF in epilepsy ani-
mal models may lead to novel anticonvulsant or
antiepileptic therapies (Binder et al., 2001).
Of course, simple up- or downregulation of NTs
may lead to many nonspecific effects. For ultimate
clinical application in specific conditions, it will be
very helpful to elucidate the mechanisms of action
of each of the effects of NT receptor activation.
Therefore, much recent research has focused on
downstream targets of the NT signaling pathways
responsible for specific phenotypic effects. For
example, BDNF activation of trkB down-regulates
hippocampal KCC2, a K+-Cl cotransporter
(Rivera et al., 2002); this suppresses chloride-
dependent fast GABAergic inhibition and may
partially account for BDNF modulation of GAB-
Aergic synapses (Wardle and Poo, 2003). In addi-
tion, BDNF phosphorylates specific subunits of
both, the NMDA receptor and the GABAA
receptor, altering their function (Suen et al.,
1997; Lin et al., 1998; Jovanovic et al., 2004).
Long-term effects of BDNF must take into ac-
count the fact that it upregulates many other plas-
ticity-related genes, such as NPY (Croll et al.,
1994; Nawa et al., 1994). NPY, for example, may
not only modulate excitability (Baraban et al.,
1997; Reibel et al., 2000a) but also other phenom-
ena such as neurogenesis (Howell et al., 2003).

New methods of modulation/upregulation of
NTs may be required to achieve translational con-
trol of diseases of NT deficiency. ‘‘Ampakines’’
represent a new class of compounds that have been
shown to upregulate BDNF over a long period of
time (Lauterborn et al., 2003). Of course, proper
dosing so as not to trigger downregulation of im-
portant NT signaling pathways will be critical to
avoiding deleterious side effects of these potential
new therapies. Nevertheless, these and similar
compounds are under active clinical investigation
as cognitive and memory enhancing drugs
(Danysz, 1999; Johnson and Simmon, 2002).
Acknowledgment
I would like to thank Helen Scharfman for her
detailed analysis and constructive criticism.
References
Acheson, A., Conover, J.C., Fandl, J.P., DeChiara, T.M., Rus-
sell, M., Thadani, A., Squinto, S.P., Yancopoulos, G.D. and
Lindsay, R.M. (1995) A BDNF autocrine loop in adult sen-
sory neurons prevents cell death. Nature, 374: 450–453.
Acsady, L., Kamondi, A., Sik, A., Freund, T. and Buzsaki, G.
(1998) GABAergic cells are the major postsynaptic targets of
mossy fibers in the rat hippocampus. J. Neurosci., 18:
3386–3403.
Adams, B., Sazgar, M., Osehobo, P., Van der Zee, C.E.E.M.,
Diamond, J., Fahnestock, M. and Racine, R.J. (1997) Nerve
growth factor accelerates seizure development, enhances
mossy fiber sprouting, and attenuates seizure-induced de-
creases in neuronal density in the kindling model of epilepsy.
J. Neurosci., 17: 5288–5296.
Adams, B., Vaccarella, L., Fahnestock, M. and Racine, R.J.
(2002) The cholinergic system modulates kindling and kin-
dling-induced mossy fiber sprouting. Synapse, 44: 132–138.
Alderson, R.F., Alterman, A.L., Barde, Y.A. and Lindsay,
R.M. (1990) Brain-derived neurotrophic factor increases sur-
vival and differentiated functions of rat septal cholinergic
neurons in culture. Neuron, 5: 297–306.
Alderson, R.F., Curtis, R., Alterman, A.L., Lindsay, R.M. and
DiStefano, P.S. (2000) Truncated TrkB mediates the end-
ocytosis and release of BDNF and neurotrophin-4/5 by rat
astrocytes and schwann cells in vitro. Brain Res., 871:
210–222.
Alonso, M., Vianna, M.R., Depino, A.M., Mello e Souza, T.,
Pereira, P., Szapiro, G., Viola, H., Pitossi, F., Izquierdo, I.
and Medina, J.H. (2002) BDNF-triggered events in the rat
hippocampus are required for both short- and long-term
memory formation. Hippocampus, 12: 551–560.
387
Altar, C.A. (1999) Neurotrophins and depression. Trends
Pharmacol. Sci., 20: 59–61.
Altar, C.A., Cai, N., Bliven, T., Juhasz, M., Conner, J.M.,
Acheson, A.L., Lindsay, R.M. and Wiegand, S.J. (1997) An-
terograde transport of brain-derived neurotrophic factor and
its role in the brain. Nature, 389: 856–860.
Altar, C.A. and DiStefano, P.S. (1998) Neurotrophin traffick-
ing by anterograde transport. Trends Neurosci., 21: 433–437.
Altar, C.A., Siuciak, J.A., Wright, P., Ip, N.Y., Lindsay, R.M.
and Wiegand, S.J. (1994) In situ hybridization of trkB and
trkC receptor mRNA in rat forebrain and association with
high-affinity binding of [125I]BDNF, [125I]NT-4/5 and
[125I]NT-3. Eur. J. Neurosci., 6: 1389–1405.
Altar, C.A., Whitehead, R.E., Chen, R., Wortwein, G. and
Madsen, T.M. (2003) Effects of electroconvulsive seizures
and antidepressant drugs on brain-derived neurotrophic fac-
tor protein in rat brain. Biol. Psychiatry, 54: 703–709.
Amir, R.E., Van den Veyver, I.B., Wan, M., Tran, C.Q.,
Francke, U. and Zoghbi, H.Y. (1999) Rett syndrome is
caused by mutations in X-linked MECP2, encoding methyl-
CpG-binding protein 2. Nat. Genet., 23: 185–188.
Anderson, K.D., Alderson, R.F., Altar, C.A., DiStefano, P.S.,
Corcoran, T.L., Lindsay, R.M. and Wiegand, S.J. (1995)
Differential distribution of exogenous BDNF, NGF, and
NT-3 in the brain corresponds to the relative abundance and
distribution of high-affinity and low-affinity neurotrophin
receptors. J. Comp. Neurol., 357: 296–317.
Asztely, F., Kokaia, M., Olofsdotter, K., Ortegren, U. and
Lindvall, O. (2000) Afferent-specific modulation of short-
term synaptic plasticity by neurotrophins in dentate gyrus.
Eur. J. Neurosci., 12: 662–669.
Baraban, S.C., Hollopeter, G., Erickson, J.C., Schwartzkroin,
P.A. and Palmiter, R.D. (1997) Knock-out mice reveal a
critical antiepileptic role for neuropeptide Y. J. Neurosci., 17:
8927–8936.
Barbacid, M. (1994) The trk family of neurotrophin receptors.
J. Neurobiol., 25: 1386–1403.
Barde, Y.A., Edgar, D. and Thoenen, H. (1982) Purification of
a new neurotrophic factor from mammalian brain. EMBO J.,
1: 549–553.
Barnabe-Heider, F. and Miller, F.D. (2003) Endogenously pro-
duced neurotrophins regulate survival and differentiation of
cortical progenitors via distinct signaling pathways. J. Ne-
urosci., 23: 5149–5160.
Bengzon, J., Kokaia, Z., Ernfors, P., Kokaia, M., Leanza, G.,
Nilsson, O.G., Persson, H. and Lindvall, O. (1993) Regula-
tion of neurotrophin and trkA, trkB and trkC tyrosine kinase
receptor messenger RNA expression in kindling. Neurosci-
ence, 53: 433–446.
Bengzon, J., Soderstrom, S., Kokaia, Z., Kokaia, M., Ernfors,
P., Persson, H., Ebendal, T. and Lindvall, O. (1992) Wide-
spread increase of nerve growth factor protein in the rat
forebrain after kindling-induced seizures. Brain Res., 587:
338–342.
Benraiss, A., Chmielnicki, E., Lerner, K., Roh, D. and Gold-
man, S.A. (2001) Adenoviral brain-derived neurotrophic fac-
tor induces both neostriatal and olfactory neuronal
388
recruitment from endogenous progenitor cells in the adult
forebrain. J. Neurosci., 21: 6718–6731.
Bhattacharyya, A., Watson, F.L., Bradlee, T.A., Pomeroy,
S.L., Stiles, C.D. and Segal, R.A. (1997) Trk receptors func-
tion as rapid retrograde signal carriers in the adult nervous
system. J. Neurosci., 17: 7007–7016.
Biffo, S., Offenhauser, N., Carter, B.D. and Barde, Y.A. (1995)
Selective binding and internalisation by truncated receptors
restrict the availability of BDNF during development.
Development, 121: 2461–2470.
Binder, D.K., Croll, S.D., Gall, C.M. and Scharfman, H.E.
(2001) BDNF and epilepsy: too much of a good thing?
Trends Neurosci., 24: 47–53.
Binder, D.K., Routbort, M.J. and McNamara, J.O. (1999a)
Immunohistochemical evidence of seizure-induced activation
of trk receptors in the mossy fiber pathway of adult rat
hippocampus. J. Neurosci., 19: 4616–4626.
Binder, D.K., Routbort, M.J., Ryan, T.E., Yancopoulos, G.D.
and McNamara, J.O. (1999b) Selective inhibition of kindling
development by intraventricular administration of TrkB
receptor body. J. Neurosci., 19: 1424–1436.
Black, I.B. (1999) Trophic regulation of synaptic plasticity. J.
Neurobiol., 41: 108–118.
Bramham, C.R., Southard, T., Sarvey, J.M., Herkenham, M.
and Brady, L.S. (1996) Unilateral LTP triggers bilateral in-
creases in hippocampal neurotrophin and trk receptor
mRNA expression in behaving rats: evidence for interhem-
ispheric communication. J. Comp. Neurol., 368: 371–382.
Bregola, G., Frigati, L., Zucchini, S. and Simonato, M. (2000)
Different patterns of induction of fibroblast growth factor-2
and brain-derived neurotrophic factor messenger RNAs dur-
ing kindling epileptogenesis, and development of a herpes
simplex vector for fibroblast growth factor-2 gene transfer in
vivo. Epilepsia, 41(Suppl 6): S122–S126.
Brigadski, T., Hartmann, M. and Lessmann, V. (2005) Differ-
ential vesicular targeting and time course of synaptic secre-
tion of the mammalian neurotrophins. J. Neurosci., 25:
7601–7614.
Burbach, G.J., Hellweg, R., Haas, C.A., Del Turco, D., Deicke,
U., Abramowski, D., Jucker, M., Staufenbiel, M. and Deller,
T. (2004) Induction of brain-derived neurotrophic factor in
plaque-associated glial cells of aged APP23 transgenic mice.
J. Neurosci., 24: 2421–2430.
Cabelli, R.J., Hohn, A. and Shatz, C.J. (1995) Inhibition of
ocular dominance column formation by infusion of NT-4/5
or BDNF. Science, 267: 1662–1666.
Cabelli, R.J., Shelton, D.L., Segal, R.A. and Shatz, C.J. (1997)
Blockade of endogenous ligands of trkB inhibits formation of
ocular dominance columns. Neuron, 19: 63–76.
Casaccia-Bonnefil, P., Carter, B.D., Dobrowsky, R.T. and
Chao, M.V. (1996) Death of oligodendrocytes mediated by
the interaction of nerve growth factor with its receptor p75.
Nature, 383: 716–719.
Castrén, E., Berninger, B., Leingartner, A. and Lindholm, D.
(1998) Regulation of brain-derived neurotrophic factor
mRNA levels in hippocampus by neuronal activity. Prog.
Brain Res., 117: 57–64.
Castrén, E., Pitkanen, M., Sirvio, J., Parsadanian, A., Lind-
holm, D., Thoenen, H. and Riekkinen, P.J. (1993) The
induction of LTP increases BDNF and NGF mRNA but
decreases NT-3 mRNA in the dentate gyrus. Neuroreport,
4: 895–898.
Castrén, E., Thoenen, H. and Lindholm, D. (1995) Brain-de-
rived neurotrophic factor messenger RNA is expressed in the
septum, hypothalamus and in adrenergic brain stem nuclei of
adult rat brain and is increased by osmotic stimulation in the
paraventricular nucleus. Neuroscience, 64: 71–80.
Castrén, E., Zafra, F., Thoenen, H. and Lindholm, D. (1992)
Light regulates expression of brain-derived neurotrophic fac-
tor mRNA in rat visual cortex. Proc. Natl. Acad. Sci. U.S.A.,
89: 9444–9448.
Cellerino, A. (1996) Expression of messenger RNA coding for
the nerve growth factor receptor trkA in the hippocampus of
the adult rat. Neuroscience, 70: 613–616.
Chang, Q., Khare, G., Dani, V., Nelson, S. and Jaenisch, R.
(2006) The disease progression of Mecp2 mutant mice is
affected by the level of BDNF expression. Neuron, 49:
341–348.
Chao, M.V. and Bothwell, M. (2002) Neurotrophins: to cleave
or not to cleave. Neuron, 33: 9–12.
Chao, M.V. and Hempstead, B.L. (1995) p75 and trk: a two-
receptor system. Trends Neurosci., 18: 321–326.
Chen, B., Dowlatshahi, D., MacQueen, G.M., Wang, J.F. and
Young, L.T. (2001) Increased hippocampal BDNF immuno-
reactivity in subjects treated with antidepressant medication.
Biol. Psychiatry, 50: 260–265.
Chen, W.G., Chang, Q., Lin, Y., Meissner, A., West, A.E.,
Griffith, E.C., Jaenisch, R. and Greenberg, M.E. (2003) De-
repression of BDNF transcription involves calcium-depend-
ent phosphorylation of MeCP2. Science, 302: 885–889.
Cheng, B., Goodman, Y., Begley, J.G. and Mattson, M.P.
(1994) Neurotrophin-4/5 protects hippocampal and cortical
neurons against energy deprivation- and excitatory amino
acid-induced injury. Brain Res., 650: 331–335.
Chmielnicki, E., Benraiss, A., Economides, A.N. and Goldman,
S.A. (2004) Adenovirally expressed noggin and brain-derived
neurotrophic factor cooperate to induce new medium spiny
neurons from resident progenitor cells in the adult striatal
ventricular zone. J. Neurosci., 24: 2133–2142.
Climent, E., Sancho-Tello, M., Minana, R., Barettino, D. and
Guerri, C. (2000) Astrocytes in culture express the full-length
Trk-B receptor and respond to brain derived neurotrophic
factor by changing intracellular calcium levels: effect of
ethanol exposure in rats. Neurosci. Lett., 288: 53–56.
Conner, J.M., Lauterborn, J.C., Yan, Q., Gall, C.M. and
Varon, S. (1997) Distribution of brain-derived neurotrophic
factor (BDNF) protein and mRNA in the normal adult rat
CNS–evidence for anterograde axonal transport. J. Neuro-
sci., 17: 2295–2313.
Conover, J.C., Erickson, J.T., Katz, D.M., Bianchi, L.M., Po-
ueymirou, W.T., McClain, J., Pan, L., Helgren, M., Ip, N.Y.,
Boland, P., et al. (1995) Neuronal deficits, not involving mo-
tor neurons, in mice lacking BDNF and/or NT4. Nature,
375: 235–238.

Cotman, C.W. and Berchtold, N.C. (2002) Exercise: a behavi-
oral intervention to enhance brain health and plasticity.
Trends Neurosci., 25: 295–301.
Croll, S.D., Suri, C., Compton, D.L., Simmons, M.V.,
Yancopoulos, G.D., Lindsay, R.M., Wiegand, S.J., Rudge,
J.S. and Scharfman, H.E. (1999) Brain-derived neurotrophic
factor transgenic mice exhibit passive avoidance deficits, in-
creased seizure severity and in vitro hyperexcitability in the
hippocampus and entorhinal cortex. Neuroscience, 93:
1491–1506.
Croll, S.D., Wiegand, S.J., Anderson, K.D., Lindsay, R.M. and
Nawa, H. (1994) Regulation of neuropeptides in adult rat
forebrain by the neurotrophins BDNF and NGF. Eur. J.
Neurosci., 6: 1343–1353.
Danysz, W. (1999) CX-516 (Cortex Pharmaceuticals Inc).
IDrugs, 2: 814–822.
Dechant, G. and Barde, Y.A. (2002) The neurotrophin receptor
p75(NTR): novel functions and implications for diseases of
the nervous system. Nat. Neurosci., 5: 1131–1136.
Dias, B.G., Banerjee, S.B., Duman, R.S. and Vaidya, V.A.
(2003) Differential regulation of brain derived neurotrophic
factor transcripts by antidepressant treatments in the adult
rat brain. Neuropharmacology, 45: 553–563.
Drake, C.T., Milner, T.A. and Patterson, S.L. (1999) Ultra-
structural localization of full-length trkB immunoreactivity
in rat hippocampus suggests multiple roles in modulating
activity-dependent synaptic plasticity. J. Neurosci., 19:
8009–8026.
Dugich-Djordjevic, M.M., Tocco, G., Lapchak, P.A., Pasinetti,
G.M., Najm, I., Baudry, M. and Hefti, F. (1992a) Regionally
specific and rapid increases in brain-derived neurotrophic
factor messenger RNA in the adult rat brain following sei-
zures induced by systemic administration of kainic acid. Ne-
uroscience, 47: 303–315.
Dugich-Djordjevic, M.M., Tocco, G., Willoughby, D.A.,
Najm, I., Pasinetti, G., Thompson, R.F., Baudry, M., Lap-
chak, P.A. and Hefti, F. (1992b) BDNF mRNA expression in
the developing rat brain following kainic acid-induced seizure
activity. Neuron, 8: 1127–1138.
Egan, M.F., Kojima, M., Callicott, J.H., Goldberg, T.E., Ko-
lachana, B.S., Bertolino, A., Zaitsev, E., Gold, B., Goldman,
D., Dean, M., Lu, B. and Weinberger, D.R. (2003) The
BDNF val66met polymorphism affects activity-dependent
secretion of BDNF and human memory and hippocampal
function. Cell, 112: 257–269.
Eide, F.F., Vining, E.R., Eide, B.L., Zang, K., Wang, X.Y. and
Reichardt, L.F. (1996) Naturally occurring truncated trkB
receptors have dominant inhibitory effects on brain-derived
neurotrophic factor signaling. J. Neurosci., 16: 3123–3129.
Elmer, E., Kokaia, M., Ernfors, P., Ferencz, I., Kokaia, Z. and
Lindvall, O. (1997) Suppressed kindling epileptogenesis and
perturbed BDNF and trkB gene regulation in NT-3 mutant
mice. Exp. Neurol., 145: 93–103.
Elmer, E., Kokaia, M., Kokaia, Z., Ferencz, I. and Lindvall, O.
(1996a) Delayed kindling development after rapidly recurring
seizures: relation to mossy fiber sprouting and neurotrophin,
GAP-43 and dynorphin gene expression. Brain Res., 712: 19–34.
389
Elmer, E., Kokaia, Z., Kokaia, M., Carnahan, J., Nawa, H.,
Bengzon, J. and Lindvall, O. (1996b) Widespread increase of
brain-derived neurotrophic factor protein in the rat forebrain
after kindling-induced seizures. Soc. Neurosci. Abstr., 22: 2089.
Elmer, E., Kokaia, Z., Kokaia, M., Carnahan, J., Nawa, H.
and Lindvall, O. (1998) Dynamic changes of brain-derived
neurotrophic factor protein levels in the rat forebrain after
single and recurring kindling-induced seizures. Neuroscience,
83: 351–362.
Ernfors, P., Bengzon, J., Kokaia, Z., Persson, H. and Lindvall,
O. (1991) Increased levels of messenger RNAs for neurotro-
phic factors in the brain during kindling epileptogenesis.
Neuron, 7: 165–176.
Ernfors, P., Wetmore, C., Olson, L. and Persson, H. (1990)
Identification of cells in rat brain and peripheral tissues ex-
pressing mRNA for members of the nerve growth factor
family. Neuron, 5: 511–526.
Esposito, D., Patel, P., Stephens, R.M., Perez, P., Chao, M.V.,
Kaplan, D.R. and Hempstead, B.L. (2001) The cytoplasmic
and transmembrane domains of the p75 and Trk A receptors
regulate high affinity binding to nerve growth factor. J. Biol.
Chem., 276: 32687–32695.
Farhadi, H.F., Mowla, S.J., Petrecca, K., Morris, S.J., Seidah,
N.G. and Murphy, R.A. (2000) Neurotrophin-3 sorts to the
constitutive secretory pathway of hippocampal neurons and
is diverted to the regulated secretory pathway by coexpres-
sion with brain-derived neurotrophic factor. J. Neurosci., 20:
4059–4068.
Fawcett, J.P., Alonso-Vanegas, M.A., Morris, S.J., Miller,
F.D., Sadikot, A.F. and Murphy, R.A. (2000) Evidence that
brain-derived neurotrophic factor from presynaptic nerve
terminals regulates the phenotype of calbindin-containing
neurons in the lateral septum. J. Neurosci., 20: 274–282.
Fawcett, J.P., Aloyz, R., McLean, J.H., Pareek, S., Miller,
F.D., McPherson, P.S. and Murphy, R.A. (1997) Detection
of brain-derived neurotrophic factor in a vesicular fraction of
brain synaptosomes. J. Biol. Chem., 272: 8837–8840.
Fawcett, J.P., Bamji, S.X., Causing, C.G., Aloyz, R., Ase, A.R.,
Reader, T.A., McLean, J.H. and Miller, F.D. (1998) Func-
tional evidence that BDNF is an anterograde neuronal tro-
phic factor in the CNS. J. Neurosci., 18: 2808–2821.
Ferrer, I., Marin, C., Rey, M.J., Ribalta, T., Goutan, E.,
Blanco, R., Tolosa, E. and Marti, E. (1999) BDNF and full-
length and truncated TrkB expression in Alzheimer disease.
Implications in therapeutic strategies. J. Neuropathol. Exp.
Neurol., 58: 729–739.
Figurov, A., Pozzo-Miller, L.D., Olafsson, P., Wang, T. and
Lu, B. (1996) Regulation of synaptic responses to high-fre-
quency stimulation and LTP by neurotrophins in the hippo-
campus. Nature, 381: 706–709.
Finkbeiner, S., Tavazoie, S.F., Maloratsky, A., Jacobs, K.M.,
Harris, K.M. and Greenberg, M.E. (1997) CREB: a major
mediator of neuronal neurotrophin responses. Neuron, 19:
1031–1047.
Frade, J.M., Rodriguez-Tebar, A. and Barde, Y.A. (1996) In-
duction of cell death by endogenous nerve growth factor
through its p75 receptor. Nature, 383: 166–168.
390
Frank, L., Ventimiglia, R., Anderson, K., Lindsay, R.M. and
Rudge, J.S. (1996) BDNF downregulates neurotrophin re-
sponsiveness, trkB protein and trkB mRNA levels in cultured
rat hippocampal neurons. Eur. J. Neurosci., 8: 1220–1230.
Frerking, M., Malenka, R.C. and Nicoll, R.A. (1998) Brain-
derived neurotrophic factor (BDNF) modulates inhibitory,
but not excitatory, transmission in the CA1 region of the
hippocampus. J Neurophysiol., 80: 3383–3386.
Frisen, J., Verge, V.M., Fried, K., Risling, M., Persson, H.,
Trotter, J., Hokfelt, T. and Lindholm, D. (1993) Character-
ization of glial trkB receptors: differential response to injury
in the central and peripheral nervous systems. Proc. Natl.
Acad. Sci. U.S.A., 90: 4971–4975.
Fryer, R.H., Kaplan, D.R., Feinstein, S.C., Radeke, M.J.,
Grayson, D.R. and Kromer, L.F. (1996) Developmental and
mature expression of full-length and truncated trkB receptors
in the rat forebrain. J. Comp. Neurol., 374: 21–40.
Fryer, R.H., Kaplan, D.R. and Kromer, L.F. (1997) Truncated
trkB receptors on nonneuronal cells inhibit BDNF-induced
neurite outgrowth in vitro. Exp. Neurol., 148: 616–627.
Fukuoka, T., Kondo, E., Dai, Y., Hashimoto, N. and Noguchi,
K. (2001) Brain-derived neurotrophic factor increases in the
uninjured dorsal root ganglion neurons in selective spinal
nerve ligation model. J. Neurosci., 21: 4891–4900.
Funabashi, T., Sasaki, H. and Kimura, F. (1988) Intraven-
tricular injection of antiserum to nerve growth factor delays
the development of amygdaloid kindling. Brain Res., 458:
132–136.
Gall, C., Lauterborn, J., Bundman, M., Murray, K. and Isack-
son, P. (1991) Seizures and the regulation of neurotrophic
factor and neuropeptide gene expression in brain. Epilepsy
Res. Suppl., 4: 225–245.
Gall, C.M. and Isackson, P.J. (1989) Limbic seizures increase
neuronal production of messenger RNA for nerve growth
factor. Science, 245: 758–761.
Gall, C.M. and Lauterborn, J. (1992) In: Ribak C.E., Gall C.M.
and Mody I. (Eds.), The Dentate Gyrus and its Role in Sei-
zures. Elsevier, Amsterdam, pp. 171–185.
Gall, C.M., Lauterborn, J.C., Guthrie, K.M. and Stinis, C.T.
(1997) Seizures and the regulation of neurotrophic factor ex-
pression: associations with structural plasticity in epilepsy.
Adv. Neurol., 72: 9–24.
Ginty, D.D. and Segal, R.A. (2002) Retrograde neurotrophin
signaling: Trk-ing along the axon. Curr. Opin. Neurobiol.,
12: 268–274.
Goggi, J., Pullar, I.A., Carney, S.L. and Bradford, H.F. (2003)
The control of [125I]BDNF release from striatal rat brain
slices. Brain Res., 967: 201–209.
Gould, E. and McEwen, B.S. (1993) Neuronal birth and death.
Curr. Opin. Neurobiol., 3: 676–682.
Green, E. and Craddock, N. (2003) Brain-derived neurotrophic
factor as a potential risk locus for bipolar disorder: evidence,
limitations, and implications. Curr. Psychiatry Rep., 5:
469–476.
Griesbeck, O., Canossa, M., Campana, G., Gartner, A., Hoe-
ner, M.C., Nawa, H., Kolbeck, R. and Thoenen, H. (1999)
Are there differences between the secretion characteristics of
NGF and BDNF? Implications for the modulatory role of
neurotrophins in activity-dependent neuronal plasticity. Mi-
crosc. Res. Tech., 45: 262–275.
Grimes, M.L., Zhou, J., Beattie, E.C., Yuen, E.C., Hall, D.E.,
Valletta, J.S., Topp, K.S., LaVail, J.H., Bunnett, N.W. and
Mobley, W.C. (1996) Endocytosis of activated trkA: evidence
that nerve growth factor induces formation of singaling
endosomes. J. Neurosci., 16: 7950–7964.
Gustafsson, E., Lindvall, O. and Kokaia, Z. (2003) Intraven-
tricular infusion of TrkB-Fc fusion protein promotes is-
chemia-induced neurogenesis in adult rat dentate gyrus.
Stroke, 34: 2710–2715.
Haapasalo, A., Koponen, E., Hoppe, E., Wong, G. and Cast-
ren, E. (2001) Truncated trkB.T1 is dominant negative in-
hibitor of trkB.TK+-mediated cell survival. Biochem.
Biophys. Res. Commun., 280: 1352–1358.
Haapasalo, A., Sipola, I., Larsson, K., Akerman, K.E., Stoilov,
P., Stamm, S., Wong, G. and Castren, E. (2002) Regulation
of TRKB surface expression by brain-derived neurotrophic
factor and truncated TRKB isoforms. J. Biol. Chem., 277:
43160–43167.
Hall, J., Thomas, K.L. and Everitt, B.J. (2000) Rapid and se-
lective induction of BDNF expression in the hippocampus
during contextual learning. Nat. Neurosci., 3: 533–535.
Hallbook, F., Ibanez, C.F. and Persson, H. (1991) Evolutionary
studies of the nerve growth factor family reveal a novel
member abundantly expressed in Xenopus ovary. Neuron, 6:
845–858.
Hartmann, M., Heumann, R. and Lessmann, V. (2001) Synap-
tic secretion of BDNF after high-frequency stimulation of
glutamatergic synapses. EMBO J., 20: 5887–5897.
Hashimoto, R., Takei, N., Shimazu, K., Christ, L., Lu, B. and
Chuang, D.M. (2002) Lithium induces brain-derived ne-
urotrophic factor and activates TrkB in rodent cortical neu-
rons: an essential step for neuroprotection against glutamate
excitotoxicity. Neuropharmacology, 43: 1173–1179.
He, X.P., Butler, L., Liu, X. and McNamara, J.O. (2006) The
tyrosine receptor kinase B ligand, neurotrophin-4, is not
required for either epileptogenesis or tyrosine receptor kin-
ase B activation in the kindling model. Neuroscience, 141:
515–520.
He, X.P., Minichiello, L., Klein, R. and McNamara, J.O.
(2002) Immunohistochemical evidence of seizure-induced ac-
tivation of trkB receptors in the mossy fiber pathway of adult
mouse hippocampus. J. Neurosci., 22: 7502–7508.
Hofer, M., Pagliusi, S.R., Hohn, A., Leibrock, J. and Barde,
Y.A. (1990) Regional distribution of brain-derived neurotro-
phic factor mRNA in the adult mouse brain. EMBO J., 9:
2459–2464.
Hong, C.J., Huo, S.J., Yen, F.C., Tung, C.L., Pan, G.M. and
Tsai, S.J. (2003) Association study of a brain-derived ne-
urotrophic-factor genetic polymorphism and mood disorders,
age of onset and suicidal behavior. Neuropsychobiology, 48:
186–189.
Horch, H.W. and Katz, L.C. (2002) BDNF release from single
cells elicits local dendritic growth in nearby neurons. Nat.
Neurosci., 5: 1177–1184.

Howell, O.W., Scharfman, H.E., Herzog, H., Sundstrom, L.E.,
Beck-Sickinger, A. and Gray, W.P. (2003) Neuropeptide Y is
neuroproliferative for post-natal hippocampal precursor
cells. J. Neurochem., 86: 646–659.
Howells, D.W., Porritt, M.J., Wong, J.Y., Batchelor, P.E.,
Kalnins, R., Hughes, A.J. and Donnan, G.A. (2000) Reduced
BDNF mRNA expression in the Parkinson’s disease subst-
antia nigra. Exp. Neurol., 166: 127–135.
Huang, E.J. and Reichardt, L.F. (2001) Neurotrophins: roles in
neuronal development and function. Annu. Rev. Neurosci.,
24: 677–736.
Huang, E.J. and Reichardt, L.F. (2003) Trk receptors: roles in
neuronal signal transduction. Annu. Rev. Biochem., 72:
609–642.
Hughes, P.E., Young, D., Preston, K.M., Yan, Q. and Dragu-
now, M. (1998) Differential regulation by MK801 of imme-
diate-early genes, brain-derived neurotrophic factor and trk
receptor mRNA induced by a kindling after-discharge. Mol.
Brain Res., 53: 138–151.
Humpel, C., Wetmore, C. and Olson, L. (1993) Regulation of
brain-derived neurotrophic factor messenger RNA and pro-
tein at the cellular level in pentylenetetrazol-induced epileptic
seizures. Neuroscience, 53: 909–918.
Hyman, C., Hofer, M., Barde, Y.A., Juhasz, M., Yancopoulos,
G.D., Squinto, S.P. and Lindsay, R.M. (1991) BDNF is a
neurotrophic factor for dopaminergic neurons of the subst-
antia nigra. Nature, 350: 230–232.
Ip, N.Y., Ibanez, C.F., Nye, S.H., McClain, J., Jones, P.F., Gies,
D.R., Belluscio, L., Le Beau, M.M., Espinosa III, R., Squinto,
S.P., et al. (1992) Mammalian neurotrophin-4: structure, chro-
mosomal localization, tissue distribution, and receptor specifi-
city. Proc. Natl. Acad. Sci. U.S.A., 89: 3060–3064.
Isackson, P.J., Huntsman, M.M., Murray, K.D. and Gall,
C.M. (1991) BDNF mRNA expression is increased in adult
rat forebrain after limbic seizures: temporal patterns of in-
duction distinct from NGF. Neuron, 6: 937–948.
Ishibashi, H., Hihara, S., Takahashi, M., Heike, T., Yokota, T.
and Iriki, A. (2002) Tool-use learning induces BDNF ex-
pression in a selective portion of monkey anterior parietal
cortex. Brain Res. Mol. Brain Res., 102: 110–112.
Ito, H., Nakajima, A., Nomoto, H. and Furukawa, S. (2003)
Neurotrophins facilitate neuronal differentiation of cultured
neural stem cells via induction of mRNA expression of basic
helix-loop-helix transcription factors Mash1 and Math1. J.
Neurosci. Res., 71: 648–658.
Jankowsky, J.L. and Patterson, P.H. (2001) The role of
cytokines and growth factors in seizures and their sequelae.
Prog. Neurobiol., 63: 125–149.
Johnson, J.E., Barde, Y.A., Schwab, M. and Thoenen, H.
(1986) Brain-derived neurotrophic factor supports the sur-
vival of cultured rat retinal ganglion cells. J. Neurosci., 6:
3031–3038.
Johnson, S.A. and Simmon, V.F. (2002) Randomized, double-
blind, placebo-controlled international clinical trial of the
Ampakine CX516 in elderly participants with mild cognitive
impairment: a progress report. J. Mol. Neurosci., 19:
197–200.
391
Jovanovic, J.N., Thomas, P., Kittler, J.T., Smart, T.G. and
Moss, S.J. (2004) Brain-derived neurotrophic factor modu-
lates fast synaptic inhibition by regulating GABA(A) recep-
tor phosphorylation, activity, and cell-surface stability. J.
Neurosci., 24: 522–530.
Kafitz, K.W., Rose, C.R., Thoenen, H. and Konnerth, A.
(1999) Neurotrophin-evoked rapid excitation through TrkB
receptors. Nature, 401: 918–921.
Kang, H. and Schuman, E.M. (1995) Long-lasting neurotro-
phin-induced enhancement of synaptic transmission in the
adult hippocampus. Science, 267: 1658–1662.
Kang, H., Welcher, A.A., Shelton, D. and Schuman, E.M.
(1997) Neurotrophins and time: different roles for trkB
signaling in hippocampal long-term potentiation. Neuron,
19: 653–664.
Katoh-Semba, R., Asano, T., Ueda, H., Morishita, R., Take-
uchi, I.K., Inaguma, Y. and Kato, K. (2002) Riluzole en-
hances expression of brain-derived neurotrophic factor with
consequent proliferation of granule precursor cells in the rat
hippocampus. FASEB J., 16: 1328–1330.
Katoh-Semba, R., Ichisaka, S., Hata, Y., Tsumoto, T., Eguchi,
K., Miyazaki, N., Matsuda, M., Takeuchi, I.K. and Kato, K.
(2003) NT-4 protein is localized in neuronal cells in the brain
stem as well as the dorsal root ganglion of embryonic and
adult rats. J. Neurochem., 86: 660–668.
Katoh-Semba, R., Kaisho, Y., Shintani, A., Nagahama, M.
and Kato, K. (1996) Tissue distribution and immunocyto-
chemical localization of neurotrophin-3 in the brain and pe-
ripheral tissues of rats. J. Neurochem., 66: 330–337.
Kernie, S.G., Liebl, D.J. and Parada, L.F. (2000) BDNF reg-
ulates eating behavior and locomotor activity in mice. EMBO
J., 19: 1290–1300.
Kerr, B.J., Bradbury, E.J., Bennett, D.L., Trivedi, P.M.,
Dassan, P., French, J., Shelton, D.B., McMahon, S.B. and
Thompson, S.W. (1999) Brain-derived neurotrophic factor
modulates nociceptive sensory inputs and NMDA-evoked
responses in the rat spinal cord. J. Neurosci., 19:
5138–5148.
Kim, H.G., Wang, T., Olafsson, P. and Lu, B. (1994) Ne-
urotrophin 3 potentiates neuronal activity and inhibits
gamma-aminobutyratergic synaptic transmission in cortical
neurons. Proc. Natl. Acad. Sci. U.S.A., 91: 12341–12345.
Kim, S.Y., Smith, M.A., Post, R.M. and Rosen, J.B. (1998)
Attenuation of kindling-induced decreases in NT-3 mRNA
by thyroid hormone depletion. Epilepsy Res., 29: 211–220.
Knusel, B., Gao, H., Okazaki, T., Yoshida, T., Mori, N., Hefti,
F. and Kaplan, D.R. (1997) Ligand-induced down-regulation
of trk messenger RNA, protein and tyrosine phosphorylation
in rat cortical neurons. Neuroscience, 78: 851–862.
Knusel, B., Winslow, J.W., Rosenthal, A., Burton, L.E., Seid,
D.P., Nikolics, K. and Hefti, F. (1991) Promotion of central
cholinergic and dopaminergic neuron differentiation by
brain-derived neurotrophic factor but not neurotrophin 3.
Proc. Natl. Acad. Sci. U.S.A., 88: 961–965.
Kohara, K., Kitamura, A., Morishima, M. and Tsumoto, T.
(2001) Activity-dependent transfer of brain-derived neurotro-
phic factor to postsynaptic neurons. Science, 291: 2419–2423.
392
Kokaia, M., Asztely, F., Olofsdotter, K., Sindreu, C.B., Kull-
mann, D.M. and Lindvall, O. (1998) Endogenous neurotro-
phin-3 regulates short-term plasticity at lateral perforant
path-granule cell synapses. J. Neurosci., 18: 8730–8739.
Kokaia, M., Ernfors, P., Kokaia, Z., Elmer, E., Jaenisch, R.
and Lindvall, O. (1995) Suppressed epileptogenesis in BDNF
mutant mice. Exp. Neurol., 133: 215–224.
Korte, M., Carroll, P., Wolf, E., Brem, G., Thoenen, H. and
Bonhoeffer, T. (1995) Hippocampal long-term potentiation is
impaired in mice lacking brain-derived neurotrophic factor.
Proc. Natl. Acad. Sci. U.S.A., 92: 8856–8860.
Korte, M., Griesbeck, O., Gravel, C., Carroll, P., Staiger, V.,
Thoenen, H. and Bonhoeffer, T. (1996) Virus-mediated gene
transfer into hippocampal CA1 region restores long-term
potentiation in brain-derived neurotrophic factor mutant
mice. Proc. Natl. Acad. Sci. U.S.A., 93: 12547–12552.
Korte, M., Minichiello, L., Klein, R. and Bonhoeffer, T. (2000)
Shc-binding site in the TrkB receptor is not required for
hippocampal long-term potentiation. Neuropharmacology,
39: 717–724.
Lahteinen, S., Pitkanen, A., Saarelainen, T., Nissinen, J., Ko-
ponen, E. and Castren, E. (2002) Decreased BDNF signalling
in transgenic mice reduces epileptogenesis. Eur. J. Neurosci.,
15: 721–734.
Larmet, Y., Reibel, S., Carnahan, J., Nawa, H., Marescaux, C.
and Depaulis, A. (1995) Protective effects of brain-derived
neurotrophic factor on the development of hippocampal kin-
dling in the rat. Neuroreport, 6: 1937–1941.
Larsson, E., Mandel, R.J., Klein, R.L., Muzyczka, N., Lindv-
all, O. and Kokaia, Z. (2002) Suppression of insult-induced
neurogenesis in adult rat brain by brain-derived neurotrophic
factor. Exp. Neurol., 177: 1–8.
Lauterborn, J.C., Isackson, P.J. and Gall, C.M. (1994) Seizure-
induced increases in NGF mRNA exhibit different time
courses across forebrain regions and are biphasic in hippo-
campus. Exp. Neurol., 125: 22–40.
Lauterborn, J.C., Rivera, S., Stinis, C.T., Hayes, V.Y., Isack-
son, P.J. and Gall, C.M. (1996) Differential effects of pro-
tein synthesis inhibition on the activity-dependent
expression of BDNF transcripts: evidence for immediate-
early gene responses from specific promoters. J. Neurosci.,
16: 7428–7436.
Lauterborn, J.C., Truong, G.S., Baudry, M., Bi, X., Lynch, G.
and Gall, C.M. (2003) Chronic elevation of brain-derived
neurotrophic factor by ampakines. J. Pharmacol. Exp. Ther.,
307: 297–305.
Lee, F.S., Kim, A.H., Khursigara, G. and Chao, M.V. (2001a)
The uniqueness of being a neurotrophin receptor. Curr.
Opin. Neurobiol., 11: 281–286.
Lee, J., Duan, W. and Mattson, M.P. (2002) Evidence that
brain-derived neurotrophic factor is required for basal ne-
urogenesis and mediates, in part, the enhancement of neuro-
genesis by dietary restriction in the hippocampus of adult
mice. J. Neurochem., 82: 1367–1375.
Lee, R., Kermani, P., Teng, K.K. and Hempstead, B.L. (2001b)
Regulation of cell survival by secreted proneurotrophins.
Science, 294: 1945–1948.
Lessmann, V., Gottmann, K. and Heumann, R. (1994) BDNF
and NT-4/5 enhance glutamatergic synaptic transmission in
cultured hippocampal neurones. Neuroreport, 6: 21–25.
Levi-Montalcini, R. and Hamburger, V. (1951) Selective
growth-stimulating effects of mouse sarcoma on the sensory
and sympathetic nervous system of the chick embryo. J. Exp.
Zool., 116: 321–361.
Li, S., Saragovi, H.U., Nedev, H., Zhao, C., Racine, R.J. and
Fahnestock, M. (2005) Differential actions of nerve growth
factor receptors TrkA and p75NTR in a rat model of epile-
ptogenesis. Mol. Cell Neurosci., 29: 162–172.
Li, S., Uri Saragovi, H., Racine, R.J. and Fahnestock, M.
(2003) A ligand of the p65/p95 receptor suppresses perforant
path kindling, kindling-induced mossy fiber sprouting, and
hilar area changes in adult rats. Neuroscience, 119:
1147–1156.
Lin, S.Y., Wu, K., Levine, E.S., Mount, H.T., Suen, P.C. and
Black, I.B. (1998) BDNF acutely increases tyrosine phos-
phorylation of the NMDA receptor subunit 2B in cortical
and hippocampal postsynaptic densities. Brain Res. Mol.
Brain Res., 55: 20–27.
Lindholm, D., Castren, E., Berzaghi, M., Blochl, A. and
Thoenen, H. (1994) Activity-dependent and hormonal regu-
lation of neurotrophin mRNA levels in the brain–implica-
tions for neuronal plasticity. J. Neurobiol., 25: 1362–1372.
Lindvall, O., Kokaia, Z., Bengzon, J., Elmer, E. and Kokaia,
M. (1994) Neurotrophins and brain insults. Trends Neuro-
sci., 17: 490–496.
Linnarsson, S., Bjorklund, A. and Ernfors, P. (1997) Learning
deficit in BDNF mutant mice. Eur. J. Neurosci., 9:
2581–2587.
Liu, X., Ernfors, P., Wu, H. and Jaenisch, R. (1995) Sensory
but not motor neuron deficits in mice lacking NT4 and
BDNF. Nature, 375: 238–241.
Lohof, A.M., Ip, N.Y. and Poo, M.M. (1993) Potentiation of
developing neuromuscular synapses by the neurotrophins
NT-3 and BDNF. Nature, 363: 350–353.
Lowenstein, D.H., Seren, M.S. and Longo, F.M. (1993) Pro-
longed increases in neurotrophic activity associated with
kainate-induced hippocampal synaptic reorganization. Ne-
uroscience, 56: 597–604.
Luikart, B.W., Nef, S., Shipman, T. and Parada, L.F. (2003) In
vivo role of truncated trkb receptors during sensory ganglion
neurogenesis. Neuroscience, 117: 847–858.
Lyons, W.E., Mamounas, L.A., Ricaurte, G.A., Coppola, V.,
Reid, S.W., Bora, S.H., Wihler, C., Koliatsos, V.E. and
Tessarollo, L. (1999) Brain-derived neurotrophic factor-defi-
cient mice develop aggressiveness and hyperphagia in con-
junction with brain serotonergic abnormalities. Proc. Natl.
Acad. Sci. U.S.A., 96: 15239–15244.
Maisonpierre, P.C., Belluscio, L., Friedman, B., Alderson,
R.F., Wiegand, S.J., Furth, M.E., Lindsay, R.M. and
Yancopoulos, G.D. (1990a) NT-3, BDNF, and NGF in the
developing rat nervous system: parallel as well as reciprocal
patterns of expression. Neuron, 5: 501–509.
Maisonpierre, P.C., Belluscio, L., Squinto, S., Ip, N.Y., Furth,
M.E., Lindsay, R.M. and Yancopoulos, G.D. (1990b)

Neurotrophin-3: a neurotrophic factor related to NGF and
BDNF. Science, 247: 1446–1451.
Malcangio, M. and Lessmann, V. (2003) A common thread for
pain and memory synapses? Brain-derived neurotrophic fac-
tor and trkB receptors. Trends Pharmacol. Sci., 24: 116–121.
Mamounas, L.A., Blue, M.E., Siuciak, J.A. and Altar, C.A.
(1995) Brain-derived neurotrophic factor promotes the sur-
vival and sprouting of serotonergic axons in rat brain. J.
Neurosci., 15: 7929–7939.
Marksteiner, J., Ortler, M., Bellmann, R. and Sperk, G. (1990)
Neuropeptide Y biosynthesis is markedly induced in mossy
fibers during temporal lobe epilepsy of the rat. Neurosci.
Lett., 112: 143–148.
Martinowich, K., Hattori, D., Wu, H., Fouse, S., He, F., Hu,
Y., Fan, G. and Sun, Y.E. (2003) DNA methylation-related
chromatin remodeling in activity-dependent BDNF gene reg-
ulation. Science, 302: 890–893.
Marty, S., Berninger, B., Carroll, P. and Thoenen, H. (1996)
GABAergic stimulation regulates the phenotype of hippo-
campal interneurons through the regulation of brain-derived
neurotrophic factor. Neuron, 16: 565–570.
Mathern, G.W., Babb, T.L., Micevych, P.E., Blanco, C.E. and
Pretorius, J.K. (1997) Granule cell mRNA levels for BDNF,
NGF, and NT-3 correlate with neuron losses or supragran-
ular mossy fiber sprouting in the chronically damaged and
epileptic human hippocampus. Mol. Chem. Neuropathol.,
30: 53–76.
McAllister, A.K., Katz, L.C. and Lo, D.C. (1997) Opposing
roles for endogenous BDNF and NT-3 in regulating cortical
dendritic growth. Neuron, 18: 767–778.
McLean Bolton, M., Pittman, A.J. and Lo, D.C. (2000) Brain-
derived neurotrophic factor differentially regulates excitatory
and inhibitory synaptic transmission in hippocampal cul-
tures. J. Neurosci., 20: 3221–3232.
Merlio, J.P., Ernfors, P., Jaber, M. and Persson, H. (1992)
Molecular cloning of rat trkC and distribution of cells ex-
pressing messenger RNAs for members of the trk family in
the rat central nervous system. Neuroscience, 51: 513–532.
Merlio, J.P., Ernfors, P., Kokaia, Z., Middlemas, D.S., Ben-
gzon, J., Kokaia, M., Smith, M.L., Seisjo, B.K., Hunter, T.
and Lindvall, O. (1993) Increased production of the trkB
protein tyrosine kinase receptor after brain insults. Neuron,
10: 151–164.
Metsis, M., Timmusk, T., Arenas, E. and Persson, H. (1993)
Differential usage of multiple brain-derived neurotrophic
factor promoters in the rat brain following neuronal activa-
tion. Proc. Natl. Acad. Sci. U.S.A., 90: 8802–8806.
Miller, F.D. and Kaplan, D.R. (2001) On Trk for retrograde
signaling. Neuron, 32: 767–770.
Minichiello, L., Calella, A.M., Medina, D.L., Bonhoeffer, T.,
Klein, R. and Korte, M. (2002) Mechanism of TrkB-medi-
ated hippocampal long-term potentiation. Neuron, 36:
121–137.
Minichiello, L., Korte, M., Wolfer, D., Kuhn, R., Unsicker, K.,
Cestari, V., Rossi-Arnaud, C., Lipp, H.P., Bonhoeffer, T.
and Klein, R. (1999) Essential role for TrkB receptors in
hippocampus-mediated learning. Neuron, 24: 401–414.
393
Morimoto, K., Sato, K., Sato, S., Yamada, N. and Hayabara,
T. (1998) Time-dependent changes in neurotrophic factor
mRNA expression after kindling and long-term potentiation
in rats. Brain Res. Bull., 45: 599–605.
Mowla, S.J., Farhadi, H.F., Pareek, S., Atwal, J.K., Morris,
S.J., Seidah, N.G. and Murphy, R.A. (2001) Biosynthesis and
post-translational processing of the precursor to brain-de-
rived neurotrophic factor. J. Biol. Chem., 276: 12660–12666.
Mowla, S.J., Pareek, S., Farhadi, H.F., Petrecca, K., Fawcett,
J.P., Seidah, N.G., Morris, S.J., Sossin, W.S. and Murphy,
R.A. (1999) Differential sorting of nerve growth factor and
brain-derived neurotrophic factor in hippocampal neurons. J.
Neurosci., 19: 2069–2080.
Mudo, G., Jiang, X.H., Timmusk, T., Bindoni, M. and Bell-
uardo, N. (1996) Change in neurotrophins and their receptor
mRNAs in the rat forebrain after status epilepticus induced
by pilocarpine. Epilepsia, 37: 198–207.
Mudo, G., Salin, T., Condorelli, D.F., Jiang, X.H., Dell’Al-
bani, P., Timmusk, T., Metsis, M., Funakoshi, H. and Bell-
uardo, N. (1995) Seizures increase trkC mRNA expression in
the dentate gyrus of rat hippocampus. Role of glutamate re-
ceptor activation. J. Mol. Neurosci., 6: 11–22.
Murer, M.G., Yan, Q. and Raisman-Vozari, R. (2001) Brain-
derived neurotrophic factor in the control human brain, and
in Alzheimer’s disease and Parkinson’s disease. Prog. Ne-
urobiol., 63: 71–124.
Murphy, D.D., Cole, N.B. and Segal, M. (1998) Brain-derived
neurotrophic factor mediates estradiol-induced dendritic
spine formation in hippocampal neurons. Proc. Natl. Acad.
Sci. U.S.A., 95: 11412–11417.
Nakata, K., Ujike, H., Sakai, A., Uchida, N., Nomura, A.,
Imamura, T., Katsu, T., Tanaka, Y., Hamamura, T. and
Kuroda, S. (2003) Association study of the brain-derived ne-
urotrophic factor (BDNF) gene with bipolar disorder. Ne-
urosci. Lett., 337: 17–20.
Narisawa-Saito, M. and Nawa, H. (1996) Differential regula-
tion of hippocampal neurotrophins during aging in rats. J.
Neurochem., 67: 1124–1131.
Nawa, H., Carnahan, J. and Gall, C. (1995) BDNF protein
measured by a novel enzyme immunoassay in normal brain
and after seizure: partial disagreement with mRNA levels.
Eur. J. Neurosci., 7: 1527–1535.
Nawa, H., Pelleymounter, M.A. and Carnahan, J. (1994) In-
traventricular administration of BDNF increases neuropep-
tide expression in newborn rat brain. J. Neurosci., 14:
3751–3765.
Neeper, S.A., Gomez-Pinilla, F., Choi, J. and Cotman, C.
(1995) Exercise and brain neurotrophins. Nature, 373: 109.
Neves-Pereira, M., Mundo, E., Muglia, P., King, N., Mac-
ciardi, F. and Kennedy, J.L. (2002) The brain-derived ne-
urotrophic factor gene confers susceptibility to bipolar
disorder: evidence from a family-based association study.
Am. J. Hum. Genet., 71: 651–655.
Nibuya, M., Morinobu, S. and Duman, R.S. (1995) Regulation
of BDNF and trkB mRNA in rat brain by chronic electro-
convulsive seizure and antidepressant drug treatments. J.
Neurosci., 15: 7539–7547.
394
Nibuya, M., Nestler, E.J. and Duman, R.S. (1996) Chronic
antidepressant administration increases the expression of
cAMP response element binding protein (CREB) in rat hip-
pocampus. J. Neurosci., 16: 2365–2372.
Osehobo, P., Adams, B., Sazgar, M., Verdi, J., Racine, R. and
Fahnestock, M. (1996) Effects of in vivo BDNF infusion on
amygdala kindling, sprouting, and hilar area. Soc. Neurosci.
Abstr., 22: 995.
Patapoutian, A. and Reichardt, L.F. (2001) Trk receptors: me-
diators of neurotrophin action. Curr. Opin. Neurobiol., 11:
272–280.
Patel, M.N. and McNamara, J.O. (1995) Selective enhancement
of axonal branching of cultured dentate gyrus neurons by
neurotrophic factors. Neuroscience, 69: 763–770.
Patterson, S.L., Abel, T., Deuel, T.A., Martin, K.C., Rose, J.C.
and Kandel, E.R. (1996) Recombinant BDNF rescues deficits
in basal synaptic transmission and hippocampal LTP in
BDNF knockout mice. Neuron, 16: 1137–1145.
Patterson, S.L., Grover, L.M., Schwartzkroin, P.A. and
Bothwell, M. (1992) Neurotrophin expression in rat hip-
pocampal slices: a stimulus paradigm inducing LTP in CA1
evokes increases in BDNF and NT-3 mRNAs. Neuron, 9:
1081–1088.
Pencea, V., Bingaman, K.D., Wiegand, S.J. and Luskin, M.B.
(2001) Infusion of brain-derived neurotrophic factor into the
lateral ventricle of the adult rat leads to new neurons in the
parenchyma of the striatum, septum, thalamus, and hypo-
thalamus. J. Neurosci., 21: 6706–6717.
Pezet, S., Malcangio, M., Lever, I.J., Perkinton, M.S., Thomp-
son, S.W., Williams, R.J. and McMahon, S.B. (2002) Nox-
ious stimulation induces Trk receptor and downstream ERK
phosphorylation in spinal dorsal horn. Mol. Cell Neurosci.,
21: 684–695.
Phillips, H.S., Hains, J.M., Armanini, M., Laramee, G.R.,
Johnson, S.A. and Winslow, J.W. (1991) BDNF mRNA is
decreased in the hippocampus of individuals with Al-
zheimer’s disease. Neuron, 7: 695–702.
Pioro, E.P. and Cuello, A.C. (1990) Distribution of nerve
growth factor receptor-like immunoreactivity in the adult rat
central nervous system. Effect of colchicine and correlation
with the cholinergic system–I. Forebrain. Neuroscience, 34:
57–87.
Poo, M.M. (2001) Neurotrophins as synaptic modulators. Nat.
Rev. Neurosci., 2: 24–32.
Qiao, X., Hughes, P.E., Venero, J.L., Dugich-Djordjevic,
M.M., Nichols, N.R., Hefti, F. and Knusel, B. (1996) NT-
4/5 protects against adrenalectomy-induced apoptosis of rat
hippocampal granule cells. Neuroreport, 7: 682–686.
Qiao, X., Suri, C., Knusel, B. and Noebels, J.L. (2001) Absence
of hippocampal mossy fiber sprouting in transgenic mice
overexpressing brain-derived neurotrophic factor. J. Neuro-
sci. Res., 64: 268–276.
Rashid, K., Van der Zee, C.E., Ross, G.M., Chapman, C.A.,
Stanisz, J., Riopelle, R.J., Racine, R.J. and Fahnestock, M.
(1995) A nerve growth factor peptide retards seizure devel-
opment and inhibits neuronal sprouting in a rat model of
epilepsy. Proc. Natl. Acad. Sci. U.S.A., 92: 9495–9499.
Reibel, S., Larmet, Y., Carnahan, J., Marescaux, C. and De-
paulis, A. (2000a) Endogenous control of hippocampal epile-
ptogenesis: a molecular cascade involving brain-derived
neurotrophic factor and neuropeptide Y. Epilepsia, 41(Sup-
pl 6): S127–S133.
Reibel, S., Larmet, Y., Le, B.T., Carnahan, J., Marescaux, C.
and Depaulis, A. (2000b) Brain-derived neurotrophic factor
delays hippocampal kindling in the rat. Neuroscience, 100:
777–788.
Riccio, A., Pierchala, B.A., Ciarallo, C.L. and Ginty, D.D.
(1997) An NGF-trkA-mediated retrograde signal to tran-
scription factor CREB in sympathetic neurons. Science, 277:
1097–1100.
Rios, M., Fan, G., Fekete, C., Kelly, J., Bates, B., Kuehn, R.,
Lechan, R.M. and Jaenisch, R. (2001) Conditional deletion
of brain-derived neurotrophic factor in the postnatal brain
leads to obesity and hyperactivity. Mol. Endocrinol., 15:
1748–1757.
Rivera, C., Li, H., Thomas-Crusells, J., Lahtinen, H., Viitanen,
T., Nanobashvili, A., Kokaia, Z., Airaksinen, M.S., Voipio,
J., Kaila, K. and Saarma, M. (2002) BDNF-induced TrkB
activation down-regulates the K+-Cl- cotransporter KCC2
and impairs neuronal Cl- extrusion. J. Cell Biol., 159:
747–752.
Roback, J.D., Marsh, H.N., Downen, M., Palfrey, H.C. and
Wainer, B.H. (1995) BDNF-activated signal transduction in
rat cortical glial cells. Eur. J. Neurosci., 7: 849–862.
Rocamora, N., Welker, E., Pascual, M. and Soriano, E. (1996)
Upregulation of BDNF mRNA expression in the barrel cor-
tex of adult mice after sensory stimulation. J. Neurosci., 16:
4411–4419.
Rose, C.R., Blum, R., Pichler, B., Lepier, A., Kafitz, K.W. and
Konnerth, A. (2003) Truncated TrkB-T1 mediates neurotro-
phin-evoked calcium signalling in glia cells. Nature, 426:
74–78.
Roux, P.P., Colicos, M.A., Barker, P.A. and Kennedy, T.E.
(1999) p75 neurotrophin receptor expression is induced in
apoptotic neurons after seizure. J. Neurosci., 19: 6887–6896.
Rudge, J.S., Mather, P.E., Pasnikowski, E.M., Cai, N., Corco-
ran, T., Acheson, A., Anderson, K., Lindsay, R.M. and
Wiegand, S.J. (1998) Endogenous BDNF protein is increased
in adult rat hippocampus after a kainic acid induced excito-
toxic insult but exogenous BDNF is not neuroprotective.
Exp. Neurol., 149: 398–410.
Russo-Neustadt, A., Beard, R.C. and Cotman, C.W. (1999)
Exercise, antidepressant medications, and enhanced brain
derived neurotrophic factor expression. Neuropsychophar-
macology, 21: 679–682.
Ryden, M. and Ibanez, C.F. (1996) Binding of neurotrophin-3
to p75LNGFR, TrkA, and TrkB mediated by a single func-
tional epitope distinct from that recognized by trkC. J. Biol.
Chem., 271: 5623–5627.
Rylett, R.J. and Williams, L.R. (1994) Role of neurotrophins in
cholinergic-neurone function in the adult and aged CNS.
Trends Neurosci., 17: 486–490.
Saarelainen, T., Hendolin, P., Lucas, G., Koponen, E., Sa-
iranen, M., MacDonald, E., Agerman, K., Haapasalo, A.,

Nawa, H., Aloyz, R., Ernfors, P. and Castren, E. (2003) Ac-
tivation of the TrkB neurotrophin receptor is induced by
antidepressant drugs and is required for antidepressant-in-
duced behavioral effects. J. Neurosci., 23: 349–357.
Saarelainen, T., Lukkarinen, J.A., Koponen, S., Grohn, O.H.,
Jolkkonen, J., Koponen, E., Haapasalo, A., Alhonen, L.,
Wong, G., Koistinaho, J., Kauppinen, R.A. and Castren, E.
(2000a) Transgenic mice overexpressing truncated trkB ne-
urotrophin receptors in neurons show increased susceptibility
to cortical injury after focal cerebral ischemia. Mol. Cell Ne-
urosci., 16: 87–96.
Saarelainen, T., Pussinen, R., Koponen, E., Alhonen, L.,
Wong, G., Sirvio, J. and Castren, E. (2000b) Transgenic mice
overexpressing truncated trkB neurotrophin receptors in
neurons have impaired long-term spatial memory but nor-
mal hippocampal LTP. Synapse, 38: 102–104.
Sato, K., Kashihara, K., Morimoto, K. and Hayabara, T.
(1996) Regional increases in brain-derived neurotrophic fac-
tor and nerve growth factor mRNAs during amygdaloid
kindling, but not in acidic and basic fibroblast growth factor
mRNAs. Epilepsia, 37: 6–14.
Scharfman, H., Goodman, J., Macleod, A., Phani, S., Anton-
elli, C. and Croll, S. (2005) Increased neurogenesis and the
ectopic granule cells after intrahippocampal BDNF infusion
in adult rats. Exp. Neurol., 192: 348–356.
Scharfman, H.E. (1997) Hyperexcitability in combined ent-
orhinal/hippocampal slices of adult rat after exposure to
brain-derived neurotrophic factor. J. Neurophysiol., 78:
1082–1095.
Scharfman, H.E. (2004) Functional implications of seizure-in-
duced neurogenesis. Adv. Exp. Med. Biol., 548: 192–212.
Scharfman, H.E., Goodman, J.H. and Sollas, A.L. (1999) Ac-
tions of brain-derived neurotrophic factor in slices from rats
with spontaneous seizures and mossy fiber sprouting in the
dentate gyrus. J. Neurosci., 19: 5619–5631.
Scharfman, H.E., Goodman, J.H., Sollas, A.L. and Croll, S.D.
(2002) Spontaneous limbic seizures after intrahippocampal
infusion of brain-derived neurotrophic factor. Exp. Neurol.,
174: 201–214.
Scharfman, H.E., Mercurio, T.C., Goodman, J.H., Wilson,
M.A. and MacLusky, N.J. (2003) Hippocampal excitability
increases during the estrous cycle in the rat: a potential role
for brain-derived neurotrophic factor. J. Neurosci., 23:
11641–11652.
Schinder, A.F. and Poo, M. (2000) The neurotrophin hy-
pothesis for synaptic plasticity. Trends Neurosci., 23:
639–645.
Schmidt-Kastner, R., Humpel, C., Wetmore, C. and Olson, L.
(1996) Cellular hybridization for BDNF, trkB, and NGF
mRNAs and BDNF-immunoreactivity in rat forebrain after
pilocarpine-induced status epilepticus. Exp. Brain Res., 107:
331–347.
Schmidt-Kastner, R. and Olson, L. (1995) Decrease of ne-
urotrophin-3 mRNA in adult rat hippocampus after pilocar-
pine seizures. Exp. Neurol., 136: 199–204.
Segal, R.A. (2003) Selectivity in neurotrophin signaling: theme
and variations. Annu. Rev. Neurosci., 26: 299–330.
395
Senger, D.L. and Campenot, R.B. (1997) Rapid retrograde ty-
rosine phosphorylation of trkA and other proteins in rat
sympathetic neurons in compartmented cultures. J. Cell Biol.,
138: 411–421.
Shieh, P.B. and Ghosh, A. (1999) Molecular mechanisms un-
derlying activity-dependent regulation of BDNF expression.
J. Neurobiol., 41: 127–134.
Shieh, P.B., Hu, S.C., Bobb, K., Timmusk, T. and Ghosh, A.
(1998) Identification of a signaling pathway involved in
calcium regulation of BDNF expression. Neuron, 20:
727–740.
Shirayama, Y., Chen, A.C., Nakagawa, S., Russell, D.S. and
Duman, R.S. (2002) Brain-derived neurotrophic factor pro-
duces antidepressant effects in behavioral models of depres-
sion. J. Neurosci., 22: 3251–3261.
Simonato, M., Bregola, G., Armellin, M., Del Piccolo, P., Rodi,
D., Zucchini, S. and Tongiorgi, E. (2002) Dendritic targeting
of mRNAs for plasticity genes in experimental models of
temporal lobe epilepsy. Epilepsia, 43(Suppl 5): 153–158.
Sklar, P., Gabriel, S.B., McInnis, M.G., Bennett, P., Lim,
Y.M., Tsan, G., Schaffner, S., Kirov, G., Jones, I., Owen,
M., Craddock, N., DePaulo, J.R. and Lander, E.S. (2002)
Family-based association study of 76 candidate genes in bi-
polar disorder: BDNF is a potential risk locus. Brain-derived
neutrophic factor. Mol. Psychiatry, 7: 579–593.
Smith, M.A., Makino, S., Kvetnansky, R. and Post, R.M.
(1995) Stress and glucocorticoids affect the expression of
brain-derived neurotrophic factor and neurotrophin-3
mRNAs in the hippocampus. J. Neurosci., 15: 1768–1777.
Smith, M.A., Zhang, L.-X., Lyons, W.E. and Mamounas, L.A.
(1997) Anterograde transport of endogenous brain-derived
neurotrophic factor in hippocampal mossy fibers. Neurore-
port, 8: 1829–1834.
Sobreviela, T., Clary, D.O., Reichardt, L.F., Brandabur, M.M.,
Kordower, J.H. and Mufson, E.J. (1994) TrkA-immunore-
active profiles in the central nervous system: colocalization
with neurons containing p75 nerve growth factor receptor,
choline acetyltransferase, and serotonin. J. Comp. Neurol.,
350: 587–611.
Spires, T.L., Grote, H.E., Varshney, N.K., Cordery, P.M., van
Dellen, A., Blakemore, C. and Hannan, A.J. (2004) Envi-
ronmental enrichment rescues protein deficits in a mouse
model of Huntington’s disease, indicating a possible disease
mechanism. J. Neurosci., 24: 2270–2276.
Spruston, N., Lubke, J. and Frotscher, M. (1997) Interneurons
in the stratum lucidum of the rat hippocampus: an anatom-
ical and electrophysiological characterization. J. Comp. Ne-
urol., 385: 427–440.
Suen, P.-C., Wu, K., Levine, E.S., Mount, H.T.J., Xu, J.-L.,
Lin, S.-Y. and Black, I.B. (1997) Brain-derived neurotrophic
factor rapidly enhances phosphorylation of the postsynaptic
N-methyl-D-aspartate receptor subunit 1. Proc. Natl. Acad.
Sci. U.S.A., 94: 8191–8195.
Takahashi, M., Hayashi, S., Kakita, A., Wakabayashi, K.,
Fukuda, M., Kameyama, S., Tanaka, R., Takahashi, H. and
Nawa, H. (1999) Patients with temporal lobe epilepsy show
an increase in brain-derived neurotrophic factor protein and
396
its correlation with neuropeptide Y. Brain Res., 818:
579–582.
Tanaka, T., Saito, H. and Matsuki, N. (1997) Inhibition of
GABAa synaptic responses by brain-derived neurotrophic
factor (BDNF) in rat hippocampus. J. Neurosci., 17:
2959–2966.
Tao, X., Finkbeiner, S., Arnold, D.B., Shaywitz, A.J. and
Greenberg, M.E. (1998) Ca2+ influx regulates BDNF tran-
scription by a CREB family transcription factor-dependent
mechanism. Neuron, 20: 709–726.
Tao, X., West, A.E., Chen, W.G., Corfas, G. and Greenberg,
M.E. (2002) A calcium-responsive transcription factor,
CaRF, that regulates neuronal activity-dependent expression
of BDNF. Neuron, 33: 383–395.
Thakker-Varia, S., Alder, J., Crozier, R.A., Plummer, M.R.
and Black, I.B. (2001) Rab3A is required for brain-derived
neurotrophic factor-induced synaptic plasticity: transcrip-
tional analysis at the population and single-cell levels. J. Ne-
urosci., 21: 6782–6790.
Thoenen, H. (1995) Neurotrophins and neuronal plasticity.
Science, 270: 593–598.
Thompson, S.W., Bennett, D.L., Kerr, B.J., Bradbury, E.J. and
McMahon, S.B. (1999) Brain-derived neurotrophic factor is
an endogenous modulator of nociceptive responses in the
spinal cord. Proc. Natl. Acad. Sci. U.S.A., 96: 7714–7718.
Timmusk, T., Belluardo, N., Metsis, M. and Persson, H.
(1993a) Widespread and developmentally regulated expres-
sion of neurotrophin-4 mRNA in rat brain and peripheral
tissues. Eur. J. Neurosci., 5: 605–613.
Timmusk, T., Palm, K., Metsis, M., Reintam, T., Paalme, V.,
Saarma, M. and Persson, H. (1993b) Multiple promoters di-
rect tissue-specific expression of the rat BDNF gene. Neuron,
10: 475–489.
Tolwani, R.J., Buckmaster, P.S., Varma, S., Cosgaya, J.M.,
Wu, Y., Suri, C. and Shooter, E.M. (2002) BDNF overex-
pression increases dendrite complexity in hippocampal dent-
ate gyrus. Neuroscience, 114: 795–805.
Tønder, N., Kragh, J., Finsen, B., Bolwig, T.G. and Zimmer, J.
(1994) Kindling induces transient changes in neuronal ex-
pression of somatostatin, neuropeptide Y, and calbindin in
adult rat hippocampus and fascia dentata. Epilepsia, 35:
1299–1308.
Tonra, J.R., Curtis, R., Wong, V., Cliffer, K.D., Park, J.S.,
Timmes, A., Nguyen, T., Lindsay, R.M., Acheson, A. and
DiStefano, P.S. (1998) Axotomy upregulates the anterograde
transport and expression of brain-derived neurotrophic fac-
tor by sensory neurons. J. Neurosci., 18: 4374–4383.
Tsai, S.J. (2004) Is mania caused by overactivity of central
brain-derived neurotrophic factor? Med. Hypotheses, 62:
19–22.
Tyler, W.J., Perrett, S.P. and Pozzo-Miller, L.D. (2002) The
role of neurotrophins in neurotransmitter release. Neurosci-
entist, 8: 524–531.
Urfer, R., Tsoulfas, P., O’Connell, L., Shelton, D.L., Parada,
L.F. and Presta, L.G. (1995) An immunoglobulin-like do-
main determines the specificity of neurotrophin receptors.
EMBO J., 14: 2795–2805.
Van der Zee, C.E., Rashid, K., Le, K., Moore, K.A., Stanisz, J.,
Diamond, J., Racine, R.J. and Fahnestock, M. (1995) Intra-
ventricular administration of antibodies to nerve growth fac-
tor retards kindling and blocks mossy fiber sprouting in adult
rats. J. Neurosci., 15: 5316–5323.
Vezzani, A., Ravizza, T., Moneta, D., Conti, M., Borroni,
A., Rizzi, M., Samanin, R. and Maj, R. (1999) Brain-
derived neurotrophic factor immunoreactivity in the limbic
system of rats after acute seizures and during spontaneo
us convulsions: temporal evolution of changes as
compared to neuropeptide Y. Neuroscience, 90:
1445–1461.
Von Bartheld, C.S., Byers, M.R., Williams, R. and Bothwell,
M. (1996a) Anterograde transport of neurotrophins and
axodendritic transfer in the developing visual system. Nature,
379: 830–833.
Von Bartheld, C.S., Williams, R., Lefcort, F., Clary, D.O.,
Reichardt, L.F. and Bothwell, M. (1996b) Retrograde trans-
port of neurotrophins from the eye to the brain in chick
embryos: roles of the p75NTR and trkB receptors. J. Ne-
urosci., 16: 2995–3008.
Wardle, R.A. and Poo, M.M. (2003) Brain-derived neurotro-
phic factor modulation of GABAergic synapses by postsy-
naptic regulation of chloride transport. J. Neurosci., 23:
8722–8732.
Wetmore, C., Ernfors, P., Persson, H. and Olson, L. (1990)
Localization of brain-derived neurotrophic factor mRNA to
neurons in the brain by in situ hybridization. Exp. Neurol.,
109: 141–152.
Wu, K., Xu, J., Suen, P., Levine, E., Huang, Y., Mount, H.T.J.,
Lin, S. and Black, I.B. (1996) Functional trkB neurotrophin
receptors are intrinsic components of the adult brain postsy-
naptic density. Mol. Brain Res., 43: 286–290.
Xu, B., Gottschalk, W., Chow, A., Wilson, R.I., Schnell, E.,
Zang, K., Wang, D., Nicoll, R.A., Lu, B. and Reichardt, L.F.
(2000) The role of brain-derived neurotrophic factor recep-
tors in the mature hippocampus: modulation of long-term
potentiation through a presynaptic mechanism involving
TrkB. J. Neurosci., 20: 6888–6897.
Xu, B., Goulding, E.H., Zang, K., Cepoi, D., Cone, R.D.,
Jones, K.R., Tecott, L.H. and Reichardt, L.F. (2003) Brain-
derived neurotrophic factor regulates energy balance down-
stream of melanocortin-4 receptor. Nat. Neurosci., 6:
736–742.
Xu, B., Michalski, B., Racine, R.J. and Fahnestock, M. (2002)
Continuous infusion of neurotrophin-3 triggers sprouting,
decreases the levels of TrkA and TrkC, and inhibits epile-
ptogenesis and activity-dependent axonal growth in adult
rats. Neuroscience, 115: 1295–1308.
Xu, B., Michalski, B., Racine, R.J. and Fahnestock, M. (2004)
The effects of brain-derived neurotrophic factor (BDNF)
administration on kindling induction, Trk expression and
seizure-related morphological changes. Neuroscience, 126:
521–531.
Yacoubian, T.A. and Lo, D.C. (2000) Truncated and full-
length TrkB receptors regulate distinct modes of dendritic
growth. Nat. Neurosci., 3: 342–349.

Yamada, K. and Nabeshima, T. (2003) Brain-derived neurotro-
phic factor/TrkB signaling in memory processes. J. Pharma-
col. Sci., 91: 267–270.
Yan, Q., Matheson, C., Sun, J., Radeke, M.J., Feinstein, S.C.
and Miller, J.A. (1994) Distribution of intracerebral ven-
tricularly administered neurotrophins in rat brain and its
correlation with trk receptor expression. Exp. Neurol., 127:
23–36.
Yan, Q., Radeke, M.J., Matheson, C.R., Talvenheimo, J.,
Welcher, A.A. and Feinstein, S.C. (1997a) Immunocyto-
chemical localization of trkB in the central nervous system of
the adult rat. J. Comp. Neurol., 378: 135–157.
Yan, Q., Rosenfeld, R.D., Matheson, C.R., Hawkins, N.,
Lopez, O.T., Bennett, L. and Welcher, A.A. (1997b) Ex-
pression of brain-derived neurotrophic factor protein in the
adult rat central nervous system. Neuroscience, 78:
431–448.
Zaccaro, M.C., Ivanisevic, L., Perez, P., Meakin, S.O. and
Saragovi, H.U. (2001) p75 Co-receptors regulate ligand-
dependent and ligand-independent Trk receptor activation,
in part by altering Trk docking subdomains. J. Biol. Chem.,
276: 31023–31029.
Zhou, X.F. and Rush, R.A. (1994) Localization of neurotro-
phin-3-like immunoreactivity in the rat central nervous sys-
tem. Brain Res., 643: 162–172.
397
Zhou, X.-F. and Rush, R.A. (1996) Endogenous brain-derived
neurotrophic factor is anterogradely transported in primary
sensory neurons. Neuroscience, 74: 945–951.
Zhou, Z., Hong, E.J., Cohen, S., Zhao, W.N., Ho, H.Y., Sch-
midt, L., Chen, W.G., Lin, Y., Savner, E., Griffith, E.C., Hu,
L., Steen, J.A., Weitz, C.J. and Greenberg, M.E. (2006)
Brain-specific phosphorylation of MeCP2 regulates activity-
dependent Bdnf transcription, dendritic growth, and spine
maturation. Neuron, 52: 255–269.
Zigova, T., Pencea, V., Wiegand, S.J. and Luskin, M.B. (1998)
Intraventricular administration of BDNF increases the
number of newly generated neurons in the adult olfactory
bulb. Mol. Cell Neurosci., 11: 234–245.
Zuccato, C., Ciammola, A., Rigamonti, D., Leavitt, B.R.,
Goffredo, D., Conti, L., MacDonald, M.E., Friedlander,
R.M., Silani, V., Hayden, M.R., Timmusk, T., Sipione, S.
and Cattaneo, E. (2001) Loss of huntingtin-mediated BDNF
gene transcription in Huntington’s disease. Science, 293:
493–498.
Zuccato, C., Tartari, M., Crotti, A., Goffredo, D., Valenza, M.,
Conti, L., Cataudella, T., Leavitt, B.R., Hayden, M.R.,
Timmusk, T., Rigamonti, D. and Cattaneo, E. (2003) Hunt-
ingtin interacts with REST/NRSF to modulate the tran-
scription of NRSE-controlled neuronal genes. Nat. Genet.,
35: 76–83.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 23

Sex steroids and the dentate gyrus

Tibor Hajszan1,2,, Teresa A. Milner3,4 and Csaba Leranth1,5

1
Department of Obstetrics, Gynecology, and Reproductive Sciences, Yale University School of Medicine, 333 Cedar Street,
FMB 312, New Haven, CT 06520, USA
2
Department of Biophysics, Biological Research Center, Hungarian Academy of Sciences, Szeged, Hungary
3
Department of Neurology and Neuroscience, Division of Neurobiology, Weill-Cornell Medical College,
411 East 69th Street, New York, NY 10021, USA
4
Harold and Milliken Hatch Laboratory of Neuroendocrinology, The Rockefeller University, New York, NY, USA
5
Department of Neurobiology, Yale University School of Medicine, New Haven, CT, USA

Abstract: In the late 1980s, the finding that the dentate gyrus contains more granule cells in the male than
in the female of certain mouse strains provided the first indication that the dentate gyrus is a significant
target for the effects of sex steroids during development. Gonadal hormones also play a crucial role in
shaping the function and morphology of the adult brain. Besides reproduction-related processes, sex ster-
oids participate in higher brain operations such as cognition and mood, in which the hippocampus is a
critical mediator. Being part of the hippocampal formation, the dentate gyrus is naturally involved in these
mechanisms and as such, this structure is also a critical target for the activational effects of sex steroids.
These activational effects are the results of three major types of steroid-mediated actions. Sex steroids
modulate the function of dentate neurons under normal conditions. In addition, recent research suggests
that hormone-induced cellular plasticity may play a larger role than previously thought, particularly in the
dentate gyrus. Specifically, the regulation of dentate gyrus neurogenesis and synaptic remodeling by sex
steroids received increasing attention lately. Finally, the dentate gyrus is influenced by gonadal hormones
in the context of cellular injury, and the work in this area demonstrates that gonadal hormones have
neuroprotective potential. The expression of estrogen, progestin, and androgen receptors in the dentate
gyrus suggests that sex steroids, which could be of gonadal origin and/or synthesized locally in the dentate
gyrus, may act directly on dentate cells. In addition, gonadal hormones could also influence the dentate
gyrus indirectly, by subcortical hormone-sensitive structures such as the cholinergic septohippocampal
system. Importantly, these three sex steroid-related themes, functional effects in the normal dentate gyrus,
mechanisms involving neurogenesis and synaptic remodeling, as well as neuroprotection, have substantial
implications for understanding normal cognitive function, with clinical importance for epilepsy,
Alzheimer’s disease and mental disorders.

Keywords: androgen; estrogen; progesterone; sex difference; electrophysiology; neurogenesis; synaptic

remodeling; neuroprotection

Introduction

Corresponding author. Tel.: +1 203 785 4748; In 1989, Wimer and Wimer (1989) reported that in
Fax: +1 203 785 7684; E-mail: tibor.hajszan@yale.edu certain mouse strains, the dentate gyrus contains

DOI: 10.1016/S0079-6123(07)63023-4 399

400
more granule cells in the male than in the female.
A few years later, the dentate granule cell layer was
shown to be larger in males relative to females, both
in adult and prepubescent rats, which is well cor-
related with performance in spatial memory tasks
(Roof and Havens, 1992; Roof, 1993). Later, in the
hilus of the dentate gyrus, the number of synapses
formed by mossy fibers, the axons of dentate gran-
ule cells, was demonstrated to be higher in male rats
than in females, consistent with the idea that more
granule cells in males provide a more robust input
to CA3 pyramidal neurons (Parducz and Garcia-
Segura, 1993). Using rigorous stereology tech-
niques, however, two studies performed later have
found no sexual dimorphism in the volume of the
dentate granule cell layer in Sprague-Dawley (Isgor
and Sengelaub, 1998) and Long-Evans rats (Jones
and Watson, 2005). Paying particular attention to
rodent strains in this respect is emphasized by an-
other mouse study, demonstrating that the overall
volume of the dentate granule cell layer in A/J mice
is larger in males than in females, while there is
no such sexual dimorphism in C57Bl/6J mice
(Tabibnia et al., 1999). These key developments in
our understanding of sex differences in the dentate
gyrus were followed by other examples of sexual
dimorphism in certain aspects of dentate gyrus
function and morphology. For example, female
rats produce more newly born cells than males in
the dentate gyrus, but not in the subventricular
zone (Tanapat et al., 1999).
Expression of proteins regulated by estrogen,
such as brain-derived neurotrophic factor (BDNF)
(Sohrabji et al., 1995), also differs in males vs.
females. Using immunocytochemistry, it appears
that the mossy fibers contain the vast majority of
BDNF protein (Conner et al., 1997). When a pro-
estrous or estrous female rat with high estrogen
levels was compared to a metestrous female,
ovariectomized female, or male, which have low
estrogen blood concentrations, BDNF protein
expression in the mossy fibers was relatively low
(Scharfman et al., 2003). These studies are
consistent with BDNF expression in mossy fiber-
containing micropunches of CA3 assayed by
ELISA, demonstrating a higher level of BDNF
in the female relative to the male rat (Franklin and
Perrot-Sinal, 2006).
The finding that prepubescent female rats in-
jected neonatally with testosterone develop a male-
like dentate gyrus and perform better in the Morris
water maze (Roof, 1993) suggests that the ‘‘gen-
der’’ of the dentate gyrus and many other sexually
dimorphic brain functions and structures could be
readily manipulated via the hormonal milieu. The
window during which these hormonal manipula-
tions can affect dentate gyrus morphology seems
to be restricted to the first few postnatal days
(Roof, 1993), because no morphological responses
to sex steroid treatment in the rat dentate gyrus
were found before or after this period (Isgor and
Sengelaub, 1998). The findings of Roof (1993)
support the so-called ‘‘aromatization hypothesis,’’
i.e., the brain is masculinized by estrogen that is
produced by aromatase from testosterone, which
is readily available in the developing male but not
in the female (Naftolin et al., 1975). Also consist-
ent with this hypothesis, estrogen receptor (ER)
mRNA levels in the dentate gyrus increase signifi-
cantly between birth and postnatal day 4, and then
decline by postnatal day 10, while adult male rat
ER mRNA levels are similar to those found in
newborn and postnatal day 10 animals (O’Keefe
et al., 1995). On the other hand, it has been
reported that the dentate granule cell layer is sig-
nificantly larger in testicular feminization mutant
(Tfm) male rats than in wild-type females (Jones
and Watson, 2005). Because Tfm rats express a
dysfunctional androgen receptor (AR), this finding
implicates the AR in the development of the
dentate gyrus, which contradicts the aromatization
hypothesis. It should be noted, however, that Tfm
rats retain a considerable portion of AR activity,
whereas the dentate granule cell layer is not
sexually dimorphic in Tfm mice with complete
deletion of AR function (Tabibnia et al., 1999).
Sex steroids and dentate physiology
The findings of sexual dimorphism strongly
suggest that sex steroids, and their receptors, are
important regulators of dentate gyrus organiza-
tion, and emphasize their important role in devel-
opment. The rest of this review, however, will
focus on more recent and exciting research on the

activational effects of gonadal hormones in the
adult brain. In adulthood, sex steroids participate
not only in reproduction-related central actions,
but also in higher brain functions such as cogni-
tion and mood regulation (Korol and Kolo, 2002;
Seidman, 2003; Steiner et al., 2003; MacLusky
et al., 2006). Because the hippocampal formation
plays an essential role in declarative, spatial, and
contextual memory, as well as in the regulation
of mood and the hypothalamic-pituitary-adrenal
axis, this limbic structure is a critical target of
hormone action (McEwen and Alves, 1999;
McEwen, 2003). As a part of the hippocampal
formation, it is highly likely that the dentate gyrus
also plays an important role in cognitive function
and mood regulation. Although research has made
progress in associating hippocampal subregions
with specific functional roles in complex topics
such as cognition and mood, proving that any
specific aspect of these behaviors is dependent on
the dentate gyrus has been more difficult, probably
because complex hippocampal operations require
uncompromised signal flow throughout the entire
hippocampal formation, rather than only in select
areas.
Electrophysiological studies may provide the
most specific insights into the ways sex steroids
influence the adult dentate gyrus. Working with
rat hippocampal slices in the presence or absence
of 17b-estradiol, Kim et al. (2006) have reported
that 17b-estradiol significantly potentiates the am-
plitude and slope of field excitatory postsynaptic
potentials in dentate gyrus directly, as well as in
CA3 following mossy fiber stimulation. Repetitive
hilar stimuli frequently evoke multiple population
spikes in CA3 at proestrus and estrus, but only
rarely at other cycle stages, and never in slices of
ovariectomized rats (Scharfman et al., 2003). This
hyperexcitability in CA3 at proestrus was blocked
by exposure to the high-affinity neurotrophin re-
ceptor antagonist K252a, or by an antagonist of
the a7 nicotinic cholinergic receptor, whereas it
was induced at metestrus by the addition of BDNF
to hippocampal slices (Scharfman et al., 2003).
These findings indicate that an estrogen-induced
interaction of BDNF and a7 nicotinic receptors is
important for estrous cycle-related changes in CA3
and dentate gyrus (Scharfman et al., 2003).
401
Considering androgen, intrahippocampal mi-
croinjection of neither dehydroepiandrosterone-
sulfate (DHEAS) nor trilostane, an inhibitor of the
enzyme that metabolizes DHEAS, alters dentate
field excitatory postsynaptic potential slopes or
population spike amplitudes, but increases the
amplitude of a late component of the postsynaptic
potential. Both DHEAS and trilostane abolishes
GABA-mediated paired-pulse inhibition. In addi-
tion, both DHEAS and trilostane markedly
increases the spontaneous firing rate of dentate
hilar interneurons and synchronizes their firing
during hippocampal theta rhythm induced by tail
pinch (Steffensen, 1995).
Many studies indicate that sex steroid-induced
electrophysiological changes may be due to mod-
ulation of dentate glutamate and GABA recep-
tors. Indeed, ovariectomy in rats decreases [3H]
glutamate binding to N-methyl-D-aspartate
(NMDA) receptors in the dentate gyrus, while
hormone replacement with estradiol, tamoxifen,
or raloxifene prevents this decrease (Cyr et al.,
2000, 2001). On the other hand, [3H] MK-801
binding shows that the density of noncompetitive
NMDA antagonist sites is significantly increased
in the dentate gyrus of ovariectomized compared
to sham-operated rats (El-Bakri et al., 2004).
17b-estradiol returns [3H] MK-801 binding to the
normal levels, while progesterone has no effect
(Weiland, 1992; El-Bakri et al., 2004). In addi-
tion, estradiol treatment of ovariectomized rats
significantly increases NMDA R1 subunit protein
levels in granule cell somata, in comparison with
nontreated animals, without concomitant changes
in the corresponding mRNA hybridization signal
(Gazzaley et al., 1996). Regarding other receptor
types, ovarian steroids have no effect on the
density of kainate or a-amino-3-hydroxy-5-
methylisoxazole-4-propionic acid (AMPA) recep-
tors (Weiland, 1992; Cyr et al., 2000), while est-
radiol-benzoate increases [3H] muscimol binding
in the dentate gyrus (Schumacher et al., 1989). In
situ hybridization showed that progesterone sup-
presses mRNA levels of the a1 GABAA receptor
subunit in the dentate gyrus of animals that were
pretreated with estradiol (Weiland and Orchinik,
1995). Finally, there is a significant negative
correlation between testosterone levels and the
402
mRNA level for the a1 GABAA receptor subunit
(Orchinik et al., 1995).
Sex steroids and dentate plasticity
Although numerous findings support the view that
molecular mechanisms, including receptor
changes, play a critical role in alterations of neu-
ronal activity and functional plasticity (see above),
recent evidence suggests that sex steroids may also
influence long-term potentiation (Lynch, 2004)
and learning/memory via mediating aspects of
structural plasticity, such as neurogenesis and
synaptic remodeling (Kandel, 2001; Kasai et al.,
2003).
Sex steroids and neurogenesis
Due to the recent confirmation that neurogenesis
continues throughout the lifespan in the adult
dentate gyrus (Altman and Das, 1965), the sub-
granular zone, where progenitors are primarily
generated, has drawn considerable attention.
Translational implications are one reason: dentate
neurogenesis may be a critical factor in the patho-
physiology of depression and in the mechanism of
antidepressant action (Santarelli et al., 2003). The
first indication that sex steroids may influence
adult neurogenesis came from Tanapat et al.
(1999), who demonstrated that female rats
produce more newly born cells than males in the
dentate gyrus (but not in the subventricular zone).
They have also reported a fluctuation in cell
proliferation during the estrous cycle: females pro-
duce more newly born cells during proestrus (when
estrogen levels are highest) compared with estrus
and diestrus. Ovariectomy diminishes, while acute
treatment with estrogen rapidly increases, cell
proliferation in ovariectomized rats, an effect that
is reversed by the administration of progesterone
(Tanapat et al., 1999, 2005; Falconer and Galea,
2003). Both the prolonged absence of ovarian
hormones and chronic treatment decreases the
potential of estrogen to stimulate cell proliferation
(Tanapat et al., 2005), suggesting that both dose-
response and temporal characteristics of estrogen
treatment may critically influence its neurogenic
efficacy. Indeed, Ormerod et al. (2003) have
reported that relative to vehicle-treated rats, the
number of new cells increases following a 4-h ex-
posure but decreases following a 48-h exposure to
estrogen in ovariectomized animals. This decrease
at 48 h is abolished by adrenalectomy, suggesting a
role of adrenal activity (Ormerod et al., 2003).
In addition to being effective under normal
conditions, estrogen is capable of influencing ne-
urogenic potential when neurogenesis is examined
in other contexts. For example, a strong reduction
in cell proliferation occurs in the dentate gyrus and
subventricular zone of mice sacrificed 20 days after
streptozotocin administration, which induces a
diabetic state. This reduction is completely re-
lieved by 10 days of estradiol pellet implantation,
which increases the circulating estrogen levels
30-fold (Saravia et al., 2006). In addition, there is
a striking effect of aging on cell proliferation
that appears to be influenced by estradiol. Perez-
Martin et al. (2005) have reported that treatment
of 22-month-old ovariectomized animals for 10
weeks with a weekly subcutaneous injection of
estradiol-valerianate, or with soy extract added to
the drinking water, reverses the age-associated
decline in dentate granule cell production (Perez-
Martin et al., 2005). Similarly, estrogen also
normalizes the deficient granule cell proliferation
in the dentate gyrus of aging mice (De Nicola
et al., 2006).
Several studies have provided insight into the
mechanisms underlying the influence of estrogen
on neurogenesis. Mazzucco et al. (2006) have
shown that both diarylpropionitrile, an ERb ago-
nist, and propyl-pyrazole triol, an ERa agonist,
significantly enhances cell proliferation in the
dentate gyrus of female rats. Other findings sug-
gest that ERs are involved in the induction of
adult neurogenesis by an interaction with insulin-
like growth factor-1 (IGF-1), as estradiol and
IGF-1 have a cooperative effect to promote
neurogenesis (Perez-Martin et al., 2003). Admin-
istration of IGF-1 significantly increases dentate
neurogenesis compared to rats treated with
vehicle; and rats treated with both IGF-1 and
estradiol show a higher level of cell proliferation
than rats treated with IGF-1 or estradiol alone
(Perez-Martin et al., 2003).

Serotonin has also been linked to the effect of
estradiol on dentate neurogenesis. Administration
of 5-hydroxytryptophan, a precursor to serotonin,
restores cell proliferation that was decreased by
ovariectomy, whereas estradiol is unable to reverse
this change in ovariectomized rats treated with
p-chlorophenylalanine, an inhibitor of serotonin
synthesis (Banasr et al., 2001). These data impli-
cate the central serotonergic system in the medi-
ation of estrogen effects on dentate neurogenesis.
Indeed, several studies indicate that estrogen in-
fluences the dentate serotonergic system (Bowman
et al., 2002). In case of 5-HT1A receptors, estra-
diol treatment reduces 5-HT1A gene expression in
the dentate gyrus (Birzniece et al., 2001), while
ovariectomy increases 5-HT1A receptor stimula-
tion, which is reversed by estradiol (Le Saux and
Di Paolo, 2005).
Besides the plethora of data with respect of
estrogen and neurogenesis, there are almost no
published work that address the neurogenic effect
of androgen except an initial study in songbird
suggesting that testosterone promotes neurogene-
sis (Louissaint et al., 2002). Another study has
demonstrated that in a neuronal stem cell culture
stimulated with epidermal growth factor, nandro-
lone, a synthetic androgen reduces cell prolifera-
tion (Brannvall et al., 2005). The decrease is
abolished by flutamide, an AR antagonist. Nand-
rolone also reduces new cell production in the
dentate gyrus, an effect observed in both female
and male rats (Brannvall et al., 2005). For more
details on gonadal hormone modulation of hippo-
campal neurogenesis in the adult, some excellent
reviews are available (Gould et al., 2000; Galea
et al., 2006). For more details on adult neurogen-
esis and its role in mood regulation, we refer the
reader to other chapters within this volume.
Sex steroids and synaptic remodeling
In 1992, the discovery that estradiol mediates fluc-
tuation in hippocampal CA1 spine synapse density
during the estrous cycle in the adult rat marked the
beginning of a new era in the research of sex ster-
oids (Woolley and McEwen, 1992). Subsequent
extensive work has shown that sex steroids hold an
403
unparalleled synaptogenic power in the adult hip-
pocampus. For example, hormone replacement
induces changes on the order of 50–100% in the
number of CA1 spine synapses of rats (MacLusky
et al., 2006; Parducz et al., 2006). This synapto-
genic efficacy seems to be rivaled only by antide-
pressant drugs (Hajszan et al., 2005). Moreover,
estrogen-induced remodeling of hippocampal
spine synapses is remarkably rapid, similar to the
time course required for long-term potentiation
induction (MacLusky et al., 2005). The temporal
analogy suggests that formation of spine synapses
may be involved in sex steroid-modulated
cognitive functions. Indeed, a great deal of evi-
dence has accumulated which suggests that rapid
remodeling and stabilization of small spines
(and their associated synaptic contacts) may rep-
resent a mechanism of memory formation and
storage (Sorra and Harris, 2000; Kandel, 2001;
Kasai et al., 2003).
Unfortunately, several studies support the view
that the dentate gyrus may miss this ‘‘synaptogenic
party.’’ Woolley et al. (1990) have reported no
significant changes in dendritic spine density
across the estrous cycle in CA3 pyramidal cells
or dentate granule cells of the rat. Using a similar
Golgi-impregnation technique, Gould et al. (1990)
have shown that ovariectomy or gonadal steroid
replacement do not affect spine density of CA3
pyramidal cells or granule cells of the dentate
gyrus. In a later study, Miranda et al. (1999) have
demonstrated that there may be effects in the
dentate gyrus, but they are likely to be dependent
on age and the temporal pattern of estradiol re-
placement. In addition, Szymczak et al. (2006)
suggest that ERb expression is negatively corre-
lated with synapse formation.
Another approach to the topic of synaptic re-
modeling is to address the expression of proteins
that are associated with the pre- or postsynaptic
apparatus. This approach has also failed to lead to
a compelling body of evidence that hormonal fluc-
tuations modulate dentate synaptic remodeling.
For example, immunoreactivity for spinophilin, a
marker of dendritic spines, is increased in the hilar
region of the dentate gyrus, as well as in CA3, of
ovariectomized rats treated with estrogen for 2
days (Brake et al., 2001). However, levels of
404
syntaxin and synaptophysin (presynaptic proteins
associated with the transmitter release machinery),
as well as spinophilin, are unaltered by hormone
treatment in the dentate gyrus of rhesus monkeys
(Choi et al., 2003). Although young female rhesus
monkeys show a trend toward an estrogen-induced
increase in immunoreactive spines in the dent-
ate gyrus outer molecular layer, this effect
appears to be statistically insignificant (Hao
et al., 2003).
Glia have also been associated with synaptic
remodeling, as expansion of the dendritic tree and
spine growth may occur at the expense of shrink-
ing glial, primarily astroglial volume. As a result,
presumably, abundance of astroglial processes
and markers is negatively correlated with dendri-
tic spine density. However, the ways sex steroids
alter dentate gyrus glia do not appear to be con-
sistent. Luquin et al. (1993) have shown that the
surface density of astroglial cells is positively in-
fluenced by estrogen and progesterone. The sur-
face density of astroglial cells was significantly
increased over control values by 5 h after the in-
jection of estrogen to ovariectomized rats, and as
early as 1 h after the administration of progester-
one; it reached maximal values by 24 h and re-
turned to control levels by 48 h (Luquin et al.,
1993). In contrast, levels of glial fibrillary acidic
protein (GFAP) intron 1, a molecular marker of
adult astrocytes, shows that GFAP transcription
and mRNA are both decreased in the outer mo-
lecular layer of the dentate gyrus on the afternoon
of proestrus, when plasma estradiol levels are
highest (Stone et al., 1998b). In vitro, astrocytes
show interesting bidirectional responses, such that
estrogen treatment increases GFAP transcription
in monotypic astrocytic cultures but decreases
GFAP transcription in astrocytes cocultured with
neurons (Stone et al., 1998b). In mice, Lei et al.
(2003) have reported similar findings, i.e., long-
term 17b-estradiol treatment in aged female mice
significantly lowered the number of astrocytes
in the dentate gyrus and CA1 compared with
placebo.
What may be in the background of such vari-
able findings? Besides confounding variables such
as the age of animals, strain, dose and temporal
characteristics of hormone treatment, and/or the
dentate area examined, the chosen methodological
approach alone may be critical. What measures
most reliably the remodeling of synaptic connec-
tions, i.e., synaptogenesis or loss of synapses, is
debatable. Above, we list several light microscopic
approaches, such as Golgi-based estimation of
dendritic spine density and histochemical detection
of pre- and/or postsynaptic marker molecules,
which are widely applied, mainly due to their
relative methodological simplicity. However, it is
impossible to decide at the light microscopic level
what proportion of the measured molecular synap-
tic markers is actually associated with synapses,
and what proportion represents extrasynaptic mol-
ecules that are processed and/or stored in different
cellular compartments. Therefore, levels of synap-
tic marker molecules may and do change without
alterations in the number of synapses (Li et al.,
2004). Although Golgi-based estimation of den-
dritic spine density is a less controversial ap-
proach, it also has several limitations. The most
important is that spines may or may not form
synapses, and the Golgi method is incapable of
differentiating between spines that have synapses
and spines that do not. Thus, similar to synaptic
marker molecules, measures of dendritic spines do
not necessarily reflect the true number of spine
synapses.
The most important point in this debate is that
the number of actual spine synapses is more rel-
evant to the functional status of neurons than the
number of dendritic spines or the levels of any
molecular markers. Thus, when the true number of
synapses is questioned, one should count synapses
themselves using electron microscopic stereologi-
cal techniques, because the above-discussed light
microscopic markers are not reliable. Treating
ovariectomized rats with 10 mg/day subcutaneous
estradiol-bezoate for 2 days, a similar schedule
that has been used in previous studies (Gould
et al., 1990), estrogen increased the number of
spine synapses in the CA1 stratum radiatum by
71.6% over oil-treated control values, by 50.1% in
the CA3 stratum radiatum, and by 99.1% in the
molecular layer of the dentate gyrus (Fig. 1). It is
noteworthy that the number of granule cell spine
synapses in the dentate gyrus doubles after estro-
gen administration.

Fig. 1. Effects of ovariectomy and estrogen replacement on the
number of hippocampal spine synapses. Young adult female
Sprague-Dawley rats (250 g) were ovariectomized and one week
later, they received either 10 mg/rat/day estradiol-benzoate (EB,
solid columns) or 200 ml/rat/day sesame oil vehicle (oil, open
columns) subcutaneously for 2 days. Two days after the last
injection, the animals were sacrificed by transcardial perfusion
of fixative, and their brains were processed for electron micro-
scopic stereological analysis. Spine synapses were counted in
the CA1 and CA3 strata radiata, and in the molecular layer of
the dentate gyrus (DG). Significantly different from the cor-
responding Oil group (t-test, po0.001 in CA1 and DG, po0.01
in CA3).
Neuroprotective effects
Due to the vulnerability of some types of neurons
in the dentate gyrus to insults, this area is a com-
mon subject of neurodegenerative/neuroprotective
experiments. The work summarized below is
focused around two topics for which neuroprotec-
tion is particularly germane: epilepsy and
Alzheimer’s disease.
Effects of estrogen on seizures vary, depending
on the experimental approach, and many other
factors. Estrogen may increase neuronal excitabil-
ity and thus mediate proconvulsant effects that
have been reported in the past, but reviews of the
clinical and animal data show that estrogen may
also have no effect or even anticonvulsant effects
(Scharfman et al., 2003; Hajszan and MacLusky,
2006; Veliskova, 2006). The protective role of
estrogen in seizure-induced damage is more
straightforward. Severe seizures have been shown
405
to trigger excitotoxic cell death in the hilus of the
dentate gyrus, and patients with temporal lobe
epilepsy often exhibit neuronal loss in the hilus
also (Margerison and Corsellis, 1966). Kainic acid
is commonly used as a convulsant to elicit severe
seizures (status epilepticus) and excitotoxic dam-
age in the dentate gyrus of rats (Ben-Ari and
Cossart, 2000). Estrogen administration is capable
of preventing kainic acid-induced degeneration
(Azcoitia et al., 1998; Veliskova et al., 2000).
Estrogen may also mediate the protective actions
of other steroids. For example, the neuroster-
oids pregnenolone and DHEA showed a dose-
dependent protective effect of hilar neurons
against kainic acid. The administration of the aro-
matase inhibitor fadrozole, that blocks the con-
version of these steroids into estrogen, prevented
this effect (Veiga et al., 2003). Interestingly,
2-methoxyestradiol, an estradiol metabolite, in-
duced significant neuronal loss in the hilus, de-
tected 96 h after the treatment with this steroid.
This finding suggests that endogenous metabolism
of 17b-estradiol to 2-methoxyestradiol may coun-
terbalance the neuroprotective effects of estrogen
(Picazo et al., 2003).
Regarding mechanisms for the neuroprotective
effects of estrogen in studies of seizure-induced
neuronal damage, Veliskova et al. (2000) as well as
Haynes et al. (2003) suggest that intracellular ERs
mediate the neuroprotective effect of estrogen, be-
cause tamoxifen pretreatment effectively abolished
estrogen-induced neuroprotection. An interaction
of ER and IGF-1 receptor signaling may also be
important (Azcoitia et al., 1999a). Furthermore,
GABAB receptors are likely to play a role, because
there was a loss of GABAB receptor-mediated in-
hibition after kainic acid-induced status epilepticus
in the rat dentate gyrus, and pretreatment with
estrogen could prevent it (Velisek and Veliskova,
2002).
Other steroids besides estrogen are also likely to
reduce seizure-induced damage, and the proges-
terone metabolite 3a,5a-tetrahyroprogesterone
(allopregnanolone) has been shown to be one ex-
ample. Blocking progesterone’s metabolism to
3a,5a-tetrahyroprogesterone reduced progester-
one’s protective effects in the dentate gyrus
(Rhodes et al., 2004). In the kainate model,
406
3a,5a-tetrahyroprogesterone was able to protect
the hilus from kainic acid (Ciriza et al., 2004).
Another metabolite was also effective: 5a-hydro-
xyprogesterone (Ciriza et al., 2004).
Other models of injury, which use adrenalec-
tomy to examine neuronal loss in the dentate
gyrus, focus on the granule cells, because adrenal-
ectomy selectively kills granule cells (see chapter
by M. Joels in this volume). In this model, estra-
diol treatment reduced pyknotic cell number com-
pared to vehicle administration (Frye, 2001).
Interestingly, a synthetic glucocorticoid, dexamet-
hasone can also induce apoptosis in the dentate
gyrus, and pretreatment with estrogen substan-
tially attenuated the dexamethasone-induced neu-
ronal damage (Haynes et al., 2003). Colchicine, a
microtubule polymerization inhibitor, also selec-
tively kills granule cells, an effect that is increased
by ovariectomy and ameliorated by 17b-estradiol
(Liu et al., 2001).
Regarding other sex steroids, treatment of fe-
male or male rats with progesterone or its met-
abolites, 5a-dihydroprogesterone and 3a,5a-
tetrahyroprogesterone similarly reduced the total
number of adrenalectomy-induced pyknotic cells
in the dentate gyrus compared with vehicle ad-
ministration. In case of androgen, testosterone and
its metabolites, 5a-dihydrotestosterone and 5a-
androstane-3a,17b-diol significantly reduced the
number of pyknotic cells in the dentate gyrus
compared to vehicle-administered, adrenalectomi-
zed female rats (Frye and McCormick, 2000a, b).
Estrogen also has been found to play a critical
neuroprotective role in Azlheimer’s disease. Uni-
lateral entorhinal cortex lesion (ECX) is frequently
used as a model of Alzheimer’s disease-like deaff-
erentation in the dentate gyrus. ECX elicits sprout-
ing in the molecular layer, which is affected by
gonadectomy and hormone replacement, but only
in female rats: ovariectomy reduces fiber out-
growth and estrogen restores it (Stone et al., 2000).
However, testosterone replacement had no effect
on sprouting in castrated ECX males (Morse et al.,
1986). Sprouting in hippocampal cultures of
C57Bl/6J mice was increased by 75% after treat-
ment with 17b-estradiol, which was blocked by an
antagonist of nuclear receptors, tamoxifen (Teter
et al., 1999). In intact female mice in vivo, lesions
of the lateral part of the entorhinal cortex in-
creased axonal sprouting in the outer one-third of
the molecular layer of the dentate gyrus. Ovariec-
tomized mice receiving high and moderate estro-
gen supplementation displayed the same sprouting
response. In ovariectomized nontreated mice,
however, the sprouting response was significantly
reduced (to nearly nothing) (Kadish and van
Groen, 2002). Finally, Stone et al. (1998a) have
shown that in wild-type ECX mice, ovariectomy
decreases commissural/associational sprouting to
the inner molecular layer of the dentate gyrus,
which is reversed by estradiol replacement. In
ECX apolipoprotein E-knockout mice, however,
estradiol did not enhance sprouting, suggesting
that sprouting may be stimulated by estrogen
through its up-regulation of apolipoprotein E
expression, leading to increased recycling of mem-
brane lipids for use by sprouting neurons (Stone
et al., 1998a). Estrogen and apolipoprotein E may
therefore interact in their modulation of both
Alzheimer’s disease risk and recovery from
neuronal injury.
Mediation of sex steroid effects
Effects of sex steroids in the dentate gyrus depend
on the location of their action, and the sites of
steroid synthesis. Below we discuss the distribution
of receptors, which is summarized in Fig. 2. Sex
steroids are capable of acting directly on granule
cells, as well as indirectly via nongranule cells
within the dentate gyrus (interneurons, mossy
cells). The influence of sex steroids could also be
mediated by subcortical, hormone-sensitive struc-
tures, such as the septohippocampal cholinergic
neurons. Sites of sex steroid synthesis are usually
thought to be peripheral, but local synthesis is also
possible, and is discussed further below.
Distribution of ERa in the dentate gyrus
Estrogen binding, as well as mRNA and immuno-
reactivity for ERa, have been detected in nuclei
of scattered GABAergic interneurons, located pre-
dominantly in the subgranular region of the dent-
ate gyrus (Loy et al., 1988; Shughrue et al., 1997;
 407

Fig. 2. Subcellular localization of estrogen (ER), androgen (AR), and progestin (PR) receptors in the dentate gyrus. A subset of
GABAergic interneurons contains nuclear ERa (dark pink). Granule cells, newly born cells (identified by DCX) and some GABAergic
interneurons contain cytosolic and plasma membrane-associated ERb (blue). Dendritic spines, many originating from granule cells
contain ERa, ERb, AR (dark green), and PR (purple). A few dendritic spines in the hilus, likely originating from mossy cells, contain
ERa and ERb. ERa, ERb, AR, and PR are found in axons and axon terminals. Some ERa-containing terminals are cholinergic
(acetylcholine, orange); some ERb-containing terminals resemble monoaminergic boutons. Lot of astrocytes (stars), mostly in the
molecular layer, also contain ERa, ERb, AR, and PR. (See Color Plate 23.2 in color plate section.)

Weiland et al., 1997; Perlman et al., 2005). Nuclear
ERa-immunoreactive interneurons co-express ne-
uropeptide Y, calbindin-D28k and calretinin, but
not cholecystokinin or parvalbumin (Nakamura
and McEwen, 2005). In addition to nuclear recep-
tors, ultrastructural studies have revealed ERa
at several extranuclear sites in the dentate gyrus
(Milner et al., 2001). Specifically, ERa-immunore-
activity is affiliated with the cytoplasmic plasma-
lemma of select hilar interneurons and
with endosomes of a few granule cell perikarya.
Moreover, ERa-labeled profiles are dispersed
408
throughout the dentate gyrus. Approximately half
of these labeled profiles are unmyelinated axons
and axon terminals that contain numerous small,
synaptic vesicles. ERa-labeled terminals form both
symmetric and asymmetric synapses on dendritic
shafts and spines, suggesting that ERa-positive
axons arise from sources in addition to inhibitory
interneurons. Dual labeling revealed that ERa-
immunoreactivity is contained in axons and ter-
minals labeled with vesicular acetylcholine trans-
porter (Towart et al., 2003), suggesting that
estrogen could rapidly and directly affect the lo-
cal release and/or uptake of acetylcholine. About
one-quarter of the ERa-immunoreactive profiles
are dendritic spines, many originating from gran-
ule cells. In dendritic spines, ERa-immunoreactiv-
ity is often associated with the spine apparatus,
suggesting that estrogen might act locally through
ERa to influence protein synthesis during synaptic
remodeling. The remaining one-quarter of ERa-
labeled profiles are from glial origin that resemble
astrocytes and are often located near the spines of
granule cells. Vesicular acetylcholine transporter-
containing terminals often abut ERa-positive pre-
synaptic and glial profiles and unlabeled terminals
that contact ERa-immunoreactive spines (Towart
et al., 2003), suggesting that acetylcholine release
might play a critical role in estrogen-modulated
structural plasticity. Collectively, these results im-
ply that ERa may serve as both a genomic and
nongenomic transducer of estrogen action in the
dentate gyrus.
Distribution of ERb in the dentate gyrus
The cellular and subcellular locations of ERb-
immunoreactivity in the dentate gyrus are similar
yet distinct from ERa. In monkey, dense ERb
hybridization signal has been seen in the dentate
gyrus, CA1, CA2, CA3, CA4, and the pros-
ubiculum/subiculum areas of the hippocampus
(Gundlah et al., 2000). In rodents, cells in or near
the dentate granule cell layer transiently express
high levels of estrogen binding and ERa protein in
the nucleus during the first two postnatal weeks
(O’Keefe et al., 1995; Solum and Handa, 2001). In
adult rats and mice, ERb mRNA and protein has
been found in the perikarya of granule cells as well
as cells in the dentate subgranular layer (Li et al.,
1997; Shughrue et al., 1997; Mitra et al., 2003;
Milner et al., 2005). Szymczak et al. (2006) have
found that ERb mRNA and protein are displayed
in high levels in the estrus and in low levels in the
proestrus phase. Recently, robust mRNA expres-
sion for both the a and b subtypes of ERs has been
found in proliferating and differentiating cells of
neuronal phenotype in the subgranular zone of the
dentate gyrus (Isgor and Watson, 2005). Further-
more, ERb-immunoreactive glia has been ob-
served in the hilus of the dentate gyrus of male
and female rats. ERb-immunoreactivity has been
localized in glial processes and perikarya and, in
some cases, in glial cell nuclei. Double
immunocytochemical labeling of ERb and the
specific astroglial marker, GFAP revealed that the
ERb-immunoreactive glial cells are astrocytes
(Azcoitia et al., 1999b). Ultrastructural analysis
showed ERb-immunoreactivity at several extranu-
clear sites in the dentate gyrus (Milner et al., 2005).
ERb-immunoreactivity is affiliated with cytoplas-
mic organelles, especially endomembranes and mi-
tochondria, and with the membranes primarily of
granule cell perikarya and proximal dendrites. Re-
cent studies revealed that neuronal perikarya and
dendrites labeled with doublecortin, a marker of
newly generated cells, also contain extranuclear
ERb-immunoreactivity in both the adult and neo-
natal dentate gyrus (Herrick et al., 2006). ERb-
labeled dendritic shafts and spines have mostly
been found in the molecular layer. In dendritic
processes, ERb-immunoreactivity is near the per-
isynaptic zone adjacent to synapses formed by un-
labeled terminals. The ERb protein can also be
found in preterminal axons and axon terminals,
associated with clusters of small, synaptic vesicles.
ERb-labeled axons are particularly dense in the
hilus and outer molecular layer, forming both
asymmetric and symmetric synapses with dend-
rites. Finally, ERb-immunoreactivity has been de-
tected in glial profiles throughout the dentate
gyrus, some of which appose doublecortin-labeled
perikarya and dendrites (Herrick et al., 2006).
These results suggest that ERb may serve prima-
rily as a nongenomic transducer of estrogen ac-
tions in the dentate gyrus.

Distribution of progestin receptor (PR) in the
dentate gyrus
Cells containing PR mRNA have been detected in
the dentate subgranular zone (Hagihara et al.,
1992). By light microscopy, nuclear PR-immuno-
reactivity is undetectable in the dentate gyrus;
however, ultrastructural analysis revealed that the
PR protein is found at several extranuclear sites
(Waters et al., 2005). In the molecular layer and
hilus, PR-immunoreactivity is present in dendritic
spines, closely associated with the postsynaptic
density. The PR protein is expressed in axons and
axon terminals that contain small synaptic vesicles.
PR-positive terminals and en passant axonal
boutons form synapses with dendritic spines. PR-
immunoreactivity has also been found in glia,
many resembling astrocytes and some forming
presumed gap junctions with other astrocytic pro-
files. The considerable lack of nuclear PR labeling
may indicate that progesterone uses nongenomic
signaling mechanism in the dentate gyrus to di-
rectly affect dendritic spine morphology and
synaptic plasticity.
Distribution of AR in the dentate gyrus
Previous light microscopic studies have shown that
AR mRNA, immunoreactivity, and binding are
present in pyramidal cell nuclei but not granule
cells (Commins and Yahr, 1985; Sar et al., 1990;
Simerly et al., 1990; Kerr et al., 1995). However,
AR-immunoreactivity is present in disperse, punc-
tuate processes that are most dense in the pyram-
idal cell layer and diffusely distributed in the
mossy fiber pathway (Tabori et al., 2005). Electron
microscopic analysis revealed AR-immunoreactiv-
ity at several extranuclear sites in the dentate
gyrus. AR labeling has been found in dendritic
spines, many arising from granule cell dendrites.
AR is affiliated with clusters of small, synaptic
vesicles within preterminal axons and axon termi-
nals, the majority of these being in the central hi-
lus. AR-immunoreactive preterminal axons are
most prominent in the CA3 stratum lucidum. AR-
labeled terminals exclusively form asymmetric
synapses. Throughout the dentate gyrus,
409
AR-immunoreactivity has also been detected in
astrocytic profiles; many of them apposing termi-
nals that synapse on unlabeled dendritic spines or
forming gap junctions with other AR-positive
or unlabeled astrocytes (Tabori et al., 2005).
Together, these results suggest that ARs may serve
as both a genomic and nongenomic transducer of
androgen action in the dentate gyrus.
Role of local steroid synthesis
Recent studies (Rune et al., 2006) suggest that lo-
cally synthesized steroids may contribute to hip-
pocampal activational effects. Using slice cultures,
Rune and colleagues have reported that the
number of dentate proliferative cells decreases,
whereas the number of apoptotic cells increases
dose-dependently, in response to reduced estradiol
release into the medium after treatment with let-
rozole, an aromatase enzyme inhibitor (Fester
et al., 2006). This also holds true for cell cultures
transfected with siRNA against steroidogenic
acute regulatory protein (StAR). StAR transports
cholesterol to the inner mitochondrial membrane,
where it is converted by the cytochrome P-450 en-
zyme complex, and as such, it is the first step in the
cascade of estrogen synthesis. Application of est-
radiol to the medium had no effect on prolifera-
tion and apoptosis, whereas the antiproliferative
and pro-apoptotic effects of StAR knockdown and
letrozole administration were restored by treat-
ment of the cultures with estradiol (Fester et al.,
2006). The data of Rune and colleagues are also
supported by other studies. In situ hybridization
revealed that StAR and aromatase are highly ex-
pressed in neuronal cells of the dentate gyrus. In
addition, StAR- and aromatase-positive cells are
strictly correlated with steroidogenic factor-1, a
regulator of steroid biosynthesis, as shown by
computer-assisted confocal microscopy in double
labeling experiments (Wehrenberg et al., 2001).
Similarly, Hojo et al. (2004) have reported in
a rather comprehensive study that in the granule
cells of the adult male rat dentate gyrus, signifi-
cant colocalization is seen for both cytochromes
P45017a (dehydroepiandrosterone synthase)
and P450 aromatase by means of
410
immunohistochemical staining of slices. Only a
weak immunoreaction of these P450s has been
observed in astrocytes and oligodendrocytes
(Hojo et al., 2004). More importantly, stimulation
of hippocampal neurons with NMDA induced a
significant net production of estradiol. The anal-
ysis of radioactive metabolites demonstrated
the conversion from [3H] pregnenolone to [3H]
estradiol through dehydroepiandrosterone and
testosterone. This activity was abolished by the
application of specific inhibitors of cytochrome
P450s (Hojo et al., 2004). In summary, although
these findings seem compelling, considering the
fact that the majority of the above-mentioned
data have been obtained from cultures, which
may be quite different from conditions in vivo,
one should exercise caution in interpreting such
results.
Subcortical mediation of sex steroid effects
In addition to direct effects of sex steroids on
dentate cells, indirect actions may influence the
dentate gyrus. We already mentioned above that
the raphe serotonergic system mediates the dentate
neurogenic effect of estrogen (Banasr et al., 2001).
Consistent with this line of argument, raphe se-
rotonergic neurons express ERa (Leranth et al.,
1999). Another structure that appears to be critical
is the septohippocampal cholinergic system. Septal
cholinergic neurons express nuclear ERs (Shugh-
rue et al., 2000) and hippocampal cholinergic
axons and terminals contain extranuclear ERa
(Towart et al., 2003). Moreover, estrogen affects
septohippocampal cholinergic neurons both gen-
omically and nongenomically (Gibbs and Aggar-
wal, 1998; Rudick et al., 2003). Androgen also can
influence the expression of cholinergic markers in
the dentate gyrus. Specifically, gonadectomy re-
duced the density of choline acetyltransferase
immunoreactive fibers in the dentate gyrus, which
was reversed by the addition of testosterone pro-
pionate (Nakamura et al., 2002). Although there is
no direct data from the dentate gyrus, elsewhere in
the hippocampus estrogen modulates the inhibi-
tion by specific GABAergic interneurons, which is
partially dependent on input from basal forebrain
cholinergic neurons (Rudick et al., 2003). More
evidence is available for the cholinergic role in
dentate neurogenesis. Selective neurotoxic lesion
of the forebrain cholinergic input with 192
IgG-saporin reduced dentate cell proliferation
(Mohapel et al., 2005). Conversely, systemic ad-
ministration of the cholinergic agonist physostig-
mine increased dentate neurogenesis. The neuro-
genic effect of acetylcholine appears to involve
nicotinic receptors containing the b2 subunit
(Harrist et al., 2004), as well as m2, m3, and m4
muscarinic receptors (Ma et al., 2000; Mohapel
et al., 2005). Consistent with these findings, ova-
riectomy upregulated m4 receptors in the dentate
gyrus, whereas estrogen treatment restored
m4 binding to the level of the sham group
(El-Bakri et al., 2002). However, other results
suggest no septohippocampal involvement in the
synaptogenic action of estrogen, because rats
that received estrogen implants into the medial
septum did not exhibit changes in astroglial proc-
ess density in the dentate gyrus (Lam and Leranth,
2003).
Concluding remarks
The studies discussed in this review clearly dem-
onstrate that both the organization and function-
ing of the dentate gyrus is a significant target of sex
steroids. The gonadal hormone modulation of
physiological activity, neurogenesis, synaptic re-
modeling, and neurodegeneration/neuroprotection
mechanisms are likely to be relevant to higher
brain functions such as cognition and mood, in
addition to their clinical implications in epilepsy,
Alzheimer’s disease and mental disorders. Relative
to hippocampal subfield CA1, however, the role of
gonadal hormones in the dentate gyrus has not
received substantial attention. In particular, our
understanding of androgen effects on the dentate
gyrus is extremely limited relative to estrogen. The
available data suggest potent, complex, and po-
tentially important effects, however, and therefore
merit more attention in the future.

Abbreviations
AR androgen receptor
BDNF brain-derived neurotrophic factor
DHEAS dehydroepiandrosterone-sulfate
ECX entorhinal cortex lesion
ER estrogen receptor
GFAP glial fibrillary acidic protein
IGF-1 insulin-like growth factor-1
NMDA N-methyl-D-aspartate
PR progestin receptor
StAR steroidogenic acute regulatory protein
Tfm testicular feminization mutant
Acknowledgements
This work was supported by NIH grants
MH074021 (T.H.); DA08259, NS07080 and
HL18974 (T.A.M.); MH060858 and NS042644
(C.L.); as well as by a Hungarian National Office
for Research and Technology grant RET-08/04.
References
Altman, J. and Das, G.D. (1965) Autoradiographic and histo-
logical evidence of postnatal hippocampal neurogenesis in
rats. J. Comp. Neurol., 124(3): 319–335.
Azcoitia, I., Sierra, A. and Garcia-Segura, L.M. (1998) Estra-
diol prevents kainic acid-induced neuronal loss in the rat
dentate gyrus. Neuroreport, 9(13): 3075–3079.
Azcoitia, I., Sierra, A. and Garcia-Segura, L.M. (1999a)
Neuroprotective effects of estradiol in the adult rat hippo-
campus: interaction with insulin-like growth factor-I signal-
ling. J. Neurosci. Res., 58(6): 815–822.
Azcoitia, I., Sierra, A. and Garcia-Segura, L.M. (1999b)
Localization of estrogen receptor beta-immunoreactivity in
astrocytes of the adult rat brain. Glia, 26(3): 260–267.
Banasr, M., Hery, M., Brezun, J.M. and Daszuta, A. (2001)
Serotonin mediates oestrogen stimulation of cell proliferation
in the adult dentate gyrus. Eur. J. Neurosci., 14(9):
1417–1424.
Ben-Ari, Y. and Cossart, R. (2000) Kainate, a double agent that
generates seizures: two decades of progress. Trends Neuro-
sci., 23(11): 580–587.
Birzniece, V., Johansson, I.M., Wang, M.D., Seckl, J.R.,
Backstrom, T. and Olsson, T. (2001) Serotonin 5-HT(1A)
receptor mRNA expression in dorsal hippocampus and raphe
411
nuclei after gonadal hormone manipulation in female rats.
Neuroendocrinology, 74(2): 135–142.
Bowman, R.E., Ferguson, D. and Luine, V.N. (2002) Effects of
chronic restraint stress and estradiol on open field activity,
spatial memory, and monoaminergic neurotransmitters in
ovariectomized rats. Neuroscience, 113(2): 401–410.
Brake, W.G., Alves, S.E., Dunlop, J.C., Lee, S.J., Bulloch, K.,
Allen, P.B., Greengard, P. and McEwen, B.S. (2001) Novel
target sites for estrogen action in the dorsal hippocampus: an
examination of synaptic proteins. Endocrinology, 142(3):
1284–1289.
Brannvall, K., Bogdanovic, N., Korhonen, L. and Lindholm,
D. (2005) 19-Nortestosterone influences neural stem cell pro-
liferation and neurogenesis in the rat brain. Eur. J. Neurosci.,
21(4): 871–878.
Choi, J.M., Romeo, R.D., Brake, W.G., Bethea, C.L., Rose-
nwaks, Z. and McEwen, B.S. (2003) Estradiol increases
pre- and post-synaptic proteins in the CA1 region of the
hippocampus in female rhesus macaques (Macaca mulatta).
Endocrinology, 144(11): 4734–4738.
Ciriza, I., Azcoitia, I. and Garcia-Segura, L.M. (2004) Reduced
progesterone metabolites protect rat hippocampal neurones
from kainic acid excitotoxicity in vivo. J. Neuroendocrinol.,
16(1): 58–63.
Commins, D. and Yahr, P. (1985) Autoradiographic localiza-
tion of estrogen and androgen receptors in the sexually
dimorphic area and other regions of the gerbil brain. J.
Comp. Neurol., 231(4): 473–489.
Conner, J.M., Lauterborn, J.C., Yan, Q., Gall, C.M. and
Varon, S. (1997) Distribution of brain-derived neurotrophic
factor (BDNF) protein and mRNA in the normal adult rat
CNS: evidence for anterograde axonal transport. J. Neuro-
sci., 17(7): 2295–2313.
Cyr, M., Ghribi, O. and Di Paolo, T. (2000) Regional and
selective effects of oestradiol and progesterone on NMDA
and AMPA receptors in the rat brain. J. Neuroendocrinol.,
12(5): 445–452.
Cyr, M., Thibault, C., Morissette, M., Landry, M. and Di
Paolo, T. (2001) Estrogen-like activity of tamoxifen and
raloxifene on NMDA receptor binding and expression of its
subunits in rat brain. Neuropsychopharmacology, 25(2):
242–257.
De Nicola, A.F., Saravia, F.E., Beauquis, J., Pietranera, L. and
Ferrini, M.G. (2006) Estrogens and neuroendocrine hypo-
thalamic-pituitary-adrenal axis function. Front. Horm. Res.,
35: 157–168.
El-Bakri, N.K., Adem, A., Suliman, I.A., Mulugeta, E., Karls-
son, E., Lindgren, J.U., Winblad, B. and Islam, A. (2002)
Estrogen and progesterone treatment: effects on muscarinic
M(4) receptor subtype in the rat brain. Brain Res., 948(1–2):
131–137.
El-Bakri, N.K., Islam, A., Zhu, S., Elhassan, A., Mohammed,
A., Winblad, B. and Adem, A. (2004) Effects of estrogen and
progesterone treatment on rat hippocampal NMDA recep-
tors: relationship to Morris water maze performance. J. Cell.
Mol. Med., 8(4): 537–544.
412
Falconer, E.M. and Galea, L.A. (2003) Sex differences in cell
proliferation, cell death and defensive behavior following
acute predator odor stress in adult rats. Brain Res., 975(1–2):
22–36.
Fester, L., Ribeiro-Gouveia, V., Prange-Kiel, J., von Schassen,
C., Bottner, M., Jarry, H. and Rune, G.M. (2006) Prolifer-
ation and apoptosis of hippocampal granule cells require
local oestrogen synthesis. J. Neurochem., 97(4): 1136–1144.
Franklin, T.B. and Perrot-Sinal, T.S. (2006) Sex and ovarian
steroids modulate brain-derived neurotrophic factor (BDNF)
protein levels in rat hippocampus under stressful and non-
stressful conditions. Psychoneuroendocrinology, 31(1):
38–48.
Frye, C.A. (2001) Estradiol tends to improve inhibitory avoid-
ance performance in adrenalectomized male rats and reduces
pyknotic cells in the dentate gyrus of adrenalectomized male
and female rats. Brain Res., 889(1–2): 358–363.
Frye, C.A. and McCormick, C.M. (2000a) Androgens are
neuroprotective in the dentate gyrus of adrenalectomized
female rats. Stress, 3(3): 185–194.
Frye, C.A. and McCormick, C.M. (2000b) The neurosteroid,
3alpha-androstanediol, prevents inhibitory avoidance deficits
and pyknotic cells in the granule layer of the dentate gyrus
induced by adrenalectomy in rats. Brain Res., 855(1):
166–170.
Galea, L.A., Spritzer, M.D., Barker, J.M. and Pawluski, J.L.
(2006) Gonadal hormone modulation of hippocampal ne-
urogenesis in the adult. Hippocampus, 16(3): 225–232.
Gazzaley, A.H., Weiland, N.G., McEwen, B.S. and Morrison,
J.H. (1996) Differential regulation of NMDAR1 mRNA and
protein by estradiol in the rat hippocampus. J. Neurosci.,
16(21): 6830–6838.
Gibbs, R.B. and Aggarwal, P. (1998) Estrogen and basal fore-
brain cholinergic neurons: implications for brain aging and
Alzheimer’s disease-related cognitive decline. Horm. Behav.,
34(2): 98–111.
Gould, E., Tanapat, P., Rydel, T. and Hastings, N. (2000)
Regulation of hippocampal neurogenesis in adulthood. Biol.
Psychiatry, 48(8): 715–720.
Gould, E., Woolley, C.S., Frankfurt, M. and McEwen, B.S.
(1990) Gonadal steroids regulate dendritic spine density in
hippocampal pyramidal cells in adulthood. J. Neurosci.,
10(4): 1286–1291.
Gundlah, C., Kohama, S.G., Mirkes, S.J., Garyfallou, V.T.,
Urbanski, H.F. and Bethea, C.L. (2000) Distribution of
estrogen receptor beta (ERbeta) mRNA in hypothalamus,
midbrain and temporal lobe of spayed macaque: continued
expression with hormone replacement. Brain Res. Mol. Brain
Res., 76(2): 191–204.
Hagihara, K., Hirata, S., Osada, T., Hirai, M. and Kato, J.
(1992) Distribution of cells containing progesterone receptor
mRNA in the female rat di- and telencephalon: an in situ
hybridization study. Brain Res. Mol. Brain Res., 14(3):
239–249.
Hajszan, T. and MacLusky, N.J. (2006) Neurologic links
between epilepsy and depression in women: is hippocampal
neuroplasticity the key? Neurology, 66(6 Suppl 3): S13–S22.
Hajszan, T., MacLusky, N.J. and Leranth, C. (2005) Short-
term treatment with the antidepressant fluoxetine triggers
pyramidal dendritic spine synapse formation in rat hippo-
campus. Eur. J. Neurosci., 21(5): 1299–1303.
Hao, J., Janssen, W.G., Tang, Y., Roberts, J.A., McKay, H.,
Lasley, B., Allen, P.B., Greengard, P., Rapp, P.R., Kordow-
er, J.H., Hof, P.R. and Morrison, J.H. (2003) Estrogen
increases the number of spinophilin-immunoreactive spines
in the hippocampus of young and aged female rhesus mon-
keys. J. Comp. Neurol., 465(4): 540–550.
Harrist, A., Beech, R.D., King, S.L., Zanardi, A., Cleary,
M.A., Caldarone, B.J., Eisch, A., Zoli, M. and Picciotto,
M.R. (2004) Alteration of hippocampal cell proliferation in
mice lacking the beta 2 subunit of the neuronal nicotinic
acetylcholine receptor. Synapse, 54(4): 200–206.
Haynes, L.E., Lendon, C.L., Barber, D.J. and Mitchell, I.J.
(2003) 17 Beta-oestradiol attenuates dexamethasone-induced
lethal and sublethal neuronal damage in the striatum and
hippocampus. Neuroscience, 120(3): 799–806.
Herrick, S.P., Waters, E.M., Drake, C.T., McEwen, B.S. and
Milner, T.A. (2006) Extranuclear estrogen receptor beta
immunoreactivity is on doublecortin-containing cells in the
adult and neonatal rat dentate gyrus. Brain Res., 1121:
46–58.
Hojo, Y., Hattori, T.A., Enami, T., Furukawa, A., Suzuki, K.,
Ishii, H.T., Mukai, H., Morrison, J.H., Janssen, W.G.,
Kominami, S., Harada, N., Kimoto, T. and Kawato, S.
(2004) Adult male rat hippocampus synthesizes estradiol
from pregnenolone by cytochromes P45017alpha and P450
aromatase localized in neurons. Proc. Natl. Acad. Sci.
U.S.A., 101(3): 865–870.
Isgor, C. and Sengelaub, D.R. (1998) Prenatal gonadal steroids
affect adult spatial behavior, CA1 and CA3 pyramidal cell
morphology in rats. Horm. Behav., 34(2): 183–198.
Isgor, C. and Watson, S.J. (2005) Estrogen receptor alpha and
beta mRNA expressions by proliferating and differentiating
cells in the adult rat dentate gyrus and subventricular zone.
Neuroscience, 134(3): 847–856.
Jones, B.A. and Watson, N.V. (2005) Spatial memory perform-
ance in androgen insensitive male rats. Physiol. Behav., 85(2):
135–141.
Kadish, I. and van Groen, T. (2002) Low levels of estrogen
significantly diminish axonal sprouting after entorhinal cor-
tex lesions in the mouse. J. Neurosci., 22(10): 4095–4102.
Kandel, E.R. (2001) The molecular biology of memory storage:
a dialogue between genes and synapses. Science, 294(5544):
1030–1038.
Kasai, H., Matsuzaki, M., Noguchi, J., Yasumatsu, N. and
Nakahara, H. (2003) Structure-stability-function relation-
ships of dendritic spines. Trends Neurosci., 26(7): 360–368.
O’Keefe, J.A., Li, Y., Burgess, L.H. and Handa, R.J. (1995)
Estrogen receptor mRNA alterations in the developing rat
hippocampus. Brain Res. Mol. Brain Res., 30(1): 115–124.
Kerr, J.E., Allore, R.J., Beck, S.G. and Handa, R.J. (1995)
Distribution and hormonal regulation of androgen receptor
(AR) and AR messenger ribonucleic acid in the rat hippo-
campus. Endocrinology, 136(8): 3213–3221.

Kim, M.T., Soussou, W., Gholmieh, G., Ahuja, A., Tanguay,
A., Berger, T.W. and Brinton, R.D. (2006) 17beta-Estradiol
potentiates field excitatory postsynaptic potentials within
each subfield of the hippocampus with greatest potentiation
of the associational/commissural afferents of CA3. Neuro-
science, 141(1): 391–406.
Korol, D.L. and Kolo, L.L. (2002) Estrogen-induced changes
in place and response learning in young adult female rats.
Behav. Neurosci., 116(3): 411–420.
Lam, T.T. and Leranth, C. (2003) Gonadal hormones act
extrinsic to the hippocampus to influence the density of hip-
pocampal astroglial processes. Neuroscience, 116(2):
491–498.
Le Saux, M. and Di Paolo, T. (2005) Changes in 5-HT1A
receptor binding and G-protein activation in the rat brain
after estrogen treatment: comparison with tamoxifen and
raloxifene. J. Psychiatry Neurosci., 30(2): 110–117.
Lei, D.L., Long, J.M., Hengemihle, J., O’Neill, J., Manaye,
K.F., Ingram, D.K. and Mouton, P.R. (2003) Effects of est-
rogen and raloxifene on neuroglia number and morphology
in the hippocampus of aged female mice. Neuroscience,
121(3): 659–666.
Leranth, C., Shanabrough, M. and Horvath, T.L. (1999)
Estrogen receptor-alpha in the raphe serotonergic and sup-
ramammillary area calretinin-containing neurons of the
female rat. Exp. Brain Res., 128(3): 417–420.
Li, C., Brake, W.G., Romeo, R.D., Dunlop, J.C., Gordon, M.,
Buzescu, R., Magarinos, A.M., Allen, P.B., Greengard, P.,
Luine, V. and McEwen, B.S. (2004) Estrogen alters hippo-
campal dendritic spine shape and enhances synaptic protein
immunoreactivity and spatial memory in female mice. Proc.
Natl. Acad. Sci. U.S.A., 101(7): 2185–2190.
Li, X., Schwartz, P.E. and Rissman, E.F. (1997) Distribution of
estrogen receptor-beta-like immunoreactivity in rat fore-
brain. Neuroendocrinology, 66(2): 63–67.
Liu, Z., Gastard, M., Verina, T., Bora, S., Mouton, P.R. and
Koliatsos, V.E. (2001) Estrogens modulate experimentally
induced apoptosis of granule cells in the adult hippocampus.
J. Comp. Neurol., 441(1): 1–8.
Louissaint, A.J., Rao, S., Leventhal, C. and Goldman, S.A.
(2002) Coordinated interaction of neurogenesis and
angiogenesis in the adult songbird brain. Neuron, 34(6):
945–960.
Loy, R., Gerlach, J.L. and McEwen, B.S. (1988) Autoradio-
graphic localization of estradiol-binding neurons in the rat
hippocampal formation and entorhinal cortex. Brain Res.,
467(2): 245–251.
Luquin, S., Naftolin, F. and Garcia-Segura, L.M. (1993) Nat-
ural fluctuation and gonadal hormone regulation of astrocyte
immunoreactivity in dentate gyrus. J. Neurobiol., 24(7):
913–924.
Lynch, M.A. (2004) Long-term potentiation and memory.
Physiol. Rev., 84(1): 87–136.
Ma, W., Maric, D., Li, B.S., Hu, Q., Andreadis, J.D., Grant,
G.M., Liu, Q.Y., Shaffer, K.M., Chang, Y.H., Zhang, L.,
Pancrazio, J.J., Pant, H.C., Stenger, D.A. and Barker, J.L.
(2000) Acetylcholine stimulates cortical precursor cell
413
proliferation in vitro via muscarinic receptor activation
and MAP kinase phosphorylation. Eur. J. Neurosci., 12(4):
1227–1240.
MacLusky, N.J., Hajszan, T., Prange-Kiel, J. and Leranth, C.
(2006) Androgen modulation of hippocampal synaptic plas-
ticity. Neuroscience, 138(3): 957–965.
MacLusky, N.J., Luine, V.N., Hajszan, T. and Leranth, C.
(2005) The 17alpha and 17beta isomers of estradiol both in-
duce rapid spine synapse formation in the CA1 hippocampal
subfield of ovariectomized female rats. Endocrinology,
146(1): 287–293.
Margerison, J.H. and Corsellis, J.A. (1966) Epilepsy and the
temporal lobes. A clinical, electroencephalographic and ne-
uropathological study of the brain in epilepsy, with particular
reference to the temporal lobes. Brain, 89(3): 499–530.
Mazzucco, C.A., Lieblich, S.E., Bingham, B.I., Williamson,
M.A., Viau, V. and Galea, L.A. (2006) Both estrogen recep-
tor alpha and estrogen receptor beta agonists enhance cell
proliferation in the dentate gyrus of adult female rats.
Neuroscience, 141: 1793–1800.
McEwen, B.S. (2003) Mood disorders and allostatic load. Biol.
Psychiatry, 54(3): 200–207.
McEwen, B.S. and Alves, S.E. (1999) Estrogen actions in the
central nervous system. Endocr. Rev., 20(3): 279–307.
Milner, T.A., Ayoola, K., Drake, C.T., Herrick, S.P., Tabori,
N.E., McEwen, B.S., Warrier, S. and Alves, S.E. (2005)
Ultrastructural localization of estrogen receptor beta
immunoreactivity in the rat hippocampal formation. J.
Comp. Neurol., 491(2): 81–95.
Milner, T.A., McEwen, B.S., Hayashi, S., Li, C.J., Reagan, L.P.
and Alves, S.E. (2001) Ultrastructural evidence that hippo-
campal alpha estrogen receptors are located at extranuclear
sites. J. Comp. Neurol., 429(3): 355–371.
Miranda, P., Williams, C.L. and Einstein, G. (1999) Granule
cells in aging rats are sexually dimorphic in their response to
estradiol. J. Neurosci., 19(9): 3316–3325.
Mitra, S.W., Hoskin, E., Yudkovitz, J., Pear, L., Wilkinson,
H.A., Hayashi, S., Pfaff, D.W., Ogawa, S., Rohrer, S.P.,
Schaeffer, J.M., McEwen, B.S. and Alves, S.E. (2003)
Immunolocalization of estrogen receptor beta in the mouse
brain: comparison with estrogen receptor alpha. Endocrino-
logy, 144(5): 2055–2067.
Mohapel, P., Leanza, G., Kokaia, M. and Lindvall, O. (2005)
Forebrain acetylcholine regulates adult hippocampal neuro-
genesis and learning. Neurobiol. Aging, 26(6): 939–946.
Morse, J.K., Scheff, S.W. and DeKosky, S.T. (1986) Gonadal
steroids influence axon sprouting in the hippocampal dentate
gyrus: a sexually dimorphic response. Exp. Neurol., 94(3):
649–658.
Naftolin, F., Ryan, K.J., Davies, I.J., Reddy, V.V., Flores, F.,
Petro, Z., Kuhn, M., White, R.J., Takaoka, Y. and Wolin, L.
(1975) The formation of estrogens by central neuroendocrine
tissues. Recent Prog. Horm. Res., 31: 295–319.
Nakamura, N., Fujita, H. and Kawata, M. (2002) Effects of
gonadectomy on immunoreactivity for choline acetyltransf-
erase in the cortex, hippocampus, and basal forebrain of
adult male rats. Neuroscience, 109(3): 473–485.
414
Nakamura, N.H. and McEwen, B.S. (2005) Changes in intern-
euronal phenotypes regulated by estradiol in the adult rat
hippocampus: a potential role for neuropeptide Y. Neuro-
science, 136(1): 357–369.
Orchinik, M., Weiland, N.G. and McEwen, B.S. (1995) Chronic
exposure to stress levels of corticosterone alters GABAA
receptor subunit mRNA levels in rat hippocampus. Brain
Res. Mol. Brain Res., 34(1): 29–37.
Ormerod, B.K., Lee, T.T. and Galea, L.A. (2003) Estradiol
initially enhances but subsequently suppresses (via adrenal
steroids) granule cell proliferation in the dentate gyrus of
adult female rats. J. Neurobiol., 55(2): 247–260.
Parducz, A. and Garcia-Segura, L.M. (1993) Sexual differences
in the synaptic connectivity in the rat dentate gyrus. Neuro-
sci. Lett., 161(1): 53–56.
Parducz, A., Hajszan, T., Maclusky, N.J., Hoyk, Z., Csakvari,
E., Kurunczi, A., Prange-Kiel, J. and Leranth, C. (2006)
Synaptic remodeling induced by gonadal hormones: neuronal
plasticity as a mediator of neuroendocrine and behavioral
responses to steroids. Neuroscience, 138(3): 977–985.
Perez-Martin, M., Azcoitia, I., Trejo, J.L., Sierra, A. and
Garcia-Segura, L.M. (2003) An antagonist of estrogen
receptors blocks the induction of adult neurogenesis by
insulin-like growth factor-I in the dentate gyrus of adult
female rat. Eur. J. Neurosci., 18(4): 923–930.
Perez-Martin, M., Salazar, V., Castillo, C., Ariznavarreta, C.,
Azcoitia, I., Garcia-Segura, L.M. and Tresguerres, J.A.
(2005) Estradiol and soy extract increase the production of
new cells in the dentate gyrus of old rats. Exp. Gerontol.,
40(5): 450–453.
Perlman, W.R., Tomaskovic-Crook, E., Montague, D.M.,
Webster, M.J., Rubinow, D.R., Kleinman, J.E. and Weickert,
C.S. (2005) Alteration in estrogen receptor alpha mRNA levels
in frontal cortex and hippocampus of patients with major
mental illness. Biol. Psychiatry, 58(10): 812–824.
Picazo, O., Azcoitia, I. and Garcia-Segura, L.M. (2003)
Neuroprotective and neurotoxic effects of estrogens. Brain
Res., 990(1–2): 20–27.
Rhodes, M.E., McCormick, C.M. and Frye, C.A. (2004)
3alpha,5alpha-THP mediates progestins’ effects to protect
against adrenalectomy-induced cell death in the dentate gyrus
of female and male rats. Pharmacol. Biochem. Behav., 78(3):
505–512.
Roof, R.L. (1993) The dentate gyrus is sexually dimorphic in
prepubescent rats: testosterone plays a significant role. Brain
Res., 610(1): 148–151.
Roof, R.L. and Havens, M.D. (1992) Testosterone improves
maze performance and induces development of a male hip-
pocampus in females. Brain Res., 572(1–2): 310–313.
Rudick, C.N., Gibbs, R.B. and Woolley, C.S. (2003) A role for
the basal forebrain cholinergic system in estrogen-induced
disinhibition of hippocampal pyramidal cells. J. Neurosci.,
23(11): 4479–4490.
Rune, G.M., Lohse, C., Prange-Kiel, J., Fester, L. and Frotsc-
her, M. (2006) Synaptic plasticity in the hippocampus: effects
of estrogen from the gonads or hippocampus? Neurochem.
Res., 31(2): 145–155.
Santarelli, L., Saxe, M., Gross, C., Surget, A., Battaglia, F.,
Dulawa, S., Weisstaub, N., Lee, J., Duman, R., Arancio, O.,
Belzung, C. and Hen, R. (2003) Requirement of hippocampal
neurogenesis for the behavioral effects of antidepressants.
Science, 301(5634): 805–809.
Sar, M., Lubahn, D.B., French, F.S. and Wilson, E.M. (1990)
Immunohistochemical localization of the androgen receptor
in rat and human tissues. Endocrinology, 127(6): 3180–3186.
Saravia, F.E., Beauquis, J., Revsin, Y., Homo-Delarche, F.,
de Kloet, E.R. and De Nicola, A.F. (2006) Hippocampal
neuropathology of diabetes mellitus is relieved by estrogen
treatment. Cell. Mol. Neurobiol., 26: 943–957.
Scharfman, H.E., Mercurio, T.C., Goodman, J.H., Wilson,
M.A. and MacLusky, N.J. (2003) Hippocampal excitability
increases during the estrous cycle in the rat: a potential role
for brain-derived neurotrophic factor. J. Neurosci., 23(37):
11641–11652.
Schumacher, M., Coirini, H. and McEwen, B.S. (1989) Regu-
lation of high-affinity GABAA receptors in the dorsal hip-
pocampus by estradiol and progesterone. Brain Res., 487(1):
178–183.
Seidman, S.N. (2003) The aging male: androgens, erectile dysfunc-
tion, and depression. J. Clin. Psychiatry, 64(Suppl 10): 31–37.
Shughrue, P.J., Lane, M.V. and Merchenthaler, I. (1997) Com-
parative distribution of estrogen receptor-alpha and -beta
mRNA in the rat central nervous system. J. Comp. Neurol.,
388(4): 507–525.
Shughrue, P.J., Scrimo, P.J. and Merchenthaler, I. (2000) Est-
rogen binding and estrogen receptor characterization (ERal-
pha and ERbeta) in the cholinergic neurons of the rat basal
forebrain. Neuroscience, 96(1): 41–49.
Simerly, R.B., Chang, C., Muramatsu, M. and Swanson, L.W.
(1990) Distribution of androgen and estrogen receptor
mRNA-containing cells in the rat brain: an in situ hybrid-
ization study. J. Comp. Neurol., 294(1): 76–95.
Sohrabji, F., Miranda, R.C. and Toran-Allerand, C.D. (1995)
Identification of a putative estrogen response element in the
gene encoding brain-derived neurotrophic factor. Proc. Natl.
Acad. Sci. U.S.A., 92(24): 11110–11114.
Solum, D.T. and Handa, R.J. (2001) Localization of estrogen
receptor alpha (ER alpha) in pyramidal neurons of the de-
veloping rat hippocampus. Brain Res. Dev. Brain Res.,
128(2): 165–175.
Sorra, K.E. and Harris, K.M. (2000) Overview on the structure,
composition, function, development, and plasticity of hippo-
campal dendritic spines. Hippocampus, 10(5): 501–511.
Steffensen, S.C. (1995) Dehydroepiandrosterone sulfate sup-
presses hippocampal recurrent inhibition and synchronizes neu-
ronal activity to theta rhythm. Hippocampus, 5(4): 320–328.
Steiner, M., Dunn, E. and Born, L. (2003) Hormones and
mood: from menarche to menopause and beyond. J. Affect.
Disord., 74(1): 67–83.
Stone, D.J., Rozovsky, I., Morgan, T.E., Anderson, C.P. and
Finch, C.E. (1998a) Increased synaptic sprouting in response
to estrogen via an apolipoprotein E-dependent mechanism:
implications for Alzheimer’s disease. J. Neurosci., 18(9):
3180–3185.

Stone, D.J., Rozovsky, I., Morgan, T.E., Anderson, C.P.,
Lopez, L.M., Shick, J. and Finch, C.E. (2000) Effects of age
on gene expression during estrogen-induced synaptic sprout-
ing in the female rat. Exp. Neurol., 165(1): 46–57.
Stone, D.J., Song, Y., Anderson, C.P., Krohn, K.K., Finch,
C.E. and Rozovsky, I. (1998b) Bidirectional transcription
regulation of glial fibrillary acidic protein by estradiol in vivo
and in vitro. Endocrinology, 139(7): 3202–3209.
Szymczak, S., Kalita, K., Jaworski, J., Mioduszewska, B.,
Savonenko, A., Markowska, A., Merchenthaler, I. and
Kaczmarek, L. (2006) Increased estrogen receptor beta ex-
pression correlates with decreased spine formation in the rat
hippocampus. Hippocampus, 16(5): 453–463.
Tabibnia, G., Cooke, B.M. and Breedlove, S.M. (1999) Sex
difference and laterality in the volume of mouse dentate gyrus
granule cell layer. Brain Res., 827(1–2): 41–45.
Tabori, N.E., Stewart, L.S., Znamensky, V., Romeo, R.D.,
Alves, S.E., McEwen, B.S. and Milner, T.A. (2005) Ultra-
structural evidence that androgen receptors are located at
extranuclear sites in the rat hippocampal formation. Neuro-
science, 130(1): 151–163.
Tanapat, P., Hastings, N.B. and Gould, E. (2005) Ovarian
steroids influence cell proliferation in the dentate gyrus of
the adult female rat in a dose- and time-dependent manner.
J. Comp. Neurol., 481(3): 252–265.
Tanapat, P., Hastings, N.B., Reeves, A.J. and Gould, E. (1999)
Estrogen stimulates a transient increase in the number of new
neurons in the dentate gyrus of the adult female rat. J. Ne-
urosci., 19(14): 5792–5801.
Teter, B., Harris-White, M.E., Frautschy, S.A. and Cole, G.M.
(1999) Role of apolipoprotein E and estrogen in mossy fiber
sprouting in hippocampal slice cultures. Neuroscience, 91(3):
1009–1016.
Towart, L.A., Alves, S.E., Znamensky, V., Hayashi, S., McE-
wen, B.S. and Milner, T.A. (2003) Subcellular relationships
between cholinergic terminals and estrogen receptor-alpha in
the dorsal hippocampus. J. Comp. Neurol., 463(4): 390–401.
Veiga, S., Garcia-Segura, L.M. and Azcoitia, I. (2003) Neuro-
protection by the steroids pregnenolone and dehydroepiand-
rosterone is mediated by the enzyme aromatase. J.
Neurobiol., 56(4): 398–406.
415
Velisek, L. and Veliskova, J. (2002) Estrogen treatment protects
GABA(B) inhibition in the dentate gyrus of female rats after
kainic acid-induced status epilepticus. Epilepsia, 43(Suppl 5):
146–151.
Veliskova, J. (2006) The role of estrogens in seizures and ep-
ilepsy: the bad guys or the good guys? Neuroscience, 138(3):
837–844.
Veliskova, J., Velisek, L., Galanopoulou, A.S. and Sperber, E.F.
(2000) Neuroprotective effects of estrogens on hippocampal
cells in adult female rats after status epilepticus. Epilepsia,
41(Suppl 6): S30–S35.
Waters, E.M., Herrick, S.P., McEwen, B.S. and Milner, T.A.
(2005) Subcellular localization of progestin receptors in the
rat hippocampus. Society for Neuroscience Abstracts,
pp. 403–412.
Wehrenberg, U., Prange-Kiel, J. and Rune, G.M. (2001) Ster-
oidogenic factor-1 expression in marmoset and rat hippo-
campus: co-localization with StAR and aromatase. J.
Neurochem., 76(6): 1879–1886.
Weiland, N.G. (1992) Estradiol selectively regulates agonist
binding sites on the N-methyl-D-aspartate receptor complex
in the CA1 region of the hippocampus. Endocrinology,
131(2): 662–668.
Weiland, N.G. and Orchinik, M. (1995) Specific subunit
mRNAs of the GABAA receptor are regulated by proges-
terone in subfields of the hippocampus. Brain Res. Mol.
Brain Res., 32(2): 271–278.
Weiland, N.G., Orikasa, C., Hayashi, S. and McEwen, B.S.
(1997) Distribution and hormone regulation of estrogen
receptor immunoreactive cells in the hippocampus of male
and female rats. J. Comp. Neurol., 388(4): 603–612.
Wimer, C.C. and Wimer, R.E. (1989) On the sources of strain
and sex differences in granule cell number in the dentate area
of house mice. Brain Res. Dev. Brain Res., 48(2): 167–176.
Woolley, C.S., Gould, E., Frankfurt, M. and McEwen, B.S.
(1990) Naturally occurring fluctuation in dendritic spine den-
sity on adult hippocampal pyramidal neurons. J. Neurosci.,
10(12): 4035–4039.
Woolley, C.S. and McEwen, B.S. (1992) Estradiol mediates
fluctuation in hippocampal synapse density during the est-
rous cycle in the adult rat. J. Neurosci., 12(7): 2549–2554.
Plate 23.2. Subcellular localization of estrogen (ER), androgen (AR), and progestin (PR) receptors in the dentate gyrus. A subset of
GABAergic interneurons contains nuclear ERa (dark pink). Granule cells, newly born cells (identified by DCX) and some GABAergic
interneurons contain cytosolic and plasma membrane-associated ERb (blue). Dendritic spines, many originating from granule cells
contain ERa, ERb, AR (dark green), and PR (purple). A few dendritic spines in the hilus, likely originating from mossy cells, contain
ERa and ERb. ERa, ERb, AR, and PR are found in axons and axon terminals. Some ERa-containing terminals are cholinergic
(acetylcholine, orange); some ERb-containing terminals resemble monoaminergic boutons. Lot of astrocytes (stars), mostly in the
molecular layer, also contain ERa, ERb, AR, and PR. (For B/W version, see page 407 in the volume.)
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 24

Plastic processes in the dentate gyrus:

a computational perspective

Brian E. Derrick

Department of Biology, The Cajal Neuroscience Research Institute, The University of Texas at San Antonio,
6900 N. Loop 1604 West, San Antonio, TX 78249-0662, USA

Abstract: The dentate gyrus has the capacity for numerous types of synaptic plasticity that use diverse
mechanisms and are thought essential for the storage of information in the hippocampus. Here we review
the various forms of synaptic plasticity that involve afferents and efferents of the dentate gyrus, and, from a
computational perspective, relate how these plastic processes might contribute to sparse, orthogonal en-
coding, and the selective recall of information within the hippocampus.

Keywords: dentate gyrus; LTP; LTD; metaplasticity; pattern separation

Plastic processes of the dentate gyrus
The dentate gyrus occupies an unusual place in
both the history and anatomy of the hippocampal
formation. Historically, it was the center of excite-
ment, being the site where the phenomenon of long-
term potentiation (LTP), an activity-dependent in-
crease in synaptic efficacy, was first identified (Bliss
and Gardner-Medwin, 1973; Bliss and Lomo,
1973). However, in terms of synaptic plasticity,
the dentate gyrus has been relegated to ‘‘second
place’’ behind the CA1 region, upon which the bulk
of studies LTP have been focused for the past two
decades. Recently, the dentate has again become a
focus of much attention in that it is one of the few
neural sites that display neurogenesis after devel-
opment — dentate granule cells are continually
generated well into adulthood (Altman and Das,
1965; Bayer, 1982). This rather remarkable process,
which is unique to the dentate gyrus of the
Corresponding author. Tel.: +1 210 458 4273;
Fax: +1 210 458 5658; E-mail: bderrick@utsa.edu
hippocampal formation, has rekindled interest in
the function of the dentate gyrus, as well as indi-
cating a mechanism of plasticity besides LTP and
long-term depression (LTD) that also might con-
tribute to memory storage (Shors and Matzel,
1997).
Anatomically, the dentate occupies an unusual
place in the hippocampal formation. It is not part
of the hippocampus (the dentate gyrus, the hip-
pocampus proper — areas CA1–CA3 — together
with the hilar region — are structures of the ‘‘hip-
pocampal formation’’). In terms of evolution, it is
a recent structure, attaining its present form with
the emergence of mammals, the structure itself
appears to have been almost an afterthought,
tagged on to the hippocampus relatively late in
evolution (Treves and Roudi, 2005). From the
perspective of hippocampal function, here it has
remained — simply the first relay of a three-part
‘‘trisynaptic circuit,’’ where it has been suggested
to act as a ‘‘gate’’ to the hippocampus due prima-
rily to its modulation of output with behavioral
state (Winson and Abzug, 1978; Bramham and

DOI: 10.1016/S0079-6123(07)63024-6 417

418
Srebro, 1989). However, we know now that infor-
mation flow through the trisynaptic circuit may be
only one (and not even the primary) pathway of
input flow through the hippocampus (Mizumori et
al., 1989; Yeckel and Berger, 1990; Do et al.,
2002). Such findings offer new clues that have re-
defined the function and operation of both the
dentate gyrus and the hippocampus.
More surprising is that the dentate displays nu-
merous forms of neural plasticity. In addition to
the phenomenon of LTP, other plastic changes are
observed at synapses to and from the dentate, such
as LTD and metaplasticity. In addition, long-term
changes in granule cell function, characterized by
long-term alterations in neurotransmitter and ne-
uromodulator synthesis (White et al., 1987; Hong
et al., 1988), the formation of new synaptic con-
tacts on granule cell targets (mossy fiber synapto-
genesis, Ben-Ari and Represa, 1990; Cavazos et
al., 1991), and the addition of new granule cells
well into adulthood (adult granule cell neurogen-
esis, Altman and Das, 1965) are other plastic
processes displayed by the dentate gyrus that, to-
gether, have the potential to contribute to accurate
and efficient information storage within the hip-
pocampal formation. Because many of these plas-
tic changes involve the expression of numerous
genes, and result in long-lasting morphological
changes (such as synaptogenesis and neurogene-
sis), it seems odd that the one structure within the
hippocampal formation that shows the most pro-
nounced and sustained genetic, morphological,
and functional plastic changes would be overshad-
owed by the CA1 region of the hippocampus.
The purpose of this review is to introduce the
myriad of plastic processes that are displayed by
synapses of the dentate gyrus, and to describe their
possible contribution to information storage on the
basis of current information-theoretical views of in-
formation processing and memory storage that are
thought operative in the hippocampal formation.
The dentate gyrus: A brief description of anatomy
and architecture
As is the case for most of the hippocampal for-
mation, the dentate is essentially three-layered
archicortex (it is a surprise to many that the hip-
pocampus is actually cortex, albeit the evolution-
ary older three-layered archicortex; Amaral and
Witter, 1989). The principal cell of the dentate
gyrus is the granule cell. In the rat, the dentate
contains 1–2 million granule cells (Boss et al.,
1985), a very large number of cells given that cor-
tical input to the dentate (300,000–400,000 cells
of the entorhinal cortex in the rat) is, by compar-
ison, relatively small. The projection of a structure
to one with a larger number of neurons, referred to
as ‘‘fan out,’’ ‘‘input expansion,’’ or ‘‘input recod-
ing’’ (Rolls and Treves, 1998), offers clues to the
function of the dentate. It is an aspect of the dent-
ate gyrus thought crucial for ‘‘pattern separation,’’
a process now thought to be one of the primary
functions of the dentate, and essential for the for-
mation of discrete ‘‘memories’’ within the hippo-
campus proper (Marr, 1971; Morris and
McNaughton, 1987; Treves and Rolls, 1992, 1994).
Classically, the dentate gyrus was first charac-
terized as the initial stage of the three-stage ‘trisy-
naptic circuit’ of the hippocampal formation
(Andersen et al., 1966). As such, the dentate gyrus
represents the first stage of this circuit and the
primary target of highly processed cortical infor-
mation relayed from the entorhinal cortex via the
main synaptic input to the dentate, the perforant
pathway. The subsequent second stage of the so-
called ‘trisynaptic circuit’ is the output from the
dentate granule cells, the mossy fibers. The mossy
fibers then project to the most proximal dendritic
region of pyramidal cells within the CA3 region.
The mossy fiber axons of the granule cells are rel-
atively thin and unmyelinated, making them par-
ticularly slow as compared with the third and final
stage of the trisynaptic circuit, the CA3 projection
to CA1 mediated by CA3 ‘‘Schaffer collaterals.’’
Crucial to the more recent views of hippocampal
operation is the finding that the primary input to
CA3 pyramidal cells is not from the dentate gyrus.
Rather, the primary synaptic input to CA3 py-
ramidal cells is recurrent input arising from other
CA3 pyramidal cells (Amaral et al., 1993). This
recurrent excitatory input (termed the commis-
sural/associational CA3 pathway) is of particular
interest computationally, as such recurrent net-
works are capable of performing autoassociative
 419

Fig. 1. Schematic representation of the dentate–CA3 interface. Perforant path inputs to the CA3 region arrive directly via synapses on
the distal CA3 dendrites, and indirectly by a disynaptic relay from the dentate gyrus via the sparse but powerful mossy fiber–CA3
projections. Recurrent connections constitute the greatest input to CA3 pyramidal cells, and are thought to serve as the substrate for
autoassociative storage and recall. Adapted with permission from Amaral et al. (1990) and Rolls and Treves (1998).

function and thus represent a substrate for ‘‘con-
tent addressable’’ memories (or CAMs), a more
general term for distributed memory devices
(Kohonen, 1984).
The view of the hippocampal formation as a
‘trisynaptic circuit’ is, however, an oversimplifica-
tion of a much more elaborate and parallel input
system within the hippocampus (Yeckel and Berger,
1998). Anatomical studies indicate that there are
substantial direct perforant path projections to
both the CA3 and CA1 regions of the hippocam-
pus (Andersen et al., 1966; Amaral et al., 1990;
Claiborne et al., 1993). The dentate and CA3 re-
gions receive direct input from the same popula-
tions of neurons in layers II of the medial and
lateral entorhinal cortex (Amaral and Witter, 1989).
In addition, layer III neurons of the lateral and
medial entorhinal cortex project directly to apical
dendrites of CA1 pyramidal cells and these projec-
tions to the CA1 region constitute a separate ‘‘tem-
poroammonic pathway.’’ These inputs, once
ignored and deemed insignificant, were recognized
as potentially important by neuroanatomists
(Hjorth-Simonsen and Jeune, 1972; Steward 1976;
Steward and Scoville, 1976; Buzsaki, 1988; Witter
et al., 1988). Subsequent physiological studies in
vivo (Yeckel and Berger, 1990; Berzhanskaya et al.,
1998) demonstrated that the direct perforant path
projections to the CA3 region are not only capable
of discharging pyramidal cells, but also display LTP
(Do et al., 2002) as well as associative LTP (Marti-
nez et al., 2002), a feature crucial to many of the
models of CA3 computation (Treves and Rolls,
1994). A more realistic ‘‘schematic’’ view of dent-
ate–CA3 hippocampal circuitry is shown in Fig. 1.
An underappreciated feature of the cortical input
to the hippocampal formation is that not only can
the direct perforant path inputs activate CA3 py-
ramidal cells, but CA3 pyramidal cells are also the
first cells to respond to perforant path input (Yeckel
and Berger, 1990; Do et al., 2002). Extracellular
measures of the summed output of the CA3 region
and the dentate gyrus, as reflected in population
spike latency, indicate that the CA3 region responds
faster than the dentate gyrus, with spike discharges
in vivo preceding granule cell spikes by 0.5–3 ms
(Fig. 2). Thus, it is the CA3 region, rather than the
dentate, that is the first population of principal cells
to respond to cortical input relayed by the perforant
pathway. As noted in the original study (Yeckel and
Berger, 1990), this finding is in line with a number of
studies indicating early evoked responses in the CA3
region that precede dentate responses (Segal, 1972).
Because the CA3 region is the first region to respond
with spike output to perforant path activity, the
primacy of CA3 responses is likely relevant to the
function of the rather unusual ‘‘backward’’ projec-
tions from CA3 pyramidal cells to hilar neurons
420

Fig. 2. Comparison of perforant path responses evoked in the dentate gyrus and CA3 region. Shown are field potentials recorded in
the dentate gyrus (hatched lines) or the CA3 pyramidal layer (solid lines) following activation of the medial (top) or lateral perforant
pathway (bottom traces). Responses were recorded in vivo from the same animals under pentobarbital anesthesia. Calibration: 0.5 mV,
5 ms (CA3); 1 mV, 5 ms (dentate).

(Scharfman, 1996). Because these interneurons can
exert both excitatory and inhibitory actions on
dentate granule cells, the projection from the CA3
region to these interneurons may serve to modulate
dentate granule cell output in response to direct
perforant path–CA3 input.
The granule cells of the dentate gyrus dentate
appear, by many accounts, unremarkable. They are
small (6–10 mm) and have extensive, conical, spin-
ous dendrites that extend almost 1 mm into the
stratum moleculare (or the ‘‘molecular layer’’) of
the dentate (Steward, 1976; Amaral and Witter,
1989). It is here in the molecular layer that the
dentate gyrus receives its input from three primary
sources: the lateral entorhinal cortex (LEC), which
targets the outer one third of the dentate molecular
layer; the medial entorhinal cortex (MEC), whose
projections occupy the middle one third of the
granule cell dendrites in the molecular layer; and the
commissural/associational projections, originating
from hilar glutamatergic ‘‘mossy’’ cells and termi-
nating in the innermost one third of the molecular
layer (Amaral and Witter, 1989). The dentate asso-
ciational projections reflect yet another recurrent
excitatory pathway within the hippocampal forma-
tion, albeit one that is relayed disynaptically via
mossy cells (Amaral, 1978; Amaral and Witter,
1995). The granule cells themselves display features
that are unique as well: they have a high threshold
for activation, and granule cell bursting occurs
rarely in vivo (Jung and McNaughton, 1993), al-
though granule cells can fire sustained bursts during
theta rhythm (Munoz et al., 1990), a 4–12 Hz os-
cillation that is a prominent feature of hippocampal
activity during exploration and, presumably, during
learning (Buzsaki, 2002).
Granule cells are notoriously silent, as evidenced
by the difficulty in obtaining ‘‘place fields’’ —
single unit firing that correlates with a specific
place field in an environment (O’Keefe and Nadel,
1978) — and they often fire at very low rates
(approximately 0.5 Hz, Jung and McNaughton,
1993). As discussed further below, these tight con-
straints on granule cell activity may be viewed as

another feature crucial for ‘‘pattern separation’’
and the formation of sparse outputs by the dentate
gyrus (Skaggs and McNaughton, 1992).
Another unusual aspect of the dentate is that the
V-shaped blades of this structure appear not to be
homogeneous, actually differing in a number of
substantial ways. They appear to be under separate
genetic control, as the lower (infrapyramidal) blade
and the mossy fiber (suprapyramidal) layer differ
among strains of mice (Schwegler et al., 1990). In
addition, mossy fiber efferents arising from granule
cells in the internal (suprapyramidal) layer ramify
and extend through the stratum lucidum of the
CA3 region, whereas the mossy fibers arising from
the outer (infrapyramidal) blade of the dentate
ramify extensively in the hilus and terminate pri-
marily in the infrapyramidal layer of the CA3 re-
gion lying just below the CA3 pyramidal cell layer,
or stratum pyramidale (Claiborne et al., 1986). The
two blades may also have functional differences,
because the infrapyramidal region develops greater
mossy fiber sprouting following seizures or high-
frequency mossy fiber stimulation (Cavazos et al.,
1991; Escobar et al., 1997; Scharfman et al., 2002),
as well as spatial learning (Ramirez-Amaya et al.,
2001). In support of possible ‘‘divisions of labor’’
among the dentate blades, a greater degree of ex-
citability is observed in the lower (infrapyramidal)
blade, possibly due to differential innervation of
granule cells by inhibitory interneurons (Scharfman
et al., 2002). Moreover, dendrites on granule cells in
the infrapyramidal layer show fewer spines (Clai-
borne et al., 1986). Interestingly, while both blades
show the expression of Arc, an immediate-early
gene associated with synaptic plasticity, following
stimulation of the perforant path that induces LTP,
Arc expression in behaving animals associated with
learning is seen only in granule cells in the supra-
pyramidal (inner) blade lying between the hilus and
the CA1 region (Chawla et al., 2005). However, the
functional relevance of these differences among the
blades of the dentate gyrus is unknown.
The efferents of the granule cells, the mossy fib-
ers, also display unusual features: these granule
cell efferents are thin, unmyelinated fibers that,
upon their exit from granule cells, form many bi-
furcations, projecting primarily to the proximal
spines of CA3 pyramidal neurons. Although the
421
connections of granule cells to CA3 are sparse,
with each granule cell making contact with only
8–10 CA3 pyramidal cells, they form unusual
synapses on CA3 cells, with each mossy fiber
bouton essentially engulfing a highly elaborated
dendritic CA3 spine with multiple synaptic con-
tacts, often referred to as ‘‘thorny excrescences’’
(Gonzales et al., 2001). The mossy fiber synapse
itself is unusual in several respects, with the pri-
mary difference being that LTP observed at these
synapses is of a type that does not require the ac-
tivation of N-methyl-D-aspartate (NMDA) recep-
tors (Harris and Cotman, 1986).
In addition to the CA3 pyramidal cells, the
mossy fibers also target numerous neurons within
the hilar region. The array of neuronal subtypes
within the hilar region of the dentate is impressive
(Amaral, 1978); however the most studied neurons
in the hilus are the so-called ‘‘mossy cells.’’ These
cells are innervated by the mossy fibers and, by
virtue of their excitatory glutamatergic projections,
both ipsilaterally and contralaterally to the inner
molecular layer, comprise the commissural/
associational system of the dentate gyrus. This re-
current excitatory input has been scrutinized re-
cently (Kleschevnikov and Routtenberg, 2003), and
studies have revealed that these projections show
plasticity both at the mossy fiber–mossy cell
synapse and at their contacts with granule cells in
the inner molecular layer. Those familiar with the
major target of the dentate granule cells, the CA3
region, may notice a similarity between the dentate
and CA3 region in that both have extensive recur-
rent excitatory inputs that also project contralater-
ally. Thus, both the dentate and CA3 region have
extensive, bilateral, and recurrent excitatory feed-
back. Thus these two structures can be thought of
as a single, bilateral, highly re-entrant structure.
The functional implications of this feature remain
to be defined.
Information/theoretic views of the dentate gyrus
Although the dentate gyrus defies any single ac-
count of function based on its operation (such as its
early conceptualization as ‘‘gating’’ input to the
CA3 region (Winson and Abzug, 1978), clues to its
422
functions may be provided by current information-
theoretic models of hippocampal information
processing. To more fully appreciate the contribu-
tion of the dentate to hippocampus-based memory,
it is necessary to consider information-theoretic
models of information storage in the CA3 region of
the hippocampus.
The late David Marr’s original conceptualizat-
ion of hippocampal function remains central to
most information-theoretic theories of hippocam-
pal operation today (Marr, 1971). As suggested by
Marr, the CA3 region may serve as an ‘‘autoasso-
ciator’’ by virtue of its extensive recurrent
(CA3–CA3) excitatory connections (Morris and
McNaughton, 1987; Treves and Rolls, 1994; Rolls
and Treves, 1998). Neural network models of
autoassociators were pioneered by theoreticians
such as Tuevo Kohonen (Kohonen et al., 1977;
Kohonen, 1984) and later re-formalized by John
Hopfield (Hopfield, 1982). In such models auto-
associators are networks which by virtue of their
extensive recurrent input are capable of unique
computational operations. Representations in au-
toassociative systems involve the formation of
‘‘attractors’’ (or stable ensembles of active, recur-
rently connected cells as a result of plastic changes
among recurrent synapses). Thus autoassociative
architectures and their recurrent connections can
allow for the formation of stable ‘‘attractor’’
states. This architecture also can enable the recall
of an entire pattern of CA3 activity in response to
the presentation of only a subset of the original
pattern. This feature, termed ‘‘pattern comple-
tion,’’ is a fundamental computational feature of
autoassociators (Kohonen et al., 1977). If one
thinks of an ensemble of many recurrently con-
nected CA3 neurons with strengthened synapses
among them as a representation, one can see that
the activation of a subset of this ensemble could,
via recurrent activation of other CA3 neurons via
strengthened synapses, recruit the entire original
ensemble. Thus recurrent inputs provide the subst-
rate of CA3 ‘‘autoassociation’’ and pattern com-
pletion from partial inputs, a property that may
encode not only single stable ensembles of co-ac-
tive CA3 pyramidal cells, but also sequences of
stable states, provided that the output from the
CA3 system can, via the CA3 recurrent inputs,
become associated with subsequent CA3 inputs
over time (Levy, 1996). The view of autoassocia-
tive processes in the CA3 region, first formalized
by Marr (1971), remains a feature of most models
of hippocampal function (Morris and McNaugh-
ton, 1987; Treves and Rolls, 1994; Hasselmo et al.,
1995; McClelland et al., 1995; Rolls and Treves,
1998) and is supported by both models of the CA3
region and behavioral studies (Nakazawa et al.,
2002).
As noted above, the recurrent CA3–CA3 con-
nections constitute the majority of excitatory in-
puts to CA3 pyramidal cells, and these fibers also
display Hebbian LTP (Harris and Cotman, 1986).
It is thought that representations within the CA3
autoassociative system involve the formation of
attractor states (simply relatively stable ensembles
of active, recurrently connected CA3 cells) as a
result of plastic changes at the recurrent CA3–CA3
synapses. These plastic changes are thought to oc-
cur by potentiation of active CA3–CA3 synapses
whose activity is initiated by the direct perforant
path–CA3 inputs, and ‘‘fixed’’ by strong postsy-
naptic activity initiated by the mossy fiber
synapses from dentate granule cells. In this
scheme, the mossy fiber synapses are thought to
act as ‘‘detonator’’ synapses (Morris and McN-
aughton, 1987) that direct associative plasticity,
and thus encoding, in this system. Due to the rel-
ative proximity of these synapses to the CA3 py-
ramidal cell body, activation of these ‘‘detonator’’
synapses can initiate CA3 cell firing and, via
postsynaptic depolarization, induce LTP among
other synapses to the CA3 pyramidal cell that were
recently active (Levy and Steward, 1979, 1983).
Thus, encoding of information in the CA3 region
is thought to result from the activation of CA3
recurrent collaterals by perforant path input and
the subsequent potentiation of recently active per-
forant path–CA3 and CA3–CA3 synapses by
mossy fiber ‘‘detonator synapses.’’
Although autoassociators are capable of pattern
completion given a partial input of the original
pattern, they have their limitations. Crucial for
efficient encoding and accurate recall in autoasso-
ciative systems is a sufficient separation of patterns
that are encoded within the CA3 region, such that
stored representations in the CA3 system have

minimal overlap with other, previously encoded
ensembles (orthogonalization). This separation of
patterns is necessary because accurate encoding
and retrieval within autoassociative memory sys-
tems can occur only if correlations and redundan-
cies among discrete attractor states (ensembles of
active CA3 neurons) are reduced (see Rolls and
Treves, 1998). Thus, encoding information in an
autoassociator demands special considerations
and requires the formation of orthogonal, or un-
correlated, ensembles of CA3 attractors. How this
is done is the subject of much speculation, al-
though a feature common to many models is that
orthogonalization (sometimes referred to more
generally as ‘‘pattern separation’’) requires sparse
encoding, a process that is crucial for encoding in
the CA3 region. This sparsification is accom-
plished primarily by transformation of perforant
path inputs to the dentate into a sparse code that is
then relayed to the CA3 region (Morris and McN-
aughton, 1987). The formation of sparse dentate
outputs can increase the capacity of a recurrent
autoassociative system, and by virtue of a fewer
number of outputs (and plasticity among them)
can also serve to reduce any similarities among
ensembles formed within the CA3 region. Sparse
encoding can serve as one process that minimizes
similarity and interference among CA3 attractors,
because representations using a minimal number
of elements are less likely to overlap with other
stored patterns (Rolls and Treves, 1998).
In many models of hippocampal function, the
task of ‘‘pattern separation’’ and orthogonalizat-
ion of cortical inputs relayed to the CA3 region is
ascribed to the dentate gyrus (Morris and McN-
aughton, 1987; Treves and Rolls, 1994; Rolls and
Treves, 1998). The dentate gyrus can serve in the
formation of sparse, orthogonal CA3 ensembles
via several mechanisms. As discussed extensively
by Rolls and Treves (1998), the dentate gyrus is
thought to form sparse outputs to the CA3 region,
a process that can be facilitated in a variety of
ways by features of the dentate gyrus. Foremost
among these features are the anatomical charac-
teristics of the dentate gyrus–mossy fiber system.
This system appears ideally suited to ensure the
formation of sparse, uncorrelated input patterns
relayed to the CA3 region. The rat dentate gyrus
423
consists of 1–2 million granule cells that receive
input from only 300,000 cells in the entorhinal
cortex (Boss et al., 1985). This ‘‘fan out’’ of cor-
tical input to a structure with many more elements
is one way to enforce pattern separation (referred
to as ‘‘input expansion’’ or ‘‘expansion recoding;’’
Rolls and Treves, 1998). In addition, both the rel-
ative quiescence of granule cells in the dentate and
the sparse connectivity of the mossy fiber projec-
tions (so that each granule cell forms mossy fiber
contacts with only 10–15 CA3 pyramidal cells;
Amaral et al., 1990) together make it unlikely that
any given CA3 pyramidal cell can be activated by
more than one granule cell. Thus the anatomically
sparse dentate–CA3 connectivity by the mossy fib-
ers is an important feature for the formation of
sparse, orthogonal CA3 representations. Together,
these aspects of the dentate are ideally suited for
pattern separation, and the formation of sparse,
orthogonal outputs to the CA3 region that allow
the encoding of discrete, stable ensembles or ‘‘at-
tractors’’ in the CA3 autoassociative system.
While important, the anatomical features of the
mossy fiber output are but one way the dentate
may relay sparse transforms to the CA3 autoas-
sociative system. As we will discuss later, when
taken in a computational context, the plastic fea-
tures of the dentate gyrus, including LTP, the var-
ious forms of LTD, and metaplastic processes,
together, not only may ensure sparse, orthogonal
outputs to the CA3 autoassociative system, but
also may allow for the dentate to encode discrete
attractors for similar, or even identical inputs that
are encoded over time. In this view, the combina-
tion of plastic processes in the dentate that allow
the formation of sparse, orthogonal CA3 inputs
also may allow partial inputs to converge on dis-
tinct stable states encoded by very similar inputs
over time — a feature characteristic of episodic
memory, long thought of as the principal type
of memory involving the human hippocampus
(Tulving, 1983).
Synaptic plasticity in the dentate gyrus — LTP
Foremost among the plastic processes studied in
the dentate gyrus is LTP. The hypothesis that
424

Fig. 3. Perforant path–dentate LTP and heterosynaptic LTD. Plot of field EPSP slopes of medial (K) and lateral (J) perforant
path–dentate responses following high-frequency stimulation of the medial perforant pathway. In addition to the potentiation ob-
served at medial perforant path synapses, a sustained depression of inactive lateral perforant path synaptic responses (‘‘heterosynaptic
LTD’’) is observed. Calibration: 0.5 mV, 5 ms. Villarreal and Derrick (2006).

memory involves changes in synaptic efficacy has
remained a primary assumption for quite some
time (Ramon y Cajal, 1937; Hebb, 1949; James,
1890). LTP is a rapidly induced and long-lasting
change in synaptic strength that is observed fol-
lowing high-frequency afferent activity. LTP was
first described formally in studies of the rabbit
perforant path projections to the dentate gyrus
using an in vivo preparation (Bliss and Lomo,
1973), and the first studies that addressed LTP
longevity also were conducted using perforant
path–dentate recordings (Barnes, 1979). However,
since its discovery, and the subsequent discovery
of similar plastic processes in other brain struc-
tures (from the amygdala to spinal cord), the bulk
of studies of LTP, particularly in vitro, have fo-
cused on the Schaffer–CA1 synapse (Fig. 3).
LTP has features that make it a compelling
candidate mechanism that may contribute to in-
formation storage. LTP is induced rapidly after its
induction, which, in the laboratory, requires a
sufficient postsynaptic depolarization that usually
is provided by high frequency stimulation of a
sufficient number of afferents. Dentate LTP also is
‘‘input specific,’’ that is, LTP is confined to
synapses of afferents activated with stimulation.
This synapse specificity is in contrast to the non-
associative phenomena, such as sensitization,
wherein a general increase in responsivity of the
postsynaptic cell is observed. Importantly, input
specificity is a feature essential for any kind of
memory that requires both high capacity and high
fidelity, as this feature provides sparsity by in-
creasing the number of modifiable inputs that can
aid in encoding and retrieval in autoassociative
systems.
Another crucial feature of LTP making it an
attractive potential mechanism for information
storage is that it is remarkably persistent. Prior to
the discovery of LTP, the only activity-dependent
electrophysiological change of substantial dura-
tion was post-tetanic potentiation (PTP), a prima-
rily presynaptic phenomenon lasting from seconds
to minutes (Zucker and Regehr, 2002). By con-
trast, in the hippocampal formation, LTP can per-
sist from hours to weeks or months, depending on
the stimulation parameters. In an intact animal,
LTP usually decays within several weeks (Barnes,
1979).
A number of investigators have dismissed LTP
as a long-term memory mechanism because its
duration is far too short to account for memory

that lasts years (Shors and Matzel, 1997). While
plasticity that lasts only days to weeks is too brief
to store long-term memory, the limits of LTP lon-
gevity may be an important feature of LTP in the
hippocampal formation. One reason is that LTP,
at least within the hippocampus, need not be per-
manent. This is because, as is thought to be the
case in humans (Scoville and Milner, 1957), the rat
hippocampus also is thought to have a time-de-
limited role in memory, with persistent, ‘‘long
term’’ memory being gradually consolidated in
neocortical areas over time (Marr, 1971; Squire,
1992). Currently, the most popular view is that
memories formed by the hippocampus are trans-
ferred to and consolidated in the neocortex, pos-
sibly during slow wave (NREM) sleep states
(Wilson and McNaughton, 1993; Cartwright,
2004), a process that usually requires 2–3 weeks
in the rat, as indicated by both lesion and imaging
data (Bontempi et al., 1999). Thus, if the hippo-
campus does indeed serve as a temporary repos-
itory, or ‘‘buffer’’ for information, LTP may not
need to persist for extended periods of time. In this
scheme, any long-term retention of information
within the hippocampus beyond the a few weeks
might even be detrimental to hippocampal-based
memory, resulting in interference with previously
stored patterns of synaptic activity (Rosenzweig
et al., 2002).
How long, then, can LTP last? We reported
that, in perforant path projections to the dentate
gyrus, LTP decay is an active process mediated by
NMDA receptors, and blocking these receptors
can prevent decay (Villarreal et al., 2002). There-
fore it appears that LTP doesn’t simply ‘‘fade;’’
instead, it is actively ‘‘erased’’ via processes that
require NMDA receptor activation. A likely can-
didate mechanism for such ‘‘erasure’’ is LTD. As
discussed more fully below, dentate LTD, like
LTP, requires the activation of NMDA receptors
(Desmond et al., 1991; Christie and Abraham,
1992b). Thus, NMDA receptor activation and a
reversal of LTP by the induction of LTD (also
referred to as ‘‘depotentiation,’’ both discussed
below) may mediate LTP decay. However,
NMDA receptors on dentate granule cells can be
activated by low frequency perforant path synaptic
activity and even miniature EPSPs (Dalby and
425
Mody, 2003). Consequently, background synaptic
activity completely unrelated to learning may be
sufficient to mediate LTP decay at perforant
path–granule cell synapses, suggesting that synap-
tic activity unrelated to learning also may contrib-
ute to LTP decay. Regardless, because LTP decay
requires activation of NMDA receptors, LTP de-
cay still reflects an active process (Villarreal et al.,
2002). Thus perforant path–dentate LTP, at least
theoretically, is potentially permanent, and could
last as long as the synapse itself, provided that it is
not ‘‘erased’’ by subsequent synaptic activity and
NMDA receptor activation.
LTP also is ‘‘associative.’’ Specifically, if activity
in a set of afferents to a common postsynaptic
target induces LTP, then individual active
synapses can be recruited to express LTP, pro-
vided that the synapse is coactive within a delim-
ited window of time (Levy and Steward, 1983).
The phenomenon of associativity is derived from
the requirements for activation of the NMDA re-
ceptor (specifically, both presynaptic release of
glutamate and postsynaptic depolarization, with
the latter being essential for relieving the magne-
sium block of the NMDA channel). Associativity
is a property derived directly from, and is essen-
tially identical to, the property of ‘‘cooperativity’’
(McNaughton et al., 1978). Cooperativity indi-
cates LTP induction has a threshold, and activa-
tion of a threshold number of afferents is essential
to induce LTP. Because LTP is induced in a ‘‘co-
operative’’ manner, it follows that associativity is
an inherent property of ‘‘cooperativity.’’ Although
it has been suggested that ‘‘cooperativity’’ and
‘‘associativity’’ are distinct, with cooperativity re-
flecting a threshold of afferent activity necessary to
induce LTP and associativity involving distinct
sets of afferents, it should be noted that the first
demonstration of cooperativity used two distinct
afferent systems to the dentate gyrus, the medial
and lateral perforant pathways (McNaughton et
al., 1978). However, because interactions among
active afferent fibers could not be ruled out, the
term cooperativity, rather than associativity, may
have been chosen out of rigor, as the term ass-
ociativity, as envisaged by Hebb, tacitly implies a
postsynaptic integration of presynaptic activity,
something that could not be ruled out with
426
‘‘cooperative’’ effects. It should be noted here that
these studies (McNaughton et al., 1978; Levy and
Steward, 1983) were conducted at a time when the
NMDA receptor was unheard of. Subsequently, a
convincing demonstration of Hebbian-like ass-
ociativity was provided by studies showing that
pairing of low intensity, low frequency afferent
activity with postsynaptic depolarization is capa-
ble of inducing LTP (Gustafsson and Wigstrom,
1986; Gustafsson et al., 1987).
It is often assumed that mechanisms underlying
LTP can be generalized. However, what has
emerged from recent studies is that the mechanisms
underlying the induction, and possibly the mainte-
nance and expression, of LTP not only differ at
different synapses within the hippocampal forma-
tion, but may even involve a number of distinct
mechanisms, even at a single synapse. These differ-
ences can arise from numerous variables, including
what stimulation parameters are used, and even the
animal’s behavioral state at the time of LTP induc-
tion (Davis et al., 2004; see below). Thus there may
be a multitude of activity-dependent mechanisms
that serve to increase synaptic strength, all of which
have been labeled ‘‘LTP.’’ Yet these processes may
have distinct mechanisms of induction (conditions
necessary to initiate LTP) and/or maintenance
(mechanisms initiated by LTP induction that un-
derlie its expression for sustained periods of time).
Such differences in both induction and maintenance
mechanisms may be indicative of distinct ‘‘forms’’
of LTP.
The heterogeneity and complexity of LTP are
illustrated by the features of LTP induction in the
primary input to the dentate — the perforant
pathway. As noted above, the perforant pathway
projections originate from layer II neurons within
the medial and the lateral entorhinal cortices, and
they target, respectively, the middle one third of
the molecular layer for medial perforant path
(MPP) afferents, and the outermost third of the
molecular layer for the lateral perforant path
(LPP) inputs (Steward, 1976; Amaral and Witter,
1989).
Responses to the medial perforant path projec-
tion are, for a variety of reasons, the most robust,
and many studies that employ stimulation of the
perforant path often are likely to evoke responses
from medial perforant path afferents. This may be
due, at least in part, to the fact that the medial and
lateral projections are stratified in the angular
bundle, so that only stimulation of fibers deep in
the angular bundle evokes lateral perforant path
responses (McNaughton and Barnes, 1977). Lat-
eral perforant path–dentate responses differ from
medial responses in a number of ways, including a
slower rising phase of dentate field EPSPs (due to
the more distal regions of the granule cell dendrite
where lateral perforant path synapses terminate).
In addition, medial perforant path responses show
a pronounced frequency depression of synaptic
responses, whereas lateral perforant path re-
sponses show robust frequency facilitation (McN-
aughton and Barnes, 1977). The reasons for these
differences in afferent responses are not clear, al-
though it likely involves differences among the
afferents themselves, as depression and facilitation
are also observed when paired pulses are delivered
to medial and lateral perforant path projections to
the CA3 region, (Do et al., 2002).
It may come as no surprise that LTP in these
two pathways differ in a number of important
ways. The medial perforant pathway is a glut-
amatergic afferent system and displays LTP that
depends on activation of the NMDA glutamate
receptor. Thus LTP at medial perforant path
synapses may reflect a form of LTP induction
similar to the more widely-studied type at the
Schaffer–CA1 synapse, as well as among neocor-
tical regions (Kirkwood et al., 1993; Trepel and
Racine, 1998), and appears to be the primary re-
ceptor involved in the induction of LTD as well
(Desmond et al., 1991; Christie and Abraham,
1992b). NMDA receptors are activated by gluta-
mate released by the presynaptic terminal. NMDA
receptors are also voltage-dependent, such that
current flow is prevented at hyperpolarized mem-
brane potentials due to a cationic block of the
channel by magnesium. However, the combination
of presynaptic activity with postsynaptic depolari-
zation sufficient to remove the cationic block of
the NMDA receptor pore allows current flow (cal-
cium influx via NMDA receptor channels), allow-
ing for the induction of LTP or LTD. This
property of NMDA receptors, the requirement
for both presynaptic activity together with a

sufficient level of postsynaptic activity, is essen-
tially a physical instantiation of the synaptic learn-
ing rule formalized by Donald Hebb in the 1940s
(Hebb, 1949). It is now referred to as the ‘‘Hebb
rule,’’ this simple local learning rule requiring cor-
related pre- and postsynaptic activity has signifi-
cant computational power when employed in
various neural network architectures. Networks
with elements that use the Hebb rule display im-
portant computational features, and therefore it is
thought to be a principal ‘‘coincidence detector’’
that mediates synaptic plasticity based on corre-
lated pre- and postsynaptic activity.
As with many synapses in the hippocampal for-
mation, LTP at the medial perforant path–dentate
synapse is blocked by antagonists selective for
NMDA receptors. As with most known forms of
LTP, the influx of calcium postsynaptically as a
result of NMDA receptor activation is thought
to be the crucial first step in LTP induction
(Barrionuevo et al., 1980). The subsequent steps
following calcium entry are not yet defined in detail,
but as with LTP at other synapses, a concerted
effort has been made to identify important molec-
ular and biochemical events that eventually increase
synaptic strength. Here a number of protein kinases
have been implicated, including a-calcium/calmod-
ulin kinase II (CaMKII; Thomas et al., 1994; Wu et
al., 2006), protein kinase A (Wu et al., 2006), and
the ERK/MAPK cascade (Maguire et al., 1999;
Davis et al., 2000; Sweatt, 2001; Wu et al., 2006).
The subsequent phosphorylation of the cAMP re-
sponse element binding protein (CREB) is likely the
first transcription factor activated by these kinases
(Ying et al., 2002), and this is thought to initiate the
expression of immediate-early genes (IEGs; Schulz
et al., 1999; Davis et al., 2000). The later phase
(>3 h) of LTP maintenance in the dentate requires
the synthesis of proteins (Krug et al., 1984; Nguyen
and Kandel, 1996) and transcription of new
mRNAs (Frey et al., 1996), and appears to involve
a later activation of protein kinase C (Colley et al.,
1990; Thomas et al., 1994).
Several IEGs are implicated in the expression of
LTP. These genes are induced rapidly in a protein
synthesis-independent manner, and generate a
number of proteins that, together with other IEG
proteins, can form transcription factors that
427
regulate the expression of other ‘‘late effector’’
genes. The major IEGs implicated in LTP include
Arc (Steward et al., 1998), Homer (Kato et al.,
1997), and Zif268 (Davis et al., 2000; Jones et al.,
2001). Arc is a very interesting IEG in that the
mRNA for this gene is transported into granule
cells dendrites, where the synthesis of Arc proteins
occurs (Steward et al., 1998; Guzowski et al.,
2000). Similarly, the transport of CaMKII
mRNAs to synaptic regions, and enhanced syn-
thesis of CaMKII, also is observed (Havik et al.,
2003). Finally, there is a protracted increase in the
synthesis of NR2B subunits of the NMDA recep-
tor (Thomas et al., 1996; Williams et al., 2003).
Together, these proteins are suggested to partici-
pate in the formation of elaborate modifications to
the postsynaptic density (PSD) of granule cell
synapses, a process thought to involve NR2B
NMDA receptor subunits and CaMKII via inter-
actions with actin in PSD. Here, elaboration of the
postsynaptic element and localization of CaMKII
may mediate the insertion of new receptors and/or
receptor subunits (Lisman and McIntyre, 2001),
also referred to as ‘‘AMPA receptor trafficking’’
(see Malinow, 2003, for review). These genes also
may mediate the morphological changes reported
to occur at axospinous synapses in the molecular
layer (Trommald et al., 1996), such as changes in
postsynaptic structure (Desmond and Levy, 1983,
1986). However, few data exist for the role of
AMPA trafficking in dentate granule cells, as most
data supporting this process in LTP were obtained
in studies of LTP in the Schaffer–CA1 pathway.
Many of the genes thought to be involved in dent-
ate LTP induction (genes that also code for pro-
teins that can associate with the PSD and that are
implicated in AMPA receptor trafficking) are also
thought to be involved in LTP at the Schaf-
fer–CA1 synapses. Thus NMDA receptor-depend-
ent LTP as observed at medial perforant
path–dentate and Schaffer–CA1 synapses may re-
flect a common, prototypical ‘‘form’’ of LTP in-
duction and expression.
Induction of LTP at the lateral perforant path
synapse is quite different from the medial perfo-
rant path projections. As first demonstrated by
Bramham et al. (1988), naloxone, an antagonist of
the opioid receptor, blocks lateral perforant
428
path–dentate LTP in vivo. This was verified by
studies in vitro, where naloxone blocked lateral
perforant path–dentate LTP, and m opioid recep-
tor-selective agonists facilitated LTP induction
(Xie and Lewis, 1991). However, in these studies,
naloxone had little or no effect on low frequency
lateral perforant path–dentate responses. The lack
of effect of this opioid receptor antagonist on re-
sponses evoked at low frequencies suggests that
endogenous opioid peptides play a frequency-de-
pendent role in LTP induction at lateral perforant
path–dentate synapses (Bramham et al., 1988;
Matthies et al., 2000; Krug et al., 2001). Although
the requirement for opioid peptides in LTP induc-
tion may seem unusual, it may be less surprising
when one considers that the lateral entorhinal cor-
tex projections co-release opioid peptides with its
principal transmitter, glutamate (Chavkin et al.,
1985). These peptides, derived primarily from pro-
enkephalin, with [Met]- and [Leu]-enkephalin be-
ing the primary opioid peptides produced by pro-
enkephalin and released by the lateral perforant
path, are stored in dense-core vesicles and are re-
leased in an activity-dependent manner (Neumaier
and Chavkin, 1989; Wagner et al., 1990, 1991).
The frequency dependence of opioid peptide re-
lease by the lateral perforant path is common to
peptide co-transmitters; principally, their release
often requires intense or repetitive activity (Hong
et al., 1988). It is thought that such activity is es-
sential for the release of neuropeptides that are
stored in dense core vesicles, as they often are re-
mote from both release sites and plasma mem-
brane presynaptic terminals, and thus may require
large amounts of calcium for their release. This
aspect of peptide storage and release may also ex-
plain why opioid receptor blockade has little effect
on synaptic responses evoked in opioidergic pro-
jections at low frequencies. Tacit in this interpre-
tation is the idea that high-frequency lateral
perforant path activity is necessary for LTP in-
duction in order to achieve both postsynaptic de-
polarization and the frequency-dependent release
of opioid neuropeptides.
A similar dependence of lateral perforant path
LTP on high-frequency afferent activity is also
observed in vitro, as pairing low frequency presy-
naptic activity with postsynaptic depolarization,
while effective in inducing LTP in the medial per-
forant–dentate synapses, is not effective in induc-
ing LTP in lateral perforant path–dentate synapses
(Colino and Malenka, 1993). A similar lack of
‘‘single pulse associativity’’ also is observed at
mossy fiber synapses on CA3 pyramidal cells, a
synapse that also contains and releases opioid
peptides (Chavkin et al., 1985) that are thought
essential for Hebbian LTP induction at the mossy
fiber synapse (Derrick et al., 1994a; Williams and
Johnston, 1996). While this effect has been inter-
preted as reflecting a lack of associativity at mossy
fiber synapses (Zalutsky and Nicoll, 1990), the lack
of single pulse mossy fiber associativity may sim-
ply reflect this requirement for mossy fiber activity
necessary for the frequency-dependent release of
opioid peptides (Neumaier and Chavkin, 1989;
Wagner et al., 1990, 1991; Derrick and Martinez,
1994a). Thus the requirement for high frequency
activity for LTP induction at lateral perforant
path–dentate synapses similarly may reflect a re-
quirement for the frequency-dependent release of
opioid peptides for LTP induction (Colino and
Malenka, 1993). There are a number of other ac-
tions of both pro-enkephalin and pro-dynorphin-
derived opioid peptides released by dentate gran-
ule cells. Most notable is the rather surprising
finding that dynorphins, which act at kappa re-
ceptors and exert primarily inhibitory effects, at-
tenuate perforant path synaptic responses and
LTP induction, most likely via dendritic release of
dynorphin-derived peptides (Wagner et al., 1993),
adding a different dimension to the roles of opioid
peptides in hippocampal information processing.
The actions of opioids within the hippocampus
are varied and often depend on the receptors in-
volved as well as the site of the receptors. How-
ever, the primary effect of many opioid peptides in
the hippocampus is that they elicit excitation (Fry
et al., 1979; Corrigall, 1983), which is observed
with activation of both the m and d opioid receptor
types (Piguet and North, 1983). The application of
exogenous opioids can even elicit seizures, and re-
peated application can eventually elicit kindling
(Cain and Corcoran, 1985). These excitatory
effects are unusual for m and d opioid receptors,
as these opioid receptor types are usually coupled
to the superfamily of GTPase-binding proteins

that inhibit cellular responses (Piguet and North,
1983). Why would neuromodulators normally as-
sociated with inhibitory actions elicit excitation
within the hippocampus? This paradox was clar-
ified in subsequent studies indicating that one
principal mechanism by which opioids mediate
their excitatory effects is their attenuation of
GABA release (Cohen et al., 1992). Thus the ex-
citatory effects of proenkephalin-derived opioid
peptides result from an attenuation of GABAergic
inhibition, an effect thought to result from the m or
d opioid receptor-mediated enhancement of po-
tassium conductances in GABAergic interneurons,
which inhibit the release of GABA (North et al.,
1987).
How could these disinhibitory effects be impor-
tant in LTP induction? Agents that attenuate
GABAergic inhibition generally facilitate LTP in-
duction (Wigstrom and Gustaffson, 1985). This
effect is thought to arise from the enhanced depo-
larization, and presumably, enhanced NMDA re-
ceptor activation that follows a reduction in
GABAergic inhibition. In fact, the disinhibitory
effect of opioid peptides likely underlies their re-
quirement for the induction of lateral perforant
path–dentate LTP, as the block of lateral perfo-
rant path LTP following naloxone can be pre-
vented by co-application of a GABAA receptor
antagonist (Bramham and Sarvey, 1996). This the-
ory of the contribution of opioid peptides to LTP
induction has been longstanding (Swearengen and
Chavkin, 1987).
Although its modulation by opioids has been
clarified, the mechanisms underlying lateral perfo-
rant path LTP induction are not as simple. Studies
by Bramham and colleagues have shown that,
while the activation of both m and d-type opioid
receptors is crucial for LTP induction (Bramham
et al., 1991a; Bramham and Sarvey, 1996), the fa-
cilitation of LTP by opioid peptides is unlikely to
arise via a facilitation of NMDA receptor activa-
tion. This is because LTP at the lateral perforant
path–dentate synapse was found to be independent
from NMDA receptors in vivo, as competitive
NMDA receptor antagonists were ineffective in
blocking its induction (Bramham et al., 1991b).
Thus, the way that opioid receptor activation fa-
cilitates LTP induction at lateral perforant
429
path–dentate synapses, is not likely to be related
to NMDA receptors. Perhaps the enhanced
postsynaptic depolarization afforded by opioid
peptides permits calcium influx, which is necessary
for the induction of virtually all forms of LTP,
from other sources. For example, voltage-depend-
ent calcium channels can respond to the depolari-
zation afforded by the disinhibitory effects of
opioid peptides. In addition, d opioid receptors are
suggested to activate phospholipase C and the
formation of IP3, a second messenger that often
regulates intracellular calcium release (Bramham,
1992). Thus direct actions of opioid peptides on
principal cells may be necessary for LTP induction
by enhancing intracellular Ca++ levels by en-
hancing release of Ca++ from intracellular stores.
The suggestion of a distinct, NMDA receptor-
independent, opioid receptor-dependent form of
LTP in lateral perforant path afferents is also sup-
ported by our own studies in area CA3, where
projections from the lateral perforant path termi-
nates in the stratum lacunosum-moleculare. This
direct perforant path projection to CA3 pyramidal
cells is appreciable: the average CA3 pyramidal
cell receives the same number of perforant path
synapses as does the average granule cell (Amaral
et al., 1990). Here, the story is quite similar to the
dentate gyrus, in that while m, but not d, opioid
receptor antagonists are effective in blocking LTP
at lateral perforant path–CA3 synapses, NMDA
receptor antagonists have no effect on LTP induc-
tion (Do et al., 2002; Kosub et al., 2005). This is in
line with studies of opioid peptide receptor occu-
pation, which show a high density of m opioid re-
ceptors in the stratum lacunosum-moleculare of
the rat CA3 region (Neumaier and Chavkin, 1989).
Conversely, NMDA receptor antagonists are
effective in blocking LTP at medial perforant
path–CA3 synapses (Do et al., 2002), whereas op-
ioid receptor antagonists have little or no effect
(Breindl et al., 1994; Do et al., 2002). In addition,
stimulation of perforant path fibers often results in
a depression of responses at neighboring, inactive
synapses, a phenomenon termed ‘‘heterosynaptic
LTD.’’ In our studies, we found that when LTP is
induced in a lateral perforant pathway, the
NMDA receptor antagonist CPP was ineffective
in blocking lateral perforant path LTP induction,
430
CPP was effective in blocking the heterosynaptic
LTD of medial perforant path–CA3 responses
normally observed following lateral perforant path
stimulation (Kosub et al., 2005). Thus, like their
synapses in the dentate gyrus, synapses from the
lateral perforant pathway to the CA3 region dis-
play LTP induction that is insensitive to NMDA
receptor antagonists, but is sensitive to antagonists
of the m opioid receptor.
Although these results are consistent with the
high density of m opioid receptors in the termina-
tion site of the lateral perforant path in area CA3,
physiological data indicate that NMDA receptors
are expressed in both stratum lacunosum-molecul-
are of area CA3 and the outer molecular layer of
the dentate, although to a lesser degree than the
regions where medial perforant path afferents ter-
minate (see Monaghan and Cotman, 1985; Bram-
ham, 1992). If LTP induction in both the dentate
and the CA3 region is independent from NMDA
receptors, what is the function of NMDA recep-
tors located in their synaptic regions (Milner and
Drake, 2001)? One possibility may be the induc-
tion of heterosynaptic LTD. Supporting this view,
the NMDA receptors found in the molecular layer
of the dentate may be extrasynaptic and found on
dendrites and dendritic shafts (Monaghan and
Cotman, 1985; Bramham, 1992). Because recent
evidence implicates extrasynaptic NMDA recep-
tors, possibly NMDA receptors containing the
NR2B subunit, in LTD induction (Massey et al.,
2004), it is possible that the principal process in-
volving NMDA receptors at lateral perforant path
targets (and likely localized to extrasynaptic sites)
is the induction of various forms of perforant path
LTD (see below).
It is also important to note the independence of
LTP at the lateral perforant path synapse from
NMDA receptors is not unchallenged. Zhang and
Levy (1992) showed that lateral perforant
path–DG LTP could be blocked by systemic ad-
ministration of ketamine, an uncompetitive
NMDA receptor antagonist. A possible explana-
tion for this effect may be the type of NMDA
receptor involved in LTP induction. It is known
that not all NMDA receptor heteromers (such as
NMDA receptors that contain NR2B subunit) are
sensitive to competitive NMDA antagonists like
CPP or AP-5 (Snyder et al., 2001). As uncompeti-
tive antagonists generally are effective at most
NMDA receptors, it remains possible that the
NMDA receptors in the terminal fields of lateral
perforant path projections may be a heteromeric
variety (such as NR2B multimers, Massey et al.,
2004) that are unaffected by typical NMDA re-
ceptor antagonists. In this case, the disinhibitory
effects of opioid peptides might contribute to LTP
induction via enhanced postsynaptic depolariza-
tion and a facilitation of the activation of NMDA
receptor heteromers that may be insensitive to
traditional competitive NMDA receptor antago-
nists.
It remains unknown whether the LTP at medial
and lateral perforant path terminals reflect distinct
‘‘forms’’ of LTP that display distinct mechanisms
of both induction and maintenance, or if the LTP
in the lateral and medial perforant pathways sim-
ply reflect differences in the mechanisms of induc-
ing a common form of LTP. The issue is of
interest, as there may be as many forms of LTP as
there are forms of LTP induction. However, it is
also quite possible that perforant path–dentate
LTP may differ simply in its induction require-
ments, yet share common, downstream molecular
processes of LTP maintenance. Nonetheless, if
medial and lateral perforant path synapses display
a similar form LTP that differs only in its induc-
tion mechanism, this also is of importance, as it
suggests that distinct rules, or ‘‘constraints,’’ gov-
ern LTP induction among different synaptic pop-
ulations of the hippocampal formation. Such
difference may serve to restrict LTP in specific
synaptic population to specific types of synaptic
activity. For example, because opioid receptor ac-
tivation is crucial for the induction of lateral per-
forant path LTP, and because opioid peptides
display frequency-dependent release (Neumaier
and Chavkin, 1989; Wagner et al., 1990, 1991;
Derrick et al., 1994a, b), it would be expected that
LTP at this synapse would only occur with high
frequency presynaptic activity. Thus repetitive pre-
synaptic activity, such as is observed with bursting
(in this case, bursting in cells of the lateral ent-
orhinal cortex giving rise to the lateral perforant
pathway) may be absolutely required for LTP in-
duction in these projections. Thus, the additional

requirement for opioid peptides, and their require-
ment for high frequency presynaptic activity or
bursting for their release, may be viewed as a fea-
ture that imposes tighter ‘‘constraints’’ on the
types of synaptic activity that are effective in in-
ducing lateral perforant path LTP. Satisfaction of
this additional constraint might only be provided
by certain types of presynaptic activity (such as
bursting), thereby restricting LTP induction in
these synaptic populations to states, such as theta
states, that are associated with bursting of granule
cells (Munoz et al., 1990) or cells of the entorhinal
cortex (Jeffery et al., 1995).
While no data are available on whether or not
medial and lateral perforant path synapses display
distinct ‘‘forms’’ of LTP, it should be noted that
different ‘‘forms’’ of LTP may occur even within a
single synaptic population. Recent evidence sug-
gests that different forms of LTP, reflected in
differences in the longevity of LTP, and therefore
LTP maintenance, can be regulated by a number
of variables, such as the pattern of afferent activity
used to induce LTP, or by the behavioral state of
the animal during LTP induction. In the case of
LTP induced with different types of stimulation,
high-frequency stimulation that mimics naturally-
occurring theta rhythm is suggested to induce a
‘‘non-decremental’’ form of LTP at Schaffer–CA1
synapses (Staubli and Lynch, 1987), possibly by
inducing a number of forms of LTP (Morgan and
Teyler, 2001). In the case of different behavioral
states, the longevity of LTP also can depend on
behavioral state of the animal when LTP is in-
duced. At perforant path–dentate synapses, LTP
induced following a single stimulation session typ-
ically persists for 5–7 days (Barnes, 1979; Davis
et al., 2004). However, the duration of LTP is ex-
tended to more than 2 weeks if perforant path LTP
is induced while the animal is actively engaged in
learning (such as during the initial exploration of a
novel environment; Davis et al., 2004). Thus, the
production of LTP with a sustained time course is
highly suggestive of distinct mechanisms of LTP
maintenance, and thus distinct ‘‘forms’’ of LTP.
The mechanisms underlying these effects remain to
be elucidated, although the latter effect of novelty
on enhancing the longevity of perforant
path–dentate LTP appears to require the presence
431
of specific monoamines often associated with
arousal, such as dopamine, norepinephrine, and
serotonin (Li et al., 2003; Straube et al., 2003;
Sanberg et al., 2006).
When LTP is induced either with theta bursts or
by novel environments, LTP displays a distinct,
extended time course. The implication of these
findings is that distinct molecular mechanisms, or
the modulation or modification of a common mo-
lecular mechanism, may underlie LTP mainte-
nance when induced under different conditions.
Thus, not only can different synapses display
different ‘‘forms’’ of LTP, but a given synapse may
also display distinct ‘‘forms’’ of LTP depending on
the animal’s behavioral state during its induction.
Thus, the present challenge is two-fold: to deter-
mine the necessary and sufficient mechanisms that
may be shared by different ‘‘forms’’ of LTP, if they
exist, and to determine the functional significance
of the variations of LTP induction and mainte-
nance. This approach has now become crucial, as
differences in LTP forms or LTP induction mech-
anisms among the afferent populations and sub-
regions in the hippocampal formation are likely to
reveal important clues regarding the role of these
distinct hippocampal afferents and hippocampal
subregions in the processing and storage of infor-
mation in the hippocampal formation.
Plasticity of mossy fibers
It should also be mentioned that opioid peptides are
also thought to play a role in LTP in efferents of the
granule cells, the mossy fibers. In the first report by
Harris and Cotman (1986) that LTP at the mossy
fiber–CA3 synapse is insensitive to NMDA recep-
tor antagonists, these investigators also suggested
that the insensitivity of mossy fiber LTP to NMDA
receptor antagonists may reflect a distinct ‘‘form’’
of LTP, although these investigators suggested that
closer scrutiny of mossy fiber LTP maintenance
might be important in order to verify this possibil-
ity. In addition, mossy fiber LTP induced in vivo
displays an unusual, incremental, or slowly devel-
oping LTP that decay little over hours following
induction, a feature not shared by commissural/
associational CA3 inputs that display NMDA
432
receptor-dependent LTP. These differences in both
the mechanism of LTP induction and the incre-
mental time course of LTP expression led us to
suggest mossy fiber–CA3 LTP reflects a distinct
form of LTP (Derrick and Martinez, 1989). Our
subsequent studies (Derrick et al., 1991, 1992)
showed that mossy fiber LTP induced in vivo re-
quires the activation of opioid receptors, replicating
an original report by Martin (1983). Subsequent
studies revealed that the activation of the m opioid
receptor is crucial for the induction of mossy fiber
LTP (Derrick et al., 1992). Interestingly, subsequent
studies revealed that induction and development of
mossy fiber LTP is blocked by protein synthesis
inhibitors (Barea-Rodriguez et al., 2000). As such
inhibitors usually only alter the late phase (>3 h) of
NMDA receptor-dependent LTP, this result offers
further proof of distinct mechanisms of mainte-
nance, and thus a distinct form of LTP. A 1991
report by Nicoll and colleagues came to a similar
conclusion, but for different reasons. In this study,
the induction of mossy fiber LTP was not only
NMDA receptor-independent, but displayed induc-
tion mechanisms that were exclusively presynaptic
(Zalutsky and Nicoll, 1990).
However, the report by Zalutsky and Nicoll
(1990) was in conflict with earlier reports indicat-
ing a dependence of mossy fiber LTP on both
postsynaptic calcium (Williams and Johnston,
1989) and depolarization (Jaffe and Johnston,
1990). While at that time, our data favored a pre-
synaptic locus for mossy fiber LTP induction
(Derrick et al., 1992), our subsequent studies re-
vealed an intensity-dependent threshold for mossy
fiber LTP induction (Derrick and Martinez,
1994b). Moreover, associative mossy fiber LTP
could be observed when weak mossy fiber synapses
were co-active with intense activation of CA3
associational/commissural inputs. Importantly,
this effect could be blocked by an NMDA recep-
tor antagonist (Derrick et al., 1994b). This was a
crucial finding, because NMDA antagonists, while
effective in blocking LTP among commissural/
associational inputs, normally do not block mossy
fiber LTP. That NMDA receptor antagonists
blocked mossy fiber LTP induced associatively
by coactivation of an NMDA receptor dependent
pathway offers evidence that a common
postsynaptic factor that apparently is provided
by NMDA receptor activation is crucial for the
induction of mossy fiber LTP.
What role might opioid peptides play in the in-
duction of mossy fiber LTP? It seems likely that, as
with lateral perforant path LTP, opioid receptor
activation may facilitate LTP induction as a result
of their disinhibitory effects. However, tacit in this
view is that mossy fiber LTP induction involves
postsynaptic depolarization. Although there is am-
ple evidence that mossy fiber LTP depends on
postsynaptic factors, including postsynaptic depo-
larization (Jaffe and Johnston, 1990; Derrick et al.,
1994b), m opioid receptors may have entirely
different actions at this synapse. Johnston and
colleagues (Williams and Johnston, 1996) demon-
strate that while naloxone blocks mossy fiber LTP
in vitro, the addition of GABAA receptor antag-
onists does not obviate the block of mossy fiber
LTP induction by naloxone. Thus, unlike the lat-
eral perforant pathway, reducing GABAergic in-
hibition does not obviate the dependence of mossy
fiber LTP on m opioid receptor activation. This
suggests another role for m opioid receptors in
mossy fiber LTP induction that is distinct from
their actions on GABAergic inhibition.
This same conclusion can be reached a priori
from previous studies in vitro; while the mossy
fiber–CA3 synapse normally does not display sin-
gle-pulse associativity, single-pulse associativity
can be observed at mossy fiber synapses when m
opioid receptor agonists (which normally would be
released with repetitive activity) are provided (Der-
rick et al., 1994a). However, single mossy fiber
pulses delivered in conjunction with postsynaptic
depolarization of CA3 pyramidal cells still fails to
induce single-pulse associative mossy fiber LTP.
This effect is difficult to reconcile from the per-
spective of opioid peptide-mediated disinhibition;
if opioid receptor activation were necessary solely
for disinhibitory effects (and enhanced postsynap-
tic depolarization), it would be expected that in-
duced postsynaptic depolarization of CA3
pyramidal cells would obviate the need for any
further depolarization afforded by the disinhibito-
ry effects of opioid peptides. However, this ap-
pears not to be the case, leaving open the
possibility that m opioid receptors may contribute

to mossy fiber LTP via other mechanisms. Among
the possible effects of mu receptors are an en-
hancement of presynaptic activity mediated by re-
duced GABAB receptor activation (Mott and
Lewis, 1992; Jin and Chavkin, 1999), or a m op-
ioid receptor-mediated activation of second mes-
senger cascades. For instance, while m opioid
receptors typically decrease calcium influx and en-
hance potassium efflux in most neuronal popula-
tions, m opioid receptors can activate the MAP and
ERK cascades in some preparations (Zimprich et
al., 1995), and are even reported to facilitate tran-
scription via the translocation of calmodulin to
nuclear sites (Wang et al., 2000). The latter effect is
of particular interest, as it appears that the slowly
developing, ‘‘incremental’’ potentiation seen with
mossy fiber LTP in vivo requires protein synthesis
(Barea-Rodriguez et al., 2000).
Nonetheless, the data suggesting exclusively pre-
synaptic mossy fiber LTP is nonetheless compel-
ling (Zalutsky and Nicoll, 1990; Weisskopf et al.,
1993; Mellor and Nicoll, 2001; Castillo et al., 2002;
Calixto et al., 2003). Computationally, if any
synapse could operate effectively with a purely
presynaptic form of LTP induction, particularly if
its role were simply a ‘‘detonator’’ synapse, then
the mossy fiber synapse would be it. It is already
sparse anatomically, and the various features of
dentate granule cells, in combination with com-
petitive learning exemplified by LTP and LTD at
perforant path–dentate synapses, together may
provide sufficient constraints for the formation of
sparse mossy fiber outputs to the CA3 region.
These effects may obviate any need for coincident
Hebbian mechanisms at mossy fiber synapses.
Moreover, given the presumed ‘‘detonator’’ role
of this synapse, mossy fiber synapses need not dis-
play ‘‘associativity;’’ rather, mossy fiber synaptic
activity and its ‘‘detonator’’ properties are thought
crucial for inducing associative LTP at other, co-
active CA3 synapses.
The view that mossy fiber activity plays a crucial
role by virtue of its ‘‘detonator’’ properties and the
associative induction of LTP at other, coactive CA3
synapses is supported by the temporal sequence of
afferent activation of CA3 pyramids following per-
forant path activity. As noted above, the perforant
path projections to the dentate gyrus and the CA3
433
region activate CA3 pyramidal cells 2–5 ms earlier
than granule cells of the dentate (Fig. 2). Direct
perforant path afferents are capable of discharging
CA3 pyramids (which occurs in vivo with the same
perforant path stimulation intensities that activate
granule cells; see Do et al., 2002; Fig. 2). Thus the
activation of CA3 region by perforant path would
initiate recurrent CA3–CA3 activity rapidly, as the
commissural/associational CA3–CA3 fibers have
conduction velocities of 1.5 m/s in vivo (Do et al.,
2002). By contrast, the parallel and simultaneous
perforant path input to the granule cells elicits
granule cell firing 1–3 ms later than the CA3 py-
ramidal cells. In addition, because the conduction
velocity of the mossy fibers is quite slow,
0.4–0.8 m/s (Derrick et al., 1994a), the disynaptic
activation of CA3 pyramidal cells by the mossy
fibers would occur after both PP–CA3 and
CA3–CA3 recurrent activity. This sequence of
CA3 afferent activation (PP–CA3, then CA3–CA3,
followed by mossy fiber–CA3 synapses) places
mossy fiber–CA3 synaptic activity in the position
of being the last CA3 synapse to be activated di-
synaptically by perforant path input. This sequence
is crucial, because the induction of associative LTP
at recently active synapses is asymmetric, and is
observed only when presynaptic activity precedes
postsynaptic depolarization (Levy and Steward,
1983; Gustafsson and Wigstrom, 1986; Gustafsson
et al., 1987). This asymmetric feature, characterized
more recently as asymmetric LTP mediated by
spike timing-dependent plasticity (STDP), suggests
that perforant path–CA3 and CA3–CA3 synapses
activated prior to, or simultaneously with postsy-
naptic depolarization resulting from mossy fib-
er–CA3 ‘‘detonators’’ would induce associative
LTP at the recently activated CA3 synapses. As
mossy fiber ‘‘detonators’’ fire after PP–CA3 and
CA3–CA3 activity, this delay in detonator activity
would be optimal for the induction of associative
LTP at recently activated PP–CA3 and CA3–CA3
synapses. A recent study suggests such an effect or
mossy fiber activity of commissural/associational
CA3 fibers, although it appears to follow rules
atypical of those usually associated with STDP
(Kobayashi and Poo, 2004).
A primary assumption in this view of encoding in
the CA3 region is that mossy fiber synaptic activity
434
can induce associative LTP at perforant path and
CA3–CA3 synapses that are located in more distal
dendritic regions of the CA3 pyramids. While this
effect is plausible in the CA1 region, given the abil-
ity of action potentials to ‘‘backpropagate’’ to den-
dritic regions following CA1 pyramidal cell
discharge (Magee and Johnston, 1997), the CA3
pyramids do not show a large apical dendrite that
extends into the stratum radiatum, as do the CA1
pyramids. However, data obtained in vitro suggest
that the mossy fibers are indeed able to induce LTP
at more distal CA3–CA3 and PP–CA3 synapses
(Chatterji et al., 1989; McMahon and Barrionuevo,
2002; Kobayashi and Poo, 2004). In addition, while
there is evidence for the mossy fibers having ‘‘det-
onator’’ aspects and being able to initiate firing of
CA3 pyramidal cell (Henze et al., 2002), this effect
is frequency dependent. Thus repetitive activation
of mossy fiber–CA3 synapses is necessary for mossy
fiber synapses to exhibit their ‘‘detonator’’ effects.
This requirement presents an additional problem
because, as noted above, repetitive activation of the
mossy fiber synapse would necessitate bursting in
granule cells, and such bursting appears to be a rare
event (Jung and McNaughton, 1993). However,
granule cell bursting is observed in the intact animal
during theta rhythm (Munoz et al., 1990). This ob-
servation suggests that granule cell bursting, essen-
tial for initiating the ‘‘detonator’’ features of the
mossy fiber synapse, may be limited to periods
during periods of theta rhythm. Thus while the
mossy fiber synapse may display detonator proper-
ties, these properties may be limited to behavioral
states that generate granule cell bursting, such as
that observed during theta rhythm which occurs
during exploratory behaviors.
What might explain the differences in results
from different laboratories regarding pre- and
postsynaptic mechanisms of mossy fiber LTP?
Barrionuevo and colleagues (Urban and Bar-
rionuevo, 1996) suggested that mossy fiber LTP
itself may express different forms of LTP. Perhaps,
as observed with novelty-induced facilitation of
LTP that occurs at medial perforant path–dentate
synapses (Davis et al., 2004), this may indicate the
induction of a different form of mossy fiber LTP
that depends on conditions that co-occur with
LTP induction. Thus, pre- or postsynaptic forms
of mossy fiber LTP may occur in distinct behavi-
oral states, during particular patterns of afferent
activity, which then may determine if pre- or
postsynaptic forms of mossy fiber LTP are in-
duced. For instance, it is possible that the postsy-
naptic, Hebbian form of mossy fiber LTP may be
essential during encoding, whereas a strictly pre-
synaptic, non-associative, and short-lived form of
mossy fiber LTP may operate during recall. This
occurrence of distinct forms of mossy fiber LTP
may serve to prevent Hebbian associativity during
recall states, an event that can have catastrophic
effects on previously encoded patterns. This prob-
lem, often referred to as the ‘‘stability plasticity
dilemma’’ (Grossberg, 1987) is a critical, but often
ignored problem inherent in most neural network
learning schemes (Hertz et al., 1991). Competitive
learning and self-organization, as is suggested to
occur in both perforant path–dentate (Hasselmo
et al., 1995; Rolls and Treves, 1998) and mossy
fiber–CA3 synapses (Derrick and Martinez, 1996),
would, over time, result in continual changing of
synaptic strength with each input, making any
stable learning or categorization impossible (Hertz
et al., 1991). Thus the expression of a distinct,
presynaptic and non-associative form of mossy
fiber LTP limited to periods of recall would limit
the associative influence of mossy fiber synapses
during recall. Thus a presynaptic, non-associative
form of mossy fiber LTP could serve as a novel
solution to the ‘‘stability–plasticity dilemma,’’
which is a problem inherent in all devices that re-
quire both encoding and retrieval (Hertz et al.,
1991).
Subsequent studies of the different types of
mossy fiber LTP nonetheless suggest that postsy-
naptic factors are operative even with these ap-
parently distinct forms of LTP induction (Yeckel
et al., 1999). Furthermore, the participation of a
number of factors thought to act transsynaptically
at mossy fiber synapses (such as the transsynaptic
activation of Eph receptors (Contractor et al.,
2002; Armstrong et al., 2006) and the possible re-
cruitment of extant, but ‘‘silent’’ mossy fiber
synapses with LTP induction (Reid et al., 2004)
also suggest a requirement for postsynaptic proc-
esses in many forms of mossy fiber plasticity, and
with other studies, together support the view that

postsynaptic factors are involved in the induction
of both mossy fiber LTP and mossy fiber LTD
(Derrick et al., 1994b; Derrick and Martinez, 1996;
Yeckel et al., 1999; Kapur et al., 2001; Contractor
et al., 2002; Lei et al., 2003; Wang et al., 2004;
Armstrong et al., 2006).
Although the above characterization of LTP in-
duction at the mossy fiber–CA3 synapse, as well as
lateral perforant path inputs to both the dentate
and CA3 regions, together suggest a mechanism of
LTP induction involving opioid peptides, there is
good evidence that a multiplicity of LTP and LTD
forms may exist at the mossy fiber synapse (Lei et
al., 2003). Perhaps distinct forms of LTP can be
induced at mossy fiber synapses, and which form is
induced may depend critically on the initial condi-
tions during its induction. As we showed earlier in
the dentate gyrus, LTP with an enhanced time
course can be observed in vivo depending on the
conditions during LTP induction (Davis et al.,
2004; Sanberg et al., 2006). In this view, multiple
forms of LTP are possible at a synapse, with the
form of LTP that is expressed depending critically
on the conditions during (and shortly after) LTP
induction. Given that the mossy fiber–CA3 synapse
displays the three known forms of LTP (homosy-
naptic, heterosynaptic and associative LTD;
Derrick and Martinez, 1996), and given that a va-
riety of cellular mechanisms (both pre- and postsy-
naptic) underlie what appears to be distinct forms
of homosynaptic mossy fiber LTD (Manabe, 1997;
Lei et al., 2003; Wang et al., 2004), the possibility
that there are different forms of mossy fiber LTP
remains the most parsimonious and compelling ex-
planation for the disparate results seen with mossy
fiber LTP. Further work hopefully will elaborate on
the differences in the molecular machinery under-
lying and the maintenance and expression that are
associated with the distinct induction mechanisms
of mossy fiber LTP and LTD. The findings from
such studies may reveal a common cascade among
the different induction mechanisms, or reveal func-
tionally distinct forms of plasticity at this synapse.
The former result would make us reconsider the
classification of LTP and its ‘‘forms’’ based simply
upon differences in their induction mechanisms,
and may also suggest that these distinct induction
mechanisms may have functional significance, being
435
regulators of the kind of mossy fiber LTP that will
be induced. Alternatively, the latter result would
suggest distinct forms of LTD that also may have
specialized functional roles in scaling, sparsificati-
on, depotentiation, or information storage.
The evidence for both Hebbian and non-
Hebbian processes at the mossy fiber synapse
gives rise to an important question: what would be
the utility of having ‘‘detonator’’ synapses that are
plastic? What function could mossy fibers have
that would make Hebbian plasticity at a detonator
synapse important? The dentate gyrus is thought
crucial for the encoding of new information
(McNaughton et al., 1989), and shows the great-
est activity in humans with novelty (Zeineh et al.,
2003). Yet most models of autoassociative recall
do not indicate that plasticity among detonator
synapses is essential (Rolls and Treves, 1998; Kali
and Dayan, 2000). Moreover, behavioral studies
indicate that the dentate gyrus is not essential for
recollection of spatial memory or the activity of
place fields (McNaughton et al., 1989; Mizumori et
al., 1989), whereas others indicate the absence of
mossy fiber LTP has little effect on spatial memory
(Huang et al., 1995). Thus current evidence sug-
gests that the dentate gyrus is crucial for the en-
coding, but not for the recall, of information.
However, these behavioral finding must be inter-
preted carefully, as plasticity in dentate–CA3 in-
puts may contribute to aspects of hippocampal
memory that may not be readily apparent when
using simple behavioral paradigms that are used to
assess only spatial memory. Given that the hippo-
campus is found to participate in more elaborate
tasks (Agster et al., 2002;), including those that
assess directly episodic-like memory (Fortin et al.,
2002; Kart-Teke et al., 2006), it remains possible
that the contribution of the mossy fibers to mem-
ory may become evident in more elaborate and
demanding tasks.
Given that the dentate gyrus itself displays place
fields (Jung and McNaughton, 1993) and elaborate
mechanisms of synaptic plasticity, as do the mossy
fiber–CA3 outputs (Derrick and Martinez, 1996), it
seems curious that elaborate plastic processes
would even exist in the dentate and their mossy
fiber efferents unless recall was an important aspect
of mossy fiber function. The possible contribution
436
of the dentate gyrus to information storage may
best be understood from a theoretical view and the
presumed role of the direct perforant path inputs to
CA3 to initiate recall (Treves and Rolls, 1992,
1994). However, such a view may be problematic,
because activation of a subset of the direct perfo-
rant path–CA3 lines, which are thought to provide
the partial input that initiates pattern completion in
the CA3 region (Treves and Rolls, 1994), might fail
to select accurately among the many potential CA3
attractors if they happen to share a critical subset of
direct perforant path–CA3 inputs. Perhaps plastic-
ity at perforant path–CA3, together with mossy
fiber LTP may cooperate during recall, particularly
recollection of more elaborate aspects of hippo-
campal memory not readily detectable by current
behavioral methods that only assess spatial learn-
ing. In this view, perforant path–CA3 and dent-
ate–CA3 inputs operate together during recall to
select among specific attractors within the CA3 au-
toassociative system, with plasticity in both the di-
rect perforant path–CA3 and mossy fiber–CA3
inputs acting as a simple pattern associator that is
embedded within the CA3 autoassociative system.
In this way, both mossy fiber and perforant path
inputs could together select a smaller subset of CA3
pyramidal cells. Thus the additional input from
modified mossy fiber–CA3 synapses thus might al-
low for a partial input arising from direct perforant
path–CA3 synapses to select among potential CA3
attractors. Such a feature would allow for the se-
lection of CA3 attractors that may share high de-
grees of similarity among their perforant path–CA3
inputs. In this view, plasticity at mossy fiber ‘‘det-
onators,’’ and subsequent mossy fiber activity dur-
ing recall, may be crucial for the selective recall of
specific patterns or sequences of events that may
share similar features. Implicit in this view is that
mossy fiber plasticity, and the dentate gyrus in
general, may be crucial for both the formation and
recall of discrete episodic-like memories.
LTD
LTP is noteworthy in that its induction follows the
rule of pre- and postsynaptic associativity as for-
malized by Hebb (1949). However, a mechanism
serving to increase synaptic strength cannot oper-
ate alone; otherwise, the strength of synapses
could only increase, eventually reaching a point of
saturation. Other mechanisms that permit either
the reversal or the inverse of LTP are likely to be
necessary. Such a phenomenon, termed LTD, is
observed at the same synapses that display LTP.
LTD was noted in early studies, although its pos-
sible role in information storage was only sug-
gested (Barrionuevo et al., 1980). As it became
apparent that any memory device that serves as a
temporary repository for information must have
some means to both increase and decrease synaptic
strength, the mechanisms underlying LTD became
a primary focus for many studies in the 1990s
(Bear and Abraham, 1996).
In contrast to LTP, individual ‘‘forms’’ of LTD
in the dentate, as evidenced by their distinct mech-
anisms of induction, were noted early. Homosy-
naptic LTD, the term used to describe LTD that
follows synaptic activity, is typically input specific
and induced experimentally in many hippocampal
synapses by repetitive low frequency (0.5–5 Hz)
stimulation (Huang et al., 1992). At most dentate
synapses, homosynaptic LTD, like LTP, is cooper-
ative (Kerr and Abraham, 1995), associative
(Christie and Abraham, 1992a), and also requires
postsynaptic calcium, although the role of NMDA
receptors is controversial; although some studies
indicate that homosynaptic LTD is NMDA recep-
tor-dependent (Thiels et al., 1996), other studies
indicate that both homosynaptic LTD and depot-
entiation appear to involve metabotropic glutamate
receptors (O’Mara et al., 1995a; Manahan-
Vaughan, 1998; Kulla et al., 1999; Wu et al.,
2004) and calcium influx from intracellular stores,
rather than NMDA receptors (O’Mara et al.,
1995b).
For afferents to the dentate gyrus, studying ho-
mosynaptic LTD is problematic, as the induction
requirements vary among laboratories, and it is
notoriously difficult to elicit, both in vivo and in
vitro in the slice preparation (Errington et al.,
1995; Abraham, 1996; Abraham et al., 1996), al-
though some laboratories have found that partic-
ular paradigms aid in its induction in vitro
(Abraham, 1996) and in vivo (Manahan-Vaughan,
1998). Not only does this characteristic make LTD

difficult to study, but it also illustrates an impor-
tant limitation of the in vitro preparation, as al-
terations in dentate function may be an inevitable
consequence of slice preparation. The viability of
studying plastic processes in dentate slices is not
too surprising when one considers that prepara-
tion of hippocampal slices involves widespread
deafferentation, activity, and such trauma can
cause sustained plastic changes in dentate func-
tion. For example, preparation of hippocampal
slices induce a number of IEGs such as c-fos
(Dragunow et al., 1989). As IEGs often show a
refractory period persisting for hours after their
induction (Sheng and Greenberg, 1990), the dent-
ate slice preparation may be of limited utility in
dissecting the molecular events that mediate the
variety of molecular events that underlie the var-
ious forms of plasticity in the dentate gyrus.
LTD can be elicited at synapses adjacent to
those that are active or potentiated. This form of
LTD is referred to as ‘‘heterosynaptic’’ and is ob-
served at synapses that were not activated by the
stimulus used to induce potentiation. Heterosy-
naptic LTD is robust at the perforant path–dent-
ate projections following the induction of LTP in
one set of afferents (such as the medial perforant
path), which then induces heterosynaptic LTD of
responses at inactive synapses (in this case, either
unstimulated medial perforant path synapses, or
lateral perforant path synapses), and vice versa.
Here, LTD induction appears to involve to both
NMDA receptors (Desmond et al., 1991) and the
L-type voltage-dependent calcium channels
(VDCCs; Wickens and Abraham, 1991; Christie
and Abraham, 1994; Camodeca et al., 1998), sug-
gesting that the low levels of calcium necessary for
heterosynaptic LTD may be provided by VDCCs
activated in response to NMDA receptors, per-
haps via distinct NMDA receptors at distinct, ex-
trasynaptic locations (such as the NR2B subunits
of the NMDA receptor, or similar receptors on
dendrites or the spine base and neck; Massey et al.,
2004).
The third type of LTD, associative LTD, was
once regarded as a less than replicable phenome-
non (Bear and Malenka, 1994); however, studies
showing homosynaptic LTD displays cooper-
ativity (Kerr and Abraham, 1995) suggest
437
homosynaptic LTD in the dentate can be induced
in an associative manner. However, associative
LTD appears to involve cellular mechanisms of
expression that are similar to heterosynaptic LTD,
although their mechanisms of induction differ in
that associative LTD in the dentate is reportedly
NMDA receptor-independent (Christie et al.,
1995; Abraham and Tate, 1997; Philpot et al.,
1999).
Depotentiation is related to LTD, but refers to a
reversal of LTP (Levy and Steward, 1979; Bar-
rionuevo et al., 1980; Fujii et al., 1991) and may be
thought of as homosynaptic LTD of potentiated
synapses. However, depotentiation is distinct from
homosynaptic LTP in two important ways. First,
depotentiation has a narrow window of induction;
in the dentate, stimulation must occur minutes fol-
lowing LTP induction, otherwise LTP becomes re-
sistant to depotentiation (Martin, 1998; Kulla et al.,
1999). Second, depotentiation and homosynaptic
LTD appear to involve distinct cellular mechanisms
(Lee et al., 2000; Soderling and Derkach, 2000).
The early (o3 h) (Otani et al., 1989) phase of LTP
is suggested to involve phosphorylation of the
GluR1 AMPA subunit at both CaMKII and PKA
sites on this receptor subunit (Soderling and
Derkach, 2000). Studies suggest that whereas de-
potentiation involves dephosphorylation at PKA
serine/threonine residues, homosynaptic LTD at
naive synapses may involve dephosphorylation of
CaMKII serine/threonine residues (Lee et al.,
2000).
Just as the mechanisms underlying the induction
of heterosynaptic LTD, homosynaptic LTD, and
depotentiation appear distinct, so are the treat-
ments that can affect their induction. For example,
5-HT4 agonists (Kulla and Manahan-Vaughan,
2002) and D1/D5 dopamine antagonists block de-
potentiation, but not LTP or heterosynaptic LTD
(Otmakhova and Lisman, 1998; Kulla and Man-
ahan-Vaughan, 2000). Conversely, drugs, such as
nimodipine, an L-type calcium channel blocker,
can block heterosynaptic LTD in the dentate gyrus
(Wickens and Abraham, 1991) without altering
LTP (Christie and Abraham, 1994). The number
of studies of LTD in the dentate gyrus are limited,
however, as LTD in the dentate is not easily in-
duced in the slice preparation (Bear and Abraham,
438
1996), and may indicate the importance of extrin-
sic neuromodulators in the induction of LTD in
the dentate gyrus (Pang et al., 1993).
The diversity of the mechanisms that induce
LTD and depotentiation suggest that different
types of LTD may serve distinct functions (Kerr
and Abraham, 1996). As noted earlier, mechanisms
that serve to increase synaptic strength would be
expected to be accompanied by other mechanisms
that can reverse increases in synaptic strength; oth-
erwise, interference would follow from an eventual
saturation of synaptic plasticity. Thus LTD may
play a role in reversing LTP (‘‘depotentiation’’). In
line with this view, heterosynaptic LTD can reverse
established dentate LTP (Doyere et al., 1997) a
process likely to be essential in structures that have
a transient role in information storage, such as the
hippocampal formation (Bontempi et al., 1999).
Normalization processes are thought to be a crucial
aspect of neuronal homeostasis — maintaining
neuronal excitability within a dynamic range of
neuronal output, which is essential to maintain
maximal variance in neuronal output. Thus LTP
induced in a set of synapses on a neuron may result
in a concomitant, equivalent net decrease in the
strength of other, inactive, synapses on the same
neuron.
Other forms of LTD may play a role in sparse
coding (Skaggs and McNaughton, 1992), ensuring
that only the most active synapses increase in
strength in response to a given input, whereas
other less active synapses are depressed, preserving
the sparse encoding essential for distributed mem-
ory systems that employ the Hebb rule (Skaggs
and McNaughton, 1992). For example, both ho-
mosynaptic LTD and associative LTD may facil-
itate sparse encoding by depressing synapses that
show ill-timed or only modest activity. In this
view, these forms of LTD may filter out back-
ground synaptic ‘‘noise,’’ reducing and optimizing
the number of modifiable elements that contribute
to a given representation. Thus LTD may con-
tribute to information storage by removing redun-
dancies, and providing a sparse inverted code, or
‘‘transform,’’ of the same perforant path input that
is relayed directly to distal dendrites of CA3 neu-
rons (Morris and McNaughton, 1987). This fea-
ture may be crucial during encoding in order to
provide sparse, orthogonal inputs to the CA3 re-
gion (Rolls and Treves, 1998).
Another interpretation is that LTD, operating
in conjunction with LTP, together serves in infor-
mation storage (Bear and Abraham, 1996; Bear,
1999; Braunewell and Manahan-Vaughan, 2001).
In this way, LTP and LTD may act in concert,
enhancing the dynamic range of plasticity at a
given synapse, as bidirectional plasticity can en-
hance the capacity, fidelity, and flexibility of the
networks (Dayan and Willshaw, 1991). Bidirec-
tional plasticity observed at the perforant
path–dentate synapse also is thought to be crucial
in competitive learning schemes (Rolls and Treves,
1998). Thus, while competitive learning is often
implicated in self-organization (Hertz et al., 1991),
it also can be viewed simply as another way to
remove redundancies and to sparsify dentate out-
puts to the CA3 region (Rolls and Treves, 1998).
In this view, heterosynaptic LTD and homosy-
naptic LTD have the potential to contribute to
normalization and sparse encoding (Morris, 1989).
Metaplasticity
Metaplasticity refers to an alteration in the thresh-
old or magnitude of LTP or LTD induction in a
neuron as a result of prior neuronal or synaptic
activity (see Abraham and Tate, 1997; Philpot
et al., 1999). Metaplastic processes are thought to
regulate subsequent changes in synaptic strength
depending on the prior neuronal ‘history’ of
synaptic activity. Thus, metaplastic effects are
used to describe activity-dependent and sustained
effects that can regulate the ability of a synapse, or
synapses on a given neuron, to change synaptic
strength in response to subsequent synaptic activ-
ity. As such, metaplasticity reflects a distinct
mechanism of information storage (Philpot et al.,
1999).
Metaplastic effects follow predictions of several
theories of information processing that account for
activity-dependent self-organization during devel-
opment. In particular, the Bienenstock, Cooper and
Munroe (BCM) theory (Bienenstock et al., 1982;
Bear et al., 1987; Shouval et al., 1997), originally
used to account for visual cortex development,
 439

employs processes that regulate synaptic strength

within an entire neuron as a consequence of prior
synaptic activity, and serves as a useful framework
with which to predict metaplastic effects. Generally,
the BCM theory predicts that if postsynaptic activ-
ity is insufficient for LTP induction, LTD will oc-
cur. This transition point where postsynaptic
activity induces LTD or LTP, in the BCM algo-
rithm, is denoted by ym (Fig. 4). Importantly, ym is
not fixed; instead, it changes as a function of
postsynaptic activity. For example, activity that is
not sufficient for LTP induction not only induces
LTD, but also ‘‘shifts’’ the threshold for LTP in-
duction (indicated by ym) to the left, so that the
threshold of postsynaptic activity necessary to in-
duce LTP at synapses on that neuron is subse-
quently reduced. Conversely, if stimulation is
sufficient to induce LTP, ym shifts to the right, in-
creasing the threshold for inducing LTP, and favo-
ring the induction of LTD with subsequent synaptic
activity.
An important aspect of changes in ym is that
this process, as envisaged by Bienenstock et al.
(1982), is ‘‘cell wide,’’ and any change in the
Fig. 4. The BCM rule and sliding modifications of LTP and
threshold for synaptic plasticity that follows ac- LTP induction thresholds. (A) The induction of LTP or LTD
tivity would affect all synapses of a neuron. Such a depends on the level of postsynaptic activity. Postsynaptic de-
feature can serve a normalization, or scaling, func- polarization (PSD) at or beyond the threshold for LTP induc-
tion by maintaining a constant net excitatory input tion (ym) results in the induction of LTP. (B) Prior neuronal
to a given neuron so that output remains within activity sufficient for LTP induction shifts ym to the right,
making subsequent postsynaptic activity more likely to induce
the neuron’s maximal dynamic range (Bienenstock LTD. (C) Prior activity insufficient to induce LTP shifts ym to
et al., 1982; Barrionuevo and Brown, 1983). It is the left, lowering the threshold for LTP induction. Note that a
important to note that metaplastic effects are dis- fixed level of postsynaptic activity that is initially the threshold
tinct from normalization or synaptic scaling, as for LTP induction (——) in (A) subsequently induces LTD (B)
metaplasticity refers specifically to the modifica- or LTP (C) depending on prior activity and the resultant shift in
ym.
tion of thresholds necessary to induce LTP or
LTD, whereas scaling reflects modifications in
synaptic drive. (shifting ym to the left). However, exposure of
When the rules of ‘‘cell wide’’ metaplasticity that these rats to light for various periods of time re-
alter LTP and LTD thresholds are employed in sulted in a shift of LTD/LTP thresholds to values
models of visual cortex, these rules can allow for similar to controls, confirming activity-based al-
self-organization, categorization, and a selective terations in LTP and LTD threshold in the visual
‘tuning’ of specific cells within a network to spe- cortex. This simple but powerful local learning
cific inputs (Shouval et al., 1997). These predic- rule, which appears operative in the visual cortex,
tions have been confirmed to some degree in the can allow for emergent properties in neuronal net-
developing visual cortex, where rats reared in dark works, including self-organization and stimulus
environments showed a greater propensity to dis- selectivity (Shouval et al., 1997).
play LTP than LTD. Thus, reduced input to the Within the hippocampus, metaplastic phenom-
visual cortex favored subsequent induction of LTP ena also are observed, as previous studies indicate
440
that low-frequency stimulation that is ineffective
in inducing either LTP or LTD (‘priming’ stimu-
lation) can subsequently lead to a cell-wide de-
crease (Holland and Wagner, 1998) in LTD
threshold and enhance stimulation-induced re-
versal of LTP (or depotentiation). However, most
hippocampal studies in vitro that utilize priming
stimulation also report input-specific effects lim-
ited to stimulated synapses (Huang et al., 1992).
Thus, the validity of the BCM theory is compli-
cated by priming studies that report metaplastic
effects that are not ‘‘cell wide,’’ but rather, are
limited to activated synapses (‘‘input specific meta-
plasticity’’).
Within the dentate gyrus, input specific effects
are also observed with priming stimulation. Christie
and Abraham (1992a) report that priming the lat-
eral perforant path with low frequency (5 Hz) theta
burst stimulation resulted in a facilitation of the
induction of LTD. At first glance, these results seem
to be in opposition to the BCM rule (as stimulation
that does not evoke LTP should slide the LTD/LTP
threshold to the left, favoring LTP). However, it
should be noted that priming stimulation itself did
not induce a change in synaptic strength. Thus
these findings are consonant with the BCM model,
as it would be expected that low frequency stimu-
lation that was not effective in inducing either LTP
or LTD may still result in a leftward shift in thresh-
olds. Thus, both the threshold for LTD and LTP
would decrease. However, in both studies of CA1
and the dentate, the threshold for LTP induction
was increased (Fujii et al., 1991; Wexler and Stan-
ton, 1993), suggesting that the threshold for LTP
made a rightward, rather than a leftward shift.
Subsequent studies by the same investigators found
that modest lateral perforant pathway stimulation
resulted in a facilitation of subsequent LTP induc-
tion (Christie et al., 1995). This still is problematic,
as stronger stimulation would be expected to shift
the BCM curve to the right, rather than the left, and
increase the threshold for LTP induction, favoring
LTD induction.
While these findings suggest metaplastic effects
governed by priming do not always follow the pat-
ter predicted by the BCM algorithm, or its cell wide
functions, it seems likely that the use of priming
stimulation may underlie these discrepancies. By
contrast, when stimulation is sufficient to induce
LTP, subsequent perforant path–dentate potentiat-
ion with further perforant path stimulation is re-
duced (Frey et al., 1995). In addition, stimulation of
lateral perforant path–dentate inputs that induces
LTP in behaving animals reveals that subsequent
stimulation at the same intensity induces LTP of a
smaller magnitude for several weeks after initial
LTP induction. Conversely, when LTD follow lat-
eral perforant path stimulation, stimulation 1–3
weeks later at this same intensity results in a robust
LTP (Villarreal and Derrick, 2000). Perhaps synap-
tic activation that does not induce plasticity, as ob-
served with priming, may lead to a modification of
metaplastic rules such that patterned synaptic ac-
tivity that fails to induce LTP will always favor
LTD (Christie and Abraham, 1992a), whereas low
frequency activity that is random or not patterned
might decrease the threshold for both LTD and
LTP induction (Christie et al., 1995).
Recent studies suggest that the dentate gyrus
also can display ‘‘cell wide’’ metaplastic effects.
This is observed following non-synaptic (anti-
dromic) activation of dentate granule cells in vivo
by hilar stimulation (Abraham et al., 2001). Per-
haps the level of postsynaptic depolarization is the
crucial factor in setting LTP and LTD thresholds,
an effect that may not be effective when priming
stimulation is used.
Given that there is now evidence for both ‘‘input
specific’’ and ‘‘cell-wide’’ metaplasticity in the
dentate gyrus, how might the two forms of meta-
plasticity be important? Perhaps ‘‘cell wide’’ and
‘‘input specific’’ forms of metaplasticity serve func-
tionally distinct roles at neuronal and synaptic
levels, respectively. For example, input specific
metaplasticity may operate during the initial stages
of synaptic plasticity to ensure sparse encoding. In
this scheme, synapses that have been recently po-
tentiated may be refractory to further potentiat-
ion, while those neurons receiving postsynaptic
input insufficient to induce LTP would be partic-
ularly sensitive to LTP induction with subsequent
synaptic activity. Such an effect would ensure that
recently potentiated synapses are ‘‘opted out’’ of
potentiation by later inputs, further aiding in

dentate pattern separation. By contrast, cell-wide
metaplastic effects that follow BCM rules may re-
quire synaptic plasticity (either LTD or LTP).
Thus a BCM-like process that occurs with synaptic
changes can allow for long-term metaplastic
effects as predicted by the BCM theory, and al-
low for sustained stimulus selectivity, a feature
that would be particularly advantageous in a com-
petitive learning device that performs categoriza-
tion functions.
Interestingly, many priming effects that appear
to follow BCM rules are observed primarily in the
lateral perforant pathway and the mossy fiber
pathway, pathways that contain and release pro-
enkephalin-derived opioid peptides. However, no
studies have addressed metaplasticity in medial
perforant path afferents, or the contribution of
opioid peptides to the induction or expression of
metaplastic phenomena in opioidergic lateral per-
forant path afferents (Christie and Abraham,
1992a; Christie et al., 1995; Francesconi et al.,
1997). The current evidence suggests it is a likely
role for opioid peptides, given that the disinhibi-
tory effects of opioid peptides can greatly influence
the levels of postsynaptic activity, which, in turn,
can influence NMDA receptor activation, as well
as ym, and thus permit the differential induction of
either LTP or LTD (Francesconi et al., 1997;
Wagner et al., 2001).
The mechanisms underlying metaplastic effects
remain unknown, although alterations in glutamate
receptors (Abraham and Tate, 1997; Castellani
et al., 2001) and activation of group II (Manahan-
Vaughan, 1998; Gisabella et al., 2003) and Group I
(Wu et al., 2004) metabotropic glutamate receptors
are suggested to mediate metaplastic effects, and
the induction of LTD and depotentiation in general
(Manahan-Vaughan, 1998). Recent theories also
implicate alterations in NMDA receptors and their
subunit composition as a likely means of regulating
postsynaptic responses to synaptic activity (Caste-
llani et al., 2001). Several studies suggest common
mechanisms of metaplasticity among lateral and
medial inputs to the dentate (Abraham et al., 2001)
that may involve alterations in NMDA receptors.
As NR2B NMDA receptors show a decreased cal-
cium conductance once bound to calmodulin
441
(Rycroft and Gibb, 2002), and the duration of
metaplastic effects in vivo (Villarreal and Derrick,
2000) parallels LTP longevity (Davis et al., 2004),
alterations in the multimeric composition of
NMDA receptors that involve the NR2B subunit
of the NMDA receptor may reflect an important
mechanism underlying metaplasticity. Subsequent
down stream mechanisms are suggested to involve
CaMKII (Zhang and Levy, 1992) and protein kin-
ase C (Gisabella et al., 2003).
Given that the primary function of the dentate is
thought to be pattern separation and the genera-
tion of a sparse, inverted code that allows for or-
thogonal outputs to the CA3 region, input specific
metaplasticity may play a role in the pattern sep-
aration processes of the dentate. For example, cells
that were recently active and potentiated will dis-
play a rightward shift in ym, and thus an increase
in LTP induction threshold, making these potenti-
ated synapses refractory to any further induction
of LTP, and even favor the induction of LTD with
subsequent synaptic activity. By contrast, synapses
that are activated to a degree that is insufficient to
induce LTP would show a leftward shift in y,
greatly reducing LTP induction threshold for sub-
sequent LTP induction. Therefore, synapses that
have been recently activated or modified may thus
be ‘‘opted out’’ and resistant LTP induction or to
any further increase in potentiation. Thus, analo-
gous to competitive learning, input-specific forms
of metaplasticity observed in perforant path–dent-
ate synapses may also contribute to the ‘‘pattern
separation’’ functions of the dentate gyrus, favo-
ring LTP induction at synapses activated below
LTP threshold, allowing for formation of sparse,
non-overlapping patterns of granule cell activity, a
function that is likely crucial for encoding infor-
mation in the CA3 autoassociative system.
Conclusions
The dentate gyrus displays plastic processes that
may underlie learning and memory in the hippo-
campus. However, the same plastic processes that
are thought to contribute to information storage
also might serve to ensure the formation of sparse,
442
orthogonal transforms that are relayed to the CA3
region that can promote the encoding of discrete
CA3 attractors, because these same plastic proc-
esses also can aid in the formation of sparse, or-
thogonal outputs to the CA3 region. First, both the
anatomically sparse mossy fiber projections and the
‘‘input expansion’’ seen in the relay from the ent-
orhinal cortex to the dentate gyrus are anatomical
features that can facilitate dentate pattern separa-
tion. In addition, both competitive learning (as ex-
emplified by LTP and LTD) and metaplasticity,
often implicated in self-organization, also may con-
tribute in ensuring sparse orthogonal dentate out-
puts relayed to the CA3 region, principally by
removing redundancies from dentate outputs and
by limiting synaptic plasticity at previously potenti-
ated synapses, and favoring LTP only at synapses
that have not been potentiated previously. Moreo-
ver, the requirement of opioid peptides for plasticity
and their requirement for specific types of
presynaptic activity for their release may constrain
conditions during which plasticity can occur,
further contributing to sparse encoding. Together,
these mechanisms may insure that the dentate
gyrus relays sparse, orthogonal transforms of the
same perforant path input arriving distally on
CA3 pyramids and aiding in the encoding of
discrete attractors within the CA3 autoassociative
matrix.
Recent evidence suggests that the dentate gyrus,
while essential for encoding, is not necessary for
accurate recall (Mizumori et al., 1989; Lee and
Kesner, 2004; Jerman et al., 2006). Why would
there be synaptic plasticity in both afferents and
efferents of granule cells, and elaborate mecha-
nisms regulating it if the output is only crucial for
encoding? One interpretation is that the ‘‘plastic’’
processes of LTP, LTD, and metaplasticity may
operate primarily to enforce pattern separation by
the dentate gyrus and sparse encoding in the CA3
system. Thus, the plastic processes in the dentate,
rather than participating in memory and informa-
tion storage per se, actually may serve primarily to
preserve pattern separation functions of the dent-
ate. These plastic processes could then provide
mechanisms that allow for sustained pattern
separation of dentate input patterns relayed to,
and directing storage in, the CA3 autoassociative
system over time. In this view, the sustained plastic
processes displayed by the dentate gyrus, rather
than subserving ‘‘information storage’’ in a formal
sense, maintains a ‘‘memory’’ for past inputs in
order to maintain a continued generation of
sparse, orthogonal patterns that are relayed to
the CA3 region, where the actual information is
both stored and recalled. Alternatively, bidirec-
tional plasticity and metaplasticity in the dentate
may play a more direct role in memory storage,
rather than simply serving to maintain separation
among patterns relayed and stored in the CA3
system. The dentate gyrus itself displays place
fields (Jung and McNaughton, 1993) and elaborate
mechanisms of long-lasting synaptic potentiation
and depression, as does the mossy fiber–CA3
synapse (Derrick and Martinez, 1996). Thus it
seems odd that such elaborate plastic processes
would exist unless recall also was an important
aspect of dentate–mossy fiber function. Perhaps
the plasticity observed in both perforant path
inputs and mossy fiber outputs of the dentate are
crucial for both encoding and recall. In this case,
sustained changes at both perforant path–dentate
and dentate–CA3 synapses, in conjunction with
activity initiated by direct perforant path–CA3 in-
puts, may allow for a more ‘‘fine tuned’’ selection
of attractors in the CA3 region. This is because
partial inputs arriving distally at perforant
path–CA3 inputs and subsequent pattern comple-
tion have the possibility of converging on any
number of possible CA3 attractors that share these
same perforant path–CA3 inputs. Thus, sustained
perforant path–dentate and mossy fiber–CA3 plas-
ticity might contribute to recall, with the mossy
fiber inputs acting as an additional ‘‘cue’’ that
allows for the selective activation of CA3 attrac-
tors among many that may share common
features. In this view, plasticity in the dentate
gyrus and its CA3 outputs are critical for accurate
recall among a number of potential CA3 attractor
states. In this case, it would be predicted that
damage to the dentate gyrus, or altering mecha-
nisms of LTP at either perforant path–dentate of
mossy fiber–CA3 synapses, would display behavi-
oral deficits, but primarily in tasks that require

finer discriminations, such as disambiguation of
encoded events that have common features (such
as context). Thus the dentate gyrus, and plastic
processes observed in the dentate CA3 relay, may
be crucial in the formation of similar, but discrete
events over time, a feature that may be crucial for
accurate encoding and recollection of episodic
memory.
References
Abraham, W.C. (1996) Induction of heterosynaptic and
homosynaptic LTD in hippocampal subregions in vivo.
J. Physiol. Paris, 90: 305–306.
Abraham, W.C., Mason-Parker, S.E., Bear, M.F., Webb, S.
and Tate, W.P. (2001) Heterosynaptic metaplasticity
in the hippocampus in vivo: a BCM-like modifiable
threshold for LTP. Proc. Natl. Acad. Sci., 98(19):
10924–10929.
Abraham, W.C., Mason-Parker, S.E. and Logan, B. (1996)
Low-frequency stimulation does not readily cause long-term
depression or depotentiation in the dentate gyrus of awake
rats. Brain Res., 722(1–2): 217–221.
Abraham, W.C. and Tate, W.P. (1997) Metaplasticity: a new
vista across the field of synaptic plasticity. Prog. Brain Res.,
52(4): 303–323.
Agster, K.L., Fortin, N.J. and Eichenbaum, H. (2002) The
hippocampus and disambiguation of overlapping sequences.
J. Neurosci., 22(13): 5760–5768.
Altman, J. and Das, G.D. (1965) Autoradiological and histo-
logical evidence of postnatal hippocampal neurogenesis in
rats. J. Comp. Neurol., 124: 319–335.
Amaral, D.G. (1978) A Golgi study of cell types in the hilar
region of the hippocampus in the rat. J. Comp. Neurol.,
182(4 Pt 2): 851–914.
Amaral, D.G., Ishizuka, N. and Claiborne, B. (1990) Neurons,
numbers and the hippocampal network. Prog. Brain Res., 83:
1–11.
Amaral, D.G. and Witter, M.P. (1989) The three-di
mensional organization of the hippocampal formation:
a review of anatomical data. Neuroscience, 31(3):
571–591.
Amaral, D.G. and Witter, M.P. (1995) Hippocampal
formation. In: Paxinos G. (Ed.), The Rat Nervous
System (4th ed.). Academic Press, San Diego, CA, pp.
443–493.
Andersen, P., Holmqvist, B. and Voorhoeve, P.E. (1966)
Excitatory synapses on hippocampal apical dendrites
activated by entorhinal stimulation. Acta Physiol. Scand.,
66: 461–472.
Armstrong, J.N., Saganich, M.J., Xu, N.J., Henkemeyer, M.,
Heinemann, S.F. and Contractor, A. (2006) B-ephrin reverse
signaling is required for NMDA-independent long-term
443
potentiation of mossy fibers in the hippocampus. J. Neuro-
sci., 26(13): 3474–3481.
Barea-Rodriguez, E.J., Rivera, D.T., Jaffe, D.B. and Martinez
Jr., J.L. (2000) Protein synthesis inhibition blocks the induc-
tion of mossy fiber long-term potentiation in vivo.
J. Neurosci., 20(22): 8528–8532.
Barnes, C.A. (1979) Memory deficits associated with senes-
cence: a neurophysiological and behavioral study in the rat.
J. Comp. Physiol. Psychol., 93(1): 74–104.
Barrionuevo, G. and Brown, T.H. (1983) Associative long-term
potentiation in hippocampal slices. Proc. Natl. Acad. Sci., 80:
7347–7351.
Barrionuevo, G., Schottler, F. and Lynch, G. (1980) The effects
of repetitive low frequency stimulation on control and
‘‘potentiated’’ synaptic responses in the hippocampus. Life
Sci., 27(24): 2385–2391.
Bayer, S.A. (1982) Changes in the total number of dentate
granule cells in juvenile and adult rats: a correlated
volumetric and thymidine study. Exp. Brain Res., 46:
315–323.
Bear, M. and Abraham, W.C. (1996) Long-term depression in
hippocampus. Annu. Rev. Neurosci., 19: 437–462.
Bear, M.F. (1999) Homosynaptic long-term depression: a
mechanism for memory? Proc. Natl. Acad. Sci. U.S.A.,
96(17): 9457–9458.
Bear, M.F., Cooper, L.N. and Ebner, F.F. (1987) A physio-
logical basis for a theory of synapse modification. Science,
237(4810): 42–48.
Bear, M.F. and Malenka, R.C. (1994) Synaptic plasticity: LTP
and LTD. Curr. Opin. Neurobiol., 4(3): 389–399.
Ben-Ari, Y. and Represa, A. (1990) Brief seizure episodes
induce long-term potentiation and mossy fibre sprou
ting in the hippocampus. Trends Neurosci., 13(8): 312–318.
Berzhanskaya, J., Urban, N. and Barrionuevo, G. (1998)
Electrophysiological and pharmacological characterization
of the direct perforant path input to hippocampal area CA3.
J. Neurophysiol., 79(4): 2111–2118.
Bienenstock, E.L., Cooper, L.N. and Munro, P.W. (1982)
Theory for the development of neuron selectivity: orientation
specificity and binocular interaction in visual cortex. J. Ne-
urosci., 2(1): 32–48.
Bliss, T. and Gardner-Medwin, A. (1973) Long lasting potent-
iation of synaptic-transmission in the dentate area of the
unanesthetized rabbit following stimulation of the perforant
path. J. Physiol., 232(2): 357–374.
Bliss, T.V.P. and Lomo, T. (1973) Long-lasting potentiation of
synaptic transmission in the dentate area of the anaesthetized
rabbit following stimulation of the perforant path.
J. Physiol., 232: 331–356.
Bontempi, B., Laurent-Demir, C., Destrade, C. and Jaffard, R.
(1999) Time-dependent reorganization of brain circuitry
underlying long-term memory storage. Nature, 400(6745):
671–675.
Boss, B.D., Peterson, G.M. and Cowan, W.M. (1985) On the
number of neurons in the dentate gyrus of the rat. Brain Res.,
338: 144–150.
444
Bramham, C.R. (1992) Opioid receptor-dependent long-
term potentiation: peptidergic regulation of synaptic
plasticity in the hippocampus. Neurochem. Int., 20(4):
441–455.
Bramham, C.R., Errington, M.L. and Bliss, T.V. (1988)
Naloxone blocks the induction of long-term potentia
tion in the lateral but not in the medial perforant
pathway in the anesthetized rat. Brain Res., 449(1–2):
352–356.
Bramham, C.R., Milgram, N.W. and Srebro, B. (1991a) Delta
opioid receptor activation is required to induce LTP of
synaptic transmission in the lateral perforant path in vivo.
Brain Res., 567(1): 42–50.
Bramham, C.R., Milgram, N.W. and Srebro, B. (1991b)
Activation of AP5-sensitive NMDA receptors is not required
to induce LTP of synaptic transmission in the lateral perfo-
rant path. Eur. J. Neurosci., 3: 1300–1308.
Bramham, C.R. and Sarvey, J.M. (1996) Endogenous activa-
tion of mu and delta-1 opioid receptors is required for long-
term potentiation induction in the lateral perforant path:
dependence on GABAergic inhibition. J. Neurosci., 16:
8123–8131.
Bramham, C.R. and Srebro, B. (1989) Synaptic plasticity in the
hippocampus is modulated by behavioral state. Brain Res.,
493(1): 74–86.
Braunewell, K.H. and Manahan-Vaughan, D. (2001) Long-
term depression: a cellular basis for learning? Rev. Neurosci.,
12(2): 121–140.
Breindl, A., Derrick, B.E., Rodriguez, S.B. and Martinez Jr.,
J.L. (1994) Opioid receptor-dependent long-term potentiat-
ion at the lateral perforant path-CA3 synapse in rat hippo-
campus. Brain Res. Bull., 33(1): 17–24.
Buzsaki, G. (1988) Polysynaptic long-term potentiation: a
physiological role of the perforant path-CA3/CA1 pyrami-
dal cell synapse. Brain Res., 455: 192–195.
Buzsaki, G. (2002) Theta oscillations in the hippocampus.
Neuron, 33: 325–340.
Cain, D.P. and Corcoran, M.E. (1985) Epileptiform effects of
met-enkephalin, beta-endorphin and morphine: kindling of
generalized seizures and potentiation of epileptiform effects
by handling. Brain Res., 338(2): 327–336.
Calixto, E., Thiels, E., Klann, E. and Barrionuevo, G. (2003)
Early maintenance of hippocampal mossy fiber — long-term
potentiation depends on protein and RNA synthesis and
presynaptic granule cell integrity. J. Neurosci., 23(12):
4842–4849.
Camodeca, N., Rowan, M.J. and Anwyl, R. (1998) Induction
of LTD by increasing extracellular Ca2+ from a low
level in the dentate gyrus in vitro. Neurosci. Lett., 255(1):
53–56.
Cartwright, R.D. (2004) The role of sleep in changing our
minds: a psychologist’s discussion of papers on memory re-
activation and consolidation in sleep. Learn. Mem., 11(6):
660–663.
Castellani, G.C., Quinlan, E.M., Cooper, L.N. and Shouval,
H.Z. (2001) A biophysical model of bidirectional synaptic
plasticity: dependence on AMPA and NMDA receptors.
Proc. Natl. Acad. Sci., 98(22): 12772–12777.
Castillo, P.E., Schoch, S., Schmitz, F., Sudhof, T.C. and
Malenka, RC. (2002) RIM1alpha is required for pre-
synaptic long-term potentiation. Nature, 415(6869): 327–330.
Cavazos, J.E., Golarai, G. and Sutula, T.P. (1991) Mossy fiber
synaptic reorganization induced by kindling: time course of
development, progression, and permanence. J. Neurosci.,
11(9): 2795–2803.
Chatterji, S., Stanton, P.K. and Sejnowski, T.J. (1989) Com-
missural synapses, but not mossy fiber synapses, in hippo-
campal field CA3 exhibit associative long-term potentiation.
Brain Res., 495: 145–150.
Chavkin, C., Shoemaker, W.J., McGinty, J.F., Bayon, A. and
Bloom, F.E. (1985) Characterization of pro-dynorphin and
proenkephalin neuropeptide systems in the rat hippocampus.
J. Neurosci., 5: 808–816.
Chawla, M.K., Guzowski, J.F., Ramirez-Amaya, V., Lipa, P.,
Hoffman, K.L., Marriot, L.K., Worley, P.F., McNaughton,
B.L. and Barnes, C.A. (2005) Sparse, environmentally-
selective expression of Arc RNA in the upper
blade of the rodent fascia dentate. Hippocampus, 15(5):
579–586.
Christie, B.R. and Abraham, W.C. (1992a) Priming of associ-
ative LTD by theta frequency synaptic activity. Neuron, 8:
79–84.
Christie, B.R. and Abraham, W.C. (1992b) NMDA-dependent
heterosynaptic long-term depression in the dentate gyrus of
anaesthetized rats. Synapse, 10(1): 1–6.
Christie, B.R. and Abraham, W.C. (1994) L-type voltage-
sensitive calcium channel antagonists block heterosynaptic
long-term depression in the dentate gyrus of anaesthetized
rats. Neurosci. Lett., 167(1-2): 41–45.
Christie, B.R., Stellwagen, D. and Abraham, W.C. (1995)
Evidence for common expression mechanisms underlying
heterosynaptic and associative long-term depression in the
dentate gyrus. J. Neurophysiol., 74(3): 1244–1247.
Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1986)
A light and electron microscopic analysis of the mossy
fibers of the rat dentate gyrus. J. Comp. Neurol., 246(4):
435–458.
Claiborne, B.J., Xiang, Z. and Brown, T.H. (1993) Hippo-
campal circuitry complicates analysis of long-term poten
tiation in mossy fiber synapses. Hippocampus, 3(2):
115–121.
Cohen, G.A., Doze, V.A. and Madison, D.V. (1992) Opioid
inhibition of GABA release from presynaptic terminals of rat
hippocampal interneurons. Neuron, 9: 325–335.
Colino, A. and Malenka, R.C. (1993) Mechanisms underlying
the induction of long-term potentiation in rat medial and
lateral perforant paths in vitro. J. Neurophysiol., 69:
1150–1159.
Colley, P.A., Sheu, F.S. and Routtenberg, A. (1990) Inhibition
of protein kinase C blocks two components of LTP persist-
ence, leaving initial potentiation intact. J. Neurosci., 10(10):
3353–3360.

Contractor, A., Rogers, C., Maron, C., Henkemeyer, M.,
Swanson, G.T. and Heinemann, S.F. (2002) Trans-synaptic
Eph receptor-ephrin signaling in hippocampal mossy fiber
LTP. Science, 296(5574): 1864–1869.
Corrigall, W.A. (1983) Opiates and the hippocampus: a review
of the functional and morphological evidence. Pharmacol.
Biochem. Behav., 18: 255–262.
Dalby, N.O. and Mody, I. (2003) Activation of NMDA recep-
tors in rat dentate gyrus granule cells by spontaneous and
evoked transmitter release. J. Neurophysiol., 90(2): 786–797.
Davis, C., Jones, F. and Derrick, B.E. (2004) Novelty enhances
the induction and maintenance of LTP in the in the dentate
gyrus. J. Neurosci., 24(29): 6497–6506.
Davis, S., Vanhoutte, P., Pages, C., Caboche, J. and Laroche, S.
(2000) The MAPK/ERK cascade targets both Elk-1 and
cAMP response element-binding protein to control long-term
potentiation-dependent gene expression in the dentate gyrus
in vivo. J. Neurosci., 20(12): 4563–4572.
Dayan, P. and Willshaw, D.J. (1991) Optimising synaptic
learning rules in linear associative memories. Biol. Cybern.,
65(4): 253–265.
Derrick, B.E. and Martinez Jr., J.L. (1989) A unique opioid
peptide-dependent form of long-term potentiation is found in
the CA3 region of the rat hippocampus. Adv. Biosci., 75:
213–216.
Derrick, B.E. and Martinez Jr., J.L. (1994a) Opioid receptor
activation is one factor underlying the frequency depen-
dence of mossy fiber LTP induction. J. Neurosci., 14(7):
4359–4367.
Derrick, B.E. and Martinez Jr., J.L. (1994b) Frequency-
dependent associative long-term potentiation at the hippo-
campal mossy fiber-CA3 synapse. Proc. Natl. Acad. Sci.
U.S.A., 91: 10290–10294.
Derrick, B.E. and Martinez Jr., J.L. (1996) Associative, bidi-
rectional modifications at the hippocampal mossy fibre-CA3
synapse. Nature, 381: 429–434.
Derrick, B.E., Rodriguez, S.B., Lieberman, D.N. and Martinez
Jr., J.L. (1992) Mu opioid receptors are associated with
the induction of LTP at hippocampal mossy fiber synapses.
J. Pharmacol. Exp. Ther., 263: 725–733.
Desmond, N.L., Colbert, C.M., Zhang, D.X. and Levy, W.B.
(1991) NMDA receptor antagonists block the induction of
long-term depression in the hippocampal dentate gyrus of the
anesthetized rat. Brain Res., 552: 93–98.
Desmond, N.L. and Levy, W.B. (1983) Synaptic correlates of
associative potentiation/depression: an ultrastructural study
in the hippocampus. Brain Res., 265(1): 21–30.
Desmond, N.L. and Levy, W.B. (1986) Changes in the postsy-
naptic density with long-term potentiation in the dentate
gyrus. J. Comp. Neurol., 253(4): 476–482.
Do, V.H., Martinez, C.O., Martinez Jr., J.L. and Derrick, B.E.
(2002) Long-term potentiation in direct perforant path pro-
jections to hippocampal area CA3 in vivo. J. Neurophysiol.,
87: 669–674.
Doyere, V., Srebro, B. and Laroche, S. (1997) Heterosynaptic
LTD and depotentiation in the medial perforant path of the
445
dentate gyrus in the freely moving rat. J. Neurophysiol.,
77(2): 571–578.
Dragunow, M., Abraham, W.C., Goulding, M., Mason, S.E.,
Robertson, H.A. and Faull, R.L. (1989) Long-term potent-
iation and the induction of c-fos mRNA and proteins in the
dentate gyrus of unanesthetized rats. Neurosci. Lett., 101(3):
274–280.
Errington, M.L., Bliss, T.V., Richter-Levin, G., Yenk, K.,
Doyere, V. and Laroche, S. (1995) Stimulation at 1–5 Hz
does not produce long-term depression or depotentiation in
the hippocampus of the adult rat in vivo. J. Neurophysiol.,
74(4): 1793–1799.
Escobar, M., Barea-Rodriguez, E.J., Derrick, B.E. and Martinez
Jr., J.L. (1997) Mossy fiber synaptogenesis is induced by
high frequency stimulation of mossy fibers. Brain Res., 751:
330–335.
Fortin, N.J., Agster, K.L. and Eichenbaum, H.B. (2002)
Critical role of the hippocampus in memory for sequences
of events. Nat. Neurosci., 5(5): 458–462.
Francesconi, W., Berton, F., Demuro, A., Madamba, S.G. and
Siggins, G.R. (1997) Naloxone blocks long-term depression
of excitatory transmission in rat CA1 hippocampus in vitro.
Arch. Ital. Biol., 135(1): 37–48.
Frey, U., Frey, S., Schollmeier, F. and Krug, M. (1996)
Influence of actinomycin D, a RNA synthesis inhibitor, on
long-term potentiation in rat hippocampal neurons in vivo
and in vitro. J. Physiol., 490(Pt 3): 703–711.
Frey, U., Schollmeier, K., Reymann, K.G. and Seidenbecher,
T. (1995) Asymptotic hippocampal long-term potentiation in
rats does not preclude additional potentiation at later phases.
Neuroscience, 67(4): 799–807.
Fry, J.P., Zieglgansberger, W. and Herz, A. (1979) Specific
versus non-specific actions of opioids on hippocampal neu-
rons in the rat brain. Brain Res., 163: 295–305.
Fujii, S., Saito, K., Miyakawa, H., Ito, K. and Kato, H. (1991)
Reversal of long-term potentiation (depotentiation)
induced by tetanus stimulation of the input to CA1 neurons
of guinea pig hippocampal slices. Brain Res., 555(1):
112–122.
Gisabella, B., Rowan, M.J. and Anwyl, R. (2003) Mechanisms
underlying the inhibition of long-term potentiation by
preconditioning stimulation in the hippocampus in vitro.
Neuroscience, 121(2): 297–305.
Gonzales, R.B., DeLeon Galvan, C.J., Rangel, Y.M. and
Claiborne, B.J. (2001) Distribution of thorny excrescences on
CA3 pyramidal neurons in the rat hippocampus. J. Comp.
Neurol., 430(3): 357–368.
Grossberg, S. (1987) Competitive learning: from interactive
activation to adaptive resonance. Cogn. Sci., 11: 23–63.
Gustafsson, B. and Wigstrom, H. (1986) Hippocampal long-
lasting potentiation produced by pairing single volleys
and brief conditioning tetani evoked in separate afferents.
J. Neurosci., 6(6): 1575–1582.
Gustafsson, B., Wigstrom, H., Abraham, W.C. and Huang,
Y.Y. (1987) Long-term potentiation in the hippocampus us-
ing depolarizing current pulses as the conditioning stimulus
446
to single volley synaptic potentials. J. Neurosci., 7(3):
774–780.
Guzowski, J.F., Lyford, G.L., Stevenson, G.D., Houston, F.P.,
McGaugh, J.L., Worley, P.F. and Barnes, C.A. (2000)
Inhibition of activity-dependent arc protein expression in
the rat hippocampus impairs the maintenance of long-term
potentiation and the consolidation of long-term memory.
J. Neurosci., 20(11): 3993–4001.
Harris, E.W. and Cotman, C.W. (1986) Long-term potentiation
of guinea pig mossy fiber responses is not blocked by
N-methyl D-aspartate antagonists. Neurosci. Lett., 70:
132–137.
Hasselmo, M.E., Schnell, E. and Barkai, E. (1995) Dynamics of
learning and recall at excitatory recurrent synapses and
cholinergic modulation in rat hippocampal region CA3.
J. Neurosci., 15(7 Pt 2): 5249–5262.
Havik, B., Rokke, H., Bardsen, K., Davanger, S. and
Bramham, C.R. (2003) Bursts of high-frequency stimulation
trigger rapid delivery of pre-existing alpha-CaMKII mRNA
to synapses: a mechanism in dendritic protein synthesis
during long-term potentiation in adult awake rats. Eur.
J. Neurosci., 17(12): 2679–2689.
Hebb, D.O. (1949) The Organization of Behavior. Wiley,
New York.
Henze, D.A., Wittner, L. and Buzsaki, G. (2002) Single granule
cells reliably discharge targets in the hippocampal CA3
network in vivo. Nat. Neurosci., 5(8): 790–795.
Hertz, J., Krogh, A. and Palmer, R.G. (1991) Introduction
to the Theory of Neural Computation. Addison-Wesley
Publishing, Redwood City, CA.
Hjorth-Simonsen, A. and Jeune, B. (1972) Origin and ter-
mination of the hippocampal perforant path in the rat
studied by silver impregnation. J. Comp. Neurol., 144:
215–232.
Holland, L.L. and Wagner, J.J. (1998) Primed facilitation of
homosynaptic long-term depression and depotentiation in rat
hippocampus. J. Neurosci., 18(3): 887–894.
Hong, J.S., McGinty, J.F., Grimes, L., Kanamatsu, T., Obie, J.
and Mitchell, C.L. (1988) Seizure-induced alterations in the
metabolism of hippocampal opioid peptides suggest opioid
modulation of seizure-induced behaviors. In: McGinty, J.F.,
Friedman, D.P. (Eds.), Opioids in the Hippocampus. NIDA,
Rockville, MD, pp. 48–66.
Hopfield, J.J. (1982) Neural networks and physical systems with
emergent collective computational abilities. Proc. Natl. Acad.
Sci. U.S.A., 79(8): 2554–2558.
Huang, Y.Y., Colino, A., Selig, D.K. and Malenka, R.C.
(1992) The influence of prior synaptic activity on the
induction of long-term potentiation. Science, 255(5045):
730–733.
Huang, Y.Y., Kandel, E.R., Varshavsky, L., Brandon,
E.P., Qi, M., Idzerda, R.L., McKnight, G.S. and
Bourtchouladze, R. (1995) A genetic test of the effects
of mutations in PKA on mossy fiber LTP and its
relation to spatial and contextual learning. Cell, 83(7):
1211–1222.
Jaffe, D. and Johnston, D. (1990) Induction of long-term
potentiation at hippocampal mossy fibers follow a Hebbian
rule. J. Neurophysiol., 64: 948–960.
James, W. (1890) Association. In: Psychology: A Briefer
Course. Holt, New York, pp. 253–279.
Jeffery, K.J., Donnett, J.G. and O’Keefe, J. (1995)
Medial septal control of theta-correlated unit firing in the
entorhinal cortex of awake rats. Neuroreport, 6(16):
2166–2170.
Jerman, T., Kesner, R.P. and Hunsaker, M.R. (2006) Discon-
nection analysis of CA3 and DG in mediating encoding but
not retrieval in a spatial maze learning task. Learn. Mem.,
13(4): 458–464.
Jin, W. and Chavkin, C. (1999) Mu opioids enhance mossy fiber
synaptic transmission indirectly by reducing GABAB recep-
tor activation. Brain Res., 821(2): 286–293.
Jones, M.W., Errington, M.L., French, P.J., Fine, A., Bliss,
T.V., Garel, S., Charnay, P., Bozon, B., Laroche, S. and
Davis, S. (2001) A requirement for the immediate early gene
Zif268 in the expression of late LTP and long-term memories.
Nat. Neurosci., 4(3): 289–296.
Jung, M.W. and McNaughton, B.L. (1993) Spatial selectivity of
unit activity in the hippocampal granular layer. Hippocam-
pus, 3: 165–182.
Kali, S. and Dayan, P. (2000) The involvement of
recurrent connections in area CA3 in establishing the
properties of place fields: a model. J. Neurosci., 20(19):
7463–7477.
Kapur, A., Yeckel, M.F. and Johnston, D. (2001) Hippocampal
mossy fiber activity evokes Ca2+ release in CA3 pyramidal
neurons via a metabotropic glutamate receptor. J. Ne-
urophysiol., 79(4): 2181–2190.
Kart-Teke, E., De Souza Silva, M.A., Huston, J.P. and
Dere, E. (2006) Wistar rats show episodic-like memory
for unique experiences. Neurobiol. Learn. Mem., 85(2):
173–182.
Kato, A., Ozawa, F., Saitoh, Y., Hirai, K. and Inokuchi, K.
(1997) vesl, a gene encoding VASP/Ena family related pro-
tein, is upregulated during seizure, long-term potentiation
and synaptogenesis. FEBS Lett., 412(1): 183–189.
O’Keefe, J. and Nadel, L. (1978) Hippocampus as a Cognitive
Map. Oxford University Press, New York.
Kerr, D.S. and Abraham, W.C. (1995) Cooperative interactions
among afferents govern the induction of homosynaptic long-
term depression in the hippocampus. Proc. Natl. Acad. Sci.,
92(25): 11637–11641.
Kerr, D.S. and Abraham, W.C. (1996) LTD: many means to
how many ends? Hippocampus, 6(1): 30–34.
Kirkwood, A., Dudek, S.M., Gold, J.T., Aizenman, C.D. and
Bear, M.F. (1993) Common forms of synaptic plasticity in
the hippocampus and neocortex in vitro. Science, 260(5113):
1518–1521.
Kleschevnikov, A.M. and Routtenberg, A. (2003) Long-term
potentiation recruits a trisynaptic excitatory associative net-
work within the mouse dentate gyrus. Eur. J. Neurosci.,
18(6): 1717.

Kobayashi, K. and Poo, M.M. (2004) Spike train timing-
dependent associative modification of hippocampal
CA3 recurrent synapses by mossy fibers. Neuron, 41(3):
445–454.
Kohonen, T. (1984) Self organization and Associative Memory.
Springer-Verlag, Berlin.
Kohonen, T., Lehtio, P., Rovamo, J., Hyvarinen, J., Bry, K.
and Vainio, L. (1977) A principle of neural associative mem-
ory. Neuroscience, 2: 1065–1076.
Kosub, K., Do, V. and Derrick, B.E. (2005) NMDA antago-
nists block heterosynaptic LTD, but not LTP, following
lateral perforant path stimulation. Neurosci. Lett., 374:
29–34.
Krug, M., Brodemann, R. and Wagner. (2001) Simultaneous
activation and opioid modulation of long-term potentiation
in the dentate gyrus and the hippocampal CA3 region after
stimulation of the perforant pathway in freely moving rats.
Brain Res., 913(1): 68–77.
Krug, M., Lossner, B. and Ott, T. (1984) Anisomycin
blocks the late phase of long-term potentiation in the
dentate gyrus of freely moving rats. Brain Res. Bull., 13(1):
39–42.
Kulla, A. and Manahan-Vaughan, D. (2000) Depotentiation
in the dentate gyrus of freely moving rats is modulated
by D1/D5 dopamine receptors. Cereb. Cortex, 10(6):
614–620.
Kulla, A. and Manahan-Vaughan, D. (2002) Modulation by
serotonin 5-HT(4) receptors of long-term potentiation and
depotentiation in the dentate gyrus of freely moving rats.
Cereb. Cortex, 12(2): 150–162.
Kulla, A., Reymann, K.G. and Manahan-Vaughan, D. (1999)
Time-dependent induction of depotentiation in the dentate
gyrus of freely moving rats: involvement of group 2 metabo-
tropic glutamate receptors. Eur. J. Neurosci., 11(11):
3864–3872.
Lee, H.K., Barbarosie, M., Kameyama, K., Bear, M.F. and
Huganir, R.L. (2000) Regulation of distinct AMPA receptor
phosphorylation sites during bidirectional synaptic plasticity.
Nature, 405(6789): 955–959.
Lee, I. and Kesner, R.P. (2004) Encoding versus retrieval of
spatial memory: double dissociation between the dentate
gyrus and the perforant path inputs into CA3 in the dorsal
hippocampus. Hippocampus, 14(1): 66–76.
Lei, S., Pelkey, K.A., Topolnik, L., Congar, P., Lacaille, J.C.
and McBain, C.J. (2003) Depolarization-induced long-term
depression at hippocampal mossy fiber-CA3 pyramidal neu-
ron synapses. J. Neurosci., 23(30): 9786–9795.
Levy, W.B. (1996) A sequence predicting CA3 is a flexible
associator that learns and uses context to solve hippocampal-
like tasks. Hippocampus, 6(6): 579–590.
Levy, W.B. and Steward, O. (1979) Synapses as associative
memory elements in the hippocampal formation. Brain Res.,
175: 233–245.
Levy, W.B. and Steward, O. (1983) Temporal contiguity re-
quirements for long-term associative potentiation/depression
in the hippocampus. Neuroscience, 8(4): 791–797.
447
Li, S., Cullen, W.K., Anwyl, R. and Rowan, M.J. (2003)
Dopamine-dependent facilitation of LTP induction in hippo-
campal CA1 by exposure to spatial novelty. Nat. Neurosci.,
6(5): 526–531.
Lisman, J.E. and McIntyre, C.C. (2001) Synaptic plasticity:
a molecular memory switch. Curr. Biol., 11(19): R788–
R791.
Magee, J.C. and Johnston, D. (1997) A synaptically-controlled,
associative signal for Hebbian plasticity in hippocampal neu-
rons. Science, 275: 209–213.
Maguire, C., Casey, M., Kelly, A., Mullany, P.M. and Lynch,
M.A. (1999) Activation of tyrosine receptor kinase plays a
role in expression of long-term potentiation in the rat dentate
gyrus. Hippocampus, 9(5): 519–526.
Malinow, R. (2003) AMPA receptor trafficking and long-term
potentiation. Philos. Trans. R. Soc. Lond. B Biol. Sci.,
358(1432): 707–714.
Manabe, T. (1997) Two forms of hippocampal long-term
depression, the counterpart of long-term potentiation. Rev.
Neurosci., 8(3–4): 179–193.
Manahan-Vaughan, D. (1998) Priming of group 2 metabotrop-
ic glutamate receptors facilitates induction of long-term de-
pression in the dentate gyrus of freely moving rats.
Neuropharmacology, 37(12): 1459–1464.
O’Mara, S.M., Rowan, M.J. and Anwyl, R. (1995a) Metabo-
tropic glutamate receptor-induced homosynaptic long-term
depression and depotentiation in the dentate gyrus of the
rat hippocampus in vitro. Neuropharmacology, 34(8):
983–989.
O’Mara, S.M., Rowan, M.J. and Anwyl, R. (1995b) Dantrolene
inhibits long-term depression and depotentiation of synaptic
transmission in the rat dentate gyrus. Neuroscience, 68(3):
621–624.
Marr, D. (1971) Simple Memory: a theory for archicorticortex.
Proc. R. Soc. Lond. B, 262: 23–81.
Martin, M.R. (1983) Naloxone and long-term potentiation of
hippocampal CA3 field potentials in vitro. Neuropeptides,
4(1): 45–50.
Martin, S.J. (1998) Time-dependent reversal of dentate LTP by
5 Hz stimulation. Neuroreport, 9(17): 3775–3781.
Martinez, C.O., Do, V.H. and Derrick, B.E. (2002) Associative
LTP Among afferents to the CA3 region in vivo. Brain Res.,
94: 86–94.
Massey, P.V., Johnson, B.E., Moult, P.R., Auberson, Y.P.,
Brown, M.W., Molnar, E., Collingridge, G.L. and Bashir,
Z.I. (2004) Differential roles of NR2A and NR2B-
containing NMDA receptors in cortical long-term pot-
entiation and long-term depression. J. Neurosci., 24:
7821–7828.
Matthies, H., Schroeder, H., Becker, A., Loh, H., Hollt, V. and
Krug, M. (2000) Lack of expression of long-term potentiat-
ion in the dentate gyrus but not in the CA1 region of the
hippocampus of mu-opioid receptor-deficient mice. Neuro-
pharmacology, 39(6): 952–960.
McClelland, J.L., McNaughton, B.L. and O’Reilly, R.C. (1995)
Why there are complementary learning systems in the
448
hippocampus and neocortex: insights from the successes and
failures of connectionist models of learning and memory.
Psychol. Rev., 102: 419–457.
McMahon, D.B. and Barrionuevo, G. (2002) Short- and long-
term plasticity of the perforant path synapse in hippocampal
area CA3. J. Neurophysiol., 88(1): 528–533.
McNaughton, B.L. and Barnes, C.A. (1977) Physiological iden-
tification and analysis of dentate granule cell responses to
stimulation of the medial and lateral perforant pathways in
the rat. J. Comp. Neurol., 175: 439–454.
McNaughton, B.L., Barnes, C.A., Meltzer, J. and Sutherland,
R.J. (1989) Hippocampal granule cells are necessary for nor-
mal spatial learning but not for spatially-selective pyramidal
cell discharge. Exp. Brain Res., 76: 485–496.
McNaughton, B.L., Douglas, R.M. and Goddard, D.V. (1978)
Synaptic enhancement in fascia dentata: cooperativity among
coactive afferents. Brain Res., 157: 277–293.
Mellor, J. and Nicoll, R.A. (2001) Hippocampal mossy fiber
LTP is independent of postsynaptic calcium. Nat. Neurosci.,
4(2): 125–126.
Milner, T.A. and Drake, C.T. (2001) Ultrastructural
evidence for presynaptic mu opioid receptor modulation
of synaptic plasticity in NMDA-receptor-containing dend-
rites in the dentate gyrus. Brain Res. Bull., 54(2): 131–140.
Mizumori, S.J., McNaughton, B.L., Barnes, C.A. and Fox,
K.B. (1989) Preserved spatial coding in hippocampal CA1
pyramidal cells during reversible suppression of CA3c
output: evidence for pattern completion in hippocampus.
J. Neurosci., 9(11): 3915–3928.
Monaghan, D.T. and Cotman, C.W. (1985) Distribution of
N-methyl-D-aspartate sensitive l-[3H]glutamate-binding sites
in rat brain. J. Neurosci., 5: 2909–2919.
Morgan, S.L. and Teyler, T.J. (2001) Electrical stimuli pat-
terned after the theta-rhythm induce multiple forms of LTP.
J. Neurophysiol., 86(3): 1289–1296.
Morris, R.G.M. (1989) Computational neuroscience: modeling
the brain. In: Morris R.G.M. (Ed.), Parallel-Distributed
Processing: Implication for Psychology and Neurobiology.
Oxford University Press, NY, pp. 203–212.
Morris, R.G.M. and McNaughton, B.L. (1987) Memory stor-
age in a distributed model of hippocampal formation. Trends
Neurosci., 10: 408–414.
Mott, D.D. and Lewis, D.V. (1992) GABAB receptors mediate
disinhibition and facilitate long-term potentiation in the
dentate gyrus. Epilepsy Res. Suppl., 7: 119–134.
Munoz, M.D., Nunez, A. and Garcia-Austt, E. (1990) In vivo
intracellular analysis of rat dentate gyrus granule cells. Brain
Res., 509: 91–98.
Nakazawa, K., Quirk, M.C., Chitwood, R.A., Watanabe, M.,
Yeckel, M.F., Sun, L.D., Kato, A., Carr, C.A., Johnston, D.,
Wilson, M.A. and Tonegawa, S. (2002) Requirement for
hippocampal CA3 NMDA receptors in associative memory
recall. Science, 297(5579): 211–218.
Neumaier, J.F. and Chavkin, C. (1989) Release of endogenous
opioid peptides displaces [3H]diprenorphine binding in rat
hippocampal slices. Brain Res., 493: 292–302.
Nguyen, P.V. and Kandel, E.R. (1996) A macromolecular
synthesis-dependent late phase of long-term potentia-
tion requiring cAMP in the medial perforant pathway
of rat hippocampal slices. J. Neurosci., 16(10):
3189–3198.
North, R.A., Williams, J.T., Surprenant, A. and Christie, M.J.
(1987) Mu and delta receptors belong to a family of receptors
that are coupled to potassium channels. Proc. Natl. Acad.
Sci. U.S.A., 84(15): 5487–5491.
Otani, S., Marshall, C.J., Tate, W.P., Goddard, G.V. and
Abraham, W.C. (1989) Maintenance of long-term potentiat-
ion in rat dentate gyrus requires protein synthesis but not
messenger RNA synthesis immediately post-tetanization.
Neuroscience, 28(3): 519–526.
Otmakhova, N.A. and Lisman, J.E. (1998) D1/D5 dopamine
receptors inhibit depotentiation at CA1 synapses via
cAMP-dependent mechanism. J. Neurosci., 18(4):
1270–1279.
Pang, K., Williams, M.J. and Olton, D.S. (1993) Activation
of the medial septal area attenuates LTP of the lateral
perforant path and enhances heterosynaptic LTD of the
medial perforant path in aged rats. Brain Res., 632(1–2):
150–160.
Philpot, B., Bear, M.F. and Abraham, W.C. (1999) Metaplas-
ticity: the plasticity of synaptic plasticity. In: Katz P.S. (Ed.),
Beyond Neurotransmission: Neuromodulation and its
Importance for Information processing. Oxford, London,
pp. 161–197.
Piguet, P. and North, R.A. (1983) Opioid actions at mu and
delta receptors in the rat dentate gyrus in vitro. J. Pharmacol.
Exp. Ther., 266(2): 1139–1149.
Ramirez-Amaya, V., Balderas, I., Sandoval, J., Escobar, M.L.
and Bermudez-Rattoni, F. (2001) Spatial long-term memory
is related to mossy fiber synaptogenesis. J. Neurosci., 21(18):
7340–7348.
Ramon y Cajal, S. (1937) Recuerdos de mi Vida. MIT Press,
Cambridge.
Reid, C.A., Dixon, D.B., Takahashi, M., Bliss, T.V. and Fine,
A. (2004) Optical quantal analysis indicates that long-term
potentiation at single hippocampal mossy fiber synapses is
expressed through increased release probability, recruitment
of new release sites, and activation of silent synapses. J. Ne-
urosci., 24(14): 3618–3626.
Rolls, E.T. and Treves, A. (1998) Neural Networks and Brain
Function. Oxford University Press, New York.
Rosenzweig, E.S., Barnes, C.A. and McNaughton, B.L.
(2002) Making room for new memories. Nat. Neurosci., 5:
6–8.
Rycroft, B.K. and Gibb, A.J. (2002) Direct effects of
calmodulin on NMDA receptor single channel gating
in rat hippocampal granule cells. J. Neurosci., 22: 8860–
8868.
Sanberg, C.-D., Jones, F., Do, V., Dieguez, D. and Derrick,
B.E. (2006) Antagonists of the 5-HT1a receptor block dent-
ate LTP induction in novel, but not familiar environments.
Learn. Mem., 13(1): 52–62.

Scharfman, H.E. (1996) Conditions required for polysynaptic
excitation of dentate granule cells by area CA3 pyra-
midal cells in rat hippocampal slices. Neuroscience, 72(3):
655–658.
Scharfman, H.E., Sollas, A.L., Smith, K.L., Jackson, M.B. and
Goodman, J.H. (2002) Structural and functional asymmetry
in the normal and epileptic rat dentate gyrus. J. Comp.
Neurol., 454(4): 424–439.
Schulz, S., Siemer, H., Krug, M. and Hollt, V. (1999) Direct
evidence for biphasic cAMP responsive element-binding
protein phosphorylation during long-term potentiation
in the rat dentate gyrus in vivo. J. Neurosci., 19(13):
5683–5692.
Schwegler, H., Crusio, W.E. and Brust, I. (1990) Hippocampal
mossy fibers and radial-maze learning in the mouse: a
correlation with spatial working memory but not with
non-spatial reference memory. Neuroscience, 34(2):
293–298.
Scoville, W.B. and Milner, B. (1957) Loss of recent memory by
bilateral hippocampal lesions. J. Neurol. Neurosurg. Psychi-
atry, 20: 11–21.
Segal, M. (1972) Hippocampal unit responses to perforant path
stimulation. Exp. Neurol., 35: 541–546.
Sheng, M. and Greenberg, M.E. (1990) The regulation and
function of c-fos and other immediate early genes in the
nervous system. Neuron, 4(4): 477–485.
Shors, T.J. and Matzel, L.D. (1997) Long-term potentiation:
what’s learning got to do with it? Behav. Brain Sci., 20(4):
597–614.
Shouval, H., Intrator, N. and Cooper, L.N. (1997) BCM
network develops orientation selectivity and ocular
dominance in natural scene environment. Vision Res.,
37(23): 3339–3342.
Skaggs, W.E. and McNaughton, B.L. (1992) Computational
approaches to hippocampal function. Curr. Opin. Neuro-
biol., 2(2): 209–211.
Snyder, J.S., Kee, N. and Wojtowicz, J.M. (2001) Effects of
adult neurogenesis on synaptic plasticity in the rat dentate
gyrus. J. Neurophysiol., 85(6): 2423–2431.
Soderling, T.R. and Derkach, V.A. (2000) Postsynaptic
protein phosphorylation and LTP. Trends Neurosci., 23(2):
75–80.
Squire, L. (1992) Memory and the hippocampus: a synthesis
from findings with rats, monkeys, and humans. Psychol.
Rev., 99: 195–231.
Staubli, U. and Lynch, G. (1987) Stable hippocampal long-term
potentiation elicited by ‘theta’ pattern stimulation. Brain
Res., 435(1–2): 227–234.
Steward, O. (1976) Topographic organization of the projections
from the entorhinal area to the hippocampal formation of the
rat. J. Comp. Neurol., 167: 285–314.
Steward, O. and Scoville, S.A. (1976) Cells of origin of ent-
orhinal cortical afferents to the hippocampus and fascia den-
tata of the rat. J. Comp. Neurol., 169: 347–370.
Steward, O., Wallace, C.S., Lyford, G.L. and Worley,
P.F. (1998) Synaptic activation causes the mRNA
449
for the IEG Arc to localize selectively near activated
postsynaptic sites on dendrites. Neuron, 21(4): 741–
751.
Straube, T., Korz, V., Balschun, D. and Frey, J.U. (2003)
Requirement of beta-adrenergic receptor activation and
protein synthesis for LTP-reinforcement by novelty in rat
dentate gyrus. J. Physiol., 552(Pt 3): 953–960.
Swearengen, E. and Chavkin, C. (1987) NMDA receptor
antagonist D-APV depresses excitatory activity produced
by normorphine in rat hippocampal slices. Neurosci. Lett.,
78: 80–84.
Sweatt, J.D. (2001) The neuronal MAP kinase cascade: a
biochemical signal integration system subserving synaptic
plasticity and memory. J. Neurochem., 76(1): 1–10.
Thiels, E., Xie, X., Yeckel, M.F., Barrionuevo, G. and Berger,
T.W. (1996) NMDA receptor-dependent LTD in different
subfields of hippocampus in vivo and in vitro. Hippocampus,
6(1): 43–51.
Thomas, K.L., Davis, S., Hunt, S.P. and Laroche, S. (1996)
Alterations in the expression of specific glutamate receptor
subunits following hippocampal LTP in vivo. Learn. Mem.,
3(2–3): 197–208.
Thomas, K.L., Laroche, S., Errington, M.L., Bliss, T.V.
and Hunt, S.P. (1994) Spatial and temporal changes in sig-
nal transduction pathways during LTP. Neuron, 13(3):
737–745.
Trepel, C. and Racine, R.J. (1998) Long-term potentiation in
the neocortex of the adult, freely moving rat. Cereb. Cortex,
8(8): 719–729.
Treves, A. and Rolls, E.T. (1992) Computational con-
straints suggest the need for two distinct input systems
to the hippocampal CA3 network. Hippocampus, 2(2):
189–199.
Treves, A. and Rolls, E.T. (1994) Computational analysis of the
role of the hippocampus in memory. Hippocampus, 4:
374–391.
Treves, A. and Roudi, Y. (2005) On the evolution of the brain.
In: Chow C.C., Gutkin B., Hansel D. and Meunier C. (Eds.),
Methods and Models in Neurophysics. Elsevier, Holland,
pp. 641–690.
Trommald, M., Hulleberg, G. and Andersen, P. (1996) Long-
term potentiation is associated with new excitatory spine
synapses on rat dentate granule cells. Learn. Mem., 3(2–3):
218–228.
Tulving, E. (1983) Elements of Episodic Memory. Clarendon
Press, Oxford.
Urban, N.N. and Barrionuevo, G. (1996) Induction of Hebbian
and non-Hebbian mossy fiber long-term potentiation by dis-
tinct patterns of high-frequency stimulation. J. Neurosci.,
16(13): 4293–4299.
Villarreal, D., Do, V., Haddad, E. and Derrick, B.E. (2002)
NMDA antagonists sustain LTP and spatial memory:
evidence for active processes underlying LTP decay. Nat.
Neurosci., 5: 48–52.
Villarreal, D.M. and Derrick, B.E. (2000) Long-term
metaplasticity: correspondence with LTP longevity
450
and the BCM Theory. Abst. Soc. for Neurosci., # 134.1
(online).
Villarreal, D.M. and Derrick, B.E. (2006) Input-specific
long-term potentiation reinduction at the rat perforant
path to dentate gyrus. Soc. for Neurosci. Abst., # 731.2
(online).
Wagner, J.J., Caudle, R.M., Neumaier, J.F. and Chavkin, C.
(1990) Stimulation of endogenous opioid release displaces
mu receptor binding in rat hippocampus. Neuroscience, 37:
45–53.
Wagner, J.J., Etemad, L.R. and Thompson, A.M. (2001)
Opioid-mediated facilitation of long-term depression
in rat hippocampus. J. Pharmacol. Exp. Ther., 296(3):
776–781.
Wagner, J.J., Evans, C.J. and Chavkin, C. (1991) Focal stim-
ulation of the mossy fibers releases endogenous dynorphins
that bind to k1-opioid receptors in guinea pig hippocampus.
J. Neurochem., 57: 333–343.
Wagner, J.J., Terman, G.W. and Chavkin, C. (1993) Endog-
enous dynorphins inhibit excitatory neurotransmission and
block LTP induction in the hippocampus. Nature, 363(6428):
451–454.
Wang, D., Tolbert, L.M., Carlson, K.W. and Sadee, W.
(2000) Nuclear Ca2+/calmodulin translocation activated
by mu-opioid (OP3) receptor. J. Neurochem., 74(4):
1418–1425.
Wang, J., Yeckel, M.F., Johnston, D. and Zucker, R.S.
(2004) Photolysis of postsynaptic caged Ca2+ can potenti-
ate and depress mossy fiber synaptic responses in rat
hippocampal CA3 pyramidal neurons. J. Neurophysiol.,
91(4): 1596–1607.
Weisskopf, M.G., Zalutsky, R.A. and Nicoll, R.A. (1993)
The opioid peptide dynorphin mediates heterosynaptic
depression of hippocampal mossy fibre synapses
and modulates long-term potentiation. Nature, 362:
423–427.
Wexler, E.M. and Stanton, P.K. (1993) Priming of homosy-
naptic long-term depression in hippocampus by previous
synaptic activity. Neuroreport, 4(5): 591–594.
White, J.D., Gall, C.M. and McKelvy, J.F. (1987) Enke-
phalin biosynthesis and enkephalin gene expression
are increased in hippocampal mossy fibers following
a unilateral lesion of the hilus. J. Neurosci., 7(3):
753–759.
Wickens, J.R. and Abraham, W.C. (1991) The involvement
of L-type calcium channels in heterosynaptic long-term
depression in the hippocampus. Neurosci. Lett., 130(1):
128–132.
Wigstrom, H. and Gustaffson, B. (1985) Facilitation of hippo-
campal long-lasting potentiation by GABA antagonists. Acta
Physiol. Scand., 125: 159–172.
Williams, J.M., Guevremont, D., Kennard, J.T., Mason-
Parker, S.E., Tate, W.P. and Abraham, W.C. (2003)
Long-term regulation of N-methyl-D-aspartate receptor
subunits and associated synaptic proteins following
hippocampal synaptic plasticity. Neuroscience, 118(4):
1003–1013.
Williams, S. and Johnston, D. (1989) Long-term potentiation
of hippocampal mossy fiber synapses is blocked by
postsynaptic injection of calcium chelators. Neuron, 3:
583–588.
Williams, S.H. and Johnston, D. (1996) Actions of endogenous
opioids on NMDA receptor-independent long-term potent-
iation in area CA3 of the hippocampus. J. Neurosci., 16:
3652–3660.
Wilson, M.A. and McNaughton, B.L. (1993) Dynamics of the
hippocampal ensemble code for space. Science, 261(5124):
1055–1058.
Winson, J. and Abzug, C. (1978) Dependence upon behavior
of neuronal transmission from perforant pathway through
entorhinal cortex. Brain Res., 147(2): 422–427.
Witter, M.P., Griffioen, A.W., Jorritsma-Byham, B. and
Krijnen, J.L. (1988) Entorhinal projections to the hippocam-
pal CA1 region in the rat: an underestimated pathway.
Neurosci. Lett., 85(2): 193–198.
Wu, J., Rowan, M.J. and Anwyl, R. (2004) Synaptically
stimulated induction of group I metabotropic glutamate
receptor-dependent long-term depression and depotentiation
is inhibited by prior activation of metabotropic glutamate
receptors and protein kinase C. Neuroscience, 123(2):
507–514.
Wu, J., Rowan, M.J. and Anwyl, R. (2006) Long-term potent-
iation is mediated by multiple kinase cascades involv-
ing CaMKII or either PKA or p42/44 MAPK in the
adult rat dentate gyrus in vitro. J. Neurophysiol., 95(6):
3519–3527.
Xie, C.-W. and Lewis, D.V. (1991) Opioid mediated facilitation
of long-term potentiation at the lateral perforant path-
dentate granule cell synapse. J. Pharmacol. Exp. Ther., 256:
289–296.
Yeckel, M.F. and Berger, T.W. (1990) Feedforward excitation
of the hippocampus by afferents from the entorhinal cortex:
redefinition of the role of the trisynaptic pathway. Proc. Natl.
Acad. Sci. U.S.A., 87: 5832–5836.
Yeckel, M.F. and Berger, T.W. (1998) Spatial distribution of
potentiated synapses in hippocampus: dependence on cellular
mechanisms and network properties. J. Neurosci., 18:
438–450.
Yeckel, M.F., Kapur, A. and Johnston, D. (1999) Multiple
forms of LTP in hippocampal CA3 neurons use a common
postsynaptic mechanism. Nat. Neurosci., 2(7): 625–633.
Ying, S.W., Futter, M., Rosenblum, K., Webber, M.J., Hunt,
S.P., Bliss, T.V. and Bramham, C.R. (2002) Brain-derived
neurotrophic factor induces long-term potentiation in intact
adult hippocampus: requirement for ERK activation coupled
to CREB and upregulation of Arc synthesis. J. Neurosci.,
22(5): 1532–1540.
Zalutsky, R.A. and Nicoll, R.A. (1990) Comparison of two
forms of long-term potentiation in single hippocampal neu-
rons. Science, 248: 1619–1624.

Zeineh, M.M., Engel, S.A., Thompson, P.M. and Bookheimer,
S.Y. (2003) Dynamics of the hippocampus during encoding
and retrieval of face-name pairs. Science, 299(5606): 577–580.
Zhang, D.X. and Levy, W.B. (1992) Ketamine blocks the
induction of LTP at the lateral entorhinal cortex-dentate
gyrus synapses. Brain Res., 593(1): 124–127.
451
Zimprich, A., Simon, T. and Hollt, V. (1995) Transfected rat
mu opioid receptors (rMOR1 and rMOR1B) stimulate
phospholipase C and Ca2+ mobilization. Neuroreport, 7(1):
54–56.
Zucker, R.S. and Regehr, W.G. (2002) Short-term synaptic
plasticity. Annu. Rev. Physiol., 64: 355–405.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 25

Control of synaptic consolidation in the dentate

gyrus: mechanisms, functions, and therapeutic
implications

Clive R. Bramham

Department of Biomedicine and Bergen Mental Health Research Center, University of Bergen, Jonas Lies vei 91,
N-5009 Bergen, Norway

Abstract: Synaptic consolidation refers to the development and stabilization of protein synthesis-depend-
ent modifications of synaptic strength as observed during long-term potentiation (LTP) and long-term
depression (LTD). Activity-dependent changes in synaptic strength are thought to underlie memory storage
and other adaptive responses of the nervous systems of importance in mood stability, reward behavior, and
pain control. This chapter focuses on the mechanisms and functions of synaptic consolidation in the
dentate gyrus, a critical structure not only in hippocampal memory function, but also in regulation of stress
responses and cognitive aspects of depression. Recent evidence suggests that synaptic consolidation at
excitatory medial perforant path-granule cell synapses requires brain-derived neurotrophic factor (BDNF)
signaling and induction of the immediate early gene activity-regulated cytoskeleton-associated protein
(Arc). Arc mRNA is strongly induced and transported to dendritic processes following high-frequency
stimulation (HFS) that induces LTP in the rat dentate gyrus in vivo. Sustained synthesis of Arc during a
surprisingly protracted time-window is required for hyperphosphorylation of actin depolymerizing factor/
cofilin and local expansion of the actin cytoskeleton in vivo. Furthermore, this process of Arc-dependent
synaptic consolidation is activated in response to brief infusion of BDNF. Microarray expression profiling
has revealed a panel of BDNF-regulated genes that may cooperate with Arc during synaptic consolidation.
In addition to regulating gene expression, BDNF signaling modulates the fine localization and biochemical
activation of the translation machinery. By modulating the spatial and temporal translation of newly
induced (Arc) and constitutively-expressed mRNA in dendrites, BDNF may effectively control the window
of synaptic consolidation. Dysregulation of BDNF synthesis and Arc function, specifically within the
dentate gyrus, is linked to behavioral symptoms and cognitive deficits in animal models of depression and
Alzheimer’s disease. Therapeutics strategies targeting synaptic consolidation hold promise for the future.

Keywords: long-term potentiation; brain-derived neurotrophic factor; BDNF; Neurotrophin; hippocampus;

Arc; gene expression; Depression; memory

Corresponding author. Tel.: +47 55 58 60 32; Fax: +47 55 58

64 10; E-mail: clive.bramham@biomed.uib.no

DOI: 10.1016/S0079-6123(07)63025-8 453

454
Control of synaptic consolidation in the dentate
gyrus
Synaptic consolidation
Persistent activity-dependent changes in synaptic
strength are believed to underlie a range of adaptive
brain responses, including memory formation,
mood stability, and drug addiction (Bliss and
Collingridge, 1993; Nestler et al., 2002). However,
our understanding of the mechanisms by which
altered activity patterns trigger lasting changes
in synaptic efficacy, exemplified by long-term
potentiation (LTP) and long-term depression
(LTD), is far from complete. A critical factor in
high-frequency stimulation (HFS)-induced LTP is
postsynaptic activation of N-methyl-D-aspartate
receptors (NMDARs) and voltage-dependent
calcium channels. The resulting activation of
calcium-sensitive kinases and other calcium-sensi-
tive effectors underlies a transient, early phase of
LTP maintained by protein phosphorylation and
membrane insertion of AMPA-type glutamate
receptors. The generation of stable, late phase
LTP requires at least one period of new gene
expression and protein synthesis (Bliss and Col-
lingridge, 1993; Frey et al., 1996; Nguyen and
Kandel, 1996). Analogous to memory consolida-
tion, the process of generating late LTP is referred
to as synaptic consolidation.
Late phase LTP is operationally defined on the
basis of sensitivity of LTP maintenance to broad
spectrum inhibitors of global gene expression and
protein synthesis. Progress in unraveling the mo-
lecular basis of synaptic consolidation has been
hampered by the lack of evidence that expression of
a specific gene product or group of gene products
mediates LTP. Alterations in gene expression in re-
sponse to changes in neuronal activity are common,
and, as pointed out by several authors (Sanes and
Lichtman, 1999; Routtenberg and Rekart, 2005), it
is likely that many regulated genes have no role or
only a subsidiary role in generating stable LTP.
BDNF as a trigger of synaptic consolidation
Several lines of evidence suggest that BDNF, act-
ing through its receptor tyrosine kinase (TrkB)
triggers synaptic consolidation at adult excitatory
synpases. BDNF, unlike other members of the neu-
rotrophin family (nerve growth factor, neurotro-
phin-3, and neurotrophin-4), is widely distributed
across the adult brain and in hippocampal princi-
pal cells. BDNF signals through catalytic TrkB
receptors present on pre- and postsynaptical
elements of glutamatergic synapses (Drake et al.,
1999). Postsynaptic TrkB receptors are found
in the postsynaptic density (PSD) and TrkB
co-immunoprecipitates with NMDAR complex
proteins (Wu et al., 1996; Drake et al., 1999; Aoki
et al., 2000; Husi et al., 2000). Postsynaptic release
of BDNF from secretogranin II-immunoreactive
vesicles occurs in response to HFS, and is sensitive
to NMDAR blockade. BDNF may therefore be
viewed as a co-neurotransmitter acting in tandem
with glutamate at excitatory synapses. Two addi-
tional features important to mention in the context
of synaptic consolidation are: (1) BDNF regulates
protein synthesis through both transcriptional and
post-transcriptional mechanisms, and (2) BDNF
is capable of stimulating its own release, possibly
allowing sustained, regenerative signaling at
synapses (Santi et al., 2006).
Genetic and pharmacological studies establish-
ing the role of BDNF-TrkB in LTP have been
reviewed in detail elsewhere (Bramham and
Messaoudi, 2005). The contributions of BDNF
signaling to LTP can be classified as permissive or
instructive. Permissive mechanisms make synapses
competent for LTP. For example, constitutive
(non-evoked) presynaptic BDNF signaling in-
creases the fraction of neurotransmitter vesicles
docked in the active zone. This increase in the
readily releasable pool of vesicles enables sustained
glutamate release during trains of action poten-
tials, and facilitates LTP induction (Figurov et al.,
1996). Instructive BDNF signaling, in contrast, is
initiated in response to HFS and is required for
LTP development. Acute release of BDNF during
HFS modulates the induction and early mainte-
nance of LTP (Kossel et al., 2001). A series of
studies based on focal application of BDNF sug-
gests a mechanism involving activation of a
tetrodotoxin (TTX)-insensitive voltage-dependent
sodium channel leading to rapid depolarization
and enhanced calcium influx into dendritic spines

during LTP induction (Blum and Konnerth,
2005). The mechanism of TrkB coupling to
the TTX-insensitive channel is unclear at the
moment.
Formation of late LTP is coupled to a period of
sustained BDNF release and TrkB receptor acti-
vation (Kang et al., 1997; Aicardi et al., 2004).
Extracellular concentrations of BDNF are ele-
vated for less than a minute following weak HFS,
leading to decremental early LTP; whereas
stimulus protocols that induce late LTP produce
BDNF elevation lasting at least 5 min. Pharmaco-
logical inhibition of TrkB receptor activation
during the first hour after HFS blocks synaptic
consolidation in hippocampal region CA1 (Kang
et al., 1997). Furthermore, bath application of
BDNF induces a protein synthesis-dependent
increase in synaptic efficacy termed BDNF-LTP
(Kang and Schuman, 1996; Messaoudi et al.,
1998). Induction of BDNF-LTP at medial perfo-
rant path (MPP)-granule cell synapses in the dent-
ate gyrus is blocked by inhibitors of RNA
synthesis and occluded by prior expression of
late phase, but not early phase, of HFS-LTP
(Messaoudi et al., 2002; Ying et al. 2002). BDNF-
LTP lasts at least 1 day in awake rats and, like
HFS-LTP, is associated with increases in both
synaptic efficacy and EPSP-spike coupling.
BDNF-LTP induction is not affected by NMDA
receptor blockade, consistent with the hypothesis
that NMDA receptor activation stimulates
BDNF release but is not involved in BDNF-LTP
induction or expression. Like HFS-LTP, BDNF-
LTP induction requires extracellular signal-
regulated protein kinase (ERK) activation and is
coupled to upregulation, dendritic transport, and
translation of the immediate early gene activity-
regulated cytoskeleton-associated protein (Arc,
aka Arg 3.1). These results suggest that endog-
enous BDNF activates synaptic consolidation and
that exogenous BDNF mimics this effect.
Arc as mediator of synapse consolidation
Persistent LTP is thought to occur when small
spines are converted to large mushroom-shaped
spines through a mechanism dependent on actin
455
polymerization (Matsuzaki et al., 2004). Insertion
of glutamate receptors at postsynaptic mem-
branes, increase in PSD diameter, and structural
remodeling of spines are all connected to regula-
tion of actin dynamics (Geinisman, 2000; Lisman
and Zhabotinsky, 2001; Weeks et al., 2001; Zhou
et al., 2001; Fukazawa et al., 2003; Harris et al.,
2003; Matsuzaki et al., 2004; Okamoto et al., 2004;
Zito et al., 2004; Oertner and Matus, 2005).
Fukazawa et al. (2003) showed that LTP at
MPP-granule cell synapses in anesthetized rats
is associated with an increase in filamentous actin
(F-actin) content at activated synapses and
enhanced phosphorylation of cofilin, a major
regulator of actin dynamics. Phosphorylation of
cofilin on Ser3 inhibits activity and promotes actin
polymerization. LIM domain kinase (LIMK), one
of the major cofilin kinases in brain, modulates the
morphology of dendritic spines through regulation
of cofilin activity. Mice lacking the LIMK1 gene
exhibit small, actin-deficient spines. Although the
initial amplitude of LTP is actually larger, LIMK
knockout mice fail to sustain late phase LTP
(Meng et al., 2002). Taken together, this suggests a
major role for cofilin regulation in actin-dependent
synapse expansion underlying late LTP.
How is gene expression coupled to actin dy-
namics and synapse expansion? LTP is associated
with the induction of a number of immediate-early
genes that encode a variety of transcription factors
and other proteins, unassociated with transcrip-
tion, and some of these have been causally linked
to late phase LTP and long-term memory. Arc is
unique because it encodes the only mRNA known
to undergo transport to distal dendritic processes
of granule cells within the first hour of LTP in-
duction (Link et al., 1995; Lyford et al., 1995). Arc
mRNA is enriched at stimulated synapses, and Arc
protein is transiently elevated in dendritic spines
following LTP induction (Steward and Worley,
2001; Moga et al., 2004; Rodriguez et al., 2005).
Arc co-precipitates with crude F-actin and is
found in the PSD of excitatory synapses (Lyford
et al., 1995; Husi et al., 2000). Arc is dynamically
expressed in principal neurons of many cortical
and limbic structures during behavioral training,
and this expression is necessary for long-term
memory in a variety of memory tasks (Guzowski
456

et al., 1999, 2000; Plath et al., 2006; Vazdarjanova
et al., 2006).
Early work of Guzowski et al. using Arc anti-
sense oligodeoxynucleotides (AS) suggested a role
for Arc in LTP maintenance in the dentate gyrus.
This issue has now been examined in more detail
(Messaoudi et al., 2007). These authors show that
Arc antisense application at 2 h (but not 4 h) after
LTP induction leads to rapid and complete re-
versal of LTP. LTP reversal is coupled to rapid
knockdown of newly synthesized Arc mRNA and
protein, dephosphorylation of hyperphosphory-
lated cofilin, and loss of a discrete band of phal-
loidin staining in the dentate molecular layer,
indicating disappearance of new F-actin at MPP
synapses. Introduction of the F-actin stabilizer
jasplakinolide during LTP maintenance blocks the
reversal of LTP by Arc AS. Interestingly, LTP is
only transiently inhibited when Arc antisense
is administered shortly before (5 min) HFS, sug-
gesting that sustained synthesis of Arc during a
time-window lasting at least 2 h is necessary to
consolidate LTP. These results suggest a pivotal
role for activity-induced Arc synthesis in control
of synaptic consolidation and local expansion of
the actin cytoskeleton. Furthermore, Arc AS
infusion blocks BDNF-LTP and reverses mainte-
nance of BDNF-LTP over a critical time-window.
Thus, Arc-dependent consolidation is directly ac-
tivated by local BDNF application. Figure 1
presents a working model of Arc function in LTP.
Although the exact role of actin polymerization
in synapse expansion is unknown, this process
must involve the addition of newly synthesized or
translocated proteins to the postsynaptic special-
ization. Rather than working alone, Arc is likely to

Fig. 1. BDNF as a trigger of synaptic consolidation. The mechanism of stable LTP formation at glutamatergic synapses is presented as
a two-stage process: translation activation and Arc-dependent consolidation. In the translation activation stage, patterned high-
frequency stimulation leads to sustained postsynaptic release of BDNF and activation of TrkB receptors pre- and postsynaptically.
Postsynaptic TrkB leads to (1) rapid activation and translocation of the translation machinery in dendritic spines, and (2) Arc
transcription in granule cell bodies. Translation activation is mediated by phosphorylation of the cap-binding protein eIF4E, and
possibly by more mRNA-specific mechanisms such as relief of miRNA-mediated translation repression. Spines activated in this way
may effectively capture and translate local mRNA pools. In this model, transcripts liberated from local RNA storage granules are
translated first, followed by dendritic transport and sustained translation of newly synthesized Arc mRNA. During Arc-dependent
consolidation, sustained translation of Arc is necessary for cofilin phosphorylation, local F-actin expansion, and formation of stable
LTP. This is a working model; several points require further experimental validation (see text). Adapted from Soule et al. (2006) with
permission from the Biochemical Society.

be part of a coordinate process of synapse growth
involving rapid activity-induced regulation of
many of genes. Wibrand et al. (2006) used cDNA
microarray expression profiling to screen for genes
that are co-upregulated with Arc mRNA 3 h after
BDNF-LTP induction in the dentate gyrus of
anesthetized rats. Of nine genes upregulated
more than fourfold, five were selected for inde-
pendent confirmation. Real-time PCR confirmed
robust upregulation of all five genes: neuronal
activity-regulated pentraxin (Narp), neuritin,
ADP-ribosylation factor-like protein-4 (ARL4L),
TGF-b-induced immediate early gene-1 (TIEG1),
and calcium/calmodulin kinase-related peptide
(CARP). In situ hybridization histochemistry
further revealed upregulation of these genes in
somata of postsynaptic granule cells following
BDNF-LTP and HFS-LTP. Consistent with a
process of synaptic growth and remodeling, Narp,
neuritin, and TIEG1 are involved in synaptogen-
esis, axon guidance, and glutamate receptor
clustering during development.
Another immediate early gene critical for late
phase LTP in the dentate gyrus and long-term
memory is zif286, one of the four members of the
early growth response (egr) family of zinc-finger
transcription factors. Mice with constitutive
deletion of zif268 (aka egr1) show impaired
LTP maintenance one day after HFS as well as
impaired long-term memory consolidation and
reconsolidation (Jones et al., 2001; Bozon et al.,
2003). Other studies examining the effects of an-
tisense injection into the hippocampus concluded
that synthesis of BDNF, but not zif268, is involved
in initial memory consolidation (Lee et al., 2004).
Conversely, zif268, but not BDNF, synthesis was
implicated in reconsolidation.
Immediate early gene responses are typically
rapid and protein synthesis-independent. In addi-
tion to the classical protein synthesis-independent
induction, Arc mRNA is induced through a
delayed protein synthesis-dependent mechanism.
Interestingly, recent work has identified Arc as a
direct transcriptional target of zif268 and egr-3
(Li et al., 2005). Protein synthesis-dependent
expression of Arc following exploration of a novel
environment or seizures is at least partly depend-
ent on egr-transcription factors including zif268.
457
This raises the intriguing possibility that zif268
stimulates the prolonged increase in Arc induction
following HFS-LTP in the dentate gyrus. The fact
that zif268 is not induced during BDNF-LTP
suggests that Arc-dependent consolidation is in-
dependent of zif268 (Messaoudi et al., 2002;
Ying et al., 2002; Wibrand et al., 2006). Endog-
enous BDNF may nonetheless contribute to zif268
expression during HFS-LTP.
BDNF gene expression is enhanced following
HFS-LTP induction in the CA1 region and dent-
ate gyrus (Patterson et al., 1992; Castren et al.,
1993; Bramham et al., 1996; Lee et al., 2005b).
LTP at MPP-granule cell synapses of awake rats
is associated with NMDA receptor-dependent in-
creases in TrkB and BDNF mRNA expression
(Bramham et al., 1996). TrkB mRNA is elevated
unilaterally in dentate granule cells 2 h after LTP
induction, whereas BDNF mRNA is elevated
bilaterally at 6 and 24 h after LTP induction. The
bilateral increase in BDNF expression is coupled
to unilateral LTP induction, as commissural re-
sponses and contralateral MPP-granule cell re-
sponses were unaffected. Conceivably, such
bilateral effects are due to altered network prop-
erties following LTP (Dragoi et al., 2003). Unilat-
eral increases in BDNF gene expression occur in
dentate granule cells 3 h after BDNF-LTP induc-
tion. Much of this response is due to regulation of
the highly calcium-responsive exon III BDNF
promoter (Tao et al., 1998; Wibrand et al.,
2006). This suggests that BDNF signaling regu-
lates BDNF transcription during long-term synap-
tic plasticity. However, the key question, whether
new BDNF transcription actually contributes to
synaptic strengthening, remains to be addressed.
It is just as likely that BDNF transcription
serves only to replenish depleted stores of protein
following intensive BDNF secretion.
BDNF and translation control in dendrites
Synaptic consolidation requires transcription,
localization, and translation of mRNA. Each of
these mechanisms is modulated by BDNF signa-
ling (Klann and Dever, 2004; Richter and
Sonenberg, 2005; Bramham and Wells, 2007).
458
The discovery of mRNA, ribosomes, and transla-
tion factors within dendrites and even within
the spine itself suggested that synapses could be
modified directly — and perhaps individually —
through regulation of local protein synthesis.
While dendritic protein synthesis is regulated by
glutamate, acetylcholine, norepinephrine, dopa-
mine, and probably other neurotransmitters, the
BDNF-TrkB system stands out as a potentially
autonomous modulator of translation at gluta-
mate synapses. In the CA1 region of adult hippo-
campal slices, BDNF-LTP is induced in synaptic
regions isolated from the pyramidal cell somata
(Kang and Schuman, 1996). In primary hippo-
campal neuronal cultures, BDNF induces expres-
sion of a reporter construct in isolated dendrites
(Aakalu et al., 2001). In synaptodendrosome (SD)
preparations, subcellular fractions containing
resealed terminal-spine contacts, BDNF rapidly
stimulates translation of several mRNAs cou-
pled to LTP or spine morphogenesis, including
a-calcium/calmodulin-dependent protein kinase II
(aCaMKII), Arc, LIMK1, and GluR1 (Yin et al.,
2002; Schratt et al., 2004). Regulation of transla-
tion by BDNF has been shown in SDs prepared
from forebrain and hippocampus of young (PN
15–22 day) rats (Yin et al., 2002; Schratt et al.,
2004; Kanhema et al., 2006), as well as dentate
gyrus from 2–4 month old rats (Kanhema et al.,
2006).
Translation factor regulation
Recent work has examined the biochemical mech-
anisms by which BDNF modulates local protein
synthesis. Phosphorylation of the eukaryotic initi-
ation factor 4E (eIF4E) is considered the rate-lim-
iting step for translation of most mRNAs (those
with a 7-methyl-guanosine ‘‘cap’’ at the 50 end).
Phosphorylation of eIF4E on Ser209 is correlated
with enhanced rates of translation, whereas hypo-
phosphorylation is associated with decreased
translation. eIF4E is phosphorylated by mitogen-
activated integrating kinase 1 (MNK1), whose
activity is regulated by ERK and p38 mitogen ac-
tivated protein kinase (MAPK). The availability of
eIF4E is controlled by several binding proteins
(BPs), most notably eIF4E-binding proteins
(4E-BPs). Signaling through receptor-coupled
phosphoinositide 3-kinase (PI3K) kinase and
mammalian target of rapamycin (mTOR) leads
to phosphorylation of 4E-BP and liberation of
eIF4E. BDNF stimulates cap-dependent transla-
tion in dendrites through TrkB-coupled PI3K-
mTOR and Ras-ERK (Takei et al., 2001,
2002; Kelleher et al., 2004; Schratt et al., 2004;
Takei et al., 2004; Kanhema et al., 2006). Genetic
or pharmacological inhibition of mTOR or
ERK impairs LTP maintenance and abolishes
BDNF-induced LTP (Rosenblum et al., 2002;
Tang et al., 2002; Ying et al., 2002; Cammalleri
et al., 2003; Kelleher et al., 2004). BDNF-LTP
in the dentate gyrus is coupled to transient ERK-
dependent phosphorylation of eIF4E and en-
hanced expression of eIF4E protein (Kanhema
et al., 2006).
Protein synthesis in synaptic plasticity is also
controlled at the level of peptide chain elongation.
Eukaryotic elongation factor 2 (eEF2) is a GTP-
binding protein that mediates translocation of
peptidyl-tRNAs from the A-site to the P-site on
the ribosome. Phosphorylation of eEF2 on Thr56
inhibits eEF2-ribosome binding and arrests elon-
gation (Browne and Proud, 2002). eEF2 phos-
phorylation observed during LTP could therefore
explain decreased synthesis of some proteins in 2D
gel studies. Paradoxically, however, some tran-
scripts important for synaptic plasticity such as
Arc and aCaMKII undergo enhanced translation
under conditions of eEF2 phosphorylation and
global suppression of protein synthesis (Scheetz
et al., 2000; Chotiner et al., 2003; Kanhema et al.,
2006). BDNF-LTP in vivo is coupled to ERK-
dependent phosphorylation of eEF2 and enhanced
Arc expression in whole dentate gyrus homogen-
ates (Kanhema et al., 2006). Consolidation of taste
memory in the neocortex is also associated with
ERK activation and transient eEF2 phosphorylat-
ion. (Belelovsky et al., 2005). In constrast, BDNF
treatment of SDs does not alter eEF2 phosphor-
ylation state, but leads to eIF4E phosphorylation
and rapid synthesis of Arc, aCaMKII, LIMK, and
other proteins (Yin et al., 2002; Schratt et al., 2004;

Kanhema et al., 2006). This suggests a compart-
ment-specific regulation in which initiation is
selectively enhanced by BDNF at synapses, while
both initiation and elongation are modulated at
synaptic sites.
eEF2 is phosphorylated by eEF2 kinase, which
itself is subject to tight regulation by calcium/
calmodulin, mTOR, ERK, and PKA through
multiple phosphorylation sites. Recent studies
have shown bidirectional effects of BDNF on
eEF2 phosphorylation depending on the prepara-
tion studied (Inamura et al., 2005). At synapses,
eEF2 is phosphorylated in response to NMDA
(Scheetz et al., 2000), but not BDNF treatment
(Kanhema et al., 2006). In primary cortical neu-
rons, BDNF induces dephosphorylation of eEF2
and increases elongation rates (Inamura et al.,
2005). Serotonin similarly reduces eEF2 phos-
phorylation in Aplysia synaptosomes. In this case,
serotonin appears to offset an increase in eEF2
phosphorylation triggered by calcium influx
(Carroll et al., 2004). Thus, the direction of
eEF2 phosphorylation is context-dependent and
controlled by multiple transmitters. While eEF2
appears to be important in transcript-specific and
compartment-specific translation control in synap-
tic plasticity, further work is needed to resolve
these mechanisms.
In addition to cap-dependent mechanisms,
mRNA translation may be initiated at internal
ribosomal entry sites (IRES). A large percentage
of total translation persists even when cap-
dependent mechanisms are almost completely
abolished. Evidence suggests that IRES-depend-
ent translation is enhanced or maintained in the
context of eIF4E hyperphosphorylation and
saturation of cap-dependent mechanisms, and
during conditions such as ischemia when eIF2a
and eEF2 are phosphorylated (Dyer et al.,
2003). A number of dendritic mRNA species,
including Arc and aCAMKII, contain IRES se-
quences and are capable of IRES-dependent
translation in cultured hippocampal neurons and
non-neuronal cell lines (Pinkstaff et al., 2001).
However, IRES-mediated translation has not
been explored in the context of synaptic
plasticity (Fig. 2).
459
miRNA
Another area bound to be important for regu-
lated protein synthesis in neuronal plasticity is
microRNAs (miRNAs). miRNAs are small, non-
coding RNAs that bind to the 30 UTR of target
mRNAs and either block translation or induce
transcript degradation (Schratt et al., 2006). As
miRNAs generally bind numerous target
mRNAs, and most mRNAs are regulated by
more than one species of miRNA, the possibilities
for fine regulation of translation are enormous.
Schratt et al. showed that miR-134, a brain-
specific miRNA found in the synaptodendritic
compartment, negatively regulates dendritic spine
morphogenesis in cultured hippocampal neurons
by repressing translation of LIMK1 mRNA. Ap-
plication of BDNF relieves miR134-mediated re-
pression of LIMK1 translation and promotes
spine morphogenesis. Interestingly, the RNA in-
terference complex (RISC) proteins that mediate
the effects of miRNA are themselves subject to
regulation. Ashraf et al. recently demonstrated
that RISC pathway proteins regulate dendritic
mRNA transport and local protein synthesis
during acquisition of long-term memory in flies
(Ashraf et al., 2006). According to their model,
training in olfactory/electric shock paradigm trig-
gers proteosomal degradation of the SDE3 heli-
case Armitage, leading to relief of RISC-mediated
suppression of several mRNAs, including aCaM-
KII, kinesin heavy chain, and staufen. Adding to
this complexity, regulation of miRNA transcrip-
tion may be coupled to specific physiological
effects. Impey et al. showed that calcium/cyclic
AMP responsive element binding protein
(CREB)-dependent transcription of miR132 pro-
motes neuronal morphogenesis by decreasing
synthesis of the GTPase activating protein,
p250GAP. Like many CREB-dependent effects
in neurons, this mechanism was activated by
BDNF (Vo et al., 2005). Furthermore, a recent
study employing microarray expression profiling
and PCR analysis has revealed specific upregula-
tion and downregulation of miRNAs following
LTP induction in the dentate gyrus in vivo
(Wibrand et al., 2006).
460

Fig. 2. TrkB and translation control in dendritic spines. The cartoon depicts some of the major signaling pathways coupling TrkB to
regulation of eIF4E and eEF2. TrkB activation of PI3K-mTOR and Ras-ERK promotes eIF4E phosphorylation and enhances
translation initiation. Phosphorylation of eEF2 stalls ribosomes and arrests peptide chain elongation. BDNF-TrkB signaling has
bidirectional effects on eEF2 phosphorylation. Adapted from Soule et al. (2006) with permission from the Biochemical Society.

Translocation also induces a redistribution of eIF4E to an

Translocation and positioning of the translation
machinery is another feature of activity-dependent
regulation. RNA storage granules in dendrites
discharge mRNA in response to strong depolari-
zation (Krichevsky and Kosik, 2001; Kanai et al.,
2004). During LTP, ribosomes move from dendri-
tic shafts to spines (Ostroff et al., 2002), and
pre-existing a-CaMKII mRNA moves to the
synaptodendritic compartment (Havik et al.,
2003). BDNF treatment of synaptoneurosomes
mRNA granule-rich cytoskeletal fraction (Smart
et al., 2003).
Model of BDNF-controlled synaptic consolidation
To summarize, current evidence supports a model
for synaptic consolidation triggered by postsynap-
tic BDNF signaling and involving regulation
of gene expression and local mRNA translation
(Fig. 1). According to this model, bursts of

excitatory synaptic activity trigger sustained re-
lease of BDNF and activation of postsynaptic
TrkB receptors. TrkB signaling rapidly activates
translation in spines and induces transcription of
Arc mRNA in cell bodies. Translational activation
in spines consists of global (eIF4E), and probably
more mRNA-specific (miRNA), mechanisms.
Spines activated in this way may effectively
capture and translate mRNAs liberated from
local RNA storage granules, as well as newly in-
duced Arc mRNA transport from the soma.
Translation of Arc during a critical time-window
is necessary for cofilin phosphorylation, local
F-actin expansion, and formation of stable LTP.
The model predicts that translation of Arc under-
lies actin polymerization-dependent enlargement
of synapses.
Stages in synaptic consolidation: from Arc
to PKCz?
Arc is likely to interact with other core mecha-
nisms of synaptic consolidation. The list of critical
players includes N-cadherin, members of the inte-
grin receptor family, matrix metalloproteinase-9,
and the atypical protein kinase C isoform PKMz,
(Bahr et al., 1997; Bozdagi et al., 2000; Ling et al.,
2002; Chan et al., 2003; Nagy et al., 2006).
Of these, PKMz warrants special note. PKMz is
constitutively active because it consists of a PKMz
catalytic domain, but lacks the autoinhibitory
domain of most PKC isoforms. Evidence suggests
that persistent PKMz activity is sufficient and
necessary for a major component of late phase
LTP. Thus, bath application of the myristoylated
z-pseudosubstrate inhibititory peptide (ZIP) to
hippocampal slices blocks potentiation produced
by intracellular perfusion of PKMz and reverses
the maintenance of LTP in the CA1 region (Ling
et al., 2002; Serrano et al., 2005). A recent study
further reported that maintenance of LTP in the
dentate gyrus of awake rats is rapidly reversed by
intrahippocampal injection of ZIP 24 h after
LTP induction (Pastalkova et al., 2006). More-
over, injection of ZIP produces persistent loss of
1-day-old spatial information. LTP is associated
with transcription and enhanced dendritic
461
translation of PKMz, which is thought to sustain
LTP by increasing the number of active postsy-
naptic a-amino-3-hydroxy-5-methyl-4-isoxazole-
propionic acid receptors (AMPAR) (Ling et al.,
2006). Interestingly, the F-actin destabilizing agent
latrunculin B blocks new synthesis of PKMz and
attenuates LTP maintenance (Kelly et al., 2007).
Taken together, this suggests a sequential mecha-
nism of LTP maintenance in the dentate gyrus in
which Arc-dependent consolidation couples to
PKCz-dependent expression at the level of actin
polymerization.
Extrinsic modulation of synaptic consolidation
Modulatory transmitters such as norepinephrine,
serotonin, dopamine, and acetylcholine modulate
LTP induction and/or maintenance (Stanton and
Sarvey, 1985a, b; Bramham and Srebro, 1989;
Frey et al., 1991; Bramham et al., 1997; Swanson-
Park et al., 1999; Graves et al., 2001; Kulla and
Manahan-Vaughan, 2002; Straube and Frey, 2003;
Harley et al., 2005). These extrinsic neurotrans-
mitter systems have diffuse, global patterns of
innervation, and communicate through spatially
dispersed, volume transmission. Neuronal firing of
modulatory inputs is a function of the animal’s
behavioral or attentional state, with changes in
activity dictating the functional modes of target
networks (i.e., local rhythmic activity, timing of
synaptic events, frequency and duration of action
potential firing). The classical modulatory trans-
mitters may also set the biochemical tone in target
neurons by modulating PKA and CREB activity.
Acetylcholine, dopamine, and norepinephrine
have all been found to modulate local protein
synthesis in mammalian principal neurons
(Feig and Lipton, 1993; Gelinas and Nguyen,
2005; Smith et al., 2005), and serotonergic regula-
tion of local protein synthesis is well known from
studies of Aplysia sensory neurons (Casadio et al.,
1999).
While these actions are spatially dispersed, they
are likely to exert broad influence and act in
concert with specific mechanisms triggered by
glutamate and BDNF signaling. b-adrenergic re-
ceptor activation facilitates late LTP through a
462
mechanism involving local protein synthesis, but
not transcription (Straube and Frey, 2003; Gelinas
and Nguyen, 2005). Locus coeruleus activation
induces a delayed protein synthesis-dependent
LTP in the dentate gyrus (Walling and Harley,
2004). b-Adrenergic receptor activation is also
necessary for protein synthesis underlying novelty-
induced facilitation of late LTP (Straube et al.,
2003), and PKA activity is necessary for synaptic
tagging and input specificity of late LTP in the
CA1 region (Young et al., 2006). One intriguing
possibility is that modulatory inputs are able to
contract or expand the time-window of synaptic
consolidation through effects on local mRNA
translation.
Presynaptic mechanisms in synaptic consolidation
Quantal analysis and biochemical studies support
a contribution of enhanced glutamate release to
LTP expression at MPP-granule cell synapses
(Min et al., 1998; Errington et al., 2003). Evidence
supporting a role for BDNF in enhanced gluta-
mate transmitter release during LTP includes the
following: (1) the maintenance phase of BDNF-
LTP, like HFS-LTP, is associated with a lasting
increase in potassium-evoked glutamate release
from synaptosomes (Gooney et al., 2004), (2) TrkB
receptors are autophosphorylated in synaptosomes
collected during the maintenance phase of both
HFS- and BDNF-induced LTP (Gooney et al.,
2002, 2004), (3) the Trk inhibitor, K252a, blocks
the sustained enhancement in neurotransmitter
release. Finally, LTP maintenance is associated
with enhanced, depolarization-evoked release of
BDNF from dentate gyrus tissue (Gooney and
Lynch, 2001).
The kinetics of these events suggest that BDNF
acts rapidly on terminals to trigger a lasting in-
crease in glutamate release. However, more de-
layed mechanisms involving retrograde nuclear
signaling could contribute. The classic hypothesis
of target-derived trophic support involves signa-
ling from the nerve terminal to the nucleus (Ginty
and Segal, 2002; Delcroix et al., 2003; Campenot
and MacInnis, 2004). In sympathetic and sensory
neurons, neurotrophin binding to presynaptic Trk
receptors activates retrograde signaling pathways
in axons leading to activation of nuclear sub-
strates, such as CREB, and modulation of gene
expression. HFS of the perforant pathway leads to
CREB phosphorylation in the entorhinal cortex
and this effect is blocked by intracerebroventricu-
lar application of the Trk inhibitor K252a (Kelly
et al., 2000; Gooney and Lynch, 2001). BDNF-
LTP, induced by local infusion into the dentate
gyrus, similarly leads to retrograde activation of
CREB in entorhinal cortex located some 4 mm
away (Gooney et al., 2004). Such retrograde nu-
clear signaling could contribute to sustained
expansion of presynaptic specializations and
enhanced neurotransmitter release. BDNF treat-
ment of cell cultures is associated with enhanced
vesicle docking, expression of Rab3a, a small
GTP-binding proteins involved in vesicle traffick-
ing, and enhanced expression of SNARE proteins
involved in vesicle exocytosis.
Lateral perforant path
The lateral perforant input (LPP) projecting from
layer II stellate neurons in the lateral entorhinal
cortex to the outer-third of the dentate gyrus mo-
lecular layer has been less extensively studied in
the context of LTP mechanisms. The LPP contains
proenkephalin-derived peptides that are released
in a frequency-dependent manner. Pharmacologi-
cal studies show that m and d opioid receptor ac-
tivation is necessary for LTP induction in the
lateral perforant pathway, while antagonists for
these receptors have no effect on LTP induction
or maintenance in the MPP (Bramham, 1992).
Several lines of evidence suggest that opioid pep-
tides facilitate LTP induction by dampening GAB-
Aergic inhibition (Madison and Nicoll, 1988; Xie
and Lewis, 1991; Hanse and Gustafsson, 1992;
Bramham and Sarvey, 1996; Milner and Drake,
2001). Relief from GABAA receptor-mediated in-
hibition facilitates activation of NMDARs and
voltage-dependent calcium channels. In transverse
hippocampal slices, LPP LTP and spike-timing-
dependent LTP and LTD is NMDAR-dependent,
whereas de-potentiation and de-depression re-
quire group 1 metabotropic glutamate receptor

activation (Xie and Lewis, 1991; Hanse and
Gustafsson, 1992; Colino and Malenka, 1993;
Bramham and Sarvey, 1996; Lin et al., 2006).
In anesthetized rats, however, normal LPP LTP is
obtained in the presence of NMDAR antagonists
that abolish MPP LTP (Bramham et al., 1991).
In vivo LTP induction in the LPP input to CA3
and is also opioid receptor-dependent and
NMDAR-independent (Do et al., 2002; Kosub
et al., 2005). The discrepancy between the hippo-
campal slice and in vivo preparations with regard
to the NMDAR dependence of LPP LTP remains
unresolved, although loss of fine inhibitory control
in the slice preparation due to transection of
the perpendicularly oriented interneurons during
slice preparation may favor NMDA receptor-
dependent mechanisms.
The role of BDNF and other neurotrophic
factors in the LPP is not well known. Studies of
paired-pulse plasticity have revealed a remarkable
pathway-specificity, such that NT-3 modulates
only LPP responses and BDNF modulates only
MPP responses (Kokaia et al., 1998; Asztely et al.,
2000). Conditional NT-3 mutant mice have deficits
in the LPP LTP, but not MPP, LTP induction
(Shimazu et al., 2006).
The molecular mechanisms of late phase LTP in
the LPP have similarly received scant attention.
However, the available evidence reveals certain
similarities with MPP LTP mechanisms. For ex-
ample: (1) Arc mRNA localizes selectively in the
LPP termination zone following long (>15 min)
sessions of repetitive HFS of the LPP (Steward
et al., 1998), (2) LPP LTP induced by standard
short bursts of HFS is associated with input-
specific accumulation of F-actin (Fukazawa et al.,
2003), and (3) Like MPP LTP (Villarreal et al.,
2002), stable LPP LTP in awake rats is degraded
by NMDA receptor activation (Abraham et al.,
2006). This NMDAR-dependent decay of LPP
LTP is induced by high-frequency heterosynaptic
stimulation of the MPP. Electron microscopic
studies of synapse morphometry show that LTP
of both the lateral and medial peforant path are
associated with increases in the number of unper-
forated axospinous synapses. However, differences
at the ultrastructural level probably exist as the
number of perforated axospinous synapses is
463
elevated 24 h after LTP induction of the MPP,
but not LPP (Mezey et al., 2004).
Functions and clinical implications of synaptic
consolidation in the dentate gyrus
Perturbations in dentate gyrus synaptic plasticity
are thought to contribute to a range of clinical
conditions including memory loss, Alzheimer’s
disease, and depression. There are many potential
mechanisms involved, including those described
above for synaptic consolidation, including
BDNF, Arc, and other critical mediators. In
addition, the role of newly generated granule cells
and the synaptic plasticity of these new cells are
likely to contribute.
Alzheimer’s disease
Modulation of BDNF synthesis would be expected
to universally enhance the actions of BDNF.
Given the diverse effects of BDNF, even within
processes like LTP, it may be advantageous to
target specific effector mechanisms such as Arc-
dependent synaptic consolidation. Recent animal
studies connect Arc expression in dentate granule
cells to Alzheimer’s disease. In transgenic mice
expressing human amyloid precursor protein
(hAPP) and hAPP-derived amyloid-b (Ab), Arc
expression is reduced primarily in granule cells
of the dentate gyrus (Palop et al., 2005). Other
neuronal populations expressing hAPP, including
CA1 pyramidal cells, show normal basal Arc
expression and induction following exposure to a
novel environment. In neuronal cell cultures,
BDNF-mediated induction of Arc is suppressed
even at sublethal levels of Ab (Wang et al., 2006).
Downregulation of Arc expression in granule cells
may therefore contribute to deficits in LTP and
learning observed in rodent models of Alzheimer’s
disease (Rowan et al., 2003; Gureviciene et al.,
2004; Jacobsen et al., 2006).
Although progress has been made in elucidating
the computational role of the entorhinal cortex
(which provides input to the dentate gyrus through
the perforant path) and CA3 region (which re-
ceives output from granule cells), the function of
464
the dentate gyrus remains enigmatic. Recent work
suggests that the dentate gyrus allows fine spatio-
temporal separation of novel and complex cues,
thereby disambiguating stimuli to allow sparse
encoding of information (Kesner et al., 2004;
Lee et al., 2005a). Arc mRNA levels in the dent-
ate gyrus are elevated several hours following
LTP induction or exploration of a novel environ-
ment, whereas Arc mRNA induction in CA1
region and other brain areas is only short-lived
(minutes) (Ramirez-Amaya et al., 2005). It is
tempting to speculate that the protracted time-
window of Arc-dependent consolidation is
uniquely associated with the function of the dent-
ate gyrus in disambiguating information for
encoding in the entorhinal-hippocampal circuitry.
Given the pivotal role of ongoing Arc synthesis,
it’s possible that this window has a dynamic range
that is regulated at the level of Arc mRNA trans-
lation. If so, it may prove possible to contract
or expand the window of synaptic consolidation
therapeutically.
Neurogenesis
Among the unique features of the dentate gyrus is
the lifelong production of new granule cells from
progenitor cells located in the subgranular zone
(Gould and Gross, 2002; Aimone et al., 2006).
Although the function of neurogenesis is still un-
clear, it seems likely that the unique contributions
of the dentate gyrus are tied to the availability of
new granule cells. One hypothesis is that new
granule cell populations are selectively engaged in
new information storage, the unique associations
of successive populations providing a kind of
timestamp for episodic memory and recall
(Aimone et al., 2006). Recent work indicates that
HFS-LTP promotes the survival of newborn gran-
ule cells (Bruel-Jungerman et al., 2006), and
NMDAR activation promotes integration of these
cells within the dentate gyrus circuitry (Tashiro
et al., 2006). BDNF-TrkB signaling is similarly
implicated in survival of adult born granule cells
(Sairanen et al., 2005; Scharfman et al., 2005).
Castren has suggested that BDNF promotes the
survival and maturation of new connections
through classic mechanism of target-derived sup-
port, with active synapses competing for limiting
amounts of growth factor (Castren, 2005). An-
other possibility is that BDNF, as a direct conse-
quence of NMDAR activation, drives synaptic
consolidation of nascent contacts as it does at
mature synapses.
Depression
Alterations in BDNF signaling figure prominently
in current theories of depression (Castren, 2004;
Monteggia et al., 2004; Berton and Nestler, 2006;
Kuipers and Bramham, 2006). Chronic stress
leading to depression-like behavior in animals
is associated with downregulation of BDNF
transcription and TrkB signaling, while chronic
antidepressant treatment increases BDNF synthe-
sis reverses the effects of stress on BDNF expres-
sion and behavior. Furthermore, the dentate gyrus
appears to be key structure in cognitive aspects of
depression. First, chronic mild stress leading to
anhedonia-like behavior in rats is associated with
reductions in BDNF expression and CREB activ-
ity in the dentate gyrus, but not hippocampus
proper (Gronli et al., 2006). Second, acute bilateral
infusion of the BDNF into the dentate gyrus,
using a protocol similar to that which induced
BDNF-LTP, has antidepressant-like behavioral
effects (Shirayama et al., 2002). Third, injection
of virus expressing histone deacetylase, which
modulates BDNF promoter activity, blocks the
behavioral effects of antidepressants (Tsankova
et al., 2006).
It is noteworthy that several important events
associated with BDNF-LTP are also induced
by chronic antidepressant treatment. Chronic
treatment with the selective serotonin reuptake
inhibitor fluoxetine leads to enhanced TrkB signa-
ling, CREB activation, and induction of several
BDNF-LTP associated genes including BDNF,
Arc, neuritin, TIEG1, and CARP (Alme et al.,
2006). Dagestad et al. (2006) recently reported that
chronic, but not acute, treatment with fluoxetine
results in region-specific changes in translation
factor activity. In the dentate gyrus, a pattern of
dual eIF4E and eEF2 phosphorylation is seen
 465

similar to that observed following BDNF-LTP in- References

duction (Dagestad et al., 2006). Taken together
this suggests a direct connection between BDNF- Aakalu, G., Smith, W.B., Nguyen, N., Jiang, C. and Schuman,
induced synaptic strengthening in the dentate E.M. (2001) Dynamic visualization of local protein synthesis
in hippocampal neurons. Neuron, 30(2): 489–502.
gyrus and antidepressant drug action. It will be Abraham, W.C., Mason-Parker, S.E., Irvine, G.I., Logan, B.
important to determine whether BDNF-LTP and Gill, A.I. (2006) Induction and activity-dependent re-
actually develops during antidepressant treatment versal of persistent LTP and LTD in lateral perforant path
and whether specific inhibition of BDNF-LTP synapses in vivo. Neurobiol. Learn. Mem., 86(1): 82–90.
blocks the behavioral effects of local BDNF Aicardi, G., Argilli, E., Cappello, S., Santi, S., Riccio, M.,
Thoenen, H. and Canossa, M. (2004) Induction of long-term
infusion. potentiation and depression is reflected by corresponding
changes in secretion of endogenous brain-derived neurotro-
phic factor. Proc. Natl. Acad. Sci. U.S.A., 101(44):
15788–15792.
Abbreviations Aimone, J.B., Wiles, J. and Gage, F.H. (2006) Potential role for
adult neurogenesis in the encoding of time in new memories.
Nat. Neurosci., 9(6): 723–727.
AMPAR a-amino-3-hydroxy-5-methyl-4- Alme, M., Wibrand, K., Kuipers, S., Dagestad, G. and
isoxazolepropionic acid (AMPA) Bramham, C.R. (2006) Chronic fluoxetine treatment en-
receptor hances expression of BDNF-regulated genes in prefrontal
Arc activity-regulated cytoskeleton- cortex. Soc. Neurosci., Abstr., 314.10.
Aoki, C., Wu, K., Elste, A., Len, G., Lin, S., McAuliffe, G. and
associated protein
Black, I.B. (2000) Localization of brain-derived neurotrophic
BDNF brain-derived neurotrophic factor and TrkB receptors to postsynaptic densities of adult
factor rat cerebral cortex. J. Neurosci. Res., 59(3): 454–463.
a-CaMKII calcium/calmodulin-dependent Ashraf, S.I., McLoon, A.L., Sclarsic, S.M. and Kunes, S. (2006)
protein kinase II Synaptic protein synthesis associated with memory is regu-
lated by the RISC pathway in Drosophila. Cell, 124(1):
CREB calcium/cyclic AMP responsive
191–205.
element binding protein Asztely, F., Kokaia, M., Olofsdotter, K., Ortegren, U. and
eEF2 eukaryotic elongation factor 2 Lindvall, O. (2000) Afferent-specific modulation of short-
eIF4E eukaryotic initiation factor 4E term synaptic plasticity by neurotrophins in dentate gyrus.
4E-BP eIF4E-binding protein Eur. J. Neurosci., 12(2): 662–669.
Bahr, B.A., Staubli, U., Xiao, P., Chun, D., Ji, Z.X., Esteban,
ERK extracellular signal-regulated
E.T. and Lynch, G. (1997) Arg-Gly-Asp-Ser-selective
protein kinase adhesion and the stabilization of long-term potentiation:
HFS high-frequency stimulation pharmacological studies and the characterization of a can-
LIMK LIM domain kinase didate matrix receptor. J. Neurosci., 17(4): 1320–1329.
LTP long-term potentiation Belelovsky, K., Elkobi, A., Kaphzan, H., Nairn, A.C. and
Mnk1 mitogen-activated integrating Rosenblum, K. (2005) A molecular switch for translational
control in taste memory consolidation. Eur. J. Neurosci.,
kinase 1 22(10): 2560–2568.
mTOR mammalian target of rapamycin Berton, O. and Nestler, E.J. (2006) New approaches to
NMDAR N-methyl-D-aspartate receptor antidepressant drug discovery: beyond monoamines. Nat.
PI3K phosphoinositide 3-kinase Rev. Neurosci., 7(2): 137–151.
PSD postsynaptic density Bliss, T.V. and Collingridge, G.L. (1993) A synaptic model of
memory: long-term potentiation in the hippocampus. Nature,
SD synaptodendrosome 361(6407): 31–39.
Blum, R. and Konnerth, A. (2005) Neurotrophin-mediated
rapid signaling in the central nervous system: mechanisms
and functions. Physiology (Bethesda), 20: 70–78.
Bozdagi, O., Shan, W., Tanaka, H., Benson, D.L. and Huntley,
Acknowledgments
G.W. (2000) Increasing numbers of synaptic puncta during
late-phase LTP: N-cadherin is synthesized, recruited to
Supported by the Norwegian Research Council synaptic sites, and required for potentiation. Neuron, 28(1):
and European Union grant 504231. 245–259.
466
Bozon, B., Davis, S. and Laroche, S. (2003) A requirement for
the immediate early gene zif268 in reconsolidation of recog-
nition memory after retrieval. Neuron, 40(4): 695–701.
Bramham, C.R. (1992) Opioid receptor dependent long-term
potentiation: peptidergic regulation of synaptic plasticity in
the hippocampus [see comments]. Neurochem. Int., 20(4):
441–455.
Bramham, C.R., Bacher-Svendsen, K. and Sarvey, J.M. (1997)
LTP in the lateral perforant path is beta-adrenergic receptor-
dependent. Neuroreport, 8(3): 719–724.
Bramham, C.R. and Messaoudi, E. (2005) BDNF function in
adult synaptic plasticity: the synaptic consolidation hypoth-
esis. Prog. Neurobiol., 76: 99–125.
Bramham, C.R., Milgram, N.W. and Srebro, B. (1991) Acti-
vation of AP5-sensitive NMDA receptors is not required to
induce LTP of synaptic transmission in the lateral perforant
path. Eur. J. Neurosci., 3: 1300–1308.
Bramham, C.R. and Sarvey, J.M. (1996) Endogenous activa-
tion of mu and delta-1 opioid receptors is required for long-
term potentiation induction in the lateral perforant path: de-
pendence on GABAergic inhibition. J. Neurosci., 16(24):
8123–8131.
Bramham, C.R., Southard, T., Sarvey, J.M., Herkenham, M.
and Brady, L.S. (1996) Unilateral LTP triggers bilateral in-
creases in hippocampal neurotrophin and trk receptor
mRNA expression in behaving rats: evidence for interhem-
ispheric communication. J. Comp. Neurol., 368(3): 371–382.
Bramham, C.R. and Srebro, B. (1989) Synaptic plasticity in the
hippocampus is modulated by behavioral state. Brain Res.,
493: 74–86.
Bramham, C.R. and Wells, D.G. (2007) Dendritic mRNA:
transport, translation, and function. Nat. Rev. Neurosci. (in
press).
Browne, G.J. and Proud, C.G. (2002) Regulation of peptide-
chain elongation in mammalian cells. Eur. J. Biochem.,
269(22): 5360–5368.
Bruel-Jungerman, E., Davis, S., Rampon, C. and Laroche, S.
(2006) Long-term potentiation enhances neurogenesis in the
adult dentate gyrus. J. Neurosci., 26(22): 5888–5893.
Cammalleri, M., Lutjens, R., Berton, F., King, A.R., Simpson,
C., Francesconi, W. and Sanna, P.P. (2003) Time-restricted
role for dendritic activation of the mTOR-p70S6K pathway
in the induction of late-phase long-term potentiation in the
CA1. Proc. Natl. Acad. Sci. U.S.A., 100(24): 14368–14373.
Campenot, R.B. and MacInnis, B.L. (2004) Retrograde trans-
port of neurotrophins: fact and function. J. Neurobiol., 58(2):
217–229.
Carroll, M., Warren, O., Fan, X. and Sossin, W.S. (2004) 5-HT
stimulates eEF2 dephosphorylation in a rapamycin-sensitive
manner in Aplysia neuritis. J. Neurochem., 90(6): 1464–1476.
Casadio, A., Martin, K.C., Giustetto, M., Zhu, H., Chen, M.,
Bartsch, D., Bailey, C.H. and Kandel, E.R. (1999) A tran-
sient, neuron-wide form of CREB-mediated long-term facil-
itation can be stabilized at specific synapses by local protein
synthesis. Cell, 99(2): 221–237.
Castren, E. (2004) Neurotrophic effects of antidepressant drugs.
Curr. Opin. Pharmacol., 4(1): 58–64.
Castren, E. (2005) Is mood chemistry? Nat. Rev. Neurosci.,
6(3): 241–246.
Castren, E., Pitkanen, M., Sirvio, J., Parsadanian, A., Lind-
holm, D., Thoenen, H. and Riekkinen, P.J. (1993) The in-
duction of LTP increases BDNF and NGF mRNA but
decreases NT-3 mRNA in the dentate gyrus. Neuroreport,
4(7): 895–898.
Chan, C.S., Weeber, E.J., Kurup, S., Sweatt, J.D. and Davis,
R.L. (2003) Integrin requirement for hippocampal synaptic
plasticity and spatial memory. J. Neurosci., 23(18):
7107–7116.
Chotiner, J.K., Khorasani, H., Nairn, A.C., O’Dell, T.J. and
Watson, J.B. (2003) Adenylyl cyclase-dependent form of
chemical long-term potentiation triggers translational regu-
lation at the elongation step. Neuroscience, 116(3): 743–752.
Colino, A. and Malenka, R.C. (1993) Mechanisms underlying
induction of long-term potentiation in rat medial and lateral
perforant paths in vitro. J. Neurophysiol., 69(4): 1150–1159.
Dagestad, G., Kuipers, S.D., Messaoudi, E. and Bramham,
C.R. (2006) Chronic fluoxetine induces region-specific
changes in translation factor eIF4E and eEF2 activity in
the rat brain. Eur. J. Neurosci., 23: 2814–2818.
Delcroix, J.D., Valletta, J.S., Wu, C., Hunt, S.J., Kowal, A.S.
and Mobley, W.C. (2003) NGF signaling in sensory neurons:
evidence that early endosomes carry NGF retrograde signals.
Neuron, 39(1): 69–84.
Do, V.H., Martinez, C.O., Martinez Jr., J.L. and Derrick, B.E.
(2002) Long-term potentiation in direct perforant path pro-
jections to the hippocampal CA3 region in vivo. J. Ne-
urophysiol., 87(2): 669–678.
Dragoi, G., Harris, K.D. and Buzsaki, G. (2003) Place repre-
sentation within hippocampal networks is modified by long-
term potentiation. Neuron, 39(5): 843–853.
Drake, C.T., Milner, T.A. and Patterson, S.L. (1999) Ultra-
structural localization of full-length trkB immunoreactivity
in rat hippocampus suggests multiple roles in modulating
activity-dependent synaptic plasticity. J. Neurosci., 19(18):
8009–8026.
Dyer, J.R., Michel, S., Lee, W., Castellucci, V.F., Wayne, N.L.
and Sossin, W.S. (2003) An activity-dependent switch to cap-
independent translation triggered by eIF4E dephosphorylat-
ion. Nat. Neurosci., 6(3): 219–220.
Errington, M.L., Galley, P.T. and Bliss, T.V. (2003) Long-term
potentiation in the dentate gyrus of the anaesthetized rat
is accompanied by an increase in extracellular gluta-
mate: real-time measurements using a novel dialysis elec-
trode. Philos. Trans. R. Soc. Lond. B Biol. Sci., 358(1432):
675–687.
Feig, S. and Lipton, P. (1993) Pairing the cholinergic agonist
carbachol with patterned Schaffer collateral stimulation in-
itiates protein synthesis in hippocampal CA1 pyramidal cell
dendrites via a muscarinic, NMDA-dependent mechanism. J.
Neurosci., 13(3): 1010–1021.
Figurov, A., Pozzo Miller, L.D., Olafsson, P., Wang, T. and
Lu, B. (1996) Regulation of synaptic responses to high-fre-
quency stimulation and LTP by neurotrophins in the hippo-
campus. Nature, 381(6584): 706–709.

Frey, U., Frey, S., Schollmeier, F. and Krug, M. (1996) Influ-
ence of actinomycin D, a RNA synthesis inhibitor, on long-
term potentiation in rat hippocampal neurons in vivo and in
vitro. J. Physiol. Lond., 490(Pt 3): 703–711.
Frey, U., Matthies, H. and Reymann, K.G. (1991) The effect of
dopaminergic D1 receptor blockade during tetanization on
the expression of long-term potentiation in the rat CA1 re-
gion in vitro. Neurosci. Lett., 129(1): 111–114.
Fukazawa, Y., Saitoh, Y., Ozawa, F., Ohta, Y., Mizuno, K.
and Inokuchi, K. (2003) Hippocampal LTP is accompanied
by enhanced F-actin content within the dendritic spine that is
essential for late LTP maintenance in vivo. Neuron, 38(3):
447–460.
Geinisman, Y. (2000) Structural synaptic modifications associ-
ated with hippocampal LTP and behavioral learning. Cereb.
Cortex, 10(10): 952–962.
Gelinas, J.N. and Nguyen, P.V. (2005) Beta-adrenergic receptor
activation facilitates induction of a protein synthesis-depend-
ent late phase of long-term potentiation. J. Neurosci., 25(13):
3294–3303.
Ginty, D.D. and Segal, R.A. (2002) Retrograde neurotrophin
signaling: Trk-ing along the axon. Curr. Opin. Neurobiol.,
12(3): 268–274.
Gooney, M. and Lynch, M.A. (2001) Long-term potentiation in
the dentate gyrus of the rat hippocampus is accompanied by
brain-derived neurotrophic factor-induced activation of
TrkB. J. Neurochem., 77(5): 1198–1207.
Gooney, M., Messaoudi, E., Maher, F.O., Bramham, C.R. and
Lynch, M.A. (2004) BDNF-induced LTP in dentate gyrus is
impaired with age: analysis of changes in cell signaling events.
Neurobiol. Aging, 25(10): 1323–1331.
Gooney, M., Shaw, K., Kelly, A., O’Mara, S.M. and Lynch,
M.A. (2002) Long-term potentiation and spatial learning are
associated with increased phosphorylation of TrkB and ex-
tracellular signal-regulated kinase (ERK) in the dentate
gyrus: evidence for a role for brain-derived neurotrophic
factor. Behav. Neurosci., 116(3): 455–463.
Gould, E. and Gross, C.G. (2002) Neurogenesis in adult mam-
mals: some progress and problems. J. Neurosci., 22(3):
619–623.
Graves, L., Pack, A. and Abel, T. (2001) Sleep and memory: a
molecular perspective. Trends Neurosci., 24(4): 237–243.
Gronli, J., Bramham, C.R., Murison, R., Kanhema, T., Fiske,
E., Bjorvatn, B., Sorensen, E., Ursin, R. and Portas, C.M.
(2006) Chronic mild stress inhibits BDNF expression and
CREB activation in the dentate gyrus but not in the hippo-
campus proper. Pharmacol. Biochem. Behav., 85(4):
842–849.
Gureviciene, I., Ikonen, S., Gurevicius, K., Sarkaki, A.,
van Groen, T., Pussinen, R., Ylinen, A. and Tanila, H.
(2004) Normal induction but accelerated decay of LTP
in APP+PS1 transgenic mice. Neurobiol. Dis., 15(2):
188–195.
Guzowski, J.F., Lyford, G.L., Stevenson, G.D., Houston, F.P.,
McGaugh, J.L., Worley, P.F. and Barnes, C.A. (2000) Inhi-
bition of activity-dependent arc protein expression in the rat
hippocampus impairs the maintenance of long-term
467
potentiation and the consolidation of long-term memory. J.
Neurosci., 20(11): 3993–4001.
Guzowski, J.F., Mcnaughton, B.L., Barnes, C.A. and Worley,
P.F. (1999) Environment-specific expression of the immedi-
ate-early gene Arc in hippocampal neuronal ensembles. Nat.
Neurosci., 2(12): 1120–1124.
Hanse, E. and Gustafsson, B. (1992) Long-term potentiation
and field EPSPs in the lateral and medial perforant paths in
the dentate gyrus in vitro: a comparison. Eur. J. Neurosci.,
4(11): 1191–1201.
Harley, C.W., Walling, S.G. and Brown, R.A.M. (2005) The
three faces of norepinephrine: plasticity at the perforant path-
dentate gyrus synapse. In: Stanton P., Bramham C. and
Scharfman H. (Eds.), Synaptic plasticity and transsynaptic
signaling. Spring Science+Business Media, Berlin.
Harris, K.M., Fiala, J.C. and Ostroff, L. (2003) Structural
changes at dendritic spine synapses during long-term potent-
iation. Philos. Trans. R. Soc. Lond. B Biol. Sci., 358(1432):
745–748.
Havik, B., Rokke, H., Bardsen, K., Davanger, S. and Bram-
ham, C.R. (2003) Bursts of high-frequency stimulation trig-
ger rapid delivery of pre-existing alpha-CaMKII mRNA to
synapses: a mechanism in dendritic protein synthesis during
long-term potentiation in adult awake rats. Eur. J. Neurosci.,
17(12): 2679–2689.
Husi, H., Ward, M.A., Choudhary, J.S., Blackstock, W.P. and
Grant, S.G. (2000) Proteomic analysis of NMDA receptor-
adhesion protein signaling complexes. Nat. Neurosci., 3(7):
661–669.
Inamura, N., Nawa, H. and Takei, N. (2005) Enhancement of
translation elongation in neurons by brain-derived neurotro-
phic factor: implications for mammalian target of rapamycin
signaling. J. Neurochem., 95(5): 1438–1445.
Jacobsen, J.S., Wu, C.C., Redwine, J.M., Comery, T.A., Arias,
R., Bowlby, M., Martone, R., Morrison, J.H., Pangalos,
M.N., Reinhart, P.H. and Bloom, F.E. (2006) Early-onset
behavioral and synaptic deficits in a mouse model of Al-
zheimer’s disease. Proc. Natl. Acad. Sci. U.S.A., 103(13):
5161–5166.
Jones, M.W., Errington, M.L., French, P.J., Fine, A., Bliss,
T.V., Garel, S., Charnay, P., Bozon, B., Laroche, S. and
Davis, S. (2001) A requirement for the immediate early gene
Zif268 in the expression of late LTP and long-term memories.
Nat. Neurosci., 4(3): 289–296.
Kanai, Y., Dohmae, N. and Hirokawa, N. (2004) Kinesin
transports RNA: isolation and characterization of an RNA-
transporting granule. Neuron, 43(4): 513–525.
Kang, H. and Schuman, E.M. (1996) A requirement for local
protein synthesis in neurotrophin-induced hippocampal
synaptic plasticity. Science, 273(5280): 1402–1406.
Kang, H., Welcher, A.A., Shelton, D. and Schuman, E.M.
(1997) Neurotrophins and time: different roles for TrkB
signaling in hippocampal long-term potentiation. Neuron,
19(3): 653–664.
Kanhema, T., Dagestad, G., Panja, D., Tiron, A., Messaoudi,
E., Havik, B., Ying, S.W., Nairn, A.C., Sonenberg, N. and
Bramham, C.R. (2006) Dual regulation of translation
468
initiation and peptide chain elongation during BDNF-in-
duced LTP in vivo: evidence for compartment-specific trans-
lation control. J. Neurochem., 19: 1328–1337.
Kelleher III, R.J., Govindarajan, A., Jung, H.Y., Kang, H. and
Tonegawa, S. (2004) Translational control by MAPK signa-
ling in long-term synaptic plasticity and memory. Cell,
116(3): 467–479.
Kelly, A., Mullany, P.M. and Lynch, M.A. (2000) Protein syn-
thesis in entorhinal cortex and long-term potentiation in
dentate gyrus. Hippocampus, 10(4): 431–437.
Kelly, M.T., Yao, Y., Sondhi, R. and Sacktor, T.C. (2007)
Actin polymerization regulates the synthesis of PKMzeta in
LTP. Neuropharmacology, 52(1): 41–45.
Kesner, R.P., Lee, I. and Gilbert, P. (2004) A behavioral as-
sessment of hippocampal function based on a subregional
analysis. Rev. Neurosci., 15(5): 333–351.
Klann, E. and Dever, T.E. (2004) Biochemical mechanisms for
translational regulation in synaptic plasticity. Nat. Rev. Ne-
urosci., 5(12): 931–942.
Kokaia, M., Asztely, F., Olofsdotter, K., Sindreu, C.B., Kull-
mann, D.M. and Lindvall, O. (1998) Endogenous neurotro-
phin-3 regulates short-term plasticity at lateral perforant
path-granule cell synapses. J. Neurosci., 18(21): 8730–8739.
Kossel, A.H., Cambridge, S.B., Wagner, U. and Bonhoeffer, T.
(2001) A caged Ab reveals an immediate/instructive effect of
BDNF during hippocampal synaptic potentiation. Proc.
Natl. Acad. Sci. U.S.A., 98(25): 14702–14707.
Kosub, K.A., Do, V.H. and Derrick, B.E. (2005) NMDA re-
ceptor antagonists block heterosynaptic long-term depression
(LTD) but not long-term potentiation (LTP) in the CA3 re-
gion following lateral perforant path stimulation. Neurosci.
Lett., 374(1): 29–34.
Krichevsky, A.M. and Kosik, K.S. (2001) Neuronal RNA
granules: a link between RNA localization and stimulation-
dependent translation. Neuron, 32(4): 683–696.
Kuipers, S.D. and Bramham, C.R. (2006) Brain-derived ne-
urotrophic factor mechanisms and function in adult synaptic
plasticity: new insights and implications for therapy. Curr.
Opin. Drug Discov. Devel., 9(5): 580–586.
Kulla, A. and Manahan-Vaughan, D. (2002) Modulation by
serotonin 5-HT(4) receptors of long-term potentiation and
depotentiation in the dentate gyrus of freely moving rats.
Cereb. Cortex, 12(2): 150–162.
Lee, I., Hunsaker, M.R. and Kesner, R.P. (2005a) The role of
hippocampal subregions in detecting spatial novelty. Behav.
Neurosci., 119(1): 145–153.
Lee, J.L., Everitt, B.J. and Thomas, K.L. (2004) Independent
cellular processes for hippocampal memory consolidation
and reconsolidation. Science., 304(5672): 839–843.
Lee, P.R., Cohen, J.E., Becker, K.G. and Fields, R.D. (2005b)
Gene expression in the conversion of early-phase to late-
phase long-term potentiation. Ann. N.Y. Acad. Sci., 1048:
259–271.
Li, L., Carter, J., Gao, X., Whitehead, J. and Tourtellotte,
W.G. (2005) The neuroplasticity-associated arc gene is a di-
rect transcriptional target of early growth response (Egr)
transcription factors. Mol. Cell Biol., 25(23): 10286–10300.
Lin, Y.W., Yang, H.W., Wang, H.J., Gong, C.L., Chiu, T.H.
and Min, M.Y. (2006) Spike-timing-dependent plasticity at
resting and conditioned lateral perforant path synapses on
granule cells in the dentate gyrus: different roles of N-methyl-
D-aspartate and group I metabotropic glutamate receptors.
Eur. J. Neurosci., 23(9): 2362–2374.
Ling, D.S., Benardo, L.S. and Sacktor, T.C. (2006) Protein
kinase Mzeta enhances excitatory synaptic transmission by
increasing the number of active postsynaptic AMPA recep-
tors. Hippocampus, 16(5): 443–452.
Ling, D.S., Benardo, L.S., Serrano, P.A., Blace, N., Kelly,
M.T., Crary, J.F. and Sacktor, T.C. (2002) Protein kinase
Mzeta is necessary and sufficient for LTP maintenance. Nat.
Neurosci., 5(4): 295–296.
Link, W., Konietzko, U., Kauselmann, G., Krug, M., Schwa-
nke, B., Frey, U. and Kuhl, D. (1995) Somatodendritic ex-
pression of an immediate early gene is regulated by synaptic
activity. Proc. Natl. Acad. Sci. U.S.A., 92(12): 5734–5738.
Lisman, J.E. and Zhabotinsky, A.M. (2001) A model of synap-
tic memory: a CaMKII/PP1 switch that potentiates trans-
mission by organizing an AMPA receptor anchoring
assembly. Neuron, 31(2): 191–201.
Lyford, G.L., Yamagata, K., Kaufmann, W.E., Barnes, C.A.,
Sanders, L.K., Copeland, N.G., Gilbert, D.J., Jenkins, N.A.,
Lanahan, A.A. and Worley, P.F. (1995) Arc, a growth factor
and activity-regulated gene, encodes a novel cytoskeleton-
associated protein that is enriched in neuronal dendrites.
Neuron, 14(2): 433–445.
Madison, D.V. and Nicoll, R.A. (1988) Enkephalin hyperpo-
larizes interneurones in the rat hippocampus. J. Physiol., 398:
123–130.
Matsuzaki, M., Honkura, N., Ellis-Davies, G.C. and Kasai, H.
(2004) Structural basis of long-term potentiation in single
dendritic spines. Nature, 429(6993): 761–766.
Meng, Y., Zhang, Y., Tregoubov, V., Janus, C., Cruz, L.,
Jackson, M., Lu, W.Y., MacDonald, J.F., Wang, J.Y., Falls,
D.L. and Jia, Z. (2002) Abnormal spine morphology and
enhanced LTP in LIMK-1 knockout mice. Neuron, 35(1):
121–133.
Messaoudi, E., Bardsen, K., Srebro, B. and Bramham, C.R.
(1998) Acute intrahippocampal infusion of BDNF induces
lasting potentiation of synaptic transmission in the rat dent-
ate gyrus. J. Neurophysiol., 79(1): 496–499.
Messaoudi, E., Kanhema, T., Soule, J., Tiron, A., Dagyte, G.,
da Silva, B. and Bramham, C.R. (2007) Sustained Arc
synthesis controls LTP consolidation through regulation
of local actin polymerization in the dentate gyrus in vivo.
Submitted.
Messaoudi, E., Ying, S.W., Kanhema, T., Croll, S.D. and
Bramham, C.R. (2002) BDNF triggers transcription-depend-
ent, late phase LTP in vivo. J. Neurosci., 22: 7453–7461.
Mezey, S., Doyere, V., De Souza, I., Harrison, E., Cambon, K.,
Kendal, C.E., Davies, H., Laroche, S. and Stewart, M.G.
(2004) Long-term synaptic morphometry changes after in-
duction of long-term potentiation and long-term depression
in the dentate gyrus of awake rats are not simply mirror
phenomena. Eur. J. Neurosci., 19(8): 2310–2318.

Milner, T.A. and Drake, C.T. (2001) Ultrastructural evidence
for presynaptic mu opioid receptor modulation of synaptic
plasticity in NMDA-receptor-containing dendrites in the
dentate gyrus. Brain Res. Bull., 54(2): 131–140.
Min, M.Y., Asztely, F., Kokaia, M. and Kullmann, D.M.
(1998) Long-term potentiation and dual-component quantal
signaling in the dentate gyrus. Proc. Natl. Acad. Sci. U.S.A.,
95(8): 4702–4707.
Moga, D.E., Calhoun, M.E., Chowdhury, A., Worley, P.,
Morrison, J.H. and Shapiro, M.L. (2004) Activity-regulated
cytoskeletal-associated protein is localized to recently acti-
vated excitatory synapses. Neuroscience, 125(1): 7–11.
Monteggia, L.M., Barrot, M., Powell, C.M., Berton, O., Ga-
lanis, V., Gemelli, T., Meuth, S., Nagy, A., Greene, R.W. and
Nestler, E.J. (2004) Essential role of brain-derived neurotro-
phic factor in adult hippocampal function. Proc. Natl. Acad.
Sci. U.S.A., 101(29): 10827–10832.
Nagy, V., Bozdagi, O., Matynia, A., Balcerzyk, M., Okulski, P.,
Dzwonek, J., Costa, R.M., Silva, A.J., Kaczmarek, L. and
Huntley, G.W. (2006) Matrix metalloproteinase-9 is required
for hippocampal late-phase long-term potentiation and mem-
ory. J. Neurosci., 26(7): 1923–1934.
Nestler, E.J., Barrot, M., DiLeone, R.J., Eisch, A.J., Gold, S.J.
and Monteggia, L.M. (2002) Neurobiology of depression.
Neuron, 34(1): 13–25.
Nguyen, P.V. and Kandel, E.R. (1996) A macromolecular syn-
thesis-dependent late phase of long-term potentiation requir-
ing cAMP in the medial perforant pathway of rat
hippocampal slices. J. Neurosci., 16(10): 3189–3198.
Oertner, T.G. and Matus, A. (2005) Calcium regulation of actin
dynamics in dendritic spines. Cell Calcium, 37(5): 477–482.
Okamoto, K., Nagai, T., Miyawaki, A. and Hayashi, Y. (2004)
Rapid and persistent modulation of actin dynamics regulates
postsynaptic reorganization underlying bidirectional plastic-
ity. Nat. Neurosci., 7(10): 1104–1112.
Ostroff, L.E., Fiala, J.C., Allwardt, B. and Harris, K.M. (2002)
Polyribosomes redistribute from dendritic shafts into spines
with enlarged synapses during LTP in developing rat hippo-
campal slices. Neuron, 35(3): 535–545.
Palop, J.J., Chin, J., Bien-Ly, N., Massaro, C., Yeung, B.Z.,
Yu, G.Q. and Mucke, L. (2005) Vulnerability of dentate
granule cells to disruption of arc expression in human am-
yloid precursor protein transgenic mice. J. Neurosci., 25(42):
9686–9693.
Pastalkova, E., Serrano, P., Pinkhasova, D., Wallace, E., Fen-
ton, A.A. and Sacktor, T.C. (2006) Storage of spatial infor-
mation by the maintenance mechanism of LTP. Science,
313(5790): 1141–1144.
Patterson, S.L., Grover, L.M., Schwartzkroin, P.A. and Both-
well, M. (1992) Neurotrophin expression in rat hippocampal
slices: a stimulus paradigm inducing LTP in CA1 evokes in-
creases in BDNF and NT-3 mRNAs. Neuron, 9(6):
1081–1088.
Pinkstaff, J.K., Chappell, S.A., Mauro, V.P., Edelman, G.M.
and Krushel, L.A. (2001) Internal initiation of translation of
five dendritically localized neuronal mRNAs. Proc. Natl.
Acad. Sci. U.S.A., 98(5): 2770–2775.
469
Plath, N., Ohana, O., Dammermann, B., Errington, M.L.,
Schmitz, D., Gross, C., Mao, X., Engelsberg, A., Mahlke, C.,
Welzl, H., Kobalz, U., Stawrakakis, A., Fernandez, E.,
Waltereit, R., Bick-Sander, A., Therstappen, E., Cooke, S.F.,
Blanquet, V., Wurst, W., Salmen, B., Bosl, M.R., Lipp, H.P.,
Grant, S.G., Bliss, T.V., Wolfer, D.P. and Kuhl, D. (2006)
Arc/Arg3.1 is essential for the consolidation of synaptic
plasticity and memories. Neuron, 52(3): 437–444.
Ramirez-Amaya, V., Vazdarjanova, A., Mikhael, D., Rosi, S.,
Worley, P.F. and Barnes, C.A. (2005) Spatial exploration-
induced Arc mRNA and protein expression: evidence for se-
lective, network-specific reactivation. J. Neurosci., 25(7):
1761–1768.
Richter, J.D. and Sonenberg, N. (2005) Regulation of cap-de-
pendent translation by eIF4E inhibitory proteins. Nature,
433(7025): 477–480.
Rodriguez, J.J., Davies, H.A., Silva, A.T., De Souza, I.E.,
Peddie, C.J., Colyer, F.M., Lancashire, C.L., Fine, A., Err-
ington, M.L., Bliss, T.V. and Stewart, M.G. (2005) Long-
term potentiation in the rat dentate gyrus is associated with
enhanced Arc/Arg3.1 protein expression in spines, dendrites
and glia. Eur. J. Neurosci., 21(9): 2384–2396.
Rosenblum, K., Futter, M., Voss, K., Erent, M., Skehel, P.A.,
French, P., Obosi, L., Jones, M.W. and Bliss, T.V. (2002) The
role of extracellular regulated kinases I/II in late-phase long-
term potentiation. J. Neurosci., 22(13): 5432–5441.
Routtenberg, A. and Rekart, J.L. (2005) Post-translational
protein modification as the substrate for long-lasting mem-
ory. Trends Neurosci., 28(1): 12–19.
Rowan, M.J., Klyubin, I., Cullen, W.K. and Anwyl, R. (2003)
Synaptic plasticity in animal models of early Alzheimer’s
disease. Philos. Trans. R. Soc. Lond. B Biol. Sci., 358(1432):
821–828.
Sairanen, M., Lucas, G., Ernfors, P., Castren, M. and Castren,
E. (2005) Brain-derived neurotrophic factor and antidepres-
sant drugs have different but coordinated effects on neuronal
turnover, proliferation, and survival in the adult dentate
gyrus. J. Neurosci., 25(5): 1089–1094.
Sanes, J.R. and Lichtman, J.W. (1999) Can molecules explain
long-term potentiation? Nat. Neurosci., 2(7): 597–604.
Santi, S., Cappello, S., Riccio, M., Bergami, M., Aicardi, G.,
Schenk, U., Matteoli, M. and Canossa, M. (2005)
Hippocampal neurons recycle BDNF for activity-dependent
secretion and LTP maintenance. EMBO J., 25(18):
4372–4380.
Scharfman, H., Goodman, J., Macleod, A., Phani, S.,
Antonelli, C. and Croll, S. (2005) Increased neurogenesis
and the ectopic granule cells after intrahippocampal BDNF
infusion in adult rats. Exp. Neurol., 192(2): 348–356.
Scheetz, A.J., Nairn, A.C. and Constantine-Paton, M. (2000)
NMDA receptor-mediated control of protein synthesis at
developing synapses. Nat. Neurosci., 3(3): 211–216.
Schratt, G.M., Nigh, E.A., Chen, W.G., Hu, L. and Greenberg,
M.E. (2004) BDNF regulates the translation of a select group
of mRNAs by a mammalian target of rapamycin-phospha-
tidylinositol 3-kinase-dependent pathway during neuronal
development. J. Neurosci., 24(33): 7366–7377.
470
Schratt, G.M., Tuebing, F., Nigh, E.A., Kane, C.G., Sabatini,
M.E., Kiebler, M. and Greenberg, M.E. (2006) A brain-
specific microRNA regulates dendritic spine development.
Nature, 439(7074): 283–289.
Serrano, P., Yao, Y. and Sacktor, T.C. (2005) Persistent
phosphorylation by protein kinase Mzeta maintains
late-phase long-term potentiation. J. Neurosci., 25(8):
1979–1984.
Shirayama, Y., Chen, A.C., Nakagawa, S., Russell, D.S. and
Duman, R.S. (2002) Brain-derived neurotrophic factor pro-
duces antidepressant effects in behavioral models of depres-
sion. J. Neurosci., 22(8): 3251–3261.
Shimazu, K., Zhao, M., Sakata, K., Akbarian, S., Bates, B.,
Jaenisch, R. and Lu, B. (2006) NT-3 facilitates hippocampal
plasticity and learning and memory by regulating neurogen-
esis. Learn. Mem., 13(3): 307–315.
Smart, F.M., Edelman, G.M. and Vanderklish, P.W. (2003)
BDNF induces translocation of initiation factor 4E to
mRNA granules: evidence for a role of synaptic microfila-
ments and integrins. Proc. Natl. Acad. Sci. U.S.A., 100(24):
14403–14408.
Smith, W.B., Starck, S.R., Roberts, R.W. and Schuman, E.M.
(2005) Dopaminergic stimulation of local protein synthe-
sis enhances surface expression of GluR1 and synaptic
transmission in hippocampal neurons. Neuron, 45(5):
765–779.
Soule, J., Messaoudi, E. and Bramham, C.R. (2006) Brain-
derived neurotrophic factor and control of synaptic consol-
idation in the adult brain. Biochem. Soc. Trans., 34(Pt 4):
600–604.
Stanton, P.K. and Sarvey, J.M. (1985a) Blockade of nor-
epinephrine-induced long-lasting potentiation in the hippo-
campal dentate gyrus by an inhibitor of protein synthesis.
Brain Res., 361(1–2): 276–283.
Stanton, P.K. and Sarvey, J.M. (1985b) Depletion of
norepinephrine, but not serotonin, reduces long- term po-
tentiation in the dentate gyrus of rat hippocampal slices.
J. Neurosci., 5(8): 2169–2176.
Steward, O., Wallace, C.S., Lyford, G.L. and Worley, P.F.
(1998) Synaptic activation causes the mRNA for the IEG Arc
to localize selectively near activated postsynaptic sites on
dendrites. Neuron, 21(4): 741–751.
Steward, O. and Worley, P.F. (2001) A cellular mechanism
for targeting newly synthesized mRNAs to synaptic sites
on dendrites. Proc. Natl. Acad. Sci. U.S.A., 98(13):
7062–7068.
Straube, T. and Frey, J.U. (2003) Involvement of beta-ad-
renergic receptors in protein synthesis-dependent late long-
term potentiation (LTP) in the dentate gyrus of freely moving
rats: the critical role of the LTP induction strength. Neuro-
science, 119(2): 473–479.
Straube, T., Korz, V., Balschun, D. and Frey, J.U. (2003) Re-
quirement of beta-adrenergic receptor activation and protein
synthesis for LTP-reinforcement by novelty in rat dentate
gyrus. J. Physiol., 552(Pt 3): 953–960.
Swanson-Park, J.L., Coussens, C.M., Mason-Parker, S.E.,
Raymond, C.R., Hargreaves, E.L., Dragunow, M., Cohen,
A.S. and Abraham, W.C. (1999) A double dissociation
within the hippocampus of dopamine D1/D5 receptor and
beta-adrenergic receptor contributions to the persistence of
long-term potentiation. Neuroscience, 92(2): 485–497.
Takei, N., Inamura, N., Kawamura, M., Namba, H., Hara, K.,
Yonezawa, K. and Nawa, H. (2004) Brain-derived neurotro-
phic factor induces mammalian target of rapamycin-depend-
ent local activation of translation machinery and protein
synthesis in neuronal dendrites. J. Neurosci., 24(44):
9760–9769.
Takei, N., Kawamura, M., Hara, K., Yonezawa, K. and Nawa,
H. (2001) Brain-derived neurotrophic factor enhances neu-
ronal translation by activating multiple initiation processes:
comparison with the effects of insulin. J. Biol. Chem.,
276(46): 42818–42825.
Tang, S.J., Reis, G., Kang, H., Gingras, A.C., Sonenberg, N.
and Schuman, E.M. (2002) A rapamycin-sensitive signaling
pathway contributes to long-term synaptic plasticity in the
hippocampus. Proc. Natl. Acad. Sci. U.S.A., 99(1): 467–472.
Tao, X., Finkbeiner, S., Arnold, D.B., Shaywitz, A.J. and
Greenberg, M.E. (1998) Ca2+ influx regulates BDNF tran-
scription by a CREB family transcription factor-dependent
mechanism. Neuron, 20(4): 709–726.
Tashiro, A., Sandler, V.M., Toni, N., Zhao, C. and Gage, F.H.
(2006) NMDA-receptor-mediated, cell-specific integration of
new neurons in adult dentate gyrus. Nature, 442(7105):
929–933.
Tsankova, N.M., Berton, O., Renthal, W., Kumar, A., Neve,
R.L. and Nestler, E.J. (2006) Sustained hippocampal chro-
matin regulation in a mouse model of depression and anti-
depressant action. Nat. Neurosci., 9(4): 519–525.
Vazdarjanova, A., Ramirez-Amaya, V., Insel, N., Plummer,
T.K., Rosi, S., Chowdhury, S., Mikhael, D., Worley, P.F.,
Guzowski, J.F. and Barnes, C.A. (2006) Spatial exploration
induces ARC, a plasticity-related immediate-early gene, only
in calcium/calmodulin-dependent protein kinase II-positive
principal excitatory and inhibitory neurons of the rat fore-
brain. J. Comp. Neurol., 498(3): 317–329.
Villarreal, D.M., Do, V., Haddad, E. and Derrick, B.E. (2002)
NMDA receptor antagonists sustain LTP and spatial mem-
ory: active processes mediate LTP decay. Nat. Neurosci.,
5(1): 48–52.
Vo, N., Klein, M.F., Varlamova, O., Keller, D.M., Yamamoto,
T., Goodman, R.H. and Impey, S. (2005) A cAMP-response
element binding protein-induced microRNA regulates neu-
ronal morphogenesis. Proc. Natl. Acad. Sci. U.S.A., 102(45):
16426–16431.
Walling, S.G. and Harley, C.W. (2004) Locus ceruleus activa-
tion initiates delayed synaptic potentiation of perforant path
input to the dentate gyrus in awake rats: a novel beta-ad-
renergic- and protein synthesis-dependent mammalian plas-
ticity mechanism. J. Neurosci., 24(3): 598–604.
Wang, D.C., Chen, S.S., Lee, Y.C. and Chen, T.J. (2006) Am-
yloid-beta at sublethal level impairs BDNF-induced arc ex-
pression in cortical neurons. Neurosci. Lett., 398(1–2): 78–82.
Weeks, A.C., Ivanco, T.L., LeBoutillier, J.C., Racine, R.J. and
Petit, T.L. (2001) Sequential changes in the synaptic

structural profile following long-term potentiation in the rat
dentate gyrus: III. Long-term maintenance phase. Synapse,
40(1): 74–84.
Wibrand, K., Messaoudi, E., Havik, B., Steenslid, V., Løvlie,
R., Steen, V.M. and Bramham, C.R. (2006) Identification of
genes co-upregulated with Arc during BDNF-induced long-
term potentiation in the adult rat dentate gyrus in vivo. Eur.
J. Neurosci., 23: 1501–1511.
Wu, K., Xu, J.L., Suen, P.C., Levine, E., Huang, Y.Y., Mount,
H.T., Lin, S.Y. and Black, I.B. (1996) Functional trkB ne-
urotrophin receptors are intrinsic components of the adult
brain postsynaptic density. Brain Res. Mol. Brain Res.,
43(1–2): 286–290.
Xie, C.W. and Lewis, D.V. (1991) Opioid-mediated facilitation
of long-term potentiation at the lateral perforant path-dent-
ate granule cell synapse. J. Pharmacol. Exp. Ther., 256(1):
289–296.
Yin, Y., Edelman, G.M. and Vanderklish, P.W. (2002) The
brain-derived neurotrophic factor enhances synthesis of Arc
471
in synaptoneurosomes. Proc. Natl. Acad. Sci. U.S.A., 99(4):
2368–2373.
Ying, S.W., Futter, M., Rosenblum, K., Webber, M.J., Hunt,
S.P., Bliss, T.V. and Bramham, C.R. (2002) Brain-derived
neurotrophic factor induces long-term potentiation in intact
adult hippocampus: requirement for ERK activation coupled
to CREB and upregulation of Arc synthesis. J. Neurosci.,
22(5): 1532–1540.
Young, J.Z., Isiegas, C., Abel, T. and Nguyen, P.V. (2006)
Metaplasticity of the late-phase of long-term potentiation: a
critical role for protein kinase A in synaptic tagging. Eur. J.
Neurosci., 23(7): 1784–1794.
Zhou, Q., Xiao, M. and Nicoll, R.A. (2001) Contribution of
cytoskeleton to the internalization of AMPA receptors. Proc.
Natl. Acad. Sci. U.S.A., 98(3): 1261–1266.
Zito, K., Knott, G., Shepherd, G.M., Shenolikar, S. and
Svoboda, K. (2004) Induction of spine growth and synapse
formation by regulation of the spine actin cytoskeleton.
Neuron, 44(2): 321–334.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 26

Comparison of cellular mechanisms of long-term

depression of synaptic strength at perforant
path–granule cell and Schaffer collateral–CA1
synapses

Beatrice Pöschel and Patric K. Stanton

Department of Cell Biology and Anatomy, New York Medical College, Valhalla, NY 10595, USA

Abstract: This chapter compares the cellular mechanisms that have been implicated in the induction and
expression of long-term depression (LTD) at Schaffer collateral–CA1 synapses to perforant path-dentate
gyrus (DG) synapses. In general, Schaffer collateral LTD and long-term potentiation (LTP) both appear to
be a complex combination of many alterations in synaptic transmission that occur at both presynaptic and
postsynaptic sites, while at perforant path synapses, most evidence has focused on postsynaptic long-term
alterations. Within the DG, the medial perforant path is far more studied than lateral perforant path
synapses, where most evidence relates to the induction of heterosynaptic LTD at lateral perforant path
synapses when LTP is induced in the medial perforant path. Of course, there remain many other classes of
synapses in the DG where synaptic plasticity, including LTD, have been largely neglected. It is clear that a
better understanding of the range of DG loci where long-lasting activity-dependent plasticity, both LTD and
LTP, are expressed will be essential to improve our understanding of the cognitive roles of such DG plasticity.

Keywords: CA1; dentate gyrus; hippocampus; long-term depression; perforant path; plasticity; Schaffer
collateral

Hippocampal long-term depression (LTD)
LTD is defined as a long lasting (hours to weeks),
usually synapse specific, decrease of the strength of
synaptic transmission. It appears to be the coun-
terpart of long-term potentiation (LTP), a long-
lasting increase in synaptic strength. LTD can be
induced at synapses in the hippocampal subre-
gions CA1, CA3, and the dentate gyrus (DG)
(Dudek and Bear, 1992; O’Mara et al., 1995a;
Corresponding author. Tel.: +001-914-594-4883;
Fax: +001-914-594-4653; E-mail: patric_stanton@nymc.edu
Kobayashi et al., 1996), as well as many other
synapses and regions. The cellular mechanisms for
induction and expression of LTD can vary
depending on the stimulation protocol, brain sub-
region, and stage of development (Domenici et al.,
1998; Kemp et al., 2000; Malenka and Bear, 2004).
The majority of studies elucidating the cellular
mechanisms of hippocampal LTD have been per-
formed in area CA1. Studies of LTD in area CA1
will be reviewed to develop a model of how LTD
can be induced and expressed, to be used as
a framework to compare LTD mechanisms in CA1
with those in the DG.

DOI: 10.1016/S0079-6123(07)63026-X 473

474
LTD in area CA1 (Schaffer collateral– CA1
synapses)
LTD can be divided into two phases: induction
and expression. The induction phase involves cel-
lular biochemical events which are triggered by
particular patterns of synaptic activity within a
timeframe of minutes, and which initiate longer-
lasting processes leading to the expression of LTD.
Induction
LTD can be induced by different means, including
both electrical stimulation and pharmacological
manipulations. While it is often mistakenly as-
sumed that these forms of LTD are similar mech-
anistically within and across brain regions, evidence
is substantial that multiple forms of LTD with dis-
tinct induction and expression mechanisms coexist
within single synapses, and that the profiles of
expression of these forms can differ greatly between
synapses.
Most commonly, LTD is induced by a prolonged
(10–15 min) train of low-frequency (1–2 Hz) stimu-
lation (LFS; Dudek and Bear, 1992). This form of
LTD is called homosynaptic because only the stim-
ulated pathway exhibits a long-term change in
synaptic strength. Heterosynaptic LTD involves
interactions between separate synapses or groups
of synapses that converge on the same neuron.
In CA1, heterosynaptic LTD can be induced in a
non-stimulated pathway by the delivery of LTP-
inducing stimuli to a parallel pathway that con-
verges on the same neuron (Lynch et al., 1977). One
computationally interesting form of heterosynaptic
LTD is associative LTD (Stanton and Sejnowski,
1989). For associative forms of synaptic plasticity,
the relative timing of activity in two (or more) input
pathways is a key determinant of the direction of
change in synaptic strength, i.e., whether LTP or
LTD is induced. During induction of associative
LTD, the activity of a test input is negatively cor-
related in time with activity of a conditioning train
of bursts applied to a separate pathway, which
leads to LTD in the test input. That is, if a series
of brief high-frequency bursts of stimuli are applied
to the conditioning input, and the test input is
stimulated between the bursts, then the test input
expresses LTD. In a closely-related form of LTD
called spike-timing LTD, repeated application of
single presynaptic stimuli that closely follow single
postsynaptic action potentials (APs) elicits LTD,
while reversing the pairing (i.e., pre-AP) elicits LTP
(Magee and Johnston, 1997; Markram et al., 1997).
Induction of LTD by pharmacological agents al-
lows for selective targeting of particular receptors
that activate different induction pathways. Thus,
particular signal cascades can be investigated with
regard to their contribution to the induction of po-
tentially separable forms of LTD that might never-
theless share similar expression mechanisms. In area
CA1, LTD can be induced by agonist activation of
a number of transmitter receptors (metabotropic
glutamate receptors (mGluRs): Manahan-Vaughan
and Reymann, 1995; N-methyl-D-aspartate recep-
tors (NMDAR): Lee et al., 1998; Kamal et al., 1999)
or activation/inactivation of intracellular messen-
gers (i.e., simultaneous activation of protein kinase
G (PKG) and inhibition of protein kinase A (PKA):
Santschi et al., 1999; Bailey et al., 2003).
The intracellular induction mechanisms of LTD
vary depending on the brain region, developmental
stage and triggering mechanism (Malenka and
Bear, 2004). Different forms of LTD can coexist
in one region, as in area CA1 (Kemp and Bashir,
1997; Manahan-Vaughan, 1997; Oliet et al., 1997;
Nicoll et al., 1998), as well as other brain regions.
The main forms of LTD at Schaffer collateral–CA1
synapses are NMDAR dependent and mGluR-
dependent LTD, which will be discussed below.
After LTD-inducing synaptic stimulation,
NMDAR- and mGluR-dependent LTD are in-
duced simultaneously in most cases. They can be
separated by selective pharmacological blockade of
NMDARs or mGluRs, respectively, or selectively
induced by agonist application.
NMDAR/mGluR-dependent LTD
During initial investigations of LTD in CA1, LTD
was elicited via a protocol of LFS trains (Dudek
and Bear, 1992). It turned out that this stimulation
induced a mixture of NMDAR- and mGluR-
dependent LTDs. Thus, studies of the intracellular
 475

Fig. 1. Cascades involved with NMDAR/mGluR-dependent LTD at Schaffer collateral–CA1 synapses. AA: arachidonic acid, AC:
adenylyl cyclase, Ca: Ca2+, cADPR: cyclic adenosine diphosphate ribose, CaM: calmodulin, CaMKII: calmodulin-dependent protein
kinase II, cAMP: cyclic adenosine monophosphate, cGMP: cyclic guanosine monophosphate, CN: calcineurin, DAG: diacyl glycerol,
HPETE: hydroperoxyeicosatetraenoic acid, I1: inhibitor 1, IP3: inositol-1,4,5-trisphosphate, IEG: immediate early genes, MAPK:
mitogen-activated protein kinase, mGluR: metabotropic glutamate receptor, NO: nitric oxide, NOS: NO-synthetase, P: phosphate,
PIP2: phosphatidylinositol-3,4-bisphosphate, PLA2: phospholipase A2, PKA: protein kinase A, PKC: protein kinase C, PKG: protein
kinase G, PLC: phospholipase C, Post: postsynapse, PP1: protein phosphatase 1, Pre: presynapse, PS: protein synthesis, VGCC:
voltage-gated Ca2+ channels.

mechanisms of stimulus-induced LTD did not dis-
tinguish between NMDAR- and mGluR-depend-
ent pathways. Later it was found that selective
activation of NMDARs and mGluRs can induce
LTD independently, and that NMDAR- and
mGluR-dependent LTD pathways could be stud-
ied independently. These studies will be reviewed
below (Fig. 1).
Induction. The first step in the induction of LTD
is the activation of glutamate receptors and/or
voltage-gated Ca2+ channels (VGCCs). NMDARs
have been a primary focus of attention, because
their requirement for membrane depolarization to
relieve Mg2+-dependent channel block plus
glutamate to activate the receptor allows them to
function as coincidence detectors of presynaptic
transmitter release and postsynaptic depolarization
(Cotman et al., 1988), and because they gate the
influx of Ca2+, they are the key factors in estab-
lishing associative long-term changes (both LTP
and LTD) of synaptic strength. It has been shown
that application of NMDAR antagonists block the
induction of LTD by prolonged LFS (1 Hz/15 min;
Dudek and Bear, 1992; Mulkey and Malenka,
1992).
Other studies have revealed that activation of
mGluRs (Oliet et al., 1997; Manahan-Vaughan,
1997) and VGCCs (Christie et al., 1995; Cummings
et al., 1996; Oliet et al., 1997) can also be necessary
476
for the induction of LTD in CA1, and the depend-
ence on these receptors varies with the precise
stimulus pattern and level of postsynaptic depo-
larization. Activation of NMDARs, mGluRs, or
VGCCs each lead to elevations in intracellular
[Ca2+]. Ca2+ is a key ion that triggers a variety of
kinases and phosphatases, some mediating intra-
cellular events, which lead to both short- and long-
term changes in synaptic strength. During the
induction of LTD, the rise in [Ca2+] is lower than
that occurring during induction of LTP (Hansel et
al., 1996; Yang et al., 1999), which is thought to
lead to a preferred activation of phosphatases with
lower Kd’s for activation than most Ca2+-depend-
ent kinases (Lisman, 1989). Besides phosphatases
(Mulkey et al., 1993; Thiels et al., 1998), LTD has
been shown to also depend on multiple kinases.
It has been reported that protein kinase C (PKC)
(Stanton, 1995; Hrabetova and Sacktor, 1996),
PKG (Reyes-Harde et al., 1999a, b), PKA (Bran-
don et al., 1995), Ca2+/calmodulin-dependent pro-
tein kinase II (CaMKII) (Stevens et al., 1994;
Stanton and Gage, 1996), and mitogen-activated
protein kinase (MAPK) (Thiels et al., 2002;
Gallagher et al., 2004) can all play roles in the in-
duction of LTD in area CA1. In some cases, these
roles appear to be played presynaptically, and in
others postsynaptically, reinforcing the notion of
multiple forms of LTD.
Expression. Phosphatases and kinases target sev-
eral effector molecules that could cause changes
in synaptic strength via effects on a wide range of
protein mechanisms, such as receptor number and
sensitivity, ion-channel open probability and con-
ductance, cytoskeletal machinery, mRNA and pro-
tein synthesis (PS), and vesicle-release probability.
(a) Phosphatase pathways
Phosphatases whose activation has been sug-
gested to be involved in the induction of LTD are
protein phosphatase 2A (PP2A), calcineurin (CN;
also, PP2B), and PP1. Inhibition of these phospha-
tases with the PP1/PP2A inhibitors okadaic acid
and calyculin A (Mulkey et al., 1993) and CN
inhibitors FK506 and cyclosporin A (Mulkey et al.,
1994) have all been demonstrated to block the
induction of LTD. Additionally, it has been shown
that the induction of LTD in CA1 in vivo is
associated with an increase in PP1 activity lasting
35 min after LTD induction, as well as PP2A ac-
tivity increases lasting 65 min (Thiels et al., 1998,
2000). CN is a phosphatase that is dependent
on Ca2+ levels and is activated by Ca2+/CaM
(Klee et al., 1979).
Amongst its effects, CN has been demonstrated
to dephosphorylate synapsin I (Hosaka et al., 1999),
resulting in a gradual, activity-dependent reduction
of neurotransmitter release, since vesicles are held in
the reserve pool by dephoshorylated synapsin and
are prevented from joining the readily releasa
ble vesicle pool (RRP) (Hilfiker et al., 1999). CN
also regulates the internalization of both alpha-
amino-3-hydroxy-5-methyl-4-isoxazoleproprionate
(AMPA) receptors (AMPARs) (Beattie et al., 2000;
Lin et al., 2000) and NMDARs (Shi et al., 2000).
CN has been reported to contribute to downregu-
lation of Ca2+ signaling by reducing both Ca2+
influx and Ca2+-induced Ca2+ release from intra-
cellular stores via dephosphorylation and inactiva-
tion of VGCCs (L-, N-, P/Q-, or R-type;
Armstrong, 1989) Additionally, it has been shown
that NMDA-induced reduction in dendritic spine
density in cultured hippocampal neurons is blocked
by pre-treatment with CN inhibitors and that this is
probably via preventing the CN-mediated promo-
tion of actin depolymerization (Halpain et al., 1998).
CN asserts its effects through disinhibition of
PP1 (postsynaptic serine/threonine PP1), so that
PP1 is only indirectly regulated by [Ca2+]. When
activated, CN dephosphorylates and inactivates
inhibitor-1 (Cohen, 1989; Mulkey et al., 1994),
which is normally bound to PP1. Inactivation of
inhibitor-1 releases and activates PP1. PP1 is also
regulated by PKA, which phosphorylates and in-
activates inhibitor-1 and thus inhibits PP1 (Cohen,
1989). PP1 regulates gene expression via the de-
phosphorylation and inactivation of the transcrip-
tion factor cyclic adenosine monophosphate
(cAMP) responsive element-binding protein
(CREB; Hagiwara et al., 1992; Bito et al., 1996),
which in its active form, initiates the transcription
of genes that contain a cAMP response element
(CRE) within their promoter sequence (Brindle
and Montminy, 1992). Another substrate of PP1
(and/or PP2A) is AMPARs (Lee et al., 2000)
which are dephosphorylated at the PKA site

(SER-845) (Lee et al., 1998, 2000), resulting in a
decrease in their open probability (Banke et al.,
2000). Like CN, PP1 has been implicated in activ-
ity-dependent endocytotic internalization of AM-
PARs, since CN and PP1 antagonists each inhibit
NMDA- and insulin-induced endocytosis of AM-
PARs (Beattie et al., 2000; Ehlers, 2000; Lin et al.,
2000).
PP2A, as well as PP1, have both been reported
to dephosphorylate the autophosphorylation site
Ser-657/660 and the trans-phosphorylation site
Ser-641 on the catalytic domain of PKC-a/PKC-
bII (Keranen et al., 1995; Thiels et al., 2000). PP1
can also dephosphorylate the autophosphorylat-
ion site Ser-641 on the catalytic domain of PKC-
bII (Keranen et al., 1995).
However, the ability of phosphatases to dephos-
phorylate substrates in vitro or in vivo does not
confirm that this dephosphorylation occurs during
the induction of LTD. For instance, PP1 and
PP2A have been shown to dephosphorylate
Thr286 of CaMKII in vitro (Shields et al., 1985;
Strack et al., 1997) but hippocampal LTD can still
be induced in CaMKIIT286A mice (Krezel et al.,
1999; Parsley et al., 2007), indicating that this de-
phosphorylation event is not required for at least
some forms of LTD.
Dephosphorylation events which have been
linked to decreases in synaptic strength during
LTD in area CA1 are: AMPAR dephosphorylat-
ion of SER-845 by CN and PP1/PP2A (Lee et al.,
2000, 2003), AMPAR internalization (Carroll
et al., 1999; Heynen et al., 2000), and PP1/PP2A-
mediated dephosphorylation of PKC at Ser-657/
660 (Thiels et al., 2000). CN dephosphorylation of
synapsin I (Hosaka et al., 1999) could, by reducing
transmitter release (Hilfiker et al., 1999), contrib-
ute to presynaptic LTD. Additional substrates de-
phosphorylated by CN that might play important
roles in the long-lasting remodeling of synaptic
structure are microtubule-associated protein 2
(MAP2), protein tau and tubulin (Goto et al.,
1985), though a requirement of these events for
LTD has yet to be confirmed.
(b) Kinase pathways
Many kinases can be activated by Ca2+ and
activated or inactivated by distinct mGluR and
other G protein-coupled metabotropic pathways.
477
mGluRs are divided into three groups: I, II, and
III. Group I mGluRs are positively coupled to
adenylate cyclase and the cAMP-PKA pathway,
and they also can activate PKC. mGluRs II and
mGluRs III are negatively coupled to adenylate
cyclase. NMDARs contribute to intracellular
[Ca2+] elevation via Ca2+ influx through
NMDAR channels, followed by Ca2+-induced
Ca2+ release from ryanodine-sensitive stores
(Alford et al., 1992). Group I mGluRs promote
the release of cytosolic Ca2+ from inositol-1,4,5-
trisphosphate (IP3)-sensitive intracellular stores
via positive coupling to IP3. It has been demon-
strated that induction of LFS-LTD requires Ca2+
release from presynaptic, but not postsynaptic,
ryanodine-sensitive stores, and Ca2+ release from
postsynaptic IP3-sensitive stores (Reyes and Stanton,
1996). Intracellular Ca2+ can activate PKA via
Ca2+/CaM-sensitive adenylyl cyclase (AC) (Eliot
et al., 1989) which synthesizes cAMP, CaMKII via
Ca2+/CaM (Colbran, 1992), PKG via the Ca2+/
CaM-NOS-NO-GC-cGMP pathway (Garbers,
1990; Bredt et al., 1991), and PKC by Ca2+ paired
with diacylglycerol (DAG; Huang, 1989).
During the induction of LTD, kinases are bidi-
rectionally regulated, as are their effector path-
ways. While sometimes phosphorylated and
activated kinases are necessary to induce LTD,
in other cases kinases are dephosphorylated and
inactivated to promote LTD.
Ca2+/phospholipid-dependent protein kinase
(PKC). Reports indicate that inhibition of PKC
may be required for LTD, since a selective inhibitor
of PKC, chelerythrine, induces a depression which
shows mutual occlusion with LFS-LTD (Hrabetova
and Sacktor, 1996) suggesting similar induction
and/or expression mechanisms. Indeed, LTD is as-
sociated with a decrease in PKC activity (Hrabetova
and Sacktor, 1996; Thiels et al., 2000). This decrease
in PKC activity has been shown to be mediated by
either a proteolysis-mediated decrease in PKCg and
PKMz activity (Hrabetova and Sacktor, 1996) or by
dephosphorylation of catalytically relevant auto-
phosphorylation sites on the C terminus of PKC
(Ser-657 on PKCa; Ser-660 on PKCbII) by PPs
(Thiels et al., 2000). While the decrease in activity of
PKCa, b, and g was only transient (35–65 min,
478
Thiels et al., 2000), downregulation of the autono-
mously active PKMz (Schwartz, 1993) persisted up
to 120 min post-LFS (Hrabetova and Sacktor, 2001)
making it a special candidate for the maintenance
phase of LTD. It has been demonstrated that LTD
is associated with a decrease in pre- as well
as postsynaptic PKC substrate phosphorylation
(Ramakers et al., 2000). Since it is known that
PKC activators such as phorbol esters enhance pre-
synaptic transmission (Chaki et al., 1994), it seems
plausible that inactivation of constitutively active
PKC might lead to depression of release. On the
postsynaptic side, PKC is known to target the
extracellular signal-regulated kinase (ERK) signa-
ling cascade (Roberson et al., 1999) and could thus
affect transcription processes during LTD (see
below).
On the other hand, there are also studies sug-
gesting a requirement for activation of PKC during
LTD. PKC is known to phosphorylate SER-880 of
the AMPAR subunit GluR2 (McDonald et al.,
2001), GluR2 SER-880 phosphorylation is in-
creased after LTD induction (Kim et al., 2001),
preventing the phosphorylation of SER-880 on
GluR2 by point mutation inhibits the constitutive
synaptic incorporation of GluR2 homomers, as
well as LTD (Seidenman et al., 2003) and removal
of GluR2 containing AMPARs from synapses is
one potential expression mechanism for LTD
(Seidenman et al., 2003). Thus, it is feasible that
PKC activates a pathway that includes the phos-
phorylation of SER-880 on GluR2, followed by
the endocytosis of GluR2 containing AMPARs to
express LTD. This pathway could be implemented
by the PKC isoforms (a, bI, bII, g) that have been
reported to be transiently (o15 min) activated
after induction of LTD (Hrabetova and Sacktor,
2001). Overall, multiple roles for both presynaptic
and postsynaptic isoforms of PKC make the func-
tions of these enzymes in induction and mainte-
nance of LTD potentially quite complex.
Ca2+/calmodulin-dependent protein kinase (CaM-
KII). It has been reported that mice that lack the
gene for alpha CaMKII show reduced LTD
(Stevens et al., 1994). Furthermore, induction of
LFS-LTD is blocked by extracellular application
of the CaMKII inhibitor KN-62 to hippocampal
slices, but not by postsynaptic intracellular infu-
sion of KN-62 into single CA1 pyramidal neurons,
indicating that the CaMKII that may be necessary
for LTD induction is presynaptically localized
(Stanton and Gage, 1996).
cAMP-dependent protein kinase (PKA). There
are contradictory studies concerning the role of
PKA in the induction of LTD. Brandon et al.
(1995) reported that the inhibition of PKA strongly
inhibits LFS-induced LTD at mouse Schaffer col-
lateral–CA1 synapses (Brandon et al., 1995), while
studies in knockout mice have reported that LTD
in field CA1 is dramatically reduced if either the
PKA regulatory subunit RIb (Brandon et al., 1995)
or the PKA catalytic subunit Cb1 (Qi et al., 1996)
are missing. Interestingly Cb1- and RIb-containing
PKA holoenzymes show a higher sensitivity to
cAMP than Cb- and RIb-containing PKA holoen-
zymes (Cadd et al., 1990) and could be preferen-
tially activated during LTD induction when Ca2+
influx, and perhaps, associated increases in [cAMP]
may be less than those needed to elicit LTP.
A number of studies explicitly contradict the
notion that PKA activation is needed for LTD.
Santschi et al. (1999) reported that multiple syn-
thetic PKA inhibitors enhanced the magnitude of
LTD at these same synapses in rat hippocampal
slices, while Nicholls et al. (2006) found that LTD
is enhanced in a transgenic mouse model where
adenylate cyclase is constitutively inhibited by
Gi2a and PKA stimulation prevented. Studies by
Kameyama et al. (1998) similarly support a role
for inhibition of PKA in promoting LTD; they
found that inhibition of postsynaptic PKA in-
duced a depression that occluded LFS-LTD and
that LFS-LTD was inhibited by prior activation of
postsynaptic PKA. Furthermore, LTD is blocked
by the selective mGluRII antagonist EGLU, and
by the A1 adenosine receptor blocker DPCPX, in
juvenile rats (Santschi et al., 2006). Since mGluR-
IIs are negatively coupled to cAMP, it is a rea-
sonable hypothesis that they mediate LTD
by decreasing [cAMP] and PKA activity. The con-
tradictory data suggesting a positive role for PKA
activity in LTD might indicate either a develop-
mental switch and/or dependence on genus, since
Kameyama et al. (1998) and Santschi et al. (1999)

used juvenile rats (14–21 days), while Brandon
et al. (1995) and Qi et al. (1996) studied adult mice
(4–6 weeks). Distinct PKA isoforms may also play
roles in different forms of synaptic plasticity, i.e.,
LTP or LTD (Brandon et al., 1997).
cGMP-dependent protein kinase (PKG). LTD at
Schaffer collateral–CA1 synapses can also be
blocked by PKG inhibitors (Reyes-Harde et al.,
1999a, b). It is thought that PKG is activated pre-
synaptically during LTD induction via a cascade
that includes the retrograde messenger nitric oxide
(NO), activation of guanylyl cyclase, production
of guanosine 30 ,50 -cyclic monophosphate (cGMP)
and subsequent PKG activation. Although double
knockout mice lacking both the neuronal and
endothelial forms of NO synthase (nNOS and
eNOS) did not show significantly reduced LTD
(Son et al., 1996; 3 months) and LFS-LTD was not
affected by the NOS inhibitor L-NG-nitroarginine
(L-NOArg) (Cummings et al., 1994), several other
studies have demonstrated a role for NO for LFS-
LTD. Izumi and Zorumski (1993) found a block
of LFS-LTD by the NO inhibitor L-NG-mon-
omethylargine (L-NMMA) as well as L-NOArg.
Santschi et al. (1999) partially reduced the magni-
tude of LFS-LTD by application of the compet-
itive NOS inhibitor L-nitroarginine (L-NA).
Furthermore, Stanton et al. (2003) also reported
a partial block of LFS-LTD by L-NA, as well as
by the extracellular NO scavenger hemoglobin.
Additional support for the NO-cGMP-PKG path-
way comes from studies of Gage et al. (1997) and
Reyes-Harde et al. (1999a, b); Gage et al. (1997)
reported that presynaptic NO-sensitive guanylyl
cyclase (NOGC) is necessary for the induction
of LTD, since an extracellularly applied NOGC
inhibitor prevented the induction of LTD, while
selective intracellular blockade of postsynaptic
NOGC did not. Furthermore, they also showed
that the membrane-permeable cGMP analog
8-pCPTcGMP and the NO donor S-nitroso-N-
acetylpenicillamine (SNAP) potentiated the induc-
tion of LTD by a weak submaximal LFS (1 Hz/400
stimuli). Reyes-Harde et al. (1999b) demonstrated
that the potentiation of LTD by SNAP was me-
diated via the NOGC-cGMP-PKG cascade and
required release of Ca2+ from Ca2+-mediated
479
ryanodine-sensitive stores, suggesting that this signal
cascade and Ca2+ release from ryanodine-sensitive
stores mediated a component of LFS-LTD.
Since it is known that cGMP can enhance Ca2+-
release from ryanodine-sensitive stores via activa-
tion of cADP-ribose (Galione et al., 1993; Me-
szaros et al., 1993; Lee, 1997), it has been proposed
that the PKG-ADP ribosyl cyclase-cADP ribose
cascade links cGMP to Ca2+-release from ryano-
dine-sensitive stores during LTD (Reyes-Harde
et al., 1999a). Reyes-Harde et al. (1999a) verified
that cGMP does stimulate the synthesis of cADP
ribose in hippocampal slices and found that an-
tagonists of cADP ribose receptors in presynaptic
terminals block the induction of a portion of LTD.
Presynaptic vesicular release. LFS-LTD has been
shown to be associated with a long-term reduction
in neurotransmitter release (Stanton et al., 2001,
2003; Zhang et al., 2006). There is evidence that key
opposing mediators of LTD on the presynaptic site
are both PKA and PKG. In fact, it has been shown
that a chemical LTD (cLTD) can be induced by
simultaneously raising [cGMP] while inhibiting
PKA (Santschi et al., 1999), and that this is asso-
ciated with a decrease in the evoked release of the
fluorescent marker of presynaptic activity FM1–43
from isolated rat hippocampal presynaptic termi-
nals (Bailey et al., 2003). Furthermore, cLTD is
occluded by LFS-LTD (Santschi et al., 1999) and
suppresses LFS-evoked FM1–43 release (Stanton
et al., 2001), demonstrating that it mimics the signal
transduction pathways necessary for LFS-LTD,
i.e., simultaneous increase of [cGMP] and inhibi-
tion of PKA, that lead to a presynaptic decrease in
neurotransmitter release.
It has also been shown that a selective PKG
inhibitor completely prevents induction of LTD
of presynaptic FM1–43 release from the RRP
(Stanton et al., 2003). Further upstream, activa-
tion of NMDARs and group II mGluRs have both
been indicated to play a role in the induction of
presynaptic LFS-LTD. Inhibition of NMDARs
or mGluRs (group I and II, but not group III)
have been shown to block induction of LFS-LTD
(Reyes-Harde et al., 1999b; Stanton et al., 2003;
Santschi et al., 2006). The induction of LFS-LTD
of FM1–43 release from the RRP by a 1–2 Hz
480
train is largely blocked by the NMDAR antago-
nist AP-5 (Stanton et al., 2003). The competitive
NOS inhibitor L-NA completely prevented the
LFS-induced RRP release and the extracellular
NO scavenger hemoglobin partial blocked LFS-
induced RRP release (Stanton et al., 2003) sug-
gesting that NMDARs mediate presynaptic LTD
via the retrograde messenger NO. Zhang et al.
(2006) found that group II, but not group I
mGluRs, contribute to the induction of presynap-
tic LTD by a different, paired-pulse stimulation
protocol, indicating that, depending upon the par-
ticular stimulus protocol, either NMDARs or
group I mGluRs can supply the necessary postsy-
naptic [Ca2+], while any presynaptic receptors,
such as group II mGluRs, that inhibit adenylate
cyclase can supply the necessary suppression of
[cAMP] in the presynaptic terminal.
Extracellular signal-regulated kinase (ERK). ERK/
MAPK cascade has also been implicated as a key
signal transduction mechanism for coupling neuronal
cell surface receptors to plasticity-induced transcrip-
tion. It has been demonstrated that ERK activation
is necessary for the induction of LTD, since an ERK
inhibitor blocks induction of LTD (Thiels et al.,
2002). Multiple second messengers, such as cAMP,
PKA, [Ca2+], and DAG, can all control ERK signa-
ling via the small G proteins Ras and Rap1 (Grewal
et al., 1999) and thus link both NMDARs and
mGluRs to this pathway. During induction of LTD,
ERK stimulates specific transcription pathways that
differ from LTP; in LTD, ERK activation does not
result in the increased phosphorylation of the tran-
scription factor CREB, but rather an increased phos-
phorylation of the transcription factor Elk-1 (Thiels
et al., 2002).
Phospholipase A2. It has been suggested that dur-
ing LTD Ca2+ activates Ca2+-sensitive phospholi-
pase A2 (PLA2) (Clark et al., 1995) besides
phosphatases and kinases. The activation of PLA2
leads to arachidonic acid (AA) and lipoxygenase
metabolite production. Findings which suggested
involvement of the 12-1ipoxygenase pathway in
LTD are: (1) A non-selective PLA2 inhibitor,
bromophenacylbromide, significantly reduced the
extent of Schaffer collateral–CA1 LTD in slices
from juvenile (P10–15; Fitzpatrick and Baudry,
1994) and adult rats (Normandin et al., 1996), but a
more selective PLA2 inhibitor did not block induc-
tion of LTD in adults (Stanton, 1996), (2) inhibitors
of lipoxygenase pathways of AA metabolism de-
creased the magnitude of LTD in juvenile (P20–25;
Chabot et al., 1998) and adult rats (Normandin et
al., 1996). The 12-1ipoxygenase pathway may have
cellular effects compatible with LTD expression,
such as inhibition of CaMKII (Piomelli et al., 1989)
or reduction of transmitter release (Freeman et al.,
1991) and changes in synaptic transmission through
the alteration of the affinity states of AMPARs
(Chabot et al., 1998). In detail, it was shown that
12-lipoxygenase products attenuate the potassium-
induced release of endogenous glutamate from
hippocampal mossy fiber synaptosomes (Freeman
et al., 1991) and low concentrations of PLA2 caused
a decrease in AMPA binding due to a change in
receptor affinity that was blocked by lipoxygenase
inhibitors (Chabot et al., 1998). The PLA2-induced
decrease in AMPA binding seems to be dependent
on protein phosphatase dephosphorylation, as the
PP inhibitor okadaic acid resulted in a significant
reduction in the PLA2-induced decrease in AMPA
(Chabot et al., 1998). As with many biochemical
cascades, the involvement/requirement for PLA2
activation in LTD may prove to be stimulus
protocol-specific.
Maintenance. It is suggested that long-lasting
expression of LTD of synaptic strength requires
the regulated expression of proteins, some of them
with low resting concentrations. It has been dem-
onstrated that LTD, like LTP (Stanton and
Sarvey, 1984), requires ongoing PS for stable
expression. In slice cultures of 8–9 days old rats,
the induction of LFS-LTD has been reported to be
inhibited by the PS inhibitor anisomycin, as well as
the transcriptional inhibitor actinomycin D (but
see Huber et al., 2000; Kauderer and Kandel,
2000). While anisomycin blocks the late phase of
LTD when applied prior to an LFS, it does not
affect the maintenance of LTD when applied
30 min after LFS, suggesting that PS is only nec-
essary during the induction and early phases of
LTD to synthesize proteins necessary for mainte-
nance of persistent depression.

LFS-LTD in adult rats (7–8 weeks) seems to be
only dependent on PS, but not transcription, as its
induction was blocked by anisomycin (in vivo:
Manahan-Vaughan et al., 2000; in slices: Sajikumar
and Frey, 2003) but not actinomycin D (Manahan-
Vaughan et al., 2000). The blockade of LTD by
anisomycin, applied before LFS, becomes effective
at approximately 5 h after LFS, suggesting that,
while existing protein levels can support alterations
in synaptic strength lasting many hours, the main-
tenance of longer-lasting synaptic plasticity, like
long-term memory, requires ongoing PS.
NMDAR-dependent LTD
NMDAR-dependent LTD can be induced in area
CA1 via application of NMDA (Lee et al., 1998;
Kamal et al., 1999), or LFS-stimulation under
pharmacological mGluR blockade (Oliet et al.,
1997; Zhang et al., 2006) (Fig. 1).
Brief application of NMDA (3 min, 20 mM) to
hippocampal slices of young (P21–35: Lee et al.,
1998; P14: Kamal et al., 1999) or adult rats
(6 months; Kamal et al., 1999) induces NMDAR-
dependent LTD blocked by the NMDAR antago-
nist 2-amino-5-phosphonopentanoic acid (AP5).
NMDA-LTD and LFS-LTD have been shown to
occlude one another (Lee et al., 1998; Kamal et al.,
1999), suggesting shared induction and/or expres-
sion mechanisms. Indeed, in juvenile rats, NMDA-
LTD, like LFS-LTD (Lee et al., 2000), is associated
with dephosphorylation of the GluR1 subunit of
AMPARs at the PKA-site (SER-845; Lee et al.,
1998). Additionally, as for LFS-LTD (Kameyama
et al., 1998), inhibition of PKA seems to play a role
for NMDA-LTD in juvenile rats, since both
NMDA-LTD and AMPAR dephosphorylation
are inhibited by prior activation of postsynaptic
PKA (Kameyama et al., 1998). Nevertheless, in ju-
venile rats, NMDA-LTD and the NMDA-induced
dephosphorylation of AMPARs cannot be blocked
by high concentrations of inhibitors of PP1 and
PP2A (Kameyama et al., 1998), in marked contrast
to earlier findings for LFS-LTD (Mulkey et al.,
1993; Lee et al., 2000). Even though PP2B inhibitors
did not affect NMDA-induced dephosphorylation
of AMPARs, they partially inhibited NMDA-LTD
481
in juvenile rats (Kameyama et al., 1998), suggesting
that PP2B dephosphorylates other substrates, but
not AMPARs, during NMDA-LTD. Interestingly,
the role of phosphatases for NMDA-LTD seems to
be developmentally regulated. While inhibitors of
PP1/PP2A do not affect NMDA-LTD in juvenile
rats (P21–30: Kameyama et al., 1998), they have
been reported to block NMDA-LTD in adult rats
(6 months: Kamal et al., 1999). Furthermore, inhi-
bition of PP2B completely blocks NMDA-LTD in
adult rats (6 months: Kamal et al., 1999), but only
partially blocks LTD in juvenile rats (P21–30:
Kameyama et al., 1998) and is without effect on
NMDA-LTD in even younger rats (P14; Kamal
et al., 1999).
Brief application of NMDA (20–50 mM for
1–2 min) with a concentration and application
duration that is similar to those that induce
NMDA-LTD (3 min, 20 mM), causes a rapid, nearly
complete internalization of surface AMPARs in
hippocampal cultures (2–3 weeks in vitro; Beattie
et al., 2000; Ehlers, 2000), suggesting that
NMDAR-LTD, like LFS-LTD, may be expressed
by AMPAR endocytosis. Indeed, Nosyreva and
Huber (2005) confirmed that NMDA-induced LTD
(3 min, 20 mM NMDA) in CA1 of neonatal hippo-
campal slices was associated with a decrease in sur-
face expression of GluR1 and GluR2/3 at 10 min
and 60 min, respectively. Like LFS-LTD, the
NMDA-induced AMPAR internalization was
dependent on Ca2+ influx, being almost completely
inhibited by removing extracellular Ca2+ (Beattie
et al., 2000) and by the Ca2+ chelator BAPTA-AM
(Ehlers, 2000). Like LFS-LTD, NMDA-induced
AMPAR internalization was also dependent on
CN, but not PP2A, since it was inhibited by the CN
blocker FK-506, but not by low concentrations of
okadaic acid (10 nM) that selectively inhibits PP2A
(Beattie et al., 2000; Ehlers, 2000).
There are contradictory findings concerning the
role of PP1 in AMPAR endocytosis. While Beattie
et al. (2000) could not inhibit NMDA-induced
AMPAR internalization with the PP1/PP2A inhib-
itors calyculin A or okadaic acid (1 mM), Ehlers
(2000) reported an almost complete inhibition of
NMDA-induced AMPAR internalization by these
inhibitors, as well as the PP1-selective inhibitor
tautomycin. While the reason for these divergent
482
findings is unclear, there were slightly different in-
duction protocols used; Beattie et al. (2000) pre-
pared dissociated cell cultures from areas CA1/
CA3 from P0 rat pups, while Ehlers (2000) pre-
pared hippocampal mixed cell cultures from 1 to 2
days old rats, with both maintained 2–3 weeks in
vitro before commencing experiments. Beattie et al.
(2000) induced AMPAR internalization by appli-
cation of 50 mM NMDA for 1 min in the presence
of both AMPAR and mGluR antagonists, while
Ehlers (2000) applied 20 mM NMDA alone for
1–2 min, possibly suggesting an additional role for
ambient activation of mGluRs in the activation of
phosphatases that lead to AMPAR internalization.
Interestingly, the findings of Ehlers (2000) seem
to support the hypothesis that AMPAR dephos-
phorylation is a step toward AMPAR internaliza-
tion. They observed that application of NMDA,
which induces AMPAR endocytosis, additionally
induces the rapid dephosphorylation of GluR1 at
the SER-845 PKA site. Both events are temporally
correlated, as NMDA-induced AMPAR internali-
zation peaked shortly after maximal dephos-
phorylation of GluR1. Additionally, internalized
GluR1 subunits showed no phosphorylation at the
PKA site SER-845, but were still phosphorylated at
the CaMKII site SER-831. Both AMPAR end-
ocytosis and dephosphorylation of GluR1 at the
SER-845 PKA site, were prevented by Ca2+-free
medium, FK-506, high concentrations of okadaic
acid (1 mM) or tautomycin, but not prevented by
low concentrations of okadaic acid. Ehlers (2000)
suggested that NMDAR activity results in Ca2+
influx and activation of a PP cascade (including
PP2B and PP1) that triggers AMPAR dephos-
phorylation followed by endocytosis.
There is evidence that NMDAR-dependent
LTD is expressed not only postsynaptically as
AMPAR dephosphorylation and endocytosis (Lee
et al., 1998; Beattie et al., 2000; Ehlers, 2000;
Zhang et al., 2006), but presynaptically as a long-
term reduction in neurotransmitter release (Zhang
et al., 2006). Zhang et al. (2006) showed, by eval-
uating changes in paired-pulse ratio (PPR), two-
photon imaging of FM1–43 release and variance-
mean analyses, that NMDA-LTD is associated
with reduced transmitter release probability in
slices from juvenile rats (P14–19). Additionally,
they demonstrated that NOS activity is required
for this presynaptic NMDAR-dependent LTD, as
they could block it with the NOS inhibitor L-NA,
suggesting NO as a retrograde transmitter that es-
tablishes presynaptic NMDAR-dependent LTD.
Finally, it has been shown that T-type VGCCs
and activation of PKC do not appear to play a role
in NMDAR-dependent LTD (as opposed to
mGluR-dependent LTD (Oliet et al., 1997), as
the T-type VGCC channel blocker Ni2+ and the
PKC inhibitory peptide, PKC19–36, did not affect
the induction of NMDAR-dependent LTD (Oliet
et al., 1997).
mGluR-dependent LTD
It is now clear that there exist multiple forms of
mGluR-dependent LTD, depending on develop-
mental stage and mGluR subtype involved. The
mGluRs are a family of heterotrimeric guanosine
triphosphate-binding protein (G protein)-coupled
receptors (Sugiyama et al., 1987; Tanabe et al.,
1992). They are divided into three groups, mGluR
I, mGluR II, and mGluR III, which are made up
of eight subtypes. This categorization is based on
their amino acid sequence homology, agonist
selectivity, and second messenger coupling (Conn
and Pin, 1997).
mGluR-dependent LTD in neonatal rats. In neo-
natal rats electrical stimulation (5 Hz/3 min) induces
both NMDAR-independent and mGluR-dependent
LTD (P3–7: Bolshakov and Siegelbaum, 1994).
Metabotropic GluR-dependent LTD has been
reported to require activation of group I mGluRs
(but group II mGluR blockade was not tested; Oliet
et al., 1997/P6–7; Nosyreva and Huber, 2005/
P8–15). mGluR activation is only necessary for in-
duction of LTD in neonatal rats, since maintenance
is not inhibited by the mGluRI inhibitor (+)-a-
methyl-4-carboxyphenylglycine (MCPG; Bolshakov
and Siegelbaum, 1994) (Fig. 2).
mGluR-dependent LTD, which is not affected by
NMDAR blockade (Li et al., 2002), can be also
induced by 1 Hz stimulation in neonatal rats (P6–8;
Li et al., 2002; P4–10; Feinmark et al., 2003).
mGluR-LTD induced by 1 Hz in neonatal rats has
 483

Fig. 2. Cascades involved with mGluR-dependent LTD in neonatal rats at Schaffer collateral–CA1 synaspes. AA: arachidonic acid,
Ca: Ca2+, HPETE: hydroperoxyeicosatetraenoic acid, IP3: inositol-1,4,5-trisphosphate, MAPK: mitogen-activated protein kinase,
mGluR: metabotropic glutamate receptor, PIP2: phosphatidylinositol-3,4-bisphosphate, PLA2: phospholipase A2, PLC: phospholipase
C, Post: postsynapse, VGCC: voltage-gated Ca2+ channels.

been claimed to be dependent on group II mGluRs as postsynaptic mechanisms like PS or changes in

only (as it was not affected by an mGluRI blocker;
Li et al., 2002) but other groups detected a blockade
of LTD after mGluR5 blockade (group II mGluR
blockade was not tested; Feinmark et al., 2003).
Induction of group I mGluR-LTD is prevented
by the Ca2+ chelator EGTA and by the blocker of
L-type VGCCs nitrendipine in slices from neonatal
(P3–8) rats (Bolshakov and Siegelbaum, 1994),
suggesting that a combination of Ca2+ influx via
these channels and release from intracellular stores
is needed for this form of LTD. Elevation of
postsynaptic [Ca2+] and mGluR activation (with
the specific mGluR agonist ACPD) is sufficient to
induce LTD (Bolshakov and Siegelbaum, 1995),
indicating a postsynaptic site of induction. Never-
theless, studies indicate that group I mGluR-LTD
in neonatal rats is not expressed postsynaptically,
the surface expression of AMPARs play no role
in group I mGluR-LTD (Nosyreva and Huber,
2005). Instead, group I mGluR-LTD in neonatal
rats appears to be expressed presynaptically as a
long-term decrease in transmitter release, since the
expression of LTD at this age is accompanied by
changes in PPR (Nosyreva and Huber, 2005),
coefficient of variation of EPSCs (Bolshakov and
Siegelbaum, 1994), a decrease in the frequency but
not amplitude, of miniature EPSCs (Bolshakov and
Siegelbaum, 1994), and a decreased rate of release
of the fluorescent marker FM1–43 (N-(3-triethyl-
ammoniumpropyl)-4-(4-(dibutylamino) styryl) pyr-
idinium dibromide; Zakharenko et al., 2002).
Since this form of LTD is induced postsynaptic-
ally but expressed presynaptically, a retrograde mes-
senger is needed to transfer the signal from the
484
postsynaptic to the presynaptic site (Bolshakov and
Siegelbaum, 1994). AA has been suggested to be the
retrograde transmitter involved in neonatal mGluR-
LTD (P4–7; Bolshakov and Siegelbaum, 1995), be-
cause AA application can induce LTD (Bolshakov
and Siegelbaum, 1995), an inhibitor of PLA2 that
initiates the AA signaling cascade by liberating
AA from membrane phospholipids, blocks LTD
(Bolshakov and Siegelbaum, 1995) and metabolites
of AA have been shown to mediate mGluR-LTD in
neonatal rats (Feinmark et al., 2003). Feinmark
et al. (2003) showed that 12(S)-hydroperoxyeicosa-
5Z,8Z,10E,14Z-tetraenoic acid (12-S-HPETE), a 12-
lipoxygenase metabolite of AA, is actively recruited
during the induction of mGluR-LTD, since (1) a
mouse in which the leukocyte-type 12-lipoxygenase
(the neuronal isoform) was deleted through homol-
ogous recombination was deficient in mGluR-LTD,
(2) pharmacological inhibition of 12-lipoxygenase
also blocked induction of mGluR-LTD, and (3) di-
rect application of 12-S-HPETE to hippocampal
slices induced a LTD of synaptic transmission that
mimicked and occluded mGluR-LTD induced by
synaptic stimulation.
Group I mGluR-LTD has also been reported to
require the postsynaptic activation of p38 MAPK in
neonatal rats, since introduction of the p38 MAPK
inhibitor SB203580 into a postsynaptic CA1 py-
ramidal neuron greatly diminished the induction of
this form of LTD, while the p42/44 inhibitor PD
98059 had no effect (P4–11; Bolshakov et al., 2000).
Additionally, it has been demonstrated that dial-
yzing CA1 neurons with the active, dually phos-
phorylated p38 MAPK induces a progressive
depression in the amplitude of the EPSC which
occludes LTD in response to subsequent 5 Hz elec-
trical stimulation. P38 MAPK provides a link to the
AA pathway since P38 MAPK can activate cyto-
solic PLA2 (cPLA2) by phosphorylation (Borsch-
Haubold et al., 1997). cPLA2 is also regulated by
[Ca2+]; Ca2+ promotes binding of the enzyme to
phospholipid substrates by causing the translocation
of cPLA2 to nuclear and other cellular membranes
through a Ca2+-dependent lipid-binding motif
(Clark et al., 1995).
mGluR-dependent LTD in juvenile rats. Group I
mGluR-dependent LTD has been induced in
juvenile rats by LFS stimulation (5 Hz/3 min;
P11-35; Oliet et al., 1997; Nicoll et al., 1998) and
by paired-pulse LFS (PP-LFS; paired pulses at
50 ms interstimulus intervals, applied at a 1 Hz
frequency for 15 min; P21–30; Huber et al., 2000)
in the presence of the NMDAR antagonist
AP5. Additionally, chemical activation of group
I mGluRs with (S)-3,5-dihydroxyphenylglycine
(DHPG) can induce an LTD in juvenile rats (Hub-
er et al., 2001) which is independent of NMDAR
activation (Huang and Hsu, 2006). LFS-induced
group I mGluR-dependent and DHPG-LTD show
mutual occlusion, which suggests similar induction/
expression mechanisms. Group I mGluR-dependent
LTD in juvenile rats seems to be mediated by group
I mGluRs, since it is blocked by the selective group
1 antagonist AIDA ([CRS]-1-aminoindan-1,5-di-
carboxylic acid) (Oliet et al., 1997), but not by the
selective group II mGluR antagonist EGLU (2S-a-
ethylglutamic acid), or the selective group III
mGluR antagonist MAP4 ((S)-2-amino-2-methyl-
4-phosphonobutanic acid) (Huang and Hsu, 2006).
The responsible group I mGluR subtype appears to
be mGluR5, since group I mGluR-LTD is not
affected by the mGluR1 selective antagonist 4CPG
((S)-4-carboxyphenylglycine) (Oliet et al., 1997;
Huang and Hsu, 2006), is blocked by the mGluR5
selective antagonist MPEP (2-methyl-6-(phenyl-
ethynyl)pyridine) (Huang et al., 2004; Huang and
Hsu, 2006) and is absent in homozygote mGluR5
knockout mice (Huber et al., 2001). Established
DHPG-LTD is completely reversed by application
of the mGluR antagonists MCPG or LY341495 and
LTD is reestablished once the antagonists are
washed out, suggesting that the induction of
DHPG-LTD involves either an alteration in func-
tional state of the receptors or the switching on of a
tonically active mGluR receptor (Huang and Hsu,
2006). In contrast to the requirement for persistent
mGluRI activation, the Gq-mediated downstream
pathways of group I mGluRs do not seem to be
necessary for the maintenance of mGluRI-LTD;
neither the phospholipase C (PLC) blocker U73122,
the PKC inhibitor Bis-1 nor the depletion of intra-
cellular Ca2+ pools by thapsigargin or cyclopiazonic
acid have any effect on pre-established DHPG-LTD
(Reyes and Stanton, 1996; Huang and Hsu, 2006).
Instead, a downstream protein tyrosine phosphatase
 485

(PTP) signaling cascade seems to be required for the
expression of DHPG-LTD (see below; Huang and
Hsu, 2006) (Fig. 3).
The induction of group I mGluR-LTD requires
postsynaptic activation of PKC (Oliet et al., 1997;
Nicoll et al., 1998; Huang and Hsu, 2006), but is not
affected by PP1/2A inhibitors (P11–35: Oliet et al.,
1997; Nicoll et al., 1998; P21–28: Huang and Hsu,
2006), contrasting with NMDAR-LTD, which does
appear to involve postsynaptic activation of PP2B
(P21–30: Kameyama et al., 1998). Like NMDAR-
LTD, induction of group I mGluR-LTD relies on
increases in postsynaptic [Ca2+] (Oliet et al., 1997;
Nicoll et al., 1998; but see Fitzjohn et al., 2001/
P12–18).
In contrast to group I mGluR-LTD in neonatal
rats (Bolshakov and Siegelbaum, 1994), induction
of group I mGluR-LTD in juvenile rats is
prevented by blockers of T-type VGCCs, but not
L-type VGCCs (Oliet et al., 1997; Nicoll et al.,
1998), suggesting a developmental switch. Addi-
tionally, it has been shown that group I mGluR-
LTD in juvenile rats is prevented by inhibitors of
mRNA translation but not transcription (Huber
et al., 2000), indicating that new PS from existing
mRNA species is a component of mGluR-LTD by
this developmental stage.
There are contradictory findings of the role of
p38 MARK for the induction of group I mGluR-
LTD; one study reported that group I mGluR-
LTD was not affected by p38 MARK inhibitors,
but was blocked by p42/44 MARK and ERK
inhibitors instead (Gallagher et al., 2004). Another
study found no effect of p42/44 MARK inhibition

Fig. 3. Cascades involved with mGluR-dependent LTD in juvenile rats at Schaffer collateral–CA1 synapses. Ca: Ca2+, DAG: diacyl
glycerol, IP3: inositol-1,4,5-trisphosphate, MAPK: mitogen-activated protein kinase, mGluR: metabotropic glutamate receptor, P:
phosphate, PIP2: phosphatidylinositol-3,4-bisphosphate, PKC: protein kinase C, PLC: phospholipase C, Post: postsynapse, PTP:
protein tyrosine phosphatase, VGCC: voltage-gated Ca2+ channels.
486
on group I mGluR-LTD, but a potent effect of
p38 MARK inhibition (Huang et al., 2004).
Huang et al. (2004) further showed, with specific
inhibitors, that during group I mGluR-LTD, p38
MARK is activated via a specific cascade that in-
cludes activation of mGluR5, release of G protein
bg-subunits, which, in turn, promotes the ex-
change of GDP with GTP of the small GTPase
Rap1, causing Rap1 to activate a sequential kinase
cascades that includes MAPK kinase 3/6 and p38
MAPK.
There is evidence for both pre- and postsynaptic
expression of group I mGluR-LTD in juvenile rats.
A presynaptic site of expression is suggested by
studies showing that expression of LTD is accom-
panied by a change in PPR (Fitzjohn et al., 2001;
Nosyreva and Huber, 2005), increase in EPSC fail-
ure rate (Fitzjohn et al., 2001), a change in the
coefficient of variation of EPSCs (Fitzjohn et al.,
2001), a decrease in the frequency but not the am-
plitude, of miniature EPSCs (Fitzjohn et al., 2001),
and a decrease in the frequency, but not in ampli-
tude, of Sr2+ asynchronously-evoked quantal events
(which only occurs at the subset of synapses that are
stimulated; Oliet et al., 1997). But these data are
challenged by conflicting findings from Zhang et al.
(2006) which showed a lack of presynaptic changes
for DHPG-LTD and pharmacologically isolated
group I mGluR-LTD, measured using changes in
PPR, two-photon imaging of FM1–43 release, and
variance-mean analysis. This study did, however,
indicate that coactivation of group II mGluRs could
recruit presynaptic changes in release.
There is evidence that group I mGluR-LTD is
expressed by postsynaptic NMDAR and AMPAR
internalization, since group I mGluR activation
induces a PS-dependent loss of postsynaptic AM-
PARs and NMDARs in cultured hippocampal
neurons (Snyder et al., 2001), and group I mGluR
activation induces an LTD in cultured slices which
is dependent on AMPAR endocytosis (Xiao et al.,
2001). In acute hippocampal slices from juvenile
rats, group I mGluR activation produces a de-
crease in AMPAR surface expression (GluR1 and
GluR2/3) which was completely blocked by the
broad-range mGluR antagonist LY341495, while
the late phase (60 min after LTD induction) was
blocked by the translational inhibitor anisomyin
(Huang et al., 2004; Nosyreva and Huber, 2005;
Huang and Hsu, 2006). During expression of
group I mGluR-LTD, the decrease in AMPAR
surface expression has been shown to be regulated
by p38 MARK. p38 MARK accelerates a rapid
loss of surface AMPARs by stimulating the activ-
ity of guanyl nucleotide dissociation inhibitor
(GDI), extracting Rab5 from endosomal mem-
branes and forming a GDI-Rab5 complex (Cavalli
et al., 2001; Huang et al., 2004). This complex
is required for the sequestration of the ligand-
receptor-complex into clathrin-coated pits
(McLauchlan et al., 1998). Huang and Hsu
(2006) showed that a transient activation of p38
MARK is sufficient to stimulate AMPAR end-
ocytosis during group I mGluR-dependent LTD,
since maintenance of previously-induced DHPG-
LTD is not affected by the selective p38 MAPK
inhibitor SB203580, and the increase in phosphor-
ylation of p38 MAPK elicited by DHPG is tran-
sient. In contrast to p38 MAPK, postsynaptic
PTPs appear to be necessary for the continuing
expression of group I mGluR-dependent LTD,
because DHPG-LTD is reversed by postsynaptic
loading of pyramidal neurons with a PTP inhibitor
(Huang and Hsu, 2006). Interestingly, it has been
demonstrated that levels of tyrosine phosphorylat-
ion of GluR2 AMPAR subunits (phosphorylated
under basal conditions (Ahmadian et al., 2004)),
but not GluR1 subunits, are dramatically de-
creased during DHPG-LTD, suggesting that AM-
PARs are targets of PTPs during maintenance of
group I mGluR-dependent LTD.
Another chemically induced form of LTD at
Schaffer collateral–CA1 synapses in hippocampal
slices from juvenile rats (14–24 days old) involves
the activation of group II mGluRs (Santschi et al.,
2006; Zhang et al., 2006). While the selective
mGluR II agonist DCGIV [(2S,20 R, 30 R)-2-(20 ,30 -
dicarboxycyclopropyl)glycine] alone only induces
a short-term depression which fully reverses upon
drug washout, a combination of DCGIV applica-
tion and elevation of endogenous intracellular
[cGMP] by application of the type V phosphodi-
esterase inhibitor zaprinast, is sufficient to induce
LTD (Santschi et al., 2006). It is probable that
mGluRs II promote LTD by inhibiting adenylate
cyclase, because pairing application of zaprinast

with either activation of A1 adenosine receptors,
which also inhibit adenylate cyclase, or with direct
pharmacologic inhibition of PKA activity, are
each sufficient to elicit LTD (Santschi et al., 2006).
LTD elicited by any of these methods of pairing
elevated [cGMP] with reduced [cAMP], is induced
and expressed presynaptically: selective postsynap-
tic inhibition of PKA does not enhance LTD,
while postsynaptic inhibition of PKG does not
block any of these chemical forms of LTD (San-
tschi et al., 1999). Consistent with this conclusion,
it has been shown that coapplication of zaprinast
with a synthetic PKA inhibitor (H-89) significantly
and persistently decreases evoked presynaptic ve-
sicular release of the fluorescent marker FM1–43
from Schaffer collateral terminals in hippocam-
pal slices (Stanton et al., 2001, 2003), and from
isolated rat hippocampal presynaptic terminals
(Bailey et al., 2003).
mGluR-dependent LTD in adult rats. (a) DHPG-
induced mGluRI-dependent LTD
The group I mGluR agonist DHPG also induces
LTD in slices from adult rats (8–10 weeks old), un-
der conditions of increased excitability (in Mg2+-
free medium; Palmer et al., 1997). Interestingly, this
form of LTD does not occlude stimulus-induced
(2 Hz/10 min) LTD (as opposed to the studies of
Huber and colleagues, see above).
Based on studies in immature rats, it was
expected that DHPG induces LTD via the activa-
tion of group I mGluR-coupled intracellular signal
cascades like PKC and PKA activation, and release
of Ca2+ from intracellular stores. Surprisingly,
DHPG-induced LTD has been reported not to be
affected by PKC or PKA inhibitors, or by depleting
intracellular Ca2+ stores (Schnabel et al., 1999a,
2001). Instead, induction, but not expression, of
DHPG-LTD is prevented by inhibitors of PTPs
(Moult et al., 2002), but the molecular targets of
tyrosine dephosphorylation remain to be identified.
DHPG-LTD is enhanced by treatment with
KN-62, an inhibitor of Ca2+/CaM-dependent pro-
tein kinases (CaMKs) (including CaMKII; Schna-
bel et al., 1999b), which indicates that some form
of dephosphorylation of proteins which are phos-
phorylated by CaMKs may be important to
DHPG-induced LTD. It could be that basal
487
CaMK activity may tend to inhibit DHPG-
induced LTD. Another possibility is that DHPG
simultaneously induces KN-62 insensitive LTD
and a smaller KN-62-sensitive LTP; blockade of
the latter leading to the apparent enhancement
of the former.
DHPG-LTD is also facilitated by inhibitors of
PP1 and PP2A, but not affected by inhibition of
PP2B (Schnabel et al., 2001), which indicates that
some form of phosphorylation of targets of PP1
and PP2A is important to DHPG-induced LTD
and that this phosphorylation is actively opposed
by dephosphorylation by PP1 or PP2A.
DHPG-induced LTD can be reversed by the
non-selective mGluR antagonist MCPG (Palmer
et al., 1997; Fitzjohn et al., 1998, 1999; Watabe
et al., 2002), and by the group I/II mGluR antag-
onist LY341495 (Fitzjohn et al., 1998; Watabe
et al., 2002), but LTD is reestablished following
washout of either antagonist. One explanation for
the transient effect of the mGluR antagonists
could be that endogenous glutamate tonically ac-
tivates mGluRIs. A second possibility is that mGlu
receptor agonists and mGlu receptor antagonists
switch the mode of activity of mGluRI (i.e., act as
agonists and inverse agonists) with respect to
downstream events, i.e., they activate opposing
signaling pathways. MCPG-induced reversal of
DHPG-induced LTD does not involve activation
of CaMKII, PKA, or release of Ca2+ from intra-
cellular stores (Schnabel et al., 1999a, b), but does
involve the activation of PP1/PP2A (Schnabel et
al., 2001). Thus, DHPG-induced LTD seems to
depend on the net activity of some yet to be de-
termined protein kinase(s) and PP1/PP2A. Acti-
vation of mGluRs by DHPG switches the
receptor-effector cascade into a mode that pro-
motes phosphorylation. Application of mGluRI
antagonists can temporarily reverse the phosphor-
ylation mode via transient activation of a cascade
that involves dephosphorylation via PP1/PP2A.
While DHPG-LTD is induced postsynaptically
(group I mGluRs are only located postsynaptical-
ly; Conn and Pin, 1997), it has been suggested to
be expressed presynaptically as inhibition of volt-
age-sensitive Ca2+ channels and K+ channels
(Watabe et al., 2002; Tan et al., 2003) at Schaffer
collateral–CA1 synapses. A presynaptic locus of
488
expression is supported by studies reporting that:
(1) DHPG induces a persistent depression of
NMDA, as well as AMPA, receptor-mediated
components of EPSCs (Watabe et al., 2002), (2)
the expression of LTD is accompanied by a change
in PPR (Faas et al., 2002; Rouach and Nicoll,
2003; Tan et al., 2003); and (3) a lack of change in
postsynaptic sensitivity to AMPA in CA1, meas-
ured by focally elicited AMPAR-currents evoked
by rapid uncaging of glutamate, following DHPG-
LTD (Rammes et al., 2003). The depression
induced by DHPG is dependent on postsynaptic,
GTP-dependent signaling pathways, as it is
blocked after replacement of GTP with guano-
sine-50 -O-(2-thiodiphosphate) (GDPgS; Watabe et
al., 2002).
(b) Stimulus-induced mGluR-dependent LTD
Stimulus-induced (2 Hz, 900 pulses during
GABAA-blockade) LTD in adult rats (7 weeks
old) shows different mechanisms than DHPG-
induced LTD (Otani and Connor, 1998). LTD de-
pends on mGluR, but not NMDAR, activation in
adult rats, as it is blocked by the mGluR antag-
onist MCPG, but not the NMDAR antagonist
AP5. An increase in postsynaptic [Ca2+] is needed
to induce this form of LTD, since postsynaptic
infusion of the Ca2+ chelator BAPTA blocked in-
duction of LTD. A transient (20–30 s) increase in
intracellular [Ca2+] could be measured during the
2 Hz stimulation. This [Ca2+] increase was the re-
sult of voltage-gated influx, rather than intracel-
lular Ca2+ release triggered by mGluR activation,
since it was blocked by hyperpolarization of the
neuronal soma to
110 mV, but not by MCPG.
Injection of the PKC inhibitor peptide PKC19–36
into the postsynaptic neuron completely blocked
LTD, verifying that PKC activation, probably via
mGluRI, is necessary for the induction of this
form of LTD.
Summary: CA1 LTD. Current data on LTD at
Schaffer collateral–CA1 synapses is becoming sub-
stantial and reveals a synapse where distinct pre-
and postsynaptic cascades cause long-term altera-
tions on both sides of the synapse. Because only
the postsynaptic dendritic spine would seem to be
privy to Hebbian information about both presy-
naptic glutamate release and global postsynaptic
voltage levels driven by many other synaptic con-
tacts, it appears that long-term changes are driven
by pairing of local spine [Ca2+] and postsynaptic
depolarization within a certain range. Postsynap-
tically, increases in [Ca2+] that preferentially ac-
tivate PPs and, perhaps independently, AMPAR
internalization appear to elicit postsynaptic LTD.
Presynaptically, the postsynaptic Ca2+-dependent
activation of NOS generates NO which diffuses
locally to proximal presynaptic terminals, where it
must pair with activation of G protein-coupled
receptors that inhibit adenylate cyclase to drive
a cyclic GMP/PKG/cyclic ADPribose and Ca2+-
dependent presynaptic cascade that causes LTD of
vesicular transmitter release from the rapidly-
recycling vesicle pool. NMDAR-dependent LTD
appears to involve both presynaptic and postsy-
naptic alterations, while group I mGluR-depend-
ent LTD is expressed purely postsynaptically, and
the activation of group II mGluRs recruits presy-
naptic cascades that may or may not be shared
with NMDAR-LTD. Thus, excitatory LTD in
field CA1 is a complex of multiple alterations
across the synapse. The extensive information now
available about relatively rapid biochemical events
during the induction of LTD is not, as of yet,
matched by sufficient data about whether and how
these changes lead to precise stimulation or sup-
pression of particular genes, or of retraction or
even complete elimination of synaptic contacts.
LTD in the dentate gyrus
Induction
LTD in the DG can also be induced by a variety of
stimulus protocols. The most common induction
protocol is LFS (1–2 Hz) of the medial perforant
path (MPP) to induce homosynaptic LTD (in
vitro: O’Mara et al., 1995a; in vivo: Abraham,
1996; Trommer et al., 1996; Pöschel and Man-
ahan-Vaughan, 2007). Heterosynaptic LTD has
also been demonstrated at both medial and lateral
perforant path (LPP) synapses by high-frequency
stimulation (HFS) of the other synaptic input,
(Abraham and Goddard, 1983; Doyere et al.,
1997). Associative LTD can also be induced in the

lateral perforant path, where it seems to share ex-
pression mechanisms with heterosynaptic LTD
(Christie et al., 1995). Christie et al. (1995) induced
primed-associative LTD of the lateral path by al-
ternating high-frequency conditioning bursts to
the medial path and single shocks to the lateral
path at 100 ms intervals, provided these were ad-
ministered 10 min after a homosynaptic priming
stimulation of the lateral path (5 Hz, 80 pulses).
Lin et al. (2006) elicited LTD at lateral perforant
path synapses by pairing antidromically-evoked
postsynaptic APs (aAPs) with synaptic excitatory
postsynaptic potentials (fEPSPs), providing that
the fEPSP occurred rapidly after the aAP. Agonist
activation of mGluRs can also induce multiple
forms of LTD in the DG (O’Mara et al., 1995a;
Camodeca et al., 1999; Huang et al., 1999a, b, c;
Klausnitzer et al., 2004; Pöschel and Manahan-
Vaughan, 2005).
mGluR agonist-induced LTD in the medial
perforant path
Metabotropic GluR activation is reported to be nec-
essary for the induction of LFS-LTD at medial per-
forant path-DG synapses (O’Mara et al., 1995a;
Huang et al., 1997; Camodeca et al., 1999;
Klausnitzer et al., 2004; Pöschel et al., 2005). The
mGluR family of heterotrimeric guanosine tripho-
sphate-binding protein (G protein)-coupled recep-
tors (Sugiyama et al., 1987; Tanabe et al., 1992) are
divided into three groups: mGluR I, II, and III.
Despite their different intracellular coupling, agonist
activation of all three groups has surprisingly been
shown to result in the induction of LTD (Huang
et al., 1999a, b, c; Camodeca et al., 1999; Naie and
Manahan-Vaughan, 2005; Pöschel and Manahan-
Vaughan, 2005) (Fig. 4)
Group I mGluRs include the receptor subtypes
mGluR1 and mGluR5 (and their splice variants).
Group I mGluRs are coupled to a pathway that
involves intracellular Ca2+ release from intracel-
lular stores and PKC activation. They produce Gq-
mediated activation of PLC which hydrolyzes in-
ositol-bisphosphate to IP3 and DAG. DAG, with
Ca2+, coactivates protein kinase C, while IP3
489
causes the release of cytosolic Ca2+ from intra-
cellular stores (Conn and Pin, 1997).
The mGluRI agonist DHPG induces an mGluRI-
dependent LTD at medial perforant path synapses
which does not require NMDAR activation (Camo-
deca et al., 1999). However, this LTD seems to occur
only in juvenile/young adult animals (up to 5 weeks
of age in rats; Camodeca et al., 1999; Wang et al.,
2007). There exist conflicting results concerning the
involvement of PKC activation in DHPG-LTD.
While an early study reported a strong inhibition
(but not complete block) of DHPG-LTD by the
PKC inhibitor Bisindolylmaleimide I (Bis I, 1 mM,
inhibits a, b, g, e at this concentration; Camodeca
et al., 1999), a later study by the same group did not
find an effect of Bis I on DHPG-LTD using identical
protocols (Rush et al., 2002). Thus, it remains to be
determined whether PKC activity is necessary for the
induction of DHPG-LTD. Furthermore, it has been
reported that DHPG-LTD is dependent on the ac-
tivation of p38 MAPK (Rush et al., 2002), as well as
non-receptor tyrosine kinases (Camodeca et al.,
1999), since the p38 MAPK inhibitor SB203580
(4-(4-fluorophenyl)-2-(4-methylsulfonylphenyl)-5-(4-
pyridyl) imidazol) and non-receptor tyrosine kinase
inhibitor lavendustin A each strongly inhibited the
induction of DHPG-LTD. Additional support for a
role for tyrosine kinases and p38 MAPK in DHPG-
LTD is the finding that activation of MAPK path-
ways (including p38 MAPK) by G protein-coupled
receptors requires activation of the Src family of
non-receptor tyrosine kinases (Luttrell et al., 1999).
MAPKs phosphorylate transcription and translation
factors, thereby regulating gene expression and PS
(Thomas and Huganir, 2004; Kelleher et al., 2004).
Consistent with this is a study showing that DHPG-
LTD is inhibited by PS blockade with either an-
isomycin or emetine (Naie and Manahan-Vaughan,
2005). It seems that DHPG-LTD is expressed only
postsynaptically, as it is not accompanied by any
change in paired-pulse depression (Camodeca et al.,
1999).
Some mGluRI subtypes (mGluR1a, mGluR5a
and mGluR5b) stimulate cAMP formation by po-
tentiating AC activation triggered by neurotrans-
mitters that act on GS-coupled receptors (Conn
and Pin, 1997; Moldrich et al., 2002). For instance,
in cultured striatal neurons, the ability of group I
490
PKC?
mGluRI
tyrosine kinase
p38 MAPK
mGluRII
+ +
2+
PKA
T-VGCC Ca
Fig. 4. Cascades involved with mGluR-induced LTD at medial perforant path-DG synapses. Ca: Ca2+, MAPK: mitogen-activated
protein kinase, mGluR: metabotropic glutamate receptor, PKA: protein kinase A, PKC: protein kinase C, Post: postsynapse, VGCC:
voltage-gated Ca2+ channels.
mGluRs to potentiate cAMP formation induced
by D1 dopamine receptor agonists correlates with
their ability to generate IP3, and this potentiation
requires PKC activity (Paolillo et al., 1998). Nev-
ertheless, mGluRI potentiation of the AC-cAMP-
PKA pathway does not seem to play a necessary
role in inducing DHPG-LTD, because the PKA
inhibitor H-89 does not affect/prevent induction of
DHPG-LTD (Camodeca et al., 1999).
Group II mGluRs consist of mGluR2 and
mGluR3, which inhibit forskolin and receptor-
induced increases in [cAMP]. A Gi-type of G pro-
tein is probably involved in this coupling (Conn and
Pin, 1997). It has also been reported that activation
of group II mGluRs potentiates increases in [cAMP]
triggered by neurotransmitters that act on Gs-
coupled receptors in adult hippocampal slices (Conn
and Pin, 1997; Moldrich et al., 2002), i.e., this
mRNA- mRNA-
transcription translation
?
+
PKC
Post
potentiation is dependent on the coactivation of
group II mGluRs plus Gs-coupled receptors. The
potentiation of cAMP accumulation seems to be
mediated by synergistic interaction between group
I and group II mGluRs, or group II mGluRs and
b-adrenoceptors, which induces endogenous adeno-
sine release (Moldrich et al., 2002). Adenosine sub-
sequently activates both A2A receptors, which are
positively coupled to cAMP, and A1 receptors cou-
pled negatively to adenylate cyclase.
Metabotropic GluRII-dependent LTD is induced
in the DG by selective agonist activation of
mGluRIIs with (2S,20 R,30 R)-2-(20 30 -dicarboxycy-
clopropyl) glycine (DCG-IV) or (1S,2S,5R,6S)-2-
aminobicyclo[3.1.0]hexane-2-6-dicarboxylic acid
(LY354740): Huang et al., 1999a, b, c; (S)-4-car-
boxy-3-hydroxyphenylglycine (4C3HPG): Pöschel
and Manahan-Vaughan, 2005), and also by

activation of the mGluRII subtype mGluR3 (with
N-acetylaspartylglutamate (NAAG): Huang et al.,
1999b; Pöschel and Manahan-Vaughan, 2005). Per-
fusion with mGluRII agonists in vitro induces a
rapid depression of MPP-evoked EPSPs which is
maintained in the presence of the agonist but, fol-
lowing washout, is partially reversed to a depression
level which is maintained throughout the experiment
(LTD was induced lasting at least 1 h following
washout; Huang et al., 1999a, b). Intracerebroven-
tricular injection of mGluRII agonists in vivo in-
duces a rapid (mGluRII agonist) or slowly
developing depression (mGluR3 agonist) of MPP-
evoked EPSPs which is maintained throughout the
experiment (up to 25 h after LFS; Pöschel and
Manahan-Vaughan, 2005). The rapidly evoked
reversible depression of the EPSP by mGluRII ago-
nists in vitro was associated with a change in paired-
pulse depression which was not maintained during
LTD following washout of the agonist (Huang et
al., 1999a, b), indicating that the reversible depres-
sion may be expressed via a presynaptic reduction in
the probability of transmitter release, but that LTD
is induced either by a presynaptic reduction in the
number of active release sites, and/or postsynaptic
changes.
While Huang et al. (1999a) reported a partial
block of mGluRII-induced LTD by the NMDAR
antagonist AP5, Pöschel and Manahan-Vaughan
(2005) could not confirm a dependence of
mGluR3-induced LTD on any NMDAR activa-
tion. This may mean that NMDAR blockade
affects only mGluR2-dependent LTD. In vitro,
mGluRII-induced LTD has been shown to depend
on depolarization, since the cessation of test stim-
ulation during mGluRII agonist application pre-
vented the induction of LTD (Huang et al.,
1999a). Furthermore, mGluRII-induced LTD,
but not the reversible initial depression, in vitro
requires the activation of VGCCs, and of PKA or
PKC (Huang et al., 1999a, c). The downstream
pathways of PKA and PKC have yet to be inves-
tigated, but PS does not seem to be required for
the induction of mGluRII-induced LTD, since it is
not blocked in vivo by the translation-inhibitor
anisomycin or the transcription-inhibitor act-
inomycin D (Pöschel and Manahan-Vaughan,
2005).
491
Group III mGluRs comprise the subtypes
mGlu4, 6, 7, and 8 (and their splice variants) and
are linked to the inhibition of Gs-activated ACs
(Conn and Pin, 1997). The induction pathways of
Group III mGluR-induced LTD have not been
investigated, but it has been shown that Group III
mGluR-induced LTD is expressed independently
of PS (Naie and Manahan-Vaughan, 2005).
Electrically-induced LTD at medial PP-DG
synapses
Although an early study reported that the non-
competitive NMDAR open-channel blockers keta-
mine and phencyclidine were able to block the in-
duction of medial perforant path LTD (Desmond
et al., 1991), follow-up studies, using the specific
NMDAR antagonist AP5, have not confirmed the
dependence of LFS-LTD on NMDAR activation
at medial PP-DG synapses (O’Mara et al., 1995a;
in vitro: Trommer et al., 1996; Wang et al., 1997;
in vivo: Pöschel and Manahan-Vaughan, 2005,
2007). Nevertheless, postsynaptic intracellular
[Ca2+] elevations are a necessary step in the in-
duction of LFS-LTD at medial PP-DG synapses,
and can be provided either via Ca2+ influx into
dentate granule cells through low voltage-acti-
vated Ca2+ channels (T-type), or Ca2+ release
from intracellular stores (O’Mara et al., 1995b;
Wang et al., 1997), or both. The intracellular
stores involved include both group I mGluR-acti-
vated IP3-receptor-dependent and ryanodine-
receptor-sensitive postsynaptic stores (O’Mara
et al., 1995b; Wang et al., 1997). Ryanodine-re-
ceptor-sensitive Ca2+ stores are a target of the
NO-cGMP pathway via a PKG-dependent activa-
tion of a cyclase/hydrolase that generates cyclic
ADPribose, a messenger that either enhances or
causes release from RyR-sensitive stores which
mediate Ca2+-dependent Ca2+ release (CDCR).
The NO-cGMP pathway has been shown to play a
role in the induction of LFS-LTD at PP-DG
synapses, as inhibition of NOS, guanylyl cyclase
and postsynaptic PKG all prevent the induction of
LFS-LTD (Wu et al., 1997, 1998). Furthermore,
cGMP-dependent LTD induced by the type V
phosphodiesterase inhibitor zaprinast shows
492

mutual occlusion with LFS-LTD, and zaprinast-
induced LTD is dependent on mGluR, but not
NMDAR, activation (Wu et al., 1998). Since
postsynaptic inhibition of PKG and postsynaptic
blockade of ryanodine receptors were reportedly
sufficient to strongly inhibit LFS-LTD in the DG,
the NO-cGMP pathway seems to be induced
postsynaptically in this region, with NO acting
not as retrograde messenger as in CA1, but as an
intercellular postsynaptic messenger in the DG
(Fig. 5).
LFS-LTD in the DG appears to be dependent on
all subtypes of mGluRs at medial PP-DG synapses.
It has been reported that LFS-LTD shows mutual
occlusion with mGluRI-induced LTD (Camodeca
et al., 1999), and with mGluRII-induced LTD
(Huang et al., 1999a), which suggests common in-
duction and/or maintenance of intracellular path-
ways shared between a compound LFS-LTD and
both mGluR I and II-dependent LTDs. Induction
of LFS-LTD fully blocks the subsequent induction
of either mGluRI-induced LTD or mGluRII-
induced LTD, whereas the induction of either
mGluRI-induced LTD or mGluRII-induced LTD
only partially inhibits the subsequent induction of
LFS-LTD, suggesting that mGluRIs and mGluRIIs
contribute with parallel/independent intracellular
pathways to LFS-LTD. Furthermore, LFS-LTD is
blocked by antagonists of mGluRIs ((AIDA):
Camodeca et al., 1999), mGluRIIs (2S,1S0 ,2S0 -2-

Fig. 5. Cascades involved with LFS-induced LTD at medial perforant path-DG synapses. AC: adenylyl cyclase, Ca: Ca2+, cADPR:
cyclic adenosine diphosphate ribose, CaM: calmodulin, cGMP: cyclic guanosine monophosphate, DAG: diacyl glycerol, IP3: inositol-
1,4,5-trisphosphate, MAPK: mitogen-activated protein kinase, mGluR: metabotropic glutamate receptor, NO: nitric oxide, NOS: NO-
synthetase, PIP2: phosphatidylinositol-3,4-bisphosphate, PKA: protein kinase A, PKC: protein kinase C, PKG: protein kinase G,
PLC: phospholipase C, Post: postsynapse, VGCC: voltage-gated Ca2+ channels.

methyl-2-(20 -carboxycyclopropyl)glycine (MCCG):
Huang et al., 1997; NAAG: Pöschel et al., 2005),
and mGluRIIIs ((R.S)-r-cyclopropyl-4-phosphono-
phenylglycine (CPPG): Klausnitzer et al., 2004).
Thus, all mGluR subgroups seem to contribute to
the induction of LFS-LTD, despite their different
intracellular signaling pathways. Kinases which are
part of different mGluR intracellular pathways,
such as PKA (Huang et al., 1999a, c), PKC (Wang
et al., 1998; Huang et al., 1999a, c) and MAPK
(Murray and O’Connor, 2003), have all been
reported to be necessary for the induction of LFS-
LTD. Application of a PKC activator can also
directly induce a depression which occludes subse-
quent induction of LFS-LTD (Wang et al., 1998)
supporting a role for PKC in mediating some por-
tion of LFS-LTD at MPP-DG synapses. The mon-
omeric G protein Ras activates certain MAPK
pathways (Thomas and Huganir, 2004) and the Ras
inhibitor manumycin A has also been shown to
significantly attenuate LFS-LTD (Murray and
O’Connor, 2004).
One stimulation protocol (step depolarization of
a DG granule neuron from a holding potential
of
70 mV to
50 mV for 1.1 s applied during
HFS medial perforant path stimulus trains in the
presence of an NMDAR blocker) can also induce
LTD which depends only on mGluRI activation
(Wu et al., 2001). This form of LTD required
membrane depolarization, was mediated by Ca2+
influx via L-type Ca2+ channels plus Ca2+ release
from ryanodine-receptor-sensitive intracellular
Ca2+ stores, and required activation of both
MAPK and PKC (Wu et al., 2001, 2004, Wang
et al., 2007). Wu et al. (2001) detected decreases in
potency and increases in failure rate after induc-
tion of mGluRI-dependent LTD. Although these
changes can be attributed to either presynaptic or
postsynaptic changes, Wu et al. (2001) argue in
favor of a postsynaptic site of expression, stating
that a decrease in potency could also be caused by
a decrease in the probability of release if more than
one synapse was being activated and the increase
in failure rate, according to the silent synapse hy-
pothesis of Liao et al. (1995), could also be caused
by a postsynaptic change associated with a failure
of detection of released transmitter as active
synapses are converted into silent synapses (i.e.,
493
AMPARs are completely removed postsynaptical-
ly). This interpretation is supported by the findings
of Camodeca et al. (1999) and Zhang et al. (2006),
who found that DHPG-LTD in the DG was not
accompanied by a change in paired-pulse depres-
sion (Camodeca et al., 1999) and group I mGluRs
did not contribute to the presynaptic expression of
LFS-LTD in field CA1 (Zhang et al., 2006).
Heterosynaptic LTD at lateral perforant path
synapses
Induction of heterosynaptic LTD in the lateral
perforant path of the DG (by HFS stimulation of
the medial PP) has been shown in vitro (Christie
and Abraham, 1992b) and in vivo (Abraham et al.,
2006) to depend on activation of both NMDARs
and L-type VGCCs. Furthermore, immediate early
gene expression has been associated with the per-
sistence of heterosynaptic LTD at both medial and
lateral perforant path synapses on dentate granule
cells, i.e., the greatest immediate early gene ex-
pression occurred following a stimulus protocol
which consistently gave the longest-lasting LTD
(Abraham et al., 1994).
Induction of associative LTD in the lateral per-
forant path, which is induced after priming of the
lateral path, does not depend on NMDAR activa-
tion (Christie and Abraham, 1992a). Nevertheless,
heterosynaptic LTD and prime-associative LTD of
the lateral path seem to share a common expression
mechanism, since they occlude one another (Christie
et al., 1995). It remains to be determined what non-
NMDARs are involved in either.
The one study of spike-timing-dependent LTD
at lateral perforant path synapses reported a
dependence on NMDAR activation, and block-
ade by inhibitors of protein kinase C and PP2B
(Lin et al., 2006).
Morphological changes associated with LTD
have been so far investigated in only two studies;
Mezey et al. (2004), who reported that heterosy-
naptic LTD of the medial path of the DG results in
an input-specific increase in axodendritic synapse
density, whereas LTP at the lateral path was asso-
ciated with an increase in perforated axospinous
synapses in the potentiated area, and Zhou et al.
494
(2004), who reported an NMDAR, CN, and cofilin-
dependent shrinkage of dendritic spines following
induction of LTD at Schaffer collateral–CA1
synapses. Thus, the range of morphological altera-
tions that may be associated with LTD, their func-
tional effects on transmission, and the time course
of such changes during the progression of LTD, are
still unknown.
Summary: DG LTD
This chapter makes it clear that significantly less is
known about LTD in the DG compared to field
CA1, or even the hippocampus proper in general.
The majority of this evidence points to postsynap-
tic alterations in MPP-DG LTD, and suggests that
an NO/cGMP/PKG cascade produces long-term
postsynaptic LTD here, in contrast to CA1. In this
respect, MPP-LTD may have more in common
with LTD in the striatum, than in field CA1.
However, there are enough studies suggesting
long-term presynaptic alterations in mGluR3-
dependent LTD (Pöschel et al., 2005), and in the
actions of brain-derived neurotrophic factor
(BDNF) that promote/elicit LTP (Gooney et al.,
2002), to leave the question of presynaptic long-
term plasticity at MPP-DG synapses open to fur-
ther study. Likewise, the relative lack of much
knowledge about LPP-DG synapse LTD (or LTP)
makes this a fertile area of future study. Most re-
ports have examined heterosynaptic LTD of LPP-
DG synapses, which are likely to involve postsy-
naptic alterations that depend on the spread of
postsynaptic depolarization and activation of volt-
age-dependent Ca2+ channels, and may differ sub-
stantially from homosynaptic LPP LTD. Even
more so than in field CA1, the connections be-
tween biochemical cascades necessary for the
induction of LTD, and eventual synaptic remode-
ling, are unexplored territory in the DG.
Acknowledgments
This work was supported by National Institutes of
Health Grant R01-NS44421 (to P.K.S.) and Na-
tional Institute of Drug Abuse Grant R01-HD45754
(to Dr. Peter Mundel).
References
Abraham, W.C. (1996) Induction of heterosynaptic and ho-
mosynaptic LTD in hippocampal sub-regions in vivo. J.
Physiol. Paris., 90: 305–306.
Abraham, W.C., Christie, B.R., Logan, B., Lawlor, P. and
Dragunow, M. (1994) Immediate early gene expression as-
sociated with the persistence of heterosynaptic long-term de-
pression in the hippocampus. Proc. Natl. Acad. Sci. U.S.A.,
91: 10049–10053.
Abraham, W.C. and Goddard, G.V. (1983) Assymetric rela-
tionships between homosynaptic long-term potentiation
and heterosynaptic long-term depression. Nature, 305:
717–719.
Abraham, W.C., Maxon-Parker, S.E., Irvine, G.I., Logan, B.
and Gill, A.I. (2006) Induction and activity-dependent re-
versal of persistent LTP and LTD in lateral perforant path
synapses in vivo. Neurobiol. Learn. Mem., 86: 82–90.
Ahmadian, G., Ju, W., Liu, L., Wyszynski, M., Lee, S.H.,
Dunah, A.W., Taghibiglou, C., Wang, Y., Lu, J., Wong,
T.P., Sheng, M. and Wang, Y.T. (2004) Tyrosine phosphor-
ylation of GluR2 is required for insulin-stimulated AMPA
receptor endocytosis and LTD. EMBO J., 23: 1040–1050.
Alford, S., Frenguelli, B.G. and Collingridge, G.L. (1992)
Characterisation of Ca2+ signals induced in hippocampal
CA1 neurones by the synaptic activation of NMDA recep-
tors. J. Physiol. Lond., 469: 693–716.
Armstrong, D.L. (1989) Calcium channels regulation by cal-
cineurin, a Ca2+-activated phosphatase in mammalian brain.
Trends Neurosci., 12: 117–122.
Bailey, C.P., Trejos, J.A., Schanne, F.A. and Stanton, P.K.
(2003) Pairing elevation of [cyclic GMP] with inhibition of
PKA produces long-term depression of glutamate release
from isolated rat hippocampal presynaptic terminals. Eur. J.
Neurosci., 17: 903–908.
Banke, T.G., Bowie, D., Lee, H., Huganir, R.L., Schousboe, A.
and Traynelis, S.F. (2000) Control of GluR1 AMPA receptor
function by cAMP-dependent protein kinase. J. Neurosci.,
20: 89–102.
Beattie, E.C., Carroll, R.C., Yu, X., Morishita, W., Yasuda, H.,
von Zastrow, M. and Malenka, R.C. (2000) Regulation of
AMPA receptor endocytosis by a signaling mechanism
shared with LTD. Nat. Neurosci., 3: 1291–1300.
Bito, H., Deisseroth, K. and Tsien, R.W. (1996) CREB phos-
phorylation and dephosphorylation: a Ca2+- and stimulus
duration-dependent switch for hippocampal gene expression.
Cell, 87: 1203–1214.
Bolshakov, V.Y., Carboni, L., Cobb, M.H., Siegelbaum, S.A.
and Belardetti, F. (2000) Dual MAP kinase pathways medi-
ate opposing forms of long-term plasticity at CA3-CA1
synapse. Nat. Neurosci., 3: 1107–1112.
Bolshakov, V.Y. and Siegelbaum, S.A. (1994) Postsynaptic in-
duction and presynaptic expression of hippocampal long-
term depression. Science, 264: 1148–1152.
Bolshakov, V.Y. and Siegelbaum, S.A. (1995) Hippocampal
long-term depression: arachidonic acid as a potential retro-
grade messenger. Neuropharmacology, 34: 1581–1587.

Borsch-Haubold, A.G., Kramer, R.M. and Watson, S.P. (1997)
Phosphorylation and activation of cytosolic phospholipase
A2 by 38-kDa mitogen-activated protein kinase in collagen-
stimulated human platelets. Eur. J. Biochem., 245: 751–759.
Brandon, E.P., Idzerda, R.L. and McKnight, G.S. (1997) PKA
isoforms, neural pathways, and behaviour: making the con-
nection. Curr. Opin. Neurobiol., 7: 397–403.
Brandon, E.P., Zhuo, M., Huang, Y.Y., Qi, M., Gerhold, K.A.,
Burton, K.A., Kandel, E.R., McKnight, G.S. and Idzerda,
R.L. (1995) Hippocampal long-term depression and depot-
entiation are defective in mice carrying a targeted disruption
of the gene encoding the RI beta subunit of cAMP-dependent
protein kinase. Proc. Natl. Acad. Sci. U.S.A., 92: 8851–8855.
Bredt, D.S., Hwang, P.M., Glatt, C.E., Lowenstein, C., Reed,
R.R. and Snyder, S.H. (1991) Cloned and expressed nitric
oxide synthase structurally resembles cytochrome P-450 red-
uctase. Nature, 351: 714–718.
Brindle, P.K. and Montminy, M.R. (1992) The CREB family of
transcription activators. Curr. Opin. Genet. Dev., 2: 199–204.
Cadd, G.G., Uhler, M.D. and McKnight, G.S. (1990) Holoen-
zymes of cAMP-dependent protein kinase containing the
neural form of type I regulatory subunit have an increased
sensitivity to cyclic nucleotides. J. Biol. Chem., 265:
19502–19506.
Camodeca, N., Breakwell, N.A., Rowan, M.J. and Anwyl, R.
(1999) Induction of LTD by activation of group I mGluR in
the dentate gyrus in vitro. Neuropharmacology, 38:
1597–1606.
Carroll, R.C., Lissin, D.V., von Zastrow, M., Nicoll, R.A. and
Malenka, R.C. (1999) Rapid redistribution of glutamate re-
ceptors contributes to long-term depression in hippocampal
cultures. Nat. Neurosci., 2: 454–460.
Cavalli, V., Vilbois, F., Corti, M., Marcote, M.J., Tamura, K.,
Karin, M., Arkinstall, S. and Gruenberg, J. (2001) The stress-
induced MAP kinase p38 regulates endocytic trafficking via
the GDI:Rab5 complex. Mol. Cell, 7: 421–432.
Chabot, C., Gagne, J., Giguere, C., Bernard, J., Baudry, M.
and Massicotte, G. (1998) Bidirectional modulation of
AMPA receptor properties by exogenous phospholipase A2
in the hippocampus. Hippocampus, 8: 299–309.
Chaki, S., Muramatsu, M. and Otomo, S. (1994) Involvement
of protein kinase C activation in regulation of acetylcholine
release from rat hippocampal slices by minaprine. Neuroc-
hem. Int., 24: 37–41.
Christie, B.R. and Abraham, W.C. (1992a) Priming of associ-
ative long-term depression in the dentate gyrus by theta fre-
quency synaptic activity. Neuron, 9: 79–84.
Christie, B.R. and Abraham, W.C. (1992b) NMDA-dependent
heterosynaptic long-term depression in the dentate gyrus of
anaesthetized rats. Synapse, 10: 1–6.
Christie, B.R., Stellwagen, D. and Abraham, W.C. (1995) Ev-
idence for common expression mechanisms underlying hete-
rosynaptic and associative long-term depression in the
dentate gyrus. J. Neurophysiol., 74: 1244–1247.
Clark, J.D., Schievella, A.R., Nalefski, E.A. and Lin, L.L.
(1995) Cytosolic phospholipase A2. J. Lipid. Mediat. Cell
Signal, 12: 83–117.
495
Cohen, P. (1989) The structure and regulation of protein
phosphatases. Annu. Rev. Biochem., 58: 453–508.
Colbran, R.J. (1992) Regulation and role of brain calcium/cal-
modulin-dependent protein kinase II. Neurochem. Int., 21:
469–497.
Conn, P.J. and Pin, J.P. (1997) Pharmacology and functions of
metabotropic glutamate receptors. Annu. Rev. Pharmacol.
Toxicol., 37: 205–237.
Cotman, C.W., Monaghan, D.T. and Ganong, A.H. (1988)
Excitatory amino acid neurotransmission: NMDA receptors
and Hebb-type synaptic plasticity. Ann. Rev. Neurosci., 11:
61–80.
Cummings, J.A., Mulkey, R.M., Nicoll, R.A. and Malenka,
R.C. (1996) Ca2+ signaling requirements for long-term de-
pression in the hippocampus. Neuron, 16: 825–833.
Cummings, J.A., Nicola, S.M. and Malenka, R.C. (1994) In-
duction in the rat hippocampus of long-term potentiation
(LTP) and long-term depression (LTD) in the presence of a
nitric oxide synthase inhibitor. Neurosci. Lett., 176: 110–114.
Desmond, N.L., Colbert, C.M., Zhang, D.X. and Levy, W.B.
(1991) NMDA receptor antagonists block the induction of
long-term depression in the hippocampal dentate gyrus of the
anesthetized rat. Brain Res., 552: 93–98.
Domenici, M.R., Berretta, N. and Cherubini, E. (1998) Two
distinct forms of long-term depression coexist at the mossy
fiber-CA3 synapse in the hippocampus during development.
Proc. Natl. Acad. Sci. U.S.A., 95: 8310–8315.
Doyere, V., Srebro, B. and Laroche, S. (1997) Heterosynaptic
LTD and depotentiation in the medial perforant path of the
dentate gyrus in the freely moving rat. J. Neurophysiol., 77:
571–578.
Dudek, S.M. and Bear, M.F. (1992) Homosynaptic long-term
depression in area CA1 of hippocampus and effects of N-
methyl-D-aspartate receptor blockade. Proc. Natl. Acad. Sci.
U.S.A., 89: 4363–4367.
Ehlers, M.D. (2000) Reinsertion or degradation of AMPA re-
ceptors determined by activity-dependent endocytic sorting.
Neuron, 28: 511–525.
Eliot, L.S., Dudai, Y., Kandel, E.R. and Abrams, T.W. (1989)
Ca2+/calmodulin sensitivity may be common to all forms of
neural adenylate cyclase. Proc. Natl. Acad. Sci. U.S.A., 86:
9564–9568.
Faas, G.C., Adwanikar, H., Gereau IV, R.W. and Saggau, P.
(2002) Modulation of presynaptic calcium transients by me-
tabotropic glutamate receptor activation: a differential role in
acute depression of synaptic transmission and long-term de-
pression. J. Neurosci., 22: 6885–6890.
Feinmark, S.J., Begum, R., Tsvetkov, E., Goussakov, I., Funk,
C.D., Siegelbaum, S.A. and Bolshakov, V.Y. (2003) 12-lip-
oxygenase metabolites of arachidonic acid mediate metabo-
tropic glutamate receptor-dependent long-term depression at
hippocampal CA3-CA1 synapses. J. Neurosci., 23:
11427–11435.
Fitzjohn, S.M., Bortolotto, Z.A., Palmer, M.J., Doherty, A.J.,
Ornstein, P.L., Schoepp, D.D., Kingston, A.E., Lodge, D.
and Collingridge, G.L. (1998) The potent mGlu receptor an-
tagonist LY341495 identifies roles for both cloned and novel
496
mGlu receptors in hippocampal synaptic plasticity. Neuro-
pharmacology, 37: 1445–1458.
Fitzjohn, S.M., Kingston, A.E., Lodge, D. and Collingridge,
G.L. (1999) DHPG-induced LTD in area CA1 of juvenile rat
hippocampus; characterization and sensitivity to novel
mGluR antagonists. Neuropharmacology, 38: 1577–1583.
Fitzjohn, S.M., Palmer, M.J., May, J.E., Neeson, A., Morris,
S.A. and Collingridge, G.L. (2001) A characterisation of
long-term depression induced by metabotropic glutamate re-
ceptor activation in the rat hippocampus in vitro. J. Physiol.,
537: 421–430.
Fitzpatrick, J.S. and Baudry, M. (1994) Blockade of long-term
depression in neonatal hippocampal slices by a phospholi-
pase A2 inhibitor. Dev. Brain Res., 78: 81–86.
Freeman, E.J., Damron, D.S., Terrian, D.M. and Dorman,
R.V. (1991) 12-Lipoxygenase products attenuate the gluta-
mate release and calcium accumulation evoked by depolari-
zation of hippocampal mossy fiber nerve endings. J.
Neurochem., 56: 1079–1082.
Gage, A.T., Reyes, M. and Stanton, P.K. (1997) Nitric-oxide-
guanylyl-cyclase-dependent and independent components of
multiple forms of long-term synaptic depression. Hippocam-
pus, 7: 286–295.
Galione, A., White, A., Willmott, N., Turner, M., Potter,
B.V.L. and Watson, S.P. (1993) cGMP mobilizes intracellu-
lar Ca21 in sea urchin eggs by stimulating cyclic ADP-ribose
synthesis. Nature, 365: 456–459.
Gallagher, S.M., Daly, C.A., Bear, M.F. and Huber, K.M.
(2004) Extracellular signal-regulated protein kinase activa-
tion is required for metabotropic glutamate receptor-depend-
ent long-term depression in hippocampal area CA1. J.
Neurosci., 24: 4859–4864.
Garbers, D.L. (1990) The guanylyl cyclase receptor family. New
Biol., 2: 499–504.
Gooney, M., Shaw, K., Kelly, A., O’Mara, S.M. and Lynch,
M.A. (2002) Long-term potentiation and spatial learning are
associated with increased phosphorylation of TrkB and ex-
tracellular signal-regulated kinase (ERK) in the dentate
gyrus: evidence for a role for brain-derived neurotrophic
factor. Behav. Neurosci., 116: 455–463.
Goto, S., Yamamoto, H., Fukunaga, K., Iwasa, T., Matsuk-
ado, Y. and Miyamoto, E. (1985) Dephosphorylation of mi-
crotubule-associated protein 2, tau factor, and tubulin by
calcineurin. J. Neurochem., 45: 276–283.
Grewal, S.S., York, R.D. and Stork, P.J. (1999) Extracellular-
signal-regulated kinase signalling in neurons. Curr. Opin.
Neurobiol., 9: 544–553.
Hagiwara, M., Alberts, A., Brindle, P., Meinkoth, J., Fe-
ramisco, J., Deng, T., Karin, M., Shenolikar, S. and Mont-
miny, M. (1992) Transcriptional attenuation following
cAMP induction requires PP-1-mediated dephosphorylation
of CREB. Cell, 70: 105–113.
Halpain, S., Hipolito, A. and Saffer, L. (1998) Regulation of F-
actin stability in dendritic spines by glutamate receptors and
calcineurin. J. Neurosci., 18: 9835–9844.
Hansel, C., Artola, A. and Singer, W. (1996) Different thresh-
old levels of postsynaptic [Ca2+]i have to be reached to
induce LTP and LTD in neocortical pyramidal cells. J.
Physiol. Paris, 90: 317–319.
Heynen, A.J., Quinlan, E.M., Bae, D.C. and Bear, M.F. (2000)
Bidirectional, activity-dependent regulation of glutamate re-
ceptors in the adult hippocampus in vivo. Neuron, 28:
527–536.
Hilfiker, S., Pieribone, V.A., Czernik, A.J., Kao, H.T., Augus-
tine, G.J. and Greengard, P. (1999) Synapsins as regulators
of neurotransmitter release. Philos. Trans. R. Soc. Lond. B
Biol. Sci., 354: 269–279.
Hosaka, M., Hammer, R.E. and Sudhof, T.C. (1999) A
phospho-switch controls the dynamic association of synap-
sins with synaptic vesicles. Neuron, 24: 377–387.
Hrabetova, S. and Sacktor, T.C. (1996) Bidirectional regulation
of protein kinase M zeta in the maintenance of long-term
potentiation and long-term depression. J. Neurosci., 16:
5324–5333.
Hrabetova, S. and Sacktor, T.C. (2001) Transient transloca-
tion of conventional protein kinase C isoforms and per-
sistent downregulation of atypical protein kinase Mzeta in
long-term depression. Brain Res. Mol. Brain Res., 95:
146–152.
Huang, C.C. and Hsu, K.S. (2006) Sustained activation of me-
tabotropic glutamate receptor 5 and protein tyrosine
phosphatases mediate the expression of (S)-3,5-di-
hydroxyphenylglycine-induced long-term depression in the
hippocampal CA1 region. J. Neurochem., 96: 179–194.
Huang, C.C., You, J.L., Wu, M.Y. and Hsu, K.S. (2004) Rap1-
induced p38 mitogen-activated protein kinase activation
facilitates AMPA receptor trafficking via the GDI.Rab5
complex. Potential role in (S)-3,5-dihydroxyphenylglycene-
induced long term depression. Biol. Chem., 279:
12286–12292.
Huang, K.P. (1989) The mechanism of protein kinase C acti-
vation. Trends Neurosci., 12: 425–432.
Huang, L., Killbride, J., Rowan, M.J. and Anwyl, R. (1999a)
Activation of mGluRII induces LTD via activation of pro-
tein kinase A and protein kinase C in the dentate gyrus of the
hippocampus in vitro. Neuropharmacology, 38: 73–83.
Huang, L.Q., Rowan, M.J. and Anwyl, R. (1997) mGluR II
agonist inhibition of LTP induction, and mGluR II antag-
onist inhibition of LTD induction, in the dentate gyrus in
vitro. Neuroreport, 8: 687–693.
Huang, L., Rowan, M.J. and Anwyl, R. (1999b) Induction
of long-lasting depression by (+)-alpha-methyl-4-car-
boxyphenylglycine and other group II mGlu receptor lig-
ands in the dentate gyrus of the hippocampus in vitro. Eur. J.
Pharmacol., 366: 151–158.
Huang, L.Q., Rowan, M.J. and Anwyl, R. (1999c) Role of
protein kinases A and C in the induction of mGluR depend-
ent long-term depression in the medial perforant path of the
rat dentate gyrus in vitro. Neurosci. Lett., 274: 71–74.
Huber, K.M., Kayser, M.S. and Bear, M.F. (2000) Role for
rapid dendritic protein synthesis in hippocampal mGluR-
dependent long-term depression. Science, 288: 1254–1257.
Huber, K.M., Roder, J.C. and Bear, M.F. (2001) Chemical
induction of mGluR5- and protein synthesis-dependent

long-term depression in hippocampal area CA1. J. Ne-
urophysiol., 86: 321–325.
Izumi, Y. and Zorumski, C.F. (1993) Nitric oxide and long-
term synaptic depression in the rat hippocampus. Neurore-
port, 4: 1131–1134.
Kamal, A., Ramakers, G.M., Urban, I.J., De Graan, P.N. and
Gispen, W.H. (1999) Chemical LTD in the CA1 field of the
hippocampus from young and mature rats. Eur. J. Neurosci.,
11: 3512–3516.
Kameyama, K., Lee, H.K., Bear, M.F. and Huganir, R.L.
(1998) Involvement of a postsynaptic protein kinase A subst-
rate in the expression of homosynaptic long-term depression.
Neuron, 21: 1163–1175.
Kauderer, B.S. and Kandel, E.R. (2000) Capture of a protein
synthesis-dependent component of long-term depression.
Proc. Natl. Acad. Sci. U.S.A., 97: 13342–13347.
Kelleher III, R.J., Govindarajan, A., Jung, H.Y., Kang, H. and
Tonegawa, S. (2004) Translational control by MAPK signa-
ling in long-term synaptic plasticity and memory. Cell, 116:
467–479.
Kemp, N., McQueen, J., Faulkes, S. and Bashir, Z.I. (2000)
Different forms of LTD in the CA1 region of the hippocam-
pus: role of age and stimulus protocol. Eur. J. Neurosci., 12:
360–366.
Keranen, L.M., Dutil, E.M. and Newton, A.C. (1995) Protein
kinase C is regulated in vivo by three functionally distinct
phosphorylations. Curr. Biol., 5: 1394–1403.
Kim, C.H., Chung, H.J., Lee, H.K. and Huganir, R.L. (2001)
Interaction of the AMPA receptor subunit GluR2/3 with
PDZ domains regulates hippocampal long-term depression.
Proc. Natl. Acad. Sci. U.S.A., 98: 11725–11730.
Klausnitzer, J., Kulla, A. and Manahan-Vaughan, D. (2004)
Role of the group III metabotropic glutamate receptor in
LTP, depotentiation and LTD in dentate gyrus of freely
moving rats. Neuropharmacology, 46: 160–170.
Klee, C.B., Crouch, T.H. and Krinks, M.H. (1979) Calcineurin:
a calcium and calmodulin-binding protein of the nervous
system. Proc. Natl. Acad. Sci. U.S.A., 76: 6270–6273.
Kobayashi, M., Manabe, T. and Takahashi, T. (1996) Presy-
naptic long-term depression at the hippocampal mossy fiber-
CA3 synapse. Science, 273: 648–650.
Krezel, W., Giese, K.P., Silva, A.J. and Chapman, P.F. (1999)
Long-term depression is unimpaired in hippocampus of adult
alpha CaMKIIT286A mice. Soc. Neurosci. Abst., 25: 987.
Lee, H.C. (1997) Mechanisms of calcium signaling by cyclic
ADP-ribose and NAADP. Physiol. Rev., 77: 1133–1164.
Lee, H.K., Barbarosie, M., Kameyama, K., Bear, M.F. and
Huganir, R.L. (2000) Regulation of distinct AMPA receptor
phosphorylation sites during bidirectional synaptic plasticity.
Nature, 405: 955–959.
Lee, H.K., Kameyama, K., Huganir, R.L. and Bear, M.F.
(1998) NMDA induces long-term synaptic depression and
dephosphorylation of the GluR1 subunit of AMPA receptors
in hippocampus. Neuron, 21: 1151–1162.
Lee, H.K., Takamiya, K., Han, J.S., Man, H., Kim, C.H.,
Rumbaugh, G., Yu, S., Ding, L., He, C. and Petralia, R.S.
(2003) Phosphorylation of the AMPA receptor GluR1
497
subunit is required for synaptic plasticity and retention of
spatial memory. Cell, 112: 631–643.
Li, S.T., Kato, K., Tomizawa, K., Matsushita, M., Moriwaki,
A., Matsui, H. and Mikoshiba, K. (2002) Calcineurin plays
different roles in group II metabotropic glutamate receptor-
and NMDA receptor-dependent long-term depression. J.
Neurosci., 22: 5034–5041.
Liao, D., Hessler, N.A. and Malinow, R. (1995) Activation of
postsynaptically silent synapses during pairing-induced LTP
in CA1 region of hippocampal slice. Nature, 375(6530):
400–404.
Lin, J.W., Ju, W., Foster, K., Lee, S.H., Ahmadian, G.,
Wyszynski, M., Wang, Y.T. and Sheng, M. (2000) Distinct
molecular mechanisms and divergent endocytotic pathways
of AMPA receptor internalization. Nat. Neurosci., 3:
1282–1290.
Lin, Y.W., Yang, H.W., Wang, H.J., Gong, C.L., Chiu, T.H.
and Min, M.Y. (2006) Spike-timing-dependent plasticity at
resting and conditioned lateral perforant path synapses on
granule cells in the dentate gyrus: different roles of N-methyl-
D-aspartate and group I metabotropic glutamate receptors.
Eur. J. Neurosci., 23: 2362–2374.
Lisman, J. (1989) A mechanism for the Hebb and the anti-Hebb
processes underlying learning and memory. Proc. Natl. Acad.
Sci. U.S.A., 86: 9574–9578.
Luttrell, L.M., Ferguson, S.S., Daaka, Y., Miller, W.E., Ma-
udsley, S., Della Rocca, G.J., Kawakatsu, H., Owada, K.,
Luttrell, D.K., Caron, M.G. and Lefkowitz, R.J. (1999) Beta-
arrestin-dependent formation of beta2 adrenergic receptor-
Src protein kinase complexes. Science, 283: 655–661.
Lynch, G.S., Dunwiddie, T. and Gribkoff, V. (1977) Heterosy-
naptic depression: a postsynaptic correlate of long-term po-
tentiation. Nature, 266: 737–739.
Magee, J.C. and Johnston, D. (1997) A synaptically controlled,
associative signal for Hebbian plasticity in hippocampal neu-
rons. Science, 275: 209–213.
Malenka, R.C. and Bear, M.F. (2004) LTP and LTD: an em-
barrassment of riches. Neuron, 44: 5–21.
Manahan-Vaughan, D. (1997) Group 1 and 2 metabotropic
glutamate receptors play differential roles in hippocampal
long-term depression and long-term potentiation in freely
moving rats. J. Neurosci., 17: 3303–3311.
Manahan-Vaughan, D. and Reymann, K.G. (1995) 1S,3R-
ACPD dose-dependently induces a slow-onset potentiation in
the rat hippocampal CA1 region in vivo. Neuropharm., 34:
1103–1105.
Manahan-Vaughan, D., Kulla, A. and Frey, J.U. (2000) Re-
quirement of translation but not transcription for the main-
tenance of long-term depression in the CA1 region of freely
moving rats. J. Neurosci., 20: 8572–8576.
Markram, H., Lubke, J., Frotscher, M. and Sakmann, B. (1997)
Regulation of synaptic efficacy by coincidence of postsynap-
tic APs and EPSPs. Science, 275: 213–215.
McDonald, B.J., Chung, H.J. and Huganir, R.L. (2001) Iden-
tification of protein kinase C phosphorylation sites within the
AMPA receptor GluR2 subunit. Neuropharmacology, 41:
672–679.
498
McLauchlan, H., Newell, J., Morrice, N., Osborne, A., West,
M. and Smythe, E. (1998) A novel role for Rab5-GDI in
ligand sequestration into clathrin-coated pits. Curr. Biol., 8:
34–45.
Meszaros, L.G., Bak, J. and Chu, A. (1993) Cyclic ADP-ribose
as an endogenous regulator of the non-skeletal type ryano-
dine receptor Ca2+ channel. Nature, 364: 76–79.
Mezey, S., Doyere, V., De Souza, I., Harrison, E., Cambon, K.,
Kendal, C.E., Davies, H., Laroche, S. and Stewart, M.G.
(2004) Long-term synaptic morphometry changes after in-
duction of long-term potentiation and long-term depression
in the dentate gyrus of awake rats are not simply mirror
phenomena. Eur. J. Neurosci., 19: 2310–2318.
Moult, P.R., Schnabel, R., Kilpatrick, I.C., Bashir, Z.I. and
Collingridge, G.L. (2002) Tyrosine dephosphorylation un-
derlies DHPG-induced LTD. Neuropharmacology, 43:
175–180.
Mulkey, R.M., Endo, S., Shenolikar, S. and Malenka, R.C.
(1994) Involvement of a calcineurin/inhibitor-1 phosphatase
cascade in hippocampal long-term depression. Nature, 369:
486–488.
Mulkey, R.M., Herron, C.E. and Malenka, R.C. (1993) An
essential role for protein phosphatases in hippocampal long-
term depression. Science, 261: 1051–1055.
Mulkey, R.M. and Malenka, R.C. (1992) Mechanisms under-
lying induction of homosynaptic long-term depression in area
CA1 of the hippocampus. Neuron, 9: 967–975.
Murray, H.J. and O’Connor, J.J. (2003) A role for COX-2 and
p38 mitogen activated protein kinase in long-term depression
in the rat dentate gyrus in vitro. Neuropharmacology, 44:
374–380.
Murray, H.J. and O’Connor, J.J. (2004) A role for monomeric
G-proteins in synaptic plasticity in the rat dentate gyrus in
vitro. Brain Res., 1000: 85–91.
Naie, K. and Manahan-Vaughan, D. (2005) Investigations of
the protein synthesis dependency of mGluR-induced long-
term depression in the dentate gyrus of freely moving rats.
Neuropharmacology, 49: 35–44.
Nicholls, R.E., Zhang, X.L., Bailey, C.P., Conklin, B.R., Kan-
del, E.R. and Stanton, P.K. (2006) mGluR2 acts through
inhibitory Galpha subunits to regulate transmission and
long-term plasticity at hippocampal mossy fiber-CA3
synapses. Proc. Natl. Acad. Sci. U.S.A., 103: 6380–6385.
Nicoll, R.A., Oliet, S.H. and Malenka, R.C. (1998) NMDA
receptor-dependent and metabotropic glutamate receptor-
dependent forms of long-term depression coexist in CA1
hippocampal pyramidal cells. Neurobiol. Learn. Mem., 70:
62–72.
Normandin, M., Gagne, J., Bernard, J., Lie, R., Miceli, D.,
Baudry, M. and Massicotte, G. (1996) Involvement of the 12-
lipoxygenase pathway of arachidonic acid metabolism in ho-
mosynaptic long term depression of the rat hippocampus.
Brain Res., 730: 40–46.
Nosyreva, E.D. and Huber, K.M. (2005) Developmental switch
in synaptic mechanisms of hippocampal metabotropic gluta-
mate receptor-dependent long-term depression. J. Neurosci.,
25: 2992–3001.
Oliet, S.H., Malenka, R.C. and Nicoll, R.A. (1997) Two distinct
forms of long-term depression coexist in CA1 hippocampal
pyramidal cells. Neuron, 18: 969–982.
O’Mara, S.M., Rowan, M.J. and Anwyl, R. (1995a) Metabo-
tropic glutamate receptor-induced homosynaptic long-term
depression and depotentiation in the dentate gyrus of the rat
hippocampus in vitro. Neuropharmacology, 34: 983–989.
O’Mara, S.M., Rowan, M.J. and Anwyl, R. (1995b) Dantrolene
inhibits long-term depression and depotentiation of synaptic
transmission in the rat dentate gyrus. Neuroscience, 68:
621–624.
Otani, S. and Connor, J.A. (1998) Requirement of rapid Ca2+
entry and synaptic activation of metabotropic glutamate re-
ceptors for the induction of long-term depression in adult rat
hippocampus. J. Physiol., 511: 761–770.
Palmer, M.J., Irving, A.J., Seabrook, G.R., Jane, D.E. and
Collingridge, G.L. (1997) The group I mGlu receptor agonist
DHPG induces a novel form of LTD in the CA1 region of the
hippocampus. Neuropharmacology, 36: 1517–1532.
Paolillo, M., Montecucco, A., Zanassi, P. and Schinelli, S.
(1998) Potentiation of dopamine-induced cAMP formation
by group I metabotropic glutamate receptors via protein
kinase C in cultured striatal neurons. Eur. J. Neurosci., 10:
1937–1945.
Parsley, S.L., Pilgram, S.M., Soto, F., Giese, K.P. and
Edwards, F.A. (2007) Enriching the environment of alpha-
CaMKIIT286A mutant mice reveals that LTD occurs in
memory processing but must be subsequently reversed by
LTP. Learn Mem., 14: 75–83.
Piomelli, D., Wang, J.K., Sihra, T.S., Nairn, A.C., Czernik,
A.J. and Greengard, P. (1989) Inhibition of Ca2+/calmod-
ulin-dependent protein kinase 11 by arachidonic acid and its
metabolites. Proc. Natl. Acad. Sci. U.S.A., 86: 8550–8554.
Pöschel, B. and Manahan-Vaughan, D. (2005) Group II
mGluR-induced long term depression in the dentate gyrus
in vivo is NMDA receptor-independent and does not require
protein synthesis. Neuropharmacology, 49: 1–12.
Pöschel, B. and Manahan-Vaughan, D. (2007) Persistent
(424 h) long-term depression in the dentate gyrus of freely
moving rats is not dependent on activation of NMDA re-
ceptors, L-type voltage-gated calcium channels or protein
synthesis. Neuropharmacology, 52: 46–54.
Pöschel, B., Wroblewska, B., Heinemann, U. and Manahan-
Vaughan, D. (2005) The metabotropic glutamate receptor
mGluR3 is critically required for hippocampal long-term de-
pression and modulates long-term potentiation in the dentate
gyrus of freely moving rats. Cereb. Cortex, 15: 1414–1423.
Qi, M., Zhuo, M., Skalhegg, B.S., Brandon, E.P., Kandel, E.R.,
McKnight, G.S. and Idzerda, R.L. (1996) Impaired hippo-
campal plasticity in mice lacking the Cbeta1 catalytic subunit
of cAMP-dependent protein kinase. Proc. Natl. Acad. Sci.
U.S.A., 93: 1571–1576.
Ramakers, G.M., Heinen, K., Gispen, W.H. and de Graan,
P.N. (2000) Long term depression in the CA1 field is as-
sociated with a transient decrease in pre- and postsynaptic
PKC substrate phosphorylation. J. Biol. Chem., 275:
28682–28687.

Rammes, G., Palmer, M., Eder, M., Dodt, H.U., Zieglgans-
berger, W. and Collingridge, G.L. (2003) Activation of mGlu
receptors induces LTD without affecting postsynaptic sensi-
tivity of CA1 neurons in rat hippocampal slices. J. Physiol.,
546: 455–460.
Reyes, M. and Stanton, P.K. (1996) Induction of hippocampal
long-term depression requires release of Ca2+ from separate
presynaptic and postsynaptic intracellular stores. J. Neuro-
sci., 16: 5951–5960.
Reyes-Harde, M., Empson, R., Potter, B.V., Galione, A. and
Stanton, P.K. (1999a) Evidence of a role for cyclic ADP-
ribose in long-term synaptic depression in hippocampus.
Proc. Natl. Acad. Sci. U.S.A., 96: 4061–4066.
Reyes-Harde, M., Potter, B.V., Galione, A. and Stanton, P.K.
(1999b) Induction of hippocampal LTD requires nitric-oxide-
stimulated PKG activity and Ca2+ release from cyclic ADP-
ribose-sensitive stores. J. Neurophysiol., 82: 1569–1576.
Roberson, E.D., English, J.D., Adams, J.P., Selcher, J.C., Ko-
ndratick, C. and Sweatt, J.D. (1999) The mitogen-activated
protein kinase cascade couples PKA and PKC to cAMP re-
sponse element binding protein phosphorylation in area CA1
of hippocampus. J. Neurosci., 19: 4337–4348.
Rouach, N. and Nicoll, R.A. (2003) Endocannabinoids con-
tribute to short-term but not long-term mGluR-induced de-
pression in the hippocampus. Eur. J. Neurosci., 18:
1017–1020.
Rush, A.M., Wu, J., Rowan, M.J. and Anwy, R. (2002) Group
I metabotropic glutamate receptor (mGluR)-dependent long-
term depression mediated via p38 mitogen-activated protein
kinase is inhibited by previous high-frequency stimulation
and activation of mGluRs and protein kinase C in the rat
dentate gyrus in vitro. J. Neurosci., 22: 6121–6128.
Sajikumar, S. and Frey, J.U. (2003) Anisomycin inhibits the
late maintenance of long-term depression in rat hippocampal
slices in vitro. Neurosci. Lett., 338: 147–150.
Santschi, L., Reyes-Harde, M. and Stanton, P.K. (1999) Chem-
ically induced, activity-independent LTD elicited by simul-
taneous activation of PKG and inhibition of PKA. J.
Neurophysiol., 82: 1577–1589.
Santschi, L.A., Zhang, X.L. and Stanton, P.K. (2006) Activa-
tion of receptors negatively coupled to adenylate cyclase is
required for induction of long-term synaptic depression at
Schaffer collateral-CA1 synapses. J. Neurobiol., 66: 205–219.
Schnabel, R., Kilpatrick, I.C. and Collingridge, G.L. (1999a)
An investigation into signal transduction mechanisms in-
volved in DHPG-induced LTD in the CA1 region of the
hippocampus. Neuropharmacology, 38: 1585–1596.
Schnabel, R., Kilpatrick, I.C. and Collingridge, G.L. (2001)
Protein phosphatase inhibitors facilitate DHPG-induced
LTD in the CA1 region of the hippocampus. Br. J. Pharma-
col., 132: 1095–1101.
Schnabel, R., Palmer, M.J., Kilpatrick, I.C. and Collingridge,
G.L. (1999b) A CaMKII inhibitor, KN-62, facilitates
DHPG-induced LTD in the CA1 region of the hippocam-
pus. Neuropharmacology, 38: 605–608.
Schwartz, J.H. (1993) Cognitive kinases. Proc. Natl. Acad. Sci.
U.S.A., 90: 8310–8313.
499
Seidenman, K.J., Steinberg, J.P., Huganir, R. and Malinow,
R. (2003) Glutamate receptor subunit 2 Serine 880 phos-
phorylation modulates synaptic transmission and mediates
plasticity in CA1 pyramidal cells. J. Neurosci., 23:
9220–9228.
Shi, J., Townsend, M. and Constantine-Paton, M. (2000) Ac-
tivity-dependent induction of tonic calcineurin activity me-
diates a rapid developmental downregulation of NMDA
receptor currents. Neuron, 28: 103–114.
Shields, S.M., Ingebritsen, T.S. and Kelly, P.T. (1985) Identi-
fication of protein phosphatase 1 in synaptic junctions: de-
phosphorylation of endogenous calmodulin-dependent
kinase II and synapse-enriched phosphoproteins. J. Neuro-
sci., 5: 3414–3422.
Snyder, E.M., Philpot, B.D., Huber, K.M., Dong, X., Fallon,
J.R. and Bear, M.F. (2001) Internalization of ionotropic
glutamate receptors in response to mGluR activation. Nat.
Neurosci., 4: 1079–1085.
Son, H., Hawkins, R.D., Martin, K., Kiebler, M., Huang, P.L.,
Fishman, M.C. and Kandel, E.R. (1996) Long-term potent-
iation is reduced in mice that are doubly mutant in end-
othelial and neuronal nitric oxide synthase. Cell, 87:
1015–1023.
Stanton, P.K. (1995) Transient protein kinase C activation
primes long-term depression and suppresses long-term po-
tentiation of synaptic transmission in hippocampus. Proc.
Natl. Acad. Sci. U.S.A., 92: 1724–1728.
Stanton, P.K. (1996) Phospholipase A2 activation is not re-
quired for long-term synaptic depression. Eur. J. Pharmacol.,
273: R7–R9.
Stanton, P.K. and Gage, A.T. (1996) Distinct synaptic loci of
Ca2+/calmodulin-dependent protein kinase II necessary for
long-term potentiation and depression. J. Neurophysiol., 76:
2097–2101.
Stanton, P.K., Heinemann, U. and Muller, W. (2001) FM1-43
imaging reveals cGMP-dependent long-term depression of
presynaptic transmitter release. J. Neurosci., 21: RC167.
Stanton, P.K. and Sarvey, J.M. (1984) Blockade of long-term
potentiation in rat hippocampal CA1 region by inhibitors of
protein synthesis. J. Neurosci., 4: 3080–3088.
Stanton, P.K. and Sejnowski, T.J. (1989) Associative long-term
depression in the hippocampus induced by hebbian covari-
ance. Nature, 339: 215–218.
Stanton, P.K., Winterer, J., Bailey, C.P., Kyrozis, A., Raginov,
I., Laube, G., Veh, R.W., Nguyen, C.Q. and Muller, W.
(2003) Long-term depression of presynaptic release from the
readily releasable vesicle pool induced by NMDA receptor-
dependent retrograde nitric oxide. J. Neurosci., 23:
5936–5944.
Stevens, C.F., Tonegawa, S. and Wang, Y. (1994) The role of
calcium-calmodulin kinase II in three forms of synaptic plas-
ticity. Curr. Biol., 4: 687–693.
Strack, S., Barban, M.A., Wadzinski, B.E. and Colbran, R.J.
(1997) Differential inactivation of postsynaptic density-asso-
ciated and soluble Ca2+/calmodulin-dependent protein kin-
ase II by protein phosphatases 1 and 2A. J. Neurochem., 68:
2119–2128.
500
Sugiyama, H., Ito, I. and Hirono, C. (1987) A new type of
glutamate receptor linked to inositol phospholipid metabo-
lism. Nature, 325: 531–533.
Tan, Y., Hori, N. and Carpenter, D.O. (2003) The mechanism
of presynaptic long-term depression mediated by group I
metabotropic glutamate receptors. Cell Mol. Neurobiol., 23:
187–203.
Tanabe, Y., Masu, M., Ishii, T., Shigemoto, R. and Nakanishi,
S. (1992) A family of metabotropic glutamate receptors.
Neuron, 8: 169–179.
Thiels, E., Kanterewicz, B.I., Knapp, L.T., Barrionuevo, G.
and Klann, E. (2000) Protein phosphatase-mediated regula-
tion of protein kinase C during long-term depression in the
adult hippocampus in vivo. J. Neurosci., 20: 7199–7207.
Thiels, E., Kanterewicz, B.I., Norman, E.D., Trzaskos, J.M.
and Klann, E. (2002) Long-term depression in the adult hip-
pocampus in vivo involves activation of extracellular signal-
regulated kinase and phosphorylation of Elk-1. J. Neurosci.,
22: 2054–2062.
Thiels, E., Norman, E.D., Barrionuevo, C. and Klann, E.
(1998) Transient and persistent increases in protein phos-
phatase activity during long-term depression in the adult
hippocampus in vivo. Neuroscience, 86: 1023–1029.
Thomas, G.M. and Huganir, R.L. (2004) MAPK cascade sig-
nalling and synaptic plasticity. Nat. Rev. Neurosci., 5:
173–183.
Trommer, B.L., Liu, Y.B. and Pasternak, J.F. (1996) Long-
term depression at the medial perforant path-granule cell
synapse in developing rat dentate gyrus. Brain Res. Dev.
Brain Res., 96: 97–108.
Wang, Q., Chang, L., Rowan, M.J. and Anwyl, R. (2007)
Developmental dependence, the role of the kinases p38
MAPK and PKC, and the involvement of tumor necrosis
factor-R1 in the induction of mGlu-5 LTD in the dentate
gyrus. Neuroscience, 144: 110–118.
Wang, Y., Rowan, M.J. and Anwyl, R. (1997) Induction
of LTD in the dentate gyrus in vitro is NMDA receptor
independent, but dependent on Ca2+ influx via low-voltage-
activated Ca2+ channels and release of Ca2+ from intracel-
lular stores. J. Neurophysiol., 77: 812–825.
Wang, Y., Wu, J., Rowan, M.J. and Anywl, R. (1998) Role of
protein kinase C in the induction of homosynaptic long-term
depression by brief low frequency stimulation in the dentate
gyrus of the rat hippocampus in vitro. J. Physiol., 513:
467–475.
Watabe, A.M., Carlisle, H.J. and O’Dell, T.J. (2002) Postsy-
naptic induction and presynaptic expression of group 1
mGluR-dependent LTD in the hippocampal CA1 region. J.
Neurophysiol., 87: 1395–1403.
Wu, J., Rowan, M.J. and Anwyl, R. (2004) Synaptically stim-
ulated induction of group I metabotropic glutamate receptor-
dependent long-term depression and depotentiation is inhib-
ited by prior activation of metabotropic glutamate receptors
and protein kinase C. Neuroscience, 123: 507–514.
Wu, J., Rush, A., Rowan, M.J. and Anwyl, R. (2001) NMDA
receptor- and metabotropic glutamate receptor-dependent
synaptic plasticity induced by high frequency stimulation in
the rat dentate gyrus in vitro. J. Physiol., 533: 745–755.
Wu, J., Wang, Y., Rowan, M.J. and Anwyl, R. (1997) Evidence
for involvement of the neuronal isoform of nitric oxide synt-
hase during induction of long-term potentiation and long-
term depression in the rat dentate gyrus in vitro. Neurosci-
ence, 78: 393–398.
Wu, J., Wang, Y., Rowan, M.J. and Anwyl, R. (1998) Evidence
for involvement of the cGMP-protein kinase G signaling
system in the induction of long-term depression, but not
long-term potentiation, in the dentate gyrus in vitro. J.
Neurosci., 18: 3589–3596.
Xiao, M.Y., Zhou, Q. and Nicoll, R.A. (2001) Metabotropic
glutamate receptor activation causes a rapid redistribution of
AMPA receptors. Neuropharmacology, 41: 664–671.
Yang, S.N., Tang, Y.G. and Zucker, R.S. (1999) Selective in-
duction of LTP and LTD by postsynaptic [Ca2]i elevation. J.
Neurophysiol., 81: 781–787.
Zakharenko, S.S., Zablow, L. and Siegelbaum, S.A. (2002)
Altered presynaptic vesicle release and cycling during
mGluR-dependent LTD. Neuron, 35: 1099–1110.
Zhang, X.L., Zhou, Z.Y., Winterer, J., Muller, W. and Stanton,
P.K. (2006) NMDA-dependent, but not group I metabo-
tropic glutamate receptor-dependent, long-term depression at
Schaffer collateral-CA1 synapses is associated with long-term
reduction of release from the rapidly recycling presynaptic
vesicle pool. J. Neurosci., 26: 10270–10280.
Zhou, Q., Homma, K.J. and Poo, M.M. (2004) Shrinkage
of dendritic spines associated with long-term depression of
hippocampal synapses. Neuron, 44: 749–757.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 27

Structural reorganization of the dentate gyrus

following entorhinal denervation: species differences
between rat and mouse

Thomas Deller, Domenico Del Turco, Angelika Rappert and Ingo Bechmann

Institute of Clinical Neuroanatomy, J.W. Goethe-University, Theodor-Stern-Kai 7, D-60590 Frankfurt/Main, Germany

Abstract: Deafferentation of the dentate gyrus by unilateral entorhinal cortex lesion or unilateral perforant
pathway transection is a classical model to study the response of the central nervous system (CNS) to
denervation. This model has been extensively characterized in the rat to clarify mechanisms underlying
denervation-induced gliosis, transneuronal degeneration of denervated neurons, and collateral sprouting of
surviving axons. As a result, candidate molecules have been identified which could regulate these changes,
but a causal link between these molecules and the postlesional changes has not yet been demonstrated. To
this end, mutant mice are currently studied by many groups. A tacit assumption is that data from the rat
can be generalized to the mouse, and fundamental species differences in hippocampal architecture and the
fiber systems involved in sprouting are often ignored. In this review, we will (1) provide an overview of some
of the basics and technical aspects of the entorhinal denervation model, (2) identify anatomical species
differences between rats and mice and will point out their relevance for the axonal reorganization process,
(3) describe glial and local inflammatory changes, (4) consider transneuronal changes of denervated dentate
neurons and the potential role of reactive glia in this context, and (5) summarize the differences in
the reorganization of the dentate gyrus between the two species. Finally, we will discuss the use of the
entorhinal denervation model in mutant mice.

Keywords: entorhinal cortex lesion; perforant pathway transection; sprouting; transneuronal degeneration;
glia; inflammation; regeneration

Introduction removal of the cellular debris, surviving non-

Injuries to the central nervous system (CNS) cause
local damage at the lesion site as well as denerva-
tion of connected brain regions. In response to this
denervation, complex cellular changes occur in the
denervated brain area (see Steward, 1994b, for re-
view): Glial cells react and participate in the
Corresponding author. Tel.: +49-69-6301-6361;
Fax: +49-69-6301-6425; E-mail: T.Deller@em.uni-frankfurt.de
lesioned axons sprout and reinnervate the dener-
vated target cells, and denervated neurons remodel
their dendritic arbor in response to denervation
and reinnervation. In the past, much effort has
been devoted to understanding these changes at
the molecular level. The rationale for this is that
the denervation of brain areas is a very common
and detrimental side effect of all kinds of brain
lesions (Steward, 1994b). It can occur following
traumatic brain damage, ischemia, inflammation,

DOI: 10.1016/S0079-6123(07)63027-1 501

502
and neurodegeneration and may contribute sig-
nificantly to the severity of the clinical symptoms
of these and other brain diseases. It is one of the
motivations of the scientists working in the field
today that understanding the reorganization proc-
esses in denervated brain areas and learning how
to influence them in a beneficial fashion may lead
to new treatment strategies for brain diseases.
One of the classical model systems used to study
the reorganization of denervated brain regions is
the reorganization of the rat dentate gyrus follow-
ing entorhinal denervation (Lynch and Cotman,
1975; Cotman and Nadler, 1978; Gall and Lynch,
1980; Gall et al., 1986; Steward, 1991; Deller and
Frotscher, 1997; Frotscher et al., 1997; Collazos-
Castro and Nieto-Sampedro, 2001; Deller et al.,
2001; Ramirez, 2001; Savaskan and Nitsch, 2001).
This model is well established, and structural,
functional, and molecular changes have been stud-
ied in detail over the years. A fairly large number
of candidate molecules have been identified that
could regulate the reorganization of the dentate
gyrus, including growth-associated molecules,
neurotrophic factors, cell adhesion molecules,
extracellular matrix molecules, and a variety
of cytokines (see Deller and Frotscher, 1997;
Collazos-Castro and Nieto-Sampedro, 2001;
Deller et al., 2001; Savaskan and Nitsch,
2001, for review). In the majority of studies,
however, correlations between molecular changes
and the reorganization process were reported, so
it remains to be seen that these candidate mole-
cules are causally linked to the postlesional re-
organization process. An attractive experimental
strategy to address these issues of causality is to
use genetically engineered mice and to analyze
the reorganization of the dentate gyrus following
entorhinal denervation in these mutants (Fagan
et al., 1998; Finsen et al., 1999; Jensen et al.,
1999; Jensen et al., 2000b; Del Turco et al., 2001,
2003; White et al., 2001a, b; Kadish and van
Groen, 2003; Rappert et al., 2004).
Although the entorhinal denervation model has
now frequently been used in mice, the reorganiza-
tion of the denervated dentate gyrus has only been
partially characterized in this species. In fact, in
the absence of data from the mouse, results ob-
tained after entorhinal denervation in mice have
often been interpreted solely on the basis of data
obtained in rats. It is assumed that the reorgan-
ization of the mouse dentate gyrus is more or less
similar to that of the rat dentate gyrus following
entorhinal deafferentation. However, this assump-
tion may not be true, given the important biolog-
ical differences between these two species and even
between different mouse strains.
Therefore, it may be prudent to keep data from
the two species separate. In this review, we will
compare the reorganization of the dentate gyrus
following entorhinal denervation in rats and mice.
We will first give an overview over some of the
fundamentals and technical aspects of the model
system. Second, we will identify species differences
between the non-lesioned dentate gyrus of rats and
mice and will point out their relevance for the ax-
onal reorganization process. Third, glial changes
occurring in the dentate gyrus of both species after
denervation will be described and considered as a
local inflammatory response of the brain. Fourth,
transneuronal changes of denervated dentate neu-
rons will be considered, and the potential role of
reactive glia in this context will be reviewed. Fifth,
differences in the reorganization of the dentate
gyrus between the two species will be summarized
and, finally, the use of the entorhinal lesion model
in mutant mice will be discussed.
Entorhinal denervation in rats and mice: technical
aspects
The entorhinal denervation model was established
in the 1970s, when it was first shown that unilat-
eral removal of entorhinal afferents to the rat
dentate gyrus results in degenerative and regener-
ative changes in the so-called ‘‘entorhinal’’ zone of
the molecular layer, i.e., the outer two-thirds of the
molecular layer (Lynch et al., 1972; Cotman and
Nadler, 1978). This model is considered particu-
larly useful, because of the relatively simple and
highly laminated cyto- and fiber architecture of the
dentate gyrus, which makes it possible to distin-
guish between denervated and non-denervated lay-
ers and enables researchers to differentiate
between the various cellular changes that occur
within a denervated layer or in its vicinity (Cotman

and Nadler, 1978; Steward, 1991; Deller et al.,
2001).
A confusing element of the entorhinal denerva-
tion model in rats is the fact that the method to
produce the lesion has changed over the years. In
the beginning, entorhinal cortex lesions (ECL)
were performed using electrolytic damage to the
entorhinal cortex itself, thereby removing the cells
of origin of the perforant pathway and denervat-
ing the dentate gyrus (Lynch et al., 1972; Steward
et al., 1974; Zimmer and Hjorth-Simonsen, 1975;
Rose et al., 1976; Gall et al., 1979a, b; Poirier
et al., 1990; Parent et al., 1993; Guthrie et al., 1995,
1997; Jucker et al., 1996). Later, other methods
were also used, for example, mechanical or elec-
trolytic transection of the perforant pathway
(Benowitz et al., 1990; Nitsch and Frotscher,
1992; Nitsch et al., 1992; Beck et al., 1993; Gwag
et al., 1994; Deller et al., 1995a) or chemical lesions
(Ueki et al., 1996). Often, these lesioning tech-
niques are regarded as more or less equivalent,
since all of them result in a robust denervation of
the dentate gyrus. However, they may differ in
their acute and chronic effects on the dentate
gyrus: whereas electrolytic lesions strongly stimu-
late the perforant pathway and may have an
epileptogenic effect, non-electrolytic lesions are
less likely to do so (Dasheiff and McNamara,
1982; Campbell et al., 1984; Kelley and Steward,
1996a). Since a variety of molecules and genes are
regulated by activity, gene expression — at least
during the initial phase of the reorganization proc-
ess — could be influenced by the lesioning tech-
nique (Kelley and Steward, 1996b, 1998; Steward
et al., 1997). Moreover, as the extent of necrotic
and apoptotic degeneration certainly varies be-
tween the different methods, it is likely that there
are also fundamental differences with regard to the
induced inflammatory and immune responses. In
addition to such differences relating to the lesion-
ing techniques, differences in rat strains were
rarely considered. In summary, data from the rat
ECL/perforant pathway transection model are
plentiful, yet some caution is warranted when re-
sults from two labs using different lesioning tech-
niques or different rat strains are compared.
In mice, mechanical transection of the perforant
pathway using wire and plate knifes has been the
503
preferred method to de-entorhinate the mouse
dentate gyrus (Fig. 1; Finsen et al., 1999; Jensen
et al., 1999, 2000b; Drojdahl et al., 2002, 2004;
Ying et al., 2002; Del Turco et al., 2003; Kovac
et al., 2004; Rappert et al., 2004; Nielsen et al.,
2006; Pedersen et al., 2006). Chemical lesions of
the entorhinal cortex itself were less frequently
performed (Kadish and van Groen, 2002, 2003;
Kadish et al., 2002). Thus, the lesioning techniques
differ between rats and mice; in mice the mechan-
ical perforant pathway transection is the most
common, whereas electrolytic lesions are more
typical in rats.
Finally, the unilateral ECL/perforant pathway
transection model should also be distinguished from
several related models that have also been employed
to study reactive changes in the denervated dentate
gyrus in the past. Among these models are bilateral
ECL (Lynch et al., 1976; Hardman et al., 1997;
Faillace et al., 2006), a combination of ECL
with fimbria-fornix-transection (Beck et al., 1993;
Peterson, 1994), and fimbria-fornix-transection with-
out ECL (Hjorth-Simonsen and Jeune, 1972; Lynch
et al., 1974). All of these models remove additional or
neurochemically different projection systems to the
dentate gyrus, resulting in an altered reorganization
process compared to unilateral ECL/perforant path-
way transection.
The non-denervated dentate gyrus: neuroanatomical
species differences
The normal anatomy of the dentate gyrus is the basis
for a meaningful interpretation of experimentally
induced changes in this brain region. In the rat, the
neuroanatomy of the dentate gyrus has been thor-
oughly investigated during the last 50 years. The
cellular architecture of neuronal and non-neuronal
cells, the connectivity, and the basic neurochemistry
of the dentate gyrus were studied using anatomical
techniques for cellular labeling, retro- and antero-
grade tracing, and immunohistochemistry at the
light- and electron-microscopic level. These neuro-
anatomical facts have been reviewed in numerous
papers and book chapters (e.g., Frotscher, 1988;
Amaral and Witter, 1995) as well as in other chap-
ters of this book. In contrast, the mouse dentate
504

Fig. 1. The mouse dentate gyrus after entorhinal denervation. (A, B) Fink-Heimer degeneration stain in a control mouse hippocampus
(A) and 5 days after entorhinal denervation (B). Note darkly stained degeneration product in the denervated zone of the molecular
layer (B, arrowheads). (C, D) Fluoro-Jade C degeneration stain of the contralateral (C) and ipsilateral (D) mouse dentate gyrus four
days after denervation. Note increased reactivity throughout the denervated portion of the molecular layer (D). (E, F) Acetylcho-
linesterase (AChE) staining of the control mouse dentate gyrus (E) and 6 weeks after entorhinal denervation (F). A dense AChE-
positive fiber band is present in the outer molecular layer (OML) and middle molecular layer (MML; arrowheads). EC, entorhinal
cortex; CA1, CA3, Hippocampal subfields; GCL, granule cell layer; IML, inner molecular layer; ML; molecular layer; H, hilus. Scale
bars: A, B: 300 mm; C–F: 100 mm.

gyrus has not been as thoroughly studied. Although
data exist on its general cellular and fiber architec-
ture (e.g., Stanfield and Cowan, 1979a, b; Stanfield
et al., 1979; Soriano et al., 1994; Supèr and Soriano,
1994; Deller et al., 1999a, b, 2002; Drakew et al.,
2002; Gebhardt et al., 2002; Del Turco et al., 2004),
data on the connectivity of identified cells are scarce.
Only a handful of tracing studies were performed in
mice, in which tracing was combined with electron
microscopy to demonstrate synaptic connections of
identified neurons (Blasco-Ibanez and Freund, 1997;
Deller et al., 1999b; Del Turco et al., 2003; Otal
et al., 2006). Similarly, the topographical organiza-
tion of the afferent projection systems, the cells of
origin of the various pathways, and the neurochem-
istry of the target cells have not been studied in de-
tail. Strain differences have rarely been considered.
Thus, neuroanatomical data on the non-denervated
mouse dentate gyrus are often incomplete and lack
the depth and precision of the rat data.
Nevertheless, on the basis of the available neuro-
anatomical data several anatomical differences are
evident, which are likely to result in significant
differences in the reorganization of the dentate gyrus
following entorhinal denervation. One relevant
difference concerns the ipsilateral entorhinal projec-
tion to the dentate gyrus, i. e., the projection system
that is lost as a consequence of ECL/perforant
pathway transection. This projection has been stud-
ied in rats (e.g. Blackstad, 1958; Amaral and Witter,
1989; Deller, 1998) and mice (Deller et al., 1999a;
van Groen et al., 2002, 2003) using anterograde
tracing and almost identical patterns of termination
were observed: In both species, the majority of
entorhinodentate fibers terminates in a laminar
fashion in the outer portion of the molecular
layers of the dentate gyrus and even individual
axons have a very similar morphology. However,
the portion of the molecular layer covered by ento-
rhinal fibers appears to be greater in the mouse. In
this species, entorhinal fibers cover approximately
four-fifth of the molecular layer (Deller et al., 1999a;
van Groen et al., 2002, 2003) compared to two-
thirds in rats (Blackstad, 1958; Amaral and Witter,
1989; Deller, 1998). Thus, entorhinal denervation
in mice will result in a more extensive denervation
of granule cell dendrites than in rats. A second
relevant difference concerns the commissural and
505
associational (C/A) projection to the inner molecu-
lar layer. Although this projection is formed by
mossy cells in both species, these cells contain
calretinin in mice but not in rats (Liu et al., 1996;
Blasco-Ibanez and Freund, 1997; Fujise et al., 1998).
In addition, the C/A projection, in mice covers only
one-fifth of the molecular layer in contrast to one-
third in rats, is more diffusely organized and
the border of the C/A fiber plexus towards the
entorhinal zone of the dentate gyrus is indistinct
(Figs. 2a, b), suggesting that the border between the
two layers is not as strict as in the rat (Del Turco
et al., 2003). A third and striking anatomical
difference relates to the crossed entorhinodentate
projection. This projection system seems to be non-
existent in mice (van Groen et al., 2002, 2003),
whereas it is readily detectable in rats (Steward
et al., 1974; Steward, 1976a; Deller et al., 1996a).
Thus, the entorhinodentate pathway and at least
two other fiber systems are different in rats and
mice, which could influence the pattern of axonal
reorganization in the denervated dentate gyrus.
Reorganization of axons and fiber systems in the
denervated dentate gyrus
The reorganization and collateral sprouting of
non-lesioned axons in the denervated dentate
gyrus is probably the most prominent and best-
known aspect of the entorhinal denervation model
in rats. It was also the first to be described (Lynch
et al., 1972). When discussing this aspect of the
entorhinal denervation model, we should distin-
guish between changes at the level of overall
number of synapses and terminals, and changes
observed at the level of identified fiber systems.
Changes in synapse and bouton densities following
entorhinal denervation in rat and mouse
In rats, degenerative and regenerative changes of
synapses and terminals have been studied in detail.
Using electron microscopy, synapse counts follow-
ing ECL show 80–90% of all synapses are lost in the
denervated zone of the dentate gyrus, followed by
reactive changes that replace up to 60–80% of all
506
 507

lost synapses by 4 weeks postlesion (Matthews et al.,
1976a, b; Steward and Vinsant, 1983). In a recent
stereological study, these data were largely con-
firmed, although synapse recovery observed with the
more sensitive stereological tools was somewhat less
than reported before (Marrone et al., 2004b). To
visualize total changes in presynaptic terminals,
immunohistochemical markers for presynaptic bou-
tons such as synaptophysin have also been em-
ployed in rats (Jucker et al., 1996; Miwa and Ueki,
1996; Stone et al., 1998; Kadish and van Groen,
2003). Although changes in synaptophysin-positive
boutons following entorhinal denervation seem to
correlate with the quantitative electron microscopic
data, a systematic comparison of the two techniques
has not yet been performed, and it remains to be
seen whether or not changes in synaptophysin-
positive puncta accurately reflect changes at the
ultrastructural level.
For the denervated mouse dentate gyrus, virtu-
ally no quantitative electron-microscopic data are
available. Thus, the extent of both synapse loss and
reactive synaptogenesis are unknown. Entorhinal
denervation has, however, been shown qualitatively
in mice using silver impregnation of degenerating
fibers and terminals (Fig. 1; Steward, 1992; Del
Turco et al., 2003). In addition, stereological esti-
mates of the total volume of the dentate gyrus fol-
lowing entorhinal denervation indicate that the
dentate gyrus shrinks to 60% of its original size,
suggesting that considerable degenerative changes
occur following entorhinal denervation (Phinney
et al., 2004). Similarly, entorhinal reinnervation has
to some extent been demonstrated qualitatively.
Synaptophysin-immunostaining was used to dem-
onstrate changes in presynaptic terminals and
revealed an increase in staining in the denervated
dentate gyrus compared to the contralateral side by
4 weeks survival time (van Groen et al., 2002;
Kadish and van Groen, 2002, 2003).
Sprouting of identified fiber systems following
entorhinal denervation
The recovery of synapse densities following ento-
rhinal denervation in rats and mice raises the
question which fibers contribute to this reinnerva-
tion and whether this process is, in fact, restorative
in nature. As axons from the ipsilateral entorhinal
cortex are completely removed from the dentate
gyrus and these fibers cannot regenerate in the
adult brain, reinnervation apparently occurs from
non-lesioned axons. The new connections thus
differ anatomically as well as functionally from the
ones removed by entorhinal denervation.
In rats, the sprouting fiber systems were iden-
tified using histochemical as well as retrograde
and anterograde tracing methods. Four fiber sys-
tems, which are already present in the non-
esioned dentate gyrus, are believed to contribute
to the reinnervation process (see Cotman
and Nadler, 1978; Steward, 1994b; Deller and
Frotscher, 1997, for review): (i) glutamatergic
C/A fibers to the inner molecular layer (mossy cell
axons), (ii) GABAergic C/A fibers to the outer
molecular layer (iii) glutamatergic cholinergic
septohippocampal fibers, and (iv) crossed ento-
rhinodentate fibers. In the following paragraphs,
we will consider the contribution of each of these
fiber systems to the reinnervation process and will
point out potentially relevant species differences.

Fig. 2. Sprouting of commissural/associational fibers in the mouse dentate gyrus following entorhinal denervation. (A, B) Dentate
gyrus of a control mouse shown at low (A) and high magnification (B). The commissural projections to the inner molecular layer (IML)
and the middle and outer molecular layer (OML; arrow in B) were labeled using Phaseolus vulgaris-leucoagglutinin (PHAL)-tracing.
(C, D) Dentate gyrus of a mouse 2 months after entorhinal denervation shown at low (C) and high (D) magnification. Compared with
the control, the commissural fiber plexus has increased in width and PHAL-labeled commissural fibers are regularly observed in the
MML (arrows). (E, F) Dentate gyrus of a control mouse stained for calretinin shown at low (E) and high (F) magnification. Calretinin-
immunostaining labels mossy cells in the hilus (H) and interneurons in the molecular layer. Calretinin-immunoreactive mossy cell
axons are abundant in the IML. (G, H) Dentate gyrus of a mouse 2 months following entorhinal denervation shown at low (G) and
high (H) magnification. Note that compared to the control, the density of calretinin-immunoreactive fibers has increased throughout
the molecular layer and that the calretinin-positive fiber plexus in the IML has expanded in width (H). CA3, hippocampal subfield;
GCL, granule cell layer; H, hilus; OML, outer molecular layer; IML, inner molecular layer; MML, middle molecular layer. Scale bars:
A, C, E, G: 100 mm; B, D, F, H: 25 mm.
508
Reorganization of commissural/associational mossy
cell axons after entorhinal denervation
As mentioned above, the C/A projection to the
inner molecular layer of the dentate gyrus in the
rat arises from glutamate-immunoreactive mossy
cells (Soriano and Frotscher, 1994). These neurons
are located in the hilus and project both contra-
laterally (commissural fibers) and ipsilaterally
(associational fibers). Their axons terminate in
a characteristic laminar pattern in the inner one-
third of the molecular layer, i.e., complementary to
the entorhinodentate fiber projection to the outer
two-thirds of the molecular layer (Amaral and
Witter, 1995). The contribution of this projection
to the reorganization of the rat dentate gyrus is
age-dependent (Gall and Lynch, 1978, 1980, 1981;
Stäubli et al., 1984; Gall et al., 1986): In young
rats, injured within the first two-postnatal weeks,
extensive sprouting of C/A axon collaterals into
the deafferented adjacent outer molecular layer
can be observed and some of these fibers extend
all the way to the hippocampal fissure. The degree
of this reactive growth response diminishes with
maturation, and is almost completely lost in the
adult rat brain (Gall and Lynch, 1981; Deller et al.,
1996c). In addition, a 30–40 mm outward expan-
sion of the entire C/A fiber plexus was also
observed (Cotman and Nadler, 1978; Gall and
Lynch, 1981). Although the nature of this
phenomenon was a matter of debate for many
years (see Deller and Frotscher, 1997, for review),
available data support the interpretation that the
expansion of the C/A plexus is not the result
of translaminar sprouting but rather caused by
reorganization and shrinkage processes occurring
within the non-denervated inner molecular layer
itself (Hoff et al., 1981; Caceres and Steward, 1983;
Frotscher et al., 1997; Marrone et al., 2004a,
b; Deller et al., 2006). We conclude, therefore, that
(1) C/A mossy cell axons do not leave their
home territory to invade the denervated zone
in adult rats, and (2) axonal remodeling occurs
not only in the denervated zone but also within
the adjacent non-denervated inner molecular
layer.
In adult C57BL/6 mice, the C/A projection to
the inner molecular layer also arises from mossy
cells in the hilus (Blasco-Ibanez and Freund,
1997). In mice, these cells contain the calcium-
binding protein calretinin (Figs. 2e, f), which
facilitates the identification of their axons in
the inner molecular layer (Liu et al., 1996;
Blasco-Ibanez and Freund, 1997; Fujise et al.,
1998; Deller et al., 1999a; Del Turco et al., 2003).
The contribution of this projection to the reor-
ganization of the mouse dentate gyrus following
entorhinal denervation was first studied by Shi and
Stanfield (1996), who used horseradish peroxidase
tracing and observed an ingrowth of C/A fibers
into the denervated molecular layer (Shi and Stan-
field, 1996). We confirmed and extended these data
using anterograde Phaseolus vulgaris-leucoagglutinin
(PHAL)-tracing in combination with immuno-
staining for calretinin at the light as well as the
electron-microscopic level (Del Turco et al., 2003).
In this study we showed that, in contrast to the
adult rat, many mossy cell axons leave the main
C/A fiber plexus in the inner molecular layer in
response to entorhinal denervation, invade the
denervated adjacent zone (Figs. 2c, d, g, h), and
form new synapses outside their normal terminal
zone. This pattern is striking in its distinction from
the layer-specific sprouting in adult rats and is
rather reminiscent of the sprouting seen in imma-
ture rats (Gall and Lynch, 1981).
In addition, an expansion of the entire C/A
termination zone was also observed in mice
(Del Turco et al., 2003; Phinney et al., 2004).
Phinney et al. (2004) studied this expansion using
three-dimensional stereological tools and com-
pared the expansion of the inner molecular layer
in single sections with changes of the total volume
of the inner molecular layer following denervation.
Although this study confirmed a significant
increase in the width of the C/A termination zone
using single sections, it did not detect an increase
in its total volume. In fact, shrinkage of the dent-
ate gyrus in all three dimensions, especially the
compression of the dentate gyrus along the septo-
temporal axis, seemed to underlie the expansion.
The authors concluded that the expansion of the
inner molecular layer does not indicate an increase
in the size of this lamina, and therefore should not
be used as an indicator of C/A sprouting in
C57BL/6 mice.

Reorganization of GABAergic commissural/
associational axons after entorhinal denervation
The GABAergic C/A projection in rats that ter-
minates in the outer molecular layer of the dentate
gyrus (Amaral and Witter, 1989; Deller et al.,
1995a, b, 1996b; Zappone and Sloviter, 2001) also
exists in mice (Deller et al., 1999a; Gebhardt et al.,
2002; Del Turco et al., 2004).
In rats, the reaction of this projection was studied
at the single-fiber level using anterograde tracing,
electron microscopy, and GABA-postembedding
(Deller et al., 1995a). This study demonstrated that
GABAergic commissural axons sprout new collat-
erals within the denervated zone and form synapses
there. We concluded from this observation that
(1) GABAergic fibers can sprout and reinnervate the
dentate gyrus in response to the loss of glutamatergic
afferents, and (2) that C/A fiber sprouting in the
adult rat is layer specific. Thus, both the C/A fibers
to the outer molecular layer as well as C/A fiber to
the inner molecular layer sprout within their home
territories in response to entorhinal denervation
in adult rats (Deller et al., 1995a, 1996c; Deller and
Frotscher, 1997; Frotscher et al., 1997). In mice,
a systematic analysis of the GABAergic C/A pro-
jection to the molecular layer following entorhinal
denervation has not yet been reported.
Reorganization of acetylcholinesterase (AChE):
positive fibers after entorhinal denervation
A third projection system contributing to the reor-
ganization of the dentate gyrus following entorhinal
denervation is the cholinergic septohippocampal
projection. In non-lesioned rats, this projection arises
from cholinergic neurons located in the diagonal
band of Broca (Nyakas et al., 1987; Jakab and
Leranth, 1995). In contrast to the C/A and ent-
orhinal fibers, this projection system terminates in all
layers of the rat and mouse dentate gyrus (Amaral
and Witter, 1995). To visualize the cholinergic septo-
hippocampal fibers in the rat dentate gyrus, acetyl-
cholinesterase (AChE) histochemistry has been
widely used, since it has been shown to be suffi-
ciently specific in this brain area (Eckenstein and
Sofroniew, 1983; Levey et al., 1983, 1984; Naumann
et al., 1997).
509
Following entorhinal denervation in rats,
changes in the density of AChE-fibers in the den-
ervated zone were observed and interpreted as a
sign of cholinergic sprouting (Lynch et al., 1972;
Nadler et al., 1977a, b; Zimmer et al., 1986). These
observations were later corroborated by antero-
grade tracing studies of septohippocampal fibers
(Nadler and Evenson, 1982; Stanfield and Cowan,
1982; Nyakas et al., 1988). In spite of these data,
however, doubts remained. It was pointed out that
changes in AChE-staining could also be the result
of denervation-induced changes in gene expression
(Chen et al., 1983; McKeon et al., 1989) or tissue
shrinkage (Storm-Mathisen, 1974). In addition,
biochemical studies did not reveal any evidence for
cholinergic sprouting (Aubert et al., 1994), and a
non-stereological quantitative analysis of choline
acetyltransferase (ChAT)-positive fibers did not
confirm the AChE-data (Henderson et al., 1998).
Thus, it was questioned whether AChE-histo-
chemistry is an appropriate technique to visualize
cholinergic sprouting in the denervated rat dentate
gyrus (Aubert et al., 1994) and whether cholinergic
sprouting occurs at all (Storm-Mathisen, 1974).
The first of these issues could be resolved by using
the immunotoxin 192 IgG-saporin, which selec-
tively destroys cholinergic neurons in the basal
forebrain (Wiley et al., 1991; Book et al., 1992;
Heckers et al., 1994). Following entorhinal dener-
vation, rats were treated with 192 IgG-saporin and
the AChE-positive fiber band could be completely
abolished (Naumann et al., 1997). Thus, AChE-
histochemistry is appropriate to visualize septo-
hippocampal fibers selectively in the denervated
dentate gyrus. However, the second of these issues
yet awaits clarification. Since it will be essential to
rule out shrinkage as a cause, cholinergic sprout-
ing should be reinvestigated in rats using assump-
tion-free stereological methodology.
In non-lesioned mice, the septohippocampal pro-
jection appears to be generally similar to the one
observed in rats (Linke et al., 1994). Similarly, in
denervated mice, an increase in the density of
AChE-positive fibers is observed in the denervated
outer molecular layer (Steward, 1992, 1994a; Shi
and Stanfield, 1996; Kadish and van Groen, 2002;
Phinney et al., 2004). Analogous to the situation in
rats, this increase in AChE-fiber density (Figs. 1e, f)
510
was interpreted as a sign of cholinergic sprouting
(Steward, 1992, 1994a; Shi and Stanfield, 1996;
Kadish and van Groen, 2002). To verify this finding,
and to account for tissue shrinkage that occurs in
the denervated dentate gyrus, a three-dimensional
stereological analysis of total AChE-fiber length was
performed (Phinney et al., 2004). Although this
study confirmed a significant increase in the density
of AChE-fibers in single sections, an increase in the
total length of AChE fibers could not be detected.
Stereological measurements of the volume of the
non-lesioned and denervated dentate gyrus revealed
that shrinkage of the dentate gyrus in all three di-
mensions, rather than growth of fibers was respon-
sible for the increase in AChE-fiber density. The
authors concluded that the increase in AChE fiber
density in the denervated outer molecular layer
should not be used as an indicator of lesion-induced
cholinergic sprouting in mice. In summary, there is
at present no direct evidence for sprouting of septo-
hippocampal fibers following entorhinal denerva-
tion in mice.
Reorganization of crossed entorhinodentate fibers
after entorhinal denervation
A fourth projection system, contributing to the
reorganization of the denervated dentate gyrus in
rats, is the crossed entorhinodentate (temporo-
dentate) pathway. Its cells of origin, trajectory,
and its termination pattern have been well char-
acterized using retrograde and anterograde tracing
techniques (Steward et al., 1974; Goldowitz et al.,
1975; Steward, 1976b, 1980; Steward and Scoville,
1976; Steward and Vinsant, 1978; Amaral and
Witter, 1995; Deller et al., 1996a). These studies
revealed that the crossed entorhinodentate projec-
tion is weaker but anatomically homologous to the
ipsilateral pathway. From a functional point of
view this homology makes the crossed entorhino-
dentate pathway particularly interesting, because
its sprouting could contribute to a restoration
of function following entorhinal denervation. Ac-
cordingly, morphological, electrophysiological,
and behavioral changes caused by the reorganiza-
tion of crossed entorhinodentate fibers were stud-
ied in detail (Steward et al., 1976, 1988; Steward,
1976a; Cotman et al., 1977). Morphological anal-
ysis demonstrated robust sprouting of the crossed
entorhinodentate projection: It increases its fiber
and terminal density by sixfold and of its synapse
density by approximately 100-fold. The reorgani-
zation of single fibers was also revealed using
anterograde PHAL-tracing (Deller et al., 1996a).
Whereas crossed entorhinodentate axons showed
only few branches and formed almost exclusively
en passant boutons in non-lesioned animals,
crossed entorhinodentate fibers displayed high
densities of small axonal extensions and tangle-
like formations resembling axonal and synaptic
clusters in denervated rats. In keeping with these
morphological data, electrophysiological studies
showed that the functional characteristics of the
crossed entorhinodentate projection become sim-
ilar to the ipsilateral projection following dener-
vation (Steward et al., 1973; White et al., 1976;
Harris et al., 1978; Reeves and Steward, 1986,
1988). Finally, behavioral experiments confirmed
that sprouting of crossed entorhinodentate fibers
can indeed ameliorate some, but not all, of the
behavioral deficits observed after unilateral ento-
rhinal denervation (Loesche and Steward, 1977;
Scheff and Cotman, 1977; Ramirez and Stein,
1984; Schenk and Morris, 1985). In summary,
these data indicate that (1) crossed entorhinodent-
ate sprouting contributes significantly to the
reinnervation of the rat dentate gyrus after ento-
rhinal denervation, and (2) this sprouting of a
homologous fiber system is probably functionally
restorative.
In mice, the crossed entorhinodentate pathway
seems to be essentially non-existent (van Groen
et al., 2002, 2003), and it is unlikely that the very
few fibers observed in the contralateral dentate
gyrus of mice could contribute significantly to the
reinnervation of the dentate gyrus following den-
ervation. However, it should also be pointed out
that this has not yet been systematically analyzed
and that it is unclear to what extent sprouting can
be elicited from the few crossed entorhinodentate
fibers observed in mice. In summary, we conclude
that the entorhinodentate projection is almost ab-
sent in mice and that there is presently no evidence
for crossed entorhinodentate fiber sprouting in this
species.

Glial changes following entorhinal denervation
In recent years, the role of non-neuronal cells in
the reorganization of the dentate gyrus after ento-
rhinal denervation has received considerable
attention. The reaction of resident glia as well as
the recruitment of blood cells into the zone of
denervation have been studied, and a role of these
cells in the reorganization process has been pro-
posed; these non-neuronal cells are not only im-
portant for the removal of cellular debris in the
zone of denervation, but they may also regulate
axonal growth and transneuronal degeneration
processes. Thus, non-neuronal cells may play
a central role in the reorganization of the dentate
gyrus following denervation (see below).
Microglia
In rats, microglial cells react very rapidly to ento-
rhinal denervation by proliferation, and by trans-
formation from a ramified to an ‘‘activated’’
phenotype (Gall et al., 1979b; Fagan and Gage,
1990, 1994; Gehrmann et al., 1991; Jensen et al.,
1994; Hailer et al., 1997). Using histological and
immunohistochemical techniques, microglial reac-
tions were visible as early as 24 h postlesion, were
strongest at approximately 3 days postlesion, and
returned to normal levels during the first 2 weeks
after denervation. Electron microscopy and Mini
Ruby-tracing demonstrated that reactive micro-
glial cells phagocytose axonal debris and myelin
sheaths (Gehrmann et al., 1991; Jensen et al., 1994;
Bechmann and Nitsch, 1997). Interestingly, mi-
croglial activation was also observed, though to
a much lesser extent, in adjacent non-denervated
areas, suggesting that microglial cells are actively
recruited to the denervated zone (Gall et al.,
1979b; Jensen et al., 1994). Molecules that could
regulate this highly specific recruitment are
integrin adhesion molecules, such as leukocyte
function antigen-1 (LFA-1), very late antigen-4
(VLA-4), and the ligand for LFA-1, intercellular
adhesion molecule-1 (ICAM-1), which are all
expressed by microglial cell in the zone of dener-
vation (Hailer et al., 1997). In addition to the
role of microglia as debris-removing phagocytes,
511
regulatory functions have also been proposed, and
it was suggested that microglia could initiate a
sequence of events, which could eventually induce
or regulate axon sprouting (Gage et al., 1988;
Fagan and Gage, 1990; Eyupoglu et al., 2003). In
short, it was suggested that entorhinal terminal
degeneration activates microglia. These activated
cells phagocytose degenerating terminals and
release interleukin-1, thereby activating the
astroglia. Activated astrocytes, in turn, secrete
neurotrophic factors, which induce sprouting of
axons carrying specific receptors (Gage et al.,
1988; Eyupoglu et al., 2003).
In mice, the morphologic transformation of mi-
croglia is fairly similar to the changes observed in
rats in that an accumulation of reactive microglia
occurs within the denervated parts of the molec-
ular layer (Fig. 3a) as early as 24 h after entorhinal
denervation (Schauwecker and Steward, 1997;
Jensen et al., 1999; Wirenfeldt et al., 2003;
Rappert et al., 2004; Pedersen et al., 2006). At
this time, the microglia have transformed into
what is regarded as an ‘‘amoeboid’’ phenotype,
and the relative depletion of the adjacent non-
denervated inner molecular layer suggests that
they migrate into the zone of axonal degeneration.
In fact, we have demonstrated that the microglial
accumulation is severely impaired in mice deficient
for CXCR3, a chemokine receptor crucial for the
attraction of microglia (Rappert et al., 2002, 2004).
The other known factor contributing to the mi-
croglial accumulation is proliferation, which has
been demonstrated in mice using modern stereo-
logical techniques (Wirenfeldt et al., 2003; Ladeby
et al., 2005a, b). The molecular regulation of the
microglial response following entorhinal denerva-
tion has been studied in greater detail in mice
compared to rats. Besides CXCR3 (Rappert et al.,
2004), the peripheral benzodiazepine binding site
ligand PK11195 (Pedersen et al., 2006), and inter-
feron-gamma (Jensen et al., 2000a), interleukin
1beta and the transforming growth factor beta1
(Eyupoglu et al., 2003) have been implicated. In
addition to the role of microglia as phagocytes,
however, regulatory functions were also proposed
and, recently, experimental evidence for a link be-
tween microglia activation and dendritic atrophy
of GABAergic parvalbumin-positive basket cells
512

Fig. 3. Microglial cells play a role in transneuronal degeneration of parvalbumin-positive dendrites. (A) Overview of the entorhinal-
hippocampal area of a control mouse after denervation of the middle molecular layer (MML) of the dentate gyrus. Mac-1 positive
microglia accumulate in the zone of anterograde axonal degeneration. The insert shows the morphology of the activated microglial
cells. This microglial response is less pronounced in CXCR3 / (KO) mice compared to wild-type (WT) controls. (B–E) If less
microglial cells are recruited, denervated dendrites are preserved. The loss of parvalbumin-positive dendrites in the dentate gyrus is
evident in WT mice (B, D) and clearly diminished in CXCR3 / animals (C, E). The boxes in (B) and (C) indicate the areas shown at
higher magnification in (D) and (E), respectively. Scale bars: A: 300 mm; inset: 50 mm; B, C: 100 mm; D, E: 25 mm.

in the denervated dentate gyrus was shown
(Rappert et al., 2004; see also below).
Blood-derived cells
Besides migration and proliferation of resident
microglia, a third mechanism contributing to the
accumulation of microglia/macrophages within
the zone of denervation is the recruitment of
blood-derived monocytic cells. At least in the
mouse, these cells are able to invade the denervat-
ed hippocampus, where they transform into mi-
croglia-like elements within 72 h after entorhinal
denervation (Bechmann et al., 2005; Ladeby et al.,
2005b).
As all antibodies used for microglia-detection
also bind to monocytes/macrophages, the invasion
of hematopoietic cells as a source of ‘‘new’’ mi-
croglia have been difficult to study. Fagan and
Gage (1994) analyzed rats using i.v. injection of
DiI-labeled low density lipoprotein (LDL) parti-
cles, but could not find any evidence for the infil-
tration of monocytes. In their preparations, T cells
were also not visible (Fagan and Gage, 1994).
Later, using the R73 antibody, we demonstrated
the presence of CD4 a/b in the middle molecular
layer after entorhinal denervation in the same rat
strain (Bechmann et al., 2001a), but the question
whether monocytes are also recruited has only
been addressed in mice, where CCL-2 has been
identified as a major chemoattractant molecule
(Babcock et al., 2003). As described below, recent
evidence links chemokine-mediated recruitment of
microglia to transsynaptic dendritic changes in the
dentate gyrus.
Astroglia
In rats, astroglial activation after entorhinal dener-
vation was observed at slightly later time points
than the activation of microglia and was charac-
terized by proliferation, upregulation of glial fibril-
lary acidic protein (GFAP), and migration into the
denervated zone of the dentate gyrus (Rose et al.,
1976; Gall et al., 1979b; Gage et al., 1988; Fagan
and Gage, 1990; Steward et al., 1993). Within the
denervated zone, astrocytes were shown to
513
participate in the removal of degenerating termi-
nals and spines (Bechmann and Nitsch, 1997;
Deller et al., 1997). Increases in GFAP messenger
RNA (mRNA) slightly preceded changes in GFAP
and were observed as early as 12 h postlesion
(Steward et al., 1990, 1993). Interestingly, GFAP
mRNA changes depended on postlesional seizure
activity, demonstrating that changes in GFAP
expression in the early postlesional time period
may be caused by changes in activity rather than
denervation (Kelley and Steward, 1996b). In addi-
tion, it was shown that reactive astrocytes change
the molecular composition of the denervated
molecular layer (Gall et al., 1979b; Gage et al.,
1988; Steward, 1991; Deller et al., 2000, 2001):
Astrocyte-derived neurotrophic factors (Kawaja
and Gage, 1991; Yoshida and Gage, 1991;
Gomez-Pinilla et al., 1992; Guthrie et al., 1995,
1997; Lee et al., 1997), cell-adhesion molecules
(Styren et al., 1994, 1995; Jucker et al., 1996), and
a variety of extracellular matrix molecules (Deller
et al., 2000) were studied and it was suggested that
these astrocyte-derived molecules could regulate
and pattern the sprouting response.
In mice, less information is available. Astroglial
changes were studied using GFAP-immunohisto-
chemistry, revealing astroglial migration and an
increased expression of GFAP-immunoreactivity
throughout the denervated zone (Steward and
Trimmer, 1997; Del Turco et al., 2003). Changes
in GFAP mRNA were measured using northern
blot (Steward and Trimmer, 1997; Steward et al.,
1997) and at the single-cell level using laser micro-
dissection in combination with quantitative real-
time-polymerase chain reaction (qRT-PCR)
(Burbach et al., 2004a, b). These postlesional
changes in GFAP mRNA depend, as in the rat,
on activity and protein synthesis (Steward, 1994a;
Steward et al., 1997). In contrast to rats, however,
studies analyzing changes in the expression of
astrocyte-derived molecules following entorhinal
denervation are rare (Poulsen et al., 2000; Mayer
et al., 2005; Schafer et al., 2005).
In summary, morphological evidence supports
a phagocytic role of astrocytes in the denervated
dentate gyrus of the adult rat and correlative data
implicate astrocytes in the patterning of the
sprouting response. In mice, the cellular response
514
of astrocytes appears to be similar to the one
observed in rats. However, information about
molecular changes induced by reactive astrocytes
is scarce and the role of astrocytes in this species
remains to be elucidated.
Oligodendroglia and NG2-positive cells
It is now well established that oligodendrocytes
and some of their surface molecules are potent in-
hibitors of axonal growth (Kapfhammer and
Schwab, 1992; Schwab et al., 1993). Since these
inhibitory molecules could influence sprouting,
they were studied following entorhinal denerva-
tion. In addition, the entorhinal denervation
model was used to study the myelination of newly
formed (sprouting) axons in the normal adult
CNS. Changes in putative oligodendroglial cell
precursors, NG2-cells (Dawson et al., 2000, 2003)
were also analyzed.
In rats, the first evidence for a reaction of
oligodendroglial cells to entorhinal denervation
was provided by Gall et al. (1979b), who observed
signs of oligodendroglial cell proliferation. Later,
changes in myelination and Nogo-expression were
analyzed (Meier et al., 2003, 2004). These studies
revealed correlations between the axonal sprouting
response and the time course of demyelination,
suggesting that oligodendroglial-derived inhibitory
molecules could regulate sprouting. Consistent
with the results of Gall et al. (1979a, b), we re-
cently showed that oligodendroglial precursor
cells, NG2-positive cells, also react to entorhinal
denervation in rats (Dehn et al., 2006).
In mice, an analysis of the dynamics of oligoden-
drocytes following entorhinal denervation was per-
formed by Drojdahl and colleagues (2004). In
addition, changes in oligodendroglial-derived mole-
cules, such as changes in the expression of myelin
basic protein (Jensen et al., 2000b; Drojdahl
et al., 2004) and myelin-associated glycoprotein
(Mingorance et al., 2005), were reported. Similar to
the observations in rats, proliferation of NG2-pos-
itive oligodendroglial precursors and the formation
of new oligodendrocytes was demonstrated follow-
ing entorhinal denervation in mice (Nielsen et al.,
2006), suggesting that sprouting axons could be
myelinated by newly formed oligodendroglial cells.
In summary, oligodendrocytes react to ento-
rhinal denervation in rats as well as mice. Their
surface molecules may be involved in the regula-
tion of sprouting, and newly formed oligodendro-
cytes may contribute to the myelination of
sprouting axons.
Dendritic changes following entorhinal denervation
Denervated neurons in the CNS remodel their
dendritic field. This reorganization process
involves spines and dendrites, and may, if the den-
ervation is extensive, lead to the death of the
denervated target cell (Steward, 1994b; Kovac
et al., 2004). In the rat, dentate gyrus granule cells
(Nafstad, 1967; Steward and Scoville, 1976;
Amaral and Witter, 1995) and parvalbumin-
positive GABAergic interneurons (Kosaka et al.,
1987; Zipp et al., 1989; Nitsch et al., 1990) are the
targets of entorhinal afferents. These fibers im-
pinge on their distal dendrites and form between
80 and 90% of all synapses on their distal dendritic
field (Matthews et al., 1976a; Steward and Vinsant,
1983). In non-lesioned C57BL/6 mice, the ana-
tomical situation appears to be similar (Deller,
1998; Gebhardt et al., 2002; Del Turco et al.,
2004). Thus, ECL/perforant pathway transection
will denervate both cell types in both species.
Transneuronal changes in granule cells and
parvalbumin-positive neurons
In rats, transneuronal changes in granule cell mor-
phology after entorhinal denervation have been
studied using Golgi-impregnated (Parnavelas et al.,
1974; Caceres and Steward, 1983) and intracellularly
labeled (Diekmann et al., 1996) granule cells. These
studies agree that the dendritic arbor of granule cells
is remodeled following denervation, and that gran-
ule cells lose spines and dendrites during the first
2 weeks after entorhinal denervation. By 30 days
postlesion, there was a recovery of the dendritic ar-
bor detected by the Golgi method (Caceres and
Steward, 1983), whereas permanent transneuronal
changes of granule cell dendrites were reported

using the intracellularly labeled granule cells
(Diekmann et al., 1996).
Dendritic remodeling of parvalbumin-positive
GABAergic neurons was also studied following
entorhinal denervation (Nitsch and Frotscher,
1991, 1992, 1993; Nitsch et al., 1992). Similar to
granule cells, the distal dendrites of parvalbumin-
positive neurons demonstrated degenerative
changes (Nitsch and Frotscher, 1991, 1992,
1993). These degenerative changes persisted, with
only a slight recovery by 2 months postlesion
(Nitsch and Frotscher, 1991).
In mice, morphological changes of granule cells
have not yet been studied at the level of single
cells, and at present only indirect — and contro-
versial — data are available. Following ibotenic
acid injections into the entorhinal cortex, Kadish
and van Groen (2003) did not observe atrophy of
the denervated molecular layer of the mouse dent-
ate gyrus, and suggested that granule cell dendritic
atrophy is limited in this species. In contrast, oth-
ers (Phinney et al., 2004) who used mechanical
transection and measured the width and the vol-
ume of the molecular layer using stereological
methods demonstrated a>40% reduction in vol-
ume 45 days postlesion, suggesting substantial
granule cell atrophy. A detailed temporal analysis
on the postlesional morphological changes of sin-
gle granule cells is needed to clarify this issue, and
its potential differences from the rat.
In contrast to granule cells, morphological
changes of parvalbumin-positive neurons have
been studied in more detail in mouse (Rappert
et al., 2004). A temporal analysis demonstrated
changes in parvalbumin-positive dendrites that
were quite similar to rats. Parvalbumin-positive
dendrites were rapidly lost in the denervated mo-
lecular layer of C57BL/6 mice and no recovery was
observed up to 31 days postlesion (Fig. 3). How-
ever, a detailed morphological analysis of identi-
fied parvalbumin-positive cells was not performed.
Transneuronal changes: insights into cellular
mechanisms
What are the potential mechanisms involved in the
transneuronal dendritic degeneration of granule
515
cells and parvalbumin-positive neurons following
entorhinal denervation? Although several hypoth-
eses have been proposed, including excitotoxicity,
elevation of intracellular calcium levels, and lack
of growth factors (Steward, 1991; Nitsch and
Frotscher, 1992; Sattler et al., 1998; Arundine and
Tymianski, 2003; Kovac et al., 2004), the role of
these mechanisms and molecules has not yet been
clarified. As far as granule cells are concerned, no
studies to date have explained the extensive
transneuronal degeneration of the denervated
granule cell arbor. In the case of the degeneration
of parvalbumin-positive neurons, Nitsch and
Frotscher (1992) showed that the application of
the glutamate receptor antagonist MK801 prior to
entorhinal denervation could prevent the atrophy
of parvalbumin-immunoreactive dendrites (Nitsch
and Frotscher, 1992). They suggested that an
NMDA-receptor mediated influx of calcium could
cause the degeneration of parvalbumin-positive
dendrites. In mice, degeneration of parvalbumin-
positive dendrites was recently investigated
(Rappert et al., 2004) and microglia and chemo-
kines were implicated in the regulation of trans-
neuronal degeneration. Since this study proposed
a novel mechanism for denervation-induced trans-
neuronal degeneration a more detailed review of
this study and the potential role of chemokines
appears to be warranted.
Transneuronal degeneration of parvalbumin-positive
dendrites: a role for the chemokine receptor CXCR3
Chemokines are a diverse family of pro-inflamma-
tory cytokines. These small homologous peptides
(8–14 kDa) (Rollins, 1997; Baggiolini, 1998) play
an important role in recruiting inflammatory cells
into tissues in response to infection and inflam-
mation (Karpus and Ransohoff, 1998; Mackay,
2001). Additionally, these molecules are involved
in the pathogenesis of many important neuroin-
flammatory diseases ranging from multiple sclero-
sis and stroke to HIV encephalopathy (Luster,
1998; Asensio and Campbell, 1999).
For many chemokine receptors in the adult
CNS, the corresponding ligand can only be de-
tected under pathological conditions, for example,
516
in the vicinity of an inflammatory lesion (Asensio
and Campbell, 1999; Bacon and Harrison, 2000)
suggesting that chemokine-receptors and their lig-
ands are primarily involved in the regulation of
pathophysiological processes. In line with this in-
terpretation, several studies have shown that
chemokines recruit leukocytes into inflamed brain
tissue (Fife et al., 2000; Izikson et al., 2000; Siebert
et al., 2000; Huang et al., 2001; Babcock and
Owens, 2003). Messenger RNA for CCL5, CCL2,
CXCL10, CCL3, CCL4 and CXCL2 were specifi-
cally induced in the lesion-reactive hippocampus
and at the site of axonal transection by 24 h after
axotomy, while other chemokines like XCL1,
CCL11 and CCL1 were not detected (Babcock
et al., 2003). Interestingly, CCL2 was upregulated
in the denervated hippocampus already at 3 h after
axotomy, while an increase in CCL5 was first de-
tected after 24–48 h. The different time course of
induction of these two chemokines suggests that
chemokines could have different functions follow-
ing ECL. In fact, as pointed out above, analyses
of chemokine receptor-deficient mice indicated
that the CCR2 ligand CCL2 (MCP-1) is critical
for leukocyte infiltration, whereas CCR5 ligands
such as CCL5 (RANTES) are not (Babcock et al.,
2003).
What could be the role of the chemokines and
their receptors in the reorganization of the dentate
gyrus following ECL? Since chemokine receptors
are expressed in neurons and glia, and chemokines
are mostly produced by glial cells, it has been sug-
gested that they are involved in glia-to-glia, glia-
to-neuron, and neuron-to-glia interactions
(Hesselgesser and Horuk, 1999; Ambrosini and
Aloisi, 2004; Adler et al., 2005). In the context of
ECL the chemokine receptor CXCR3 and its lig-
ands CXCL9, CXL10, CXCL11 and CCL21 are
especially interesting. For example, CCL21 is spe-
cifically expressed by damaged neurons (Biber
et al., 2001; de Jong et al., 2005) and not by any
glial cell type (Biber et al., 2001; Alt et al., 2002;
Columba-Cabezas et al., 2003). Since CCL21 in-
duces intracellular calcium signals and chemotaxis
of cultured microglia (Biber et al., 2001; Rappert
et al., 2002), a potential role of CCL21 in neuron-
microglia communication has been proposed —
but has to be proven in vivo. As a second ligand,
CXCL10 mRNA has been found in the lesioned
hippocampus of the mouse (Babcock et al., 2003)
and the protein could be localized on neurons by
our group (Rappert et al., 2004). In mice deficient
for CXCR3, the accumulation of microglia in
zones of denervation was dramatically reduced,
demonstrating a crucial role for CXCR3 in mi-
croglial recruitment and suggesting CXCL10 as
the signaling partner (Rappert et al., 2004).
The importance of CXCR3 is underscored by
the preservation of parvalbumin-positive dendrites
after entorhinal denervation in CXCR3-deficient
mice (Rappert et al., 2004). The absence of the
dendritic loss in the CXCR3 knockout animals is
not simply delayed, since a difference in the den-
dritic length compared to wild-type animals was
still evident 31 days after entorhinal denervation.
Although a reduction of dendrites developed, there
were significantly more dendrites in the KO in
comparison to the wild-type animal (Figs 3b–e).
This is consistent with previous observations that
the elimination of microglia in rat entorhinohip-
pocampal slice cultures diminishes lesion-induced
dendritic loss (Eyupoglu et al., 2003). Although
there is no mechanistic explanation for this phe-
nomenon, these data argue against the concept
that ‘‘retraction’’ is dependent only on the loss of
a specific input, and demonstrates a causal link
between microglial response and the structural
outcome of postlesional changes. Whether the mi-
croglia actively eliminate denervated dendritic seg-
ments or provide signals inducing their retraction
remains to be seen.
Summary
In the present review, we have summarized and
compared the available data on the structural
reorganization of the dentate gyrus following
entorhinal denervation in rats and mice. We con-
clude on the basis of this comparison, that in mice
neither the normal anatomy of the dentate gyrus
and nor many aspects concerning its structural re-
organization following denervation have yet been
described in depth. The available data, however,
already point to important species differences in
both aspects, anatomy and reorganization. In the

following paragraphs, we will discuss these species
differences in the context of sprouting and inflam-
mation and will make suggestions for the use of
the entorhinal denervation model in mutant mice.
How can the species difference in the sprouting
response be explained?
Our comparison of sprouting following entorhinal
denervation in rats and mice revealed major spe-
cies differences. Some of these differences may
readily be explained on the basis of differences in
normal anatomy. For example, if crossed entorhi-
nodentate fibers do not exist in mice, sprouting of
this fiber projection is not to be expected. Other
species differences, for example the response of the
C/A fibers to the inner molecular layer are more
difficult to explain. This projection system exists in
rats as well as mice and only subtle anatomical
differences are observed between the two species
(Del Turco et al., 2003). Nevertheless, in adult rats
these C/A fibers stay in their home territory
whereas they sprout across the laminar bounda-
ries in mice. In the absence of major anatomical
differences, an explanation for this phenomenon
should be sought on the molecular level.
In recent years, we have studied molecules which
could regulate the layer-specific sprouting of C/A
fibers to the inner molecular layer in rats (Frotsc-
her et al., 1997; Deller et al., 2000, 2001). Whereas
glia-derived neurotrophic factors and cytokines
are probably involved in the induction of new
axonal collaterals (Gomez-Pinilla et al., 1992;
Guthrie et al., 1995, 1997; Fagan et al., 1997;
Lee et al., 1997; Woods et al., 1998, 1999), other
molecules may be better suited for the guidance of
newly formed fibers. A group of molecules which
could regulate a layer-specific growth response by
forming a growth barrier between the non-dener-
vated inner and the denervated middle- and outer
molecular layer are extracellular matrix molecules
synthesized by reactive glia (Rose et al., 1976;
Gage et al., 1988). These molecules are thought to
form boundaries for growing axons during devel-
opment (Faissner and Steindler, 1995; Höke and
Silver, 1996; Pearlman and Sheppard, 1996;
Margolis and Margolis, 1997; Bovolenta and
517
Fernaud-Espinosa, 2000), and it could be demon-
strated that several growth inhibitory extracellular
matrix molecules such as tenascin-C (Deller et al.,
1997), DSD-1-proteoglycan (Deller et al., 1997),
neurocan (Haas et al., 1999), brevican (Thon et al.,
2000), and NG2 (Dehn et al., 2006) are upregu-
lated within the denervated zone of the molecular
layer after entorhinal denervation. Because these
molecules form a sharp border towards the adja-
cent non-denervated zones, we suggested that
these molecules delineate boundaries of axonal
growth following entorhinal lesion (Deller et al.,
2000, 2001). This hypothesis, although certainly
attractive, still awaits experimental verification.
In mice, the situation is different and C/A fibers
to the inner molecular layer sprout across laminar
boundaries. One explanation could be that in this
species, the expression pattern, the time course
of expression, or the order of assembly of growth
inhibitory extracellular matrix molecules could
be different. So far, only few of the extracellular
matrix molecules mentioned above have been
investigated: Brevican (Mayer et al., 2005) and
NG2 (Nielsen et al., 2006) were analyzed and den-
ervation-induced changes were reported which
were similar to those observed for these molecules
in rats. However, subtle differences with regard to
the time course of expression and the cellular dis-
tribution of the two molecules were also observed.
Whether or not these differences — or differences
in the expression of other extracellular matrix
molecules — could result in an absent or less
effective growth barrier between the non-dener-
vated inner molecular layer and the denervated
zone in mice remains to be seen.
Another explanation could be that mossy cells in
mice and rats differ in their ability to grow, i.e., in
their intrinsic growth competence (Benowitz and
Routtenberg, 1997; Caroni, 1997, 1998). It is well
established that the growth competence of neurons
depends on the expression of growth-associated
proteins, such as growth-associated protein 43
or cortical cytoskeleton-associated protein 23
(Aigner et al., 1995; Benowitz and Routtenberg,
1997; Caroni et al., 1997; Frey et al., 2000; Laux
et al., 2000). These molecules are strongly expressed
during brain development and are downregulated
postnatally (Caroni, 1997). Intriguingly, the ability
518
of rat mossy cells to sprout across laminar bound-
aries (Lynch et al., 1973; Gall and Lynch, 1980,
1981; Gall et al., 1986) and the downregulation of
growth-associated proteins (Caroni, 1997) occurs
during the same developmental time window, sug-
gesting that the ability of mossy cell axons to sprout
into the denervated dentate gyrus indeed depends on
their intrinsic growth competence. In contrast, in
mice, in which translaminar mossy cell sprouting
is found in the adult, mossy cell growth competence
is relatively strong. Recent observations in trans-
genic mice overexpressing growth-associated pro-
teins indicate that the intrinsic growth-competence
of mossy cells affects the strength and extent of the
translaminar commissural sprouting response in
mice following entorhinal denervation (Del Turco
et al., 2001).
Finally, competitive interactions of sprouting
C/A mossy cell axons with other sprouting fiber
systems could also play a role (Cotman and
Nadler, 1978; Nadler et al., 1980; Steward,
1994b). In rats, crossed entorhinodentate fibers
sprout within the denervated molecular layer
(Steward et al., 1974; Deller et al., 1996a). Since
recent data indicate that C/A fibers are repelled by
entorhinal fibers (Borrell et al., 1999; Deller et al.,
1999a), sprouting of this pathway would block
translaminar ingrowth of C/A fibers into the den-
ervated zone in rats. In contrast, C/A fibers could
grow into the denervated mouse fascia dentata
because virtually all entorhinal fibers are removed
from the mouse dentate gyrus after unilateral ent-
orhinal denervation. If this hypothesis were true,
then the absence of the crossed entorhinodentate
pathway in mice would suffice to explain the major
species differences between rats and mice with re-
gard to the axonal reorganization of the dentate
gyrus following denervation.
Inflammatory changes following denervation and
the immune response to entorhinal denervation
The morphologic transformation of astrocytes and
microglia and their recruitment into the zones of
axonal degeneration are similar in mice and rats. As
described above, much has been speculated with re-
gard to the contribution of glia to postlesional
reorganization, but as yet an impact on the sprout-
ing response has not been proven. However, in both
rats and mice, interfering with microglial recruit-
ment impacts on the loss of denervated dendrites
(Eyupoglu et al., 2003; Rappert et al., 2004).
Phagocytosis of distal segments by activated micro-
glia is an attractive explanation for this observation,
which is currently tested in our laboratory.
The immune system of rodents is capable of
mounting strong immune responses to myelin-
associated epitopes, but entorhinal denervation
does not induce evident autoimmune demyelina-
tion. This could be attributed to a state of immune
ignorance, as Fagan and Gage (1994) did not find
infiltration of haematogenous cells into zones of
axonal degeneration after entorhinal denervation
in the rat. In the same strain, we detected infil-
trating CD4 T cells interacting with MHC-II pos-
itive microglia. These myelin-phagocytosing cells
exhibited the phenotype of immature antigen-pre-
senting cells which is likely to cause T cell anergy
(Bechmann et al., 2001a). Moreover, astrocytes
expressed the death-ligand CD95L (FasL) which
causes apoptosis of activated T cells (Bechmann
et al., 1999, 2000, 2002). Thus, immune tolerance
seems to be actively maintained and, in fact, we
have shown that autoimmunity to myelin is di-
minished following entorhinal denervation in
mice. Strikingly, axonal antigens are found in cer-
vical lymph nodes after entorhinal denervation
and their appearance is accompanied by T cell
apoptosis (Mutlu et al., 2007). Monocytic cells are
known to pass the vascular wall to reside in peri-
vascular spaces in rats and mice under normal
conditions (Bechmann et al., 2001b, c), but
they cross the glia limitans and transform into
microglia after entorhinal denervation in mice
(Bechmann et al., 2005). CCL2 has been identified
as the key chemokine driving this recruitment
(Babcock et al., 2003), while it is unclear which
signals adopt the invading macrophages into ram-
ified microglia. One of the key questions is whether
such ‘‘microglia’’ transform into dendritic cells
capable of leaving the neuropil to present antigens
in lymphoid organs. In fact, after entorhinal den-
ervation in mice, T cells specifically target one
cervical lymph node via the cribroid plate and the
nasal mucosa (Goldmann et al., 2006).

It is also noteworthy that permeability of the
blood-brain barrier for IgG has been demon-
strated after entorhinal denervation in the rat
a decade ago (Jensen et al., 1997), but the role of
antibodies in this context and their impact on
postlesional reorganization is completely unclear.
It is now clear that entorhinal denervation in
mice and rats involves a systemic immune
response, which impacts on the structural out-
come at least at the levels of phagocytosis of
growth-inhibiting debris. T cells have also been
shown to secrete neurotrophins (Kerschensteiner
et al., 1999), but entorhinal denervation-studies in
mice bearing myelin-specific T cell receptors have
not yet been performed. It is evident from count-
less immunological studies that there are signifi-
cant species and strain differences in regard to the
strength and consequence of immune responses to
brain antigens (Gold et al., 2006). In regard to the
neurological outcome, they can be protective or
detrimental. The same type of lesion can evoke
overall protective or detrimental immune re-
sponses solely depending on strain subjected to
the damage (Kipnis et al., 2001). Thus, beyond
species differences, strain differences must also be
expected when evaluating the role of immune cells
in postlesional plasticity.
The intrinsic immune suppression of the brain
may have evolved to cope with situations bearing
high evolutionary pressure such as infection, where it
is better for the individual to tolerate certain infec-
tious agents than to eliminate all infected neurons.
Under conditions such as axonal injuries, which cer-
tainly provided low pressure, it may be beneficial for
the neurological outcome to therapeutically foster
immune responses (Bechmann, 2005). This can be
tested in mouse models of entorhinal denervation.
Conclusions for the use of the entorhinal denervation
model in rats and mice
In summary, ECL/perforant pathway transections
have been used to denervate the dentate gyrus of rats
and mice in order to study the reorganization of the
brain following injury. Under the assumption that
the reorganization of the dentate gyrus is similar in
the two related rodent species, data obtained in one,
519
usually the rat, were used to interpret data obtained
in the other, usually the mouse. Since this assump-
tion was never proven, we have compiled and com-
pared the data available on structural changes
following entorhinal denervation in both species in
this review. This comparative analysis revealed ma-
jor differences between the entorhinal model system
in rats and mice. We conclude, that a non-critical
transfer of data from one species to the other is not
warranted.
Given the difficulty in generalizing data from
the rat to mouse, all observations made in rats
need first to be verified in the context of the mouse
model before mutants should be analyzed. In ad-
dition, strain differences should also be considered
and it will probably be necessary to reinvestigate
certain aspects specifically in the background
strain of a given mutant. Only then, will it be pos-
sible to gain meaningful and fundamental biolog-
ical insights from the entorhinal denervation
model in mice. Although certainly demanding,
such an approach may yield additional scientific
benefits, since the investigation of species and
strain differences in hippocampal connectivity and
plasticity may open up new strategies for a com-
parative, yet functionally oriented analysis. This
way, a methodologically more careful approach
would be richly awarded and of considerable value
to the scientific community.
Acknowledgments
This review is dedicated to Michael Frotscher,
M.D., who is one of the pioneers of hippocampal
research and the entorhinal denervation model,
and who contributed to many of the studies
reviewed here. This study was supported by the
Deutsche Forschungsgemeinschaft (DE 551/8-1,
SFB 507 B16), German Israeli Foundation (T.D.),
and Gisela Stadelmann-Stiftung (D.D.T.)
References
Adler, M.W., Geller, E.B., Chen, X. and Rogers, T.J. (2005)
Viewing chemokines as a third major system of communica-
tion in the brain. AAPS J., 7: E865–E870.
520
Aigner, L., Arber, S., Kapfhammer, J.P., Laux, T., Schneider,
C., Botteri, F., Brenner, H.R. and Caroni, P. (1995) Over-
expression of the neural growth-associated protein GAP-43
induces nerve sprouting in the adult nervous system of
transgenic mice. Cell, 83: 269–278.
Alt, C., Laschinger, M. and Engelhardt, B. (2002) Functional
expression of the lymphoid chemokines CCL19 (ELC) and
CCL 21 (SLC) at the blood-brain barrier suggests their
involvement in G-protein-dependent lymphocyte recruit-
ment into the central nervous system during experimental
autoimmune encephalomyelitis. Eur. J. Immunol., 32:
2133–2144.
Amaral, D.G. and Witter, M.P. (1989) The three dimensional
organization of the hippocampal formation: a review of
anatomical data. Neuroscience, 31: 571–591.
Amaral, D.G. and Witter, M.P. (1995) The hippocampal for-
mation. In: Paxinos G. (Ed.), The Rat Nervous System
(2nd ed.). Academic Press, San Diego, CA, pp. 443–494.
Ambrosini, E. and Aloisi, F. (2004) Chemokines and glial cells:
a complex network in the central nervous system. Neuroc-
hem. Res., 29: 1017–1038.
Arundine, M. and Tymianski, M. (2003) Molecular mecha-
nisms of calcium-dependent neurodegeneration in excitotox-
icity. Cell Calcium, 34: 325–337.
Asensio, V.C. and Campbell, I.L. (1999) Chemokines in the
CNS: plurifunctional mediators in diverse states. Trends
Neurosci., 22: 504–512.
Aubert, I., Poirier, J., Gauthier, S. and Quirion, R. (1994)
Multiple cholinergic markers are unexpectedly not altered in
the rat dentate gyrus following entorhinal cortex lesions.
Neuroscience, 14: 2476–2484.
Babcock, A. and Owens, T. (2003) Chemokines in experimental
autoimmune encephalomyelitis and multiple sclerosis. Adv.
Exp. Med. Biol., 520: 120–132.
Babcock, A.A., Kuziel, W.A., Rivest, S. and Owens, T. (2003)
Chemokine expression by glial cells directs leukocytes to sites
of axonal injury in the CNS. J. Neurosci., 23: 7922–7930.
Bacon, K.B. and Harrison, J.K. (2000) Chemokines and their
receptors in neurobiology: perspectives in physiology and
homeostasis. J. Neuroimmunol., 104: 92–97.
Baggiolini, M. (1998) Chemokines and leukocyte traffic.
Nature, 392: 565–568.
Bechmann, I. (2005) Failed central nervous system regenera-
tion: a downside of immune privilege? Neuromolecular Med.,
7: 217–228.
Bechmann, I., Goldmann, J., Kovac, A.D., Kwidzinski, E.,
Simburger, E., Naftolin, F., Dirnagl, U., Nitsch, R. and
Priller, J. (2005) Circulating monocytic cells infiltrate layers
of anterograde axonal degeneration where they transform
into microglia. FASEB J., 19: 647–649.
Bechmann, I., Kwidzinski, E., Kovac, A.D., Simburger, E.,
Horvath, T., Gimsa, U., Dirnagl, U., Priller, J. and Nitsch,
R. (2001c) Turnover of rat brain perivascular cells. Exp. Ne-
urol., 168: 242–249.
Bechmann, I., Lossau, S., Steiner, B., Mor, G., Gimsa, U. and
Nitsch, R. (2000) Reactive astrocytes upregulate Fas (CD95)
and Fas ligand (CD95L) expression but do not undergo
programmed cell death during the course of anterograde de-
generation. Glia, 32: 25–41.
Bechmann, I., Mor, G., Nilsen, J., Eliza, M., Nitsch, R. and
Naftolin, F. (1999) FasL (CD95L, Apo1L) is expressed in the
normal rat and human brain: evidence for the existence of an
immunological brain barrier. Glia, 27: 62–74.
Bechmann, I. and Nitsch, R. (1997) Astrocytes and microglial cells
incorporate degenerating fibers following entorhinal lesion: a
light, confocal, and electron microscopical study using a
phagocytosis-dependent labeling technique. Glia, 20: 145–154.
Bechmann, I., Peter, S., Beyer, M., Gimsa, U. and Nitsch, R.
(2001a) Presence of B7-2 (CD86) and lack of B7-1 (CD80) on
myelin phagocytosing MHC-II-positive rat microglia is as-
sociated with nondestructive immunity in vivo. FASEB J.,
15: 1086–1088.
Bechmann, I., Priller, J., Kovac, A., Bontert, M., Wehner, T.,
Klett, F.F., Bohsung, J., Stuschke, M., Dirnagl, U. and
Nitsch, R. (2001b) Immune surveillance of mouse brain peri-
vascular spaces by blood-borne macrophages. Eur. J. Ne-
urosci., 14: 1651–1658.
Bechmann, I., Steiner, B., Gimsa, U., Mor, G., Wolf, S., Beyer,
M., Nitsch, R. and Zipp, F. (2002) Astrocyte-induced T cell
elimination is CD95 ligand dependent. J. Neuroimmunol.,
132: 60–65.
Beck, K.D., Lamballe, F., Klein, R., Barbacid, M., Schau-
wecker, P.E., McNeill, T.H., Finch, C.E., Hefti, F. and Day,
J.R. (1993) Induction of noncatalytic TrkB neurotrophin re-
ceptors during axonal sprouting in the adult hippocampus.
J. Neurosci., 13: 4001–4014.
Benowitz, L.I., Rodriguez, W.R. and Neve, R.L. (1990) The
pattern of GAP-43 immunostaining changes in the rat hip-
pocampal formation during reactive synaptogenesis. Mol.
Brain Res., 8: 17–23.
Benowitz, L.I. and Routtenberg, A. (1997) GAP-43: an intrinsic
determinant of neuronal development and plasticity. Trends
Neurosci., 20: 84–91.
Biber, K., Sauter, A., Brouwer, N., Copray, S.C. and Boddeke,
H.W. (2001) Ischemia-induced neuronal expression of the
microglia attracting chemokine secondary lymphoid-tissue
chemokine (SLC). Glia, 34: 121–133.
Blackstad, T.W. (1958) On the termination of some afferents to
the hippocampus and fascia dentata: an experimental study
in the rat. Acta Anat. (Basel), 35: 202–214.
Blasco-Ibanez, J.M. and Freund, T.F. (1997) Distribution,
ultrastructure and connectivity of calretinin-immunoreactive
mossy cells of the mouse dentate gyrus. Hippocampus,
7: 307–320.
Book, A.A., Wiley, R.G. and Schweitzer, J.B. (1992) Specificity
of 192 IgG-saporin for NGF receptor- positive cholinergic
basal forebrain neurons in the rat. Brain Res., 590: 350–355.
Borrell, V., Ruiz, M., del Rio, J.A. and Soriano, E. (1999)
Development of commissural connections in the hippocam-
pus of reeler mice: evidence of an inhibitory influence of
Cajal-Retzius cells. Exp. Neurol., 156: 268–282.
Bovolenta, P. and Fernaud-Espinosa, I. (2000) Nervous system
proteoglycans as modulators of neurite outgrowth. Prog.
Neurobiol., 61: 113–132.

Burbach, G.J., Dehn, D., Del Turco, D., Staufenbiel, M. and
Deller, T. (2004a) Laser microdissection reveals regional and
cellular differences in GFAP mRNA upregulation following
brain injury, axonal denervation, and amyloid plaque dep-
osition. Glia, 48: 76–84.
Burbach, G.J., Dehn, D., Nagel, B., Del Turco, D. and Deller,
T. (2004b) Laser microdissection of immunolabeled as-
trocytes allows quantification of astrocytic gene expression.
J. Neurosci. Methods, 138: 141–148.
Caceres, A. and Steward, O. (1983) Dendritic reorganization in
the denervated dentate gyrus of the rat following entorhinal
cortical lesions: a Golgi and electron microscopic analysis. J.
Comp. Neurol., 136: 287–295.
Campbell, K.A., Bank, B. and Milgram, N.W. (1984)
Epileptogenic effects of electrolytic lesions in the hippocam-
pus: role of iron deposition. Exp. Neurol., 86: 506–514.
Caroni, P. (1997) Intrinsic neuronal determinants that promote
axonal sprouting and elongation. Bio Essays, 19: 767–775.
Caroni, P. (1998) Driving the growth cone. Science, 281:
1465–1466.
Caroni, P., Aigner, L. and Schneider, C. (1997) Intrinsic neu-
ronal determinants locally regulate extrasynaptic and synap-
tic growth at the adult neuromuscular junction. J. Cell Biol.,
136: 679–692.
Chen, L.L., Van Hoesen, G.E., Barnes, C.L. and West, J.R.
(1983) Enhanced acetylcholinesterase staining in the hippo-
campal perforant pathway zone after combined lesions of the
septum and entorhinal cortex. Brain Res., 272: 354–359.
Collazos-Castro, J.E. and Nieto-Sampedro, M. (2001) Devel-
opmental and reactive growth of dentate gyrus afferents:
Cellular and molecular interactions. Restor. Neurol. Neuro-
sci., 19: 169–187.
Columba-Cabezas, S., Serafini, B., Ambrosini, E. and Aloisi, F.
(2003) Lymphoid chemokines CCL19 and CCL21 are ex-
pressed in the central nervous system during experimental
autoimmune encephalomyelitis: implications for the mainte-
nance of chronic neuroinflammation. Brain Pathol., 13:
38–51.
Cotman, C.W., Gentry, C. and Steward, O. (1977) Synaptic
replacement in the dentate gyrus after unilateral entorhinal
lesion: electron microscopic analysis of the extent of replace-
ment of synapses by the remaining entorhinal cortex. J. Ne-
urocytol., 6: 455–464.
Cotman, C.W. and Nadler, J.V. (1978) Reactive synaptogenesis
in the hippocampus. In: Cotman C.W. (Ed.), Neuronal Plas-
ticity. Raven Press, New York, pp. 227–271.
Dasheiff, R.M. and McNamara, J.O. (1982) Electrolytic
entorhinal cortex lesions cause seizures. Brain Res., 231:
444–450.
Dawson, M.R., Levine, J.M. and Reynolds, R. (2000) NG2-
expressing cells in the central nervous system: are they
oligodendroglial progenitors? J. Neurosci. Res., 61:
471–479.
Dawson, M.R., Polito, A., Levine, J.M. and Reynolds, R.
(2003) NG2-expressing glial progenitor cells: an abundant
and widespread population of cycling cells in the adult rat
CNS. Mol. Cell Neurosci., 24: 476–488.
521
Dehn, D., Burbach, G.J., Schafer, R. and Deller, T. (2006)
NG2 upregulation in the denervated rat fascia dentata fol-
lowing unilateral entorhinal cortex lesion. Glia, 53: 491–500.
Del Turco, D., Gebhardt, C., Burbach, G.J., Pleasure, S.J.,
Lowenstein, D.H. and Deller, T. (2004) Laminar organiza-
tion of the mouse dentate gyrus: insights from BETA2/Neuro
D mutant mice. J. Comp. Neurol., 477: 81–95.
Del Turco, D., Gebhardt, C., Woods, A.G., Kapfhammer, J.P.,
Naumann, T., Frotscher, M., Caroni, P. and Deller, T. (2001)
Overexpression of the growth associated protein CAP-23 re-
sults in translaminar commissural sprouting in the hippo-
campus after entorhinal cortex lesion in adult mice. Soc.
Neurosci. Abstr., 27: 698.12.
Del Turco, D., Woods, A.G., Gebhardt, C., Phinney, A.L.,
Jucker, M., Frotscher, M. and Deller, T. (2003) Comparison
of commissural sprouting in the mouse and rat fascia dentata
after entorhinal cortex lesion. Hippocampus, 13: 685–699.
Deller, T. (1998) The anatomical organization of the rat fascia
dentate — new aspects of laminar organization as revealed by
anterograde tracing with Phaseolus vulgaris-leucoagglutinin.
Anat. Embryol., 197: 89–103.
Deller, T., Bas, O.C., Vlachos, A., Merten, T., Del Turco, D.,
Dehn, D., Mundel, P. and Frotscher, M. (2006) Plasticity of
synaptopodin and the spine apparatus organelle in the rat
fascia dentata following entorhinal cortex lesion. J. Comp.
Neurol., 499: 471–484.
Deller, T., Drakew, A. and Frotscher, M. (1999a) Different
primary target cells are important for fiber lamination in the
fascia dentata: a lesson from reeler mutant mice. Exp. Ne-
urol., 156: 239–253.
Deller, T., Drakew, A., Heimrich, B., Förster, E., Tielsch, A.
and Frotscher, M. (1999b) The hippocampus of the reeler
mutant mouse: Fiber segregation in area CA1 depends on the
position of the postsynaptic target cells. Exp. Neurol., 156:
254–267.
Deller, T. and Frotscher, M. (1997) Lesion-induced plasticity of
central neurons: sprouting of single fibers in the rat hippo-
campus after unilateral entorhinal lesion. Prog. Neurobiol.,
53: 687–727.
Deller, T., Frotscher, M. and Nitsch, R. (1995a) Morphological
evidence for the sprouting of inhibitory commissural fibers in
response to the lesion of the excitatory entorhinal input to the
rat dentate gyrus. J. Neurosci., 15: 6868–6878.
Deller, T., Frotscher, M. and Nitsch, R. (1996a) Sprouting of
crossed entorhinodentate fibers after a unilateral entorhinal
lesion: anterograde tracing of fiber reorganization with
Phaseolus vulgaris-leucoagglutinin (PHAL). J. Comp. Ne-
urol., 365: 42–55.
Deller, T., Haas, C.A., Deissenrieder, K., Del Turco, D., Coulin,
C., Gebhardt, C., Drakew, A., Schwarz, K., Mundel, P. and
Frotscher, M. (2002) Laminar distribution of synaptopodin in
normal and reeler mouse brain depends on the position of
spine-bearing neurons. J. Comp. Neurol., 453: 33–44.
Deller, T., Haas, C.A. and Frotscher, M. (2000) Reorganization
of the rat fascia dentata after a unilateral entorhinal cortex
lesion: Role of the extracellular matrix. Ann. N.Y. Acad. Sci.,
911: 207–220.
522
Deller, T., Haas, C.A. and Frotscher, M. (2001) Sprouting in
the hippocampus after entorhinal cortex lesion is layer-spe-
cific but not translaminar: which molecules may be involved?
Restor. Neurol. Neurosci., 19: 159–167.
Deller, T., Haas, C.A., Naumann, T., Joester, A., Faissner, A.
and Frotscher, M. (1997) Upregulation of astrocyte-derived
tenascin-c correlates with neurite outgrowth in the rat dentate
gyrus after unilateral entorhinal cortex lesion. Neuroscience,
81: 829–846.
Deller, T., Nitsch, R. and Frotscher, M. (1995b) Phaseolus vul-
garis-leucoagglutinin (PHAL) tracing of commissural fibers
to the rat dentate gyrus: Evidence for a previously unknown
commissural projection to the outer molecular layer. J.
Comp. Neurol., 352: 55–68.
Deller, T., Nitsch, R. and Frotscher, M. (1996b) Heterogeneity
of the commissural projection to the rat dentate gyrus: a
Phaseolus vulgaris-leucoagglutinin tracing study. Neurosci-
ence, 75: 111–121.
Deller, T., Nitsch, R. and Frotscher, M. (1996c) Layer-specific
sprouting of commissural fibers to the rat fascia dentata after
unilateral entorhinal cortex lesion: a Phaseolus vulgaris-le-
ucoagglutinin tracing study. Neuroscience, 71: 651–660.
Diekmann, S., Ohm, T.G. and Nitsch, R. (1996) Long-lasting
transneuronal changes in rat dentate granule cell dendrites
after entorhinal cortex lesion-a combined intracellular injec-
tion and electron microscopy study. Brain Pathol., 6:
205–214.
Drakew, A., Deller, T., Heimrich, B., Gebhardt, C., Del Turco,
D., Tielsch, A., Forster, E., Herz, J. and Frotscher, M. (2002)
Dentate granule cells in reeler mutants and VLDLR and
ApoER2 knockout mice. Exp. Neurol., 176: 12–24.
Drojdahl, N., Fenger, C., Nielsen, H.H., Owens, T. and Finsen,
B. (2004) Dynamics of oligodendrocyte responses to antero-
grade axonal (Wallerian) and terminal degeneration in nor-
mal and TNF-transgenic mice. J. Neurosci. Res., 75:
203–217.
Drojdahl, N., Hegelund, I.V., Poulsen, F.R., Wree, A. and
Finsen, B. (2002) Perforant path lesioning induces sprouting
of CA3-associated fibre systems in mouse hippocampal for-
mation. Exp. Brain Res., 144: 79–87.
Eckenstein, F. and Sofroniew, M.V. (1983) Identification of cen-
tral cholinergic neurons containing both choline acetytransf-
erase and acetylcholinesterase and of central neurons containing
only acetylcholinesterase. J. Neurosci., 3: 2286–2291.
Eyupoglu, I.Y., Bechmann, I. and Nitsch, R. (2003) Modifica-
tion of microglia function protects from lesion-induced neu-
ronal alterations and promotes sprouting in the
hippocampus. FASEB J., 17: 1110–1111.
Fagan, A.M. and Gage, F.H. (1990) Cholinergic sprouting in
the hippocampus: a proposed role for il-1. Exp. Neurol., 110:
105–120.
Fagan, A.M. and Gage, F.H. (1994) Mechanismus of sprouting
in the adult central nervous system: cellular responses in ar-
eas of terminal degeneration and reinnervation in the rat
hippocampus. Neuroscience, 58: 705–725.
Fagan, A.M., Murphy, B.A., Patel, S.N., Kilbridge, J.F.,
Mobley, W.C., Bu, G. and Holtzman, D.M. (1998)
Evidence for normal aging of the septo-hippocampal
cholinergic system in apoE ( / ) mice but impaired clear-
ance of axonal degeneration products following injury.
Exp. Neurol., 151: 314–325.
Fagan, A.M., Suhr, S.T., Lucidi-Phillipi, C.A., Peterson, D.A.,
Holtzman, D.M. and Gage, F.H. (1997) Endogenous FGF-2
is important for cholinergic sprouting in the denervated hip-
pocampus. J. Neurosci., 17: 2499–2511.
Faillace, M.P., Zwiller, J., Di Scala, G. and Bernabeu, R. (2006)
Odor increases [3H]phorbol dibutyrate binding to protein
kinase C in olfactory structures of rat brain. Effect of
entorhinal cortex lesion. Brain Res., 1068: 16–22.
Faissner, A. and Steindler, D. (1995) Boundaries and inhibitory
molecules in developing neural tissues. Glia, 13: 233–254.
Fife, B.T., Huffnagle, G.B., Kuziel, W.A. and Karpus, W.J.
(2000) CC chemokine receptor 2 is critical for induction of
experimental autoimmune encephalomyelitis. J. Exp. Med.,
192: 899–905.
Finsen, B., Jensen, M.B., Lomholt, N.D., Hegelund, I.V., Po-
ulsen, F.R. and Owens, T. (1999) Axotomy-induced glial re-
actions in normal and cytokine transgenic mice. Adv. Exp.
Med. Biol., 468: 157–171.
Frey, D., Laux, T., Xu, L., Schneider, C. and Caroni, P. (2000)
Shared and unique roles of CAP23 and GAP43 in actin reg-
ulation, neurite outgrowth, and anatomical plasticity. J. Cell
Biol., 149: 1443–1454.
Frotscher, M. (1988) Neuronal elements in the hippocampus
and their synaptic connections. Adv. Anat. Embryol., 111:
2–17.
Frotscher, M., Heimrich, B. and Deller, T. (1997) Sprouting in
the hippocampus is layer-specific. Trends Neurosci., 20:
218–223.
Fujise, N., Liu, Y., Hori, N. and Kosaka, T. (1998) Distribu-
tion of calretinin immunoreactivity in the mouse dentate
gyrus: II. Mossy cells, with special reference to their dorso-
ventral difference in calretinin immunoreactivity. Neurosci-
ence, 82: 181–200.
Gage, F.H., Olejniczak, P. and Armstrong, D.M. (1988) As-
trocytes are important for sprouting in the septohippocampal
circuit. Exp. Neurol., 102: 2–13.
Gall, C., Ivy, G. and Lynch, G. (1986) Neuroanatomical plas-
ticity. Its role in organizing and reorganizing the central
nervous system. Hum. Growth, 2: 411–436.
Gall, C. and Lynch, G. (1978) Rapid axon sprouting in the
neonatal rat hippocampus. Brain Res., 153: 357–362.
Gall, C. and Lynch, G. (1981) Fiber architecture of the dentate
gyrus following ablation of the entorhinal cortex in rats of
different ages: evidence for two forms of axon sprouting in
the immature brain. Neuroscience, 6: 903–910.
Gall, C. and Lynch, G.S. (1980) The regulation of fiber growth
and synaptogenesis in the developing hippocampus. Curr.
Top. Dev. Biol., 1: 159–180.
Gall, C., McWilliams, R. and Lynch, G. (1979a) The effect of
collateral sprouting on the density of innervation of normal
target sites: implications for theories on the regulation of the
size of the developing synaptic domains. Brain Res., 175:
37–47.

Gall, C., Rose, G. and Lynch, G. (1979b) Proliferative and
migratory activity of glial cells in the partially deafferented
hippocampus. J. Comp. Neurol., 183: 539–550.
Gebhardt, C., Del Turco, D., Drakew, A., Tielsch, A., Herz, J.,
Frotscher, M. and Deller, T. (2002) Abnormal positioning of
granule cells alters afferent fiber distribution in the mouse
fascia dentata: morphologic evidence from reeler, apolipo-
protein E receptor 2-, and very low density lipoprotein
receptor knockout mice. J. Comp. Neurol., 445: 278–292.
Gehrmann, J., Schoen, S.W. and Kreutzberg, G.W. (1991)
Lesion of the rat entorhinal cortex leads to a rapid microglial
reaction in the dentate gyrus. Acta Neuropathol., 82:
442–455.
Gold, R., Linington, C. and Lassmann, H. (2006) Understand-
ing pathogenesis and therapy of multiple sclerosis via animal
models: 70 years of merits and culprits in experimental
autoimmune encephalomyelitis research. Brain, 129:
1953–1971.
Goldmann, J., Kwidzinski, E., Brandt, C., Mahlo, J., Richter,
D. and Bechmann, I. (2006) T cells traffic from brain to
cervical lymph nodes via the cribroid plate and the nasal
mucosa. J. Leukoc. Biol., 80: 797–801.
Goldowitz, D., White, W.F., Steward, O., Cotman, C.W. and
Lynch, G. (1975) Anatomical evidence for a projection from
the entorhinal cortex to the contralateral dentate gyrus of the
rat. Exp. Neurol., 47: 433–441.
Gomez-Pinilla, F., Lee, J.W. and Cotman, C.W. (1992) Basic
FGF in adult rat brain: cellular distribution and response to
entorhinal lesion and fimbria-fornix transection. J. Neurosci.,
12: 345–355.
van Groen, T., Kadish, I. and Wyss, J.M. (2002) Species differ-
ences in the projections from the entorhinal cortex to the
hippocampus. Brain Res. Bull., 57: 553–556.
van Groen, T., Miettinen, P. and Kadish, I. (2003) The ento-
rhinal cortex of the mouse: organization of the projection to
the hippocampal formation. Hippocampus, 13: 133–149.
Guthrie, K.M., Nguyen, T. and Gall, C.M. (1995) Insulin-like
growth factor-1 mRNA is increased in deafferented hippo-
campus: spatiotemporal correspondence of a trophic event
with axon sprouting. J. Comp. Neurol., 352: 147–160.
Guthrie, K.M., Woods, A.G., Nguyen, T. and Gall, C.M.
(1997) Astroglial ciliary neurotrophic factor mRNA expres-
sion is increased in fields of axonal sprouting in deafferented
hippocampus. J. Comp. Neurol., 386: 137–148.
Gwag, B.J., Sessler, F.M., Kimmerer, K. and Springer, J.E.
(1994) Neurotrophic factor mRNA expression in the dentate
gyrus is increased following angular bundle transection.
Brain Res., 647: 23–29.
Haas, C.A., Rauch, U., Thon, N., Merten, T. and Deller, T.
(1999) Entorhinal cortex lesion in adult rats induce the
expression of the neuronal chondroitin sulfate proteoglycan
neurocan in reactive astrocytes. J. Neurosci., 19:
9953–9963.
Hailer, N.P., Bechmann, I., Heizmann, S. and Nitsch, R. (1997)
Adhesion molecule expression on phagocytic microglial cells
following anterograde degeneration of perforant path axons.
Hippocampus, 7: 341–349.
523
Hardman, R., Evans, D.J., Fellows, L., Hayes, B., Rupniak,
H.T., Barnes, J.C. and Higgins, G.A. (1997) Evidence for
recovery of spatial learning following entorhinal cortex le-
sions in mice. Brain Res., 758: 187–200.
Harris, E.W., Lasher, S.S. and Steward, O. (1978) Habitutation
like decrements in transmission along the normal and lesion-
induced temporodentate pathways in the rat. Brain Res., 151:
623–631.
Heckers, S., Ohtake, T., Wiley, R.G., Lappi, D.A., Geula, C.
and Mesulam, M.M. (1994) Complete and selective choli-
nergic denervation of rat neocortex and hippocampus but not
amygdala by an immunotoxin against the p75 NGF receptor.
J. Neurosci., 14: 1271–1289.
Henderson, Z., Harrison, P.S., Jagger, E. and Beeby, J.H.
(1998) Density of choline acetyltransferase-immunreactive
terminals in the rat dentate gyrus after entorhinal cortex le-
sions: a quantitative light microscope study. Exp. Neurol.,
152: 50–63.
Hesselgesser, J. and Horuk, R. (1999) Chemokine and chemo-
kine receptor expression in the central nervous system. J.
Neurovirol., 5: 13–26.
Hjorth-Simonsen, A. and Jeune, B. (1972) Origin and termina-
tion of the hippocampal perforant path in the rat studied by
silver impregnation. J. Comp. Neurol., 144: 215–232.
Hoff, S.F., Scheff, S.W., Kwan, A.Y. and Cotman, C.W. (1981)
A new type of lesion-induced synaptogenesis: I. Synaptic
turnover in non-denervated zones of the dentate gyrus in
young adult rats. Brain Res., 222: 1–13.
Höke, A. and Silver, J. (1996) Proteoglycans and other repul-
sive molecules in glial boundaries during development and
regeneration of the nervous system. Prog. Brain Res., 108:
149–163.
Huang, D.R., Wang, J., Kivisakk, P., Rollins, B.J. and Ran-
sohoff, R.M. (2001) Absence of monocyte chemoattractant
protein 1 in mice leads to decreased local macrophage
recruitment and antigen-specific T helper cell type 1 immune
response in experimental autoimmune encephalomyelitis. J.
Exp. Med., 193: 713–726.
Izikson, L., Klein, R.S., Charo, I.F., Weiner, H.L. and Luster,
A.D. (2000) Resistance to experimental autoimmune encep-
halomyelitis in mice lacking the CC chemokine receptor
(CCR)2. J. Exp. Med., 192: 1075–1080.
Jakab, R.L. and Leranth, C. (1995) Septum. In: Paxinos G.
(Ed.), The Rat Nervous System (2nd ed.). Academic Press,
San Diego, CA, pp. 405–442.
Jensen, M.B., Finsen, B. and Zimmer, J. (1997) Morphological
and immunophenotypic microglial changes in the denervated
fascia dentata of adult rats: correlation with blood-brain
barrier damage and astroglial reactions. Exp. Neurol., 143:
103–116.
Jensen, M.B., Gonzàlez, B., Castellano, B. and Zimmer, J.
(1994) Microglial and astroglial reactions to anterograde ax-
onal degeneration: a histochemical and immuncytochemical
study of the adult rat fascia dentata after entorhinal perfo-
rant path lesions. Exp. Brain Res., 98: 245–260.
Jensen, M.B., Hegelund, I.V., Lomholt, N.D., Finsen, B. and
Owens, T. (2000a) IFNgamma enhances microglial reactions
524
to hippocampal axonal degeneration. J. Neurosci., 20:
3612–3621.
Jensen, M.B., Hegelund, I.V., Poulsen, F.R., Owens, T.,
Zimmer, J. and Finsen, B. (1999) Microglial reactivity cor-
relates to the density and the myelination of the anterograd-
ely degenerating axons and terminals following perforant
path denervation of the mouse fascia dentata. Neuroscience,
93: 507–518.
Jensen, M.B., Poulsen, F.R. and Finsen, B. (2000b) Axonal
sprouting regulates myelin basic protein gene expression in
denervated mouse hippocampus. Int. J. Dev. Biol., 18:
221–235.
de Jong, E.K., Dijkstra, I.M., Hensens, M., Brouwer, N., van
Amerongen, M., Liem, R.S., Boddeke, H.W. and Biber, K.
(2005) Vesicle-mediated transport and release of CCL21 in
endangered neurons: a possible explanation for microglia
activation remote from a primary lesion. J. Neurosci., 25:
7548–7557.
Jucker, M., D‘Amato, F., Mondadori, C., Mohajeri, H., Ma-
gyar, J., Bartsch, U. and Schachner, M. (1996) Expression of
the neural adhesion molecule L1 in the deafferented dentate
gyrus. Neuroscience, 75: 703–715.
Kadish, I. and van Groen, T. (2002) Low levels of estrogen
significantly diminish axonal sprouting after entorhinal cor-
tex lesions in the mouse. J. Neurosci., 22: 4095–4102.
Kadish, I. and van Groen, T. (2003) Differences in lesion-
induced hippocampal plasticity between mice and rats.
Neuroscience, 11: 499–509.
Kadish, I., Pradier, L. and van Groen, T. (2002) Transgenic
mice expressing the human presenilin 1 gene demonstrate
enhanced hippocampal reorganization following entorhinal
cortex lesion. Brain Res. Bull., 57: 587–594.
Kapfhammer, J.P. and Schwab, M.E. (1992) Modulators of
neuronal migration and neurite growth. Curr. Opin. Cell
Biol., 4: 863–868.
Karpus, W.J. and Ransohoff, R.M. (1998) Chemokine regula-
tion of experimental autoimmune encephalomyelitis: tempo-
ral and spatial expression patterns govern disease
pathogenesis. J. Immunol., 161: 2667–2671.
Kawaja, M.D. and Gage, F.H. (1991) Reactive astrocytes are
substrates for the growth of adult CNS axons in the presence
of elevated levels of nerve growth factor. Neuron, 7:
1019–1030.
Kelley, M.S. and Steward, O. (1996a) The process of reinner-
vation in the dentate gyrus of adult rats: physiological events
at the time of the lesion and during the early postlesion
period. Exp. Neurol., 139: 73–82.
Kelley, M.S. and Steward, O. (1996b) The role of postlesion
seizures and spreading depression in the upregulation of glial
fibrillary acidic protein mRNA after entorhinal cortex
lesions. Exp. Neurol., 139: 83–94.
Kelley, M.S. and Steward, O. (1998) Injury-induced physiolog-
ical events that may modulate gene expression in neurons and
glia. Rev. Neurosci., 8: 147–177.
Kerschensteiner, M., Gallmeier, E., Behrens, L., Leal, V.V.,
Misgeld, T., Klinkert, W.E., Kolbeck, R., Hoppe, E., Or-
opeza-Wekerle, R.L., Bartke, I., Stadelmann, C., Lassmann,
H., Wekerle, H. and Hohlfeld, R. (1999) Activated human T
cells, B cells, and monocytes produce brain-derived neurotro-
phic factor in vitro and in inflammatory brain lesions: a ne-
uroprotective role of inflammation? J. Exp. Med., 189:
865–870.
Kipnis, J., Yoles, E., Schori, H., Hauben, E., Shaked, I. and
Schwartz, M. (2001) Neuronal survival after CNS insult is
determined by a genetically encoded autoimmune response. J.
Neurosci., 21: 4564–4571.
Kosaka, T., Katsumaru, H., Hama, K., Wu, J.Y. and Heiz-
mann, C.W. (1987) GABAergic neurons containing CA2+ -
binding protein parvalbumin in the rat hippocampus and
dentate gyrus. Brain Res., 419: 119–130.
Kovac, A.D., Kwidzinski, E., Heimrich, B., Bittigau, P., Deller,
T., Nitsch, R. and Bechmann, I. (2004) Entorhinal cortex
lesion in the mouse induces transsynaptic death of perforant
path target neurons. Brain Pathol., 14: 249–257.
Ladeby, R., Wirenfeldt, M., Dalmau, I., Gregersen, R.,
Garcia-Ovejero, D., Babcock, A., Owens, T. and Finsen, B.
(2005a) Proliferating resident microglia express the stem
cell antigen CD34 in response to acute neural injury. Glia,
50: 121–131.
Ladeby, R., Wirenfeldt, M., Garcia-Ovejero, D., Fenger, C.,
Dissing-Olesen, L., Dalmau, I. and Finsen, B. (2005b) Mi-
croglial cell population dynamics in the injured adult central
nervous system. Brain Res. Brain Res. Rev., 48: 196–206.
Laux, T., Fukami, K., Thelen, M., Golub, T., Frey, D. and
Caroni, P. (2000) GAP43, MARCKS, and CAP23 modulate
PI(4,5)P(2) at plasmalemmal rafts, and regulate cell cortex
actin dynamics through a common mechanism. J. Cell Biol.,
149: 1455–1472.
Lee, M.Y., Deller, T., Kirsch, M., Frotscher, M. and Hofmann,
H.D. (1997) Differential regulation of CNTF and CNTF
receptor alpha expression in astrocytes and neurons of the
fascia dentata following entorhinal cortex lesion. J. Neuro-
sci., 17: 1137–1146.
Levey, A.I., Wainer, B.H., Mufson, E.J. and Mesulam, M.M.
(1983) Co-localization of acetylcholinesterase and choline
acetyltransferase in the rat cerebrum. Neuroscience, 9: 9–22.
Levey, A.I., Wainer, B.H., Rye, D.B., Mufson, E.J. and Me-
sulam, M.M. (1984) Choline acetyltransferase-immunoreac-
tive neurons intrinsic to rodent cortex and distinction from
acetylcholinesterase-positive neurons. Neuroscience, 13:
341–353.
Linke, R., Schwegler, H. and Boldyreva, M. (1994) Cholinergic
and GABAergic septo-hippocampal projection neurons in
mice: a retrograde tracing study combined with double
immunocytochemistry for choline acetyltransferase and parv-
albumin. Brain Res., 653: 73–80.
Liu, Y., Fujise, N. and Kosaka, T. (1996) Distribution of ca-
lretinin immunoreactivity in the mouse dentate gyrus. I.
General description. Exp. Brain Res., 108: 389–403.
Loesche, J. and Steward, O. (1977) Behavioral correlates of
denervation and reinnervation of the hippocampal formation
of the rat: recovery of alternation performance following
unilateral entorhinal cortex lesions. Brain Res. Bull., 2:
31–39.

Luster, A.D. (1998) Chemokines–chemotactic cytokines that
mediate inflammation. N. Engl. J. Med., 338: 436–445.
Lynch, G. and Cotman, C. (1975) The hippocampus as a model
for studying anatomical plasticity in the adult brain. In: Is-
aacson R.L. and Pribram K.H. (Eds.), The Hippocampus,
Vol. 1. Plenum Press, New York, pp. 123–154.
Lynch, G., Gall, C., Rose, G. and Cotman, C.W. (1976)
Changes in the distribution of the dentate gyrus associational
system following unilateral or bilateral entorhinal lesion in
the adult rat. Brain Res., 110: 57–71.
Lynch, G., Matthews, D.A., Mosko, S., Parks, T. and Cotman,
C.W. (1972) Induced acetylcholinesterase-rich layer in rat
dentate gyrus following entorhinal lesions. Brain Res., 42:
311–318.
Lynch, G., Stanfield, B. and Cotman, C.W. (1973) Develop-
mental differences in postlesion axonal growth in the hippo-
campus. Brain Res., 59: 155–168.
Lynch, G.S., Stanfield, B., Parks, T. and Cotman, C.W. (1974)
Evidence for selective post-lesion axonal growth in the dent-
ate gyrus of the rat. Brain Res., 69: 1–11.
Mackay, C.R. (2001) Chemokines: immunology’s high impact
factors. Nat. Immunol., 2: 95–101.
Margolis, R.U. and Margolis, R.K. (1997) Chondroitin sulfate
proteoglycans as mediators of axon growth and pathfinding.
Cell Tissue Res., 290: 343–348.
Marrone, D.F., LeBoutillier, J.C. and Petit, T.L. (2004a)
Changes in synaptic ultrastructure during reactive syna-
ptogenesis in the rat dentate gyrus. Brain Res., 1005:
124–136.
Marrone, D.F., LeBoutillier, J.C. and Petit, T.L. (2004b) Com-
parative analyses of synaptic densities during reactive syna-
ptogenesis in the rat dentate gyrus. Brain Res., 996: 19–30.
Matthews, D.A., Cotman, C.W. and Lynch, G. (1976a) An
electron microscopic study of lesion-induced synaptogenesis
in the dentate gyrus of the adult rat. I. Magnitude and time
course of degeneration. Brain Res., 115: 1–21.
Matthews, D.A., Cotman, C.W. and Lynch, G. (1976b) An
electron microscopic study of lesion-induced synaptogenesis
in the dentate gyrus of the adult rat. II. Reappearance of
morphologically normal synaptic contacts. Brain Res., 115:
23–41.
Mayer, J., Hamel, M.G. and Gottschall, P.E. (2005) Evidence
for proteolytic cleavage of brevican by the ADAMTSs in the
dentate gyrus after excitotoxic lesion of the mouse entorhinal
cortex. BMC Neurosci., 6: 52.
McKeon, R.J., Vietje, B.P. and Wells, J. (1989) Increase in
acetylcholinesterase in the molecular layer of the denate
gyrus in the absence of septal inputs following selective gran-
ule cell lesions. Brain Res., 503: 317–321.
Meier, S., Brauer, A.U., Heimrich, B., Nitsch, R. and Sa-
vaskan, N.E. (2004) Myelination in the hippocampus during
development and following lesion. Cell Mol. Life Sci., 61:
1082–1094.
Meier, S., Brauer, A.U., Heimrich, B., Schwab, M.E., Nitsch,
R. and Savaskan, N.E. (2003) Molecular analysis of Nogo
expression in the hippocampus during development and fol-
lowing lesion and seizure. FASEB J., 17: 1153–1155.
525
Mingorance, A., Fontana, X., Soriano, E. and del Rio, J.A.
(2005) Overexpression of myelin-associated glycoprotein af-
ter axotomy of the perforant pathway. Mol. Cell Neurosci.,
29: 471–483.
Miwa, C. and Ueki, A. (1996) Effects of entorhinal cortex le-
sion on learning behavior and on hippocampus in the rat.
Psychiatry Clin. Neurosci., 50: 223–230.
Mutlu, L., Brandt, C., Kwidzinski, E., Sawitzki, B., Gimsa, U.,
Mahlo, J., Aktas, O., Nitsch, R., van Zwam, M., Laman, J.
and Bechmann, I. (2007) Tolerogenic effect of fiber tract in-
jury: Peripheral apoptosis of T cells and reduction of EAE
severity following entorhinal cortex lesion. Exp. Brain Res.,
178: 542–553.
Nadler, J.V., Cotman, C., Paoletti, C. and Lynch, G.S. (1977a)
Histochemical evidence of altered development of cholinergic
fibers in the rat dentate gyrus following lesions. II. Effects of
partial entorhinal and simultaneous multiple lesions. J.
Comp. Neurol., 171: 589–604.
Nadler, J.V., Cotman, C.W. and Lynch, G.S. (1977b) Histo-
chemical evidence of altered development of cholinergic fib-
ers in the rat dentate gyrus following lesions. I. Time course
after complete unilateral entorhinal lesion at various ages. J.
Comp. Neurol., 171: 561–587.
Nadler, J.V. and Evenson, D.A. (1982) Autoradiographic ev-
idence that septohippocampal fibers reinnervate fascia den-
tata denervated by entorhinal lesion during development.
Anat. Embryol., 165: 113–123.
Nadler, J.V., Perry, B.W. and Cotman, C.W. (1980) Interaction
with CA4-derived fibers accounts for distribution of septo-
hippocampal fibers in rat fascia denata after entorhinal le-
sion. Exp. Neurol., 68: 185–194.
Nafstad, P.H.J. (1967) An electron microscope study on the
termination of the perforant path fibres in the hippocampus
and fascia denata. Z. Zellforsch. Mikrosk. Anat., 76:
532–542.
Naumann, T., Deller, T., Bender, R. and Frotscher, M. (1997)
192 IgG-saporin-induced loss of cholinergic neurons in
the septum abolishes cholinergic sprouting after unilat-
eral entorhinal lesion in the rat. Eur. J. Neurosci., 9:
1304–1313.
Nielsen, H.H., Ladeby, R., Drojdahl, N., Peterson, A.C. and
Finsen, B. (2006) Axonal degeneration stimulates the forma-
tion of NG2+ cells and oligodendrocytes in the mouse. Glia,
54: 105–115.
Nitsch, R., Bader, S. and Frotscher, M. (1992) Reorganization
of input synapses of parvalbumin-containing neurons in the
rat fascia dentata following entorhinal lesion. Neurosci.
Lett., 135: 33–36.
Nitsch, R. and Frotscher, M. (1991) Maintenance of peripheral
dendrites of GABAergic neurons requires specific input.
Brain Res., 554: 304–307.
Nitsch, R. and Frotscher, M. (1992) Reduction of posttrau-
matic transneuronal ‘‘early gene’’ activation and dendritic
atrophy by the N-methyl-D-aspartate receptor antagonist
MK-801. Proc. Natl. Acad. Sci. U.S.A., 89: 5197–5200.
Nitsch, R. and Frotscher, M. (1993) Transneuronal changes in
dendrites of GABAergic parvalbumin-containing neurons of
526
the rat fascia dentata following entorhinal lesion. Hippo-
campus, 3: 481–490.
Nitsch, R., Soriano, E. and Frotscher, M. (1990) The parval-
bumin-containing nonpyramidal neurons in the rat hippo-
campus. Anat. Embryol., 181: 413–425.
Nyakas, C., Luiten, P.G.M., Balkan, B. and Spencer, D.G.J.
(1988) Changes in septohippocampal projections after lateral
entorhinal or combined entorhinal-raphè lesions as studied
by anterograde tracing methods. Brain Res. Bull., 21:
285–293.
Nyakas, C., Luiten, P.G.M., Spencer, D.G. and Traber, J.
(1987) Detailed projection patterns of septal and diagonal
band efferents to the hippocampus in the rat with emphasis
on innervation of CA1 and dentate gyrus. Brain Res. Bull.,
18: 533–545.
Otal, R., Burgaya, F., Frisen, J., Soriano, E. and Martinez, A.
(2006) Ephrin-A5 modulates the topographic mapping and
connectivity of commissural axons in murine hippocampus.
Neuroscience, 141: 109–121.
Parent, A., Dea, D., Quirion, R. and Poirier, J. (1993)
[3H]phorbol ester binding sites and neuronal plasticity in
the hippocampus following entorhinal cortex lesions. Brain
Res., 607: 23–32.
Parnavelas, J.G., Lynch, G., Brecha, N., Cotman, C.W. and
Globus, A. (1974) Spine loss and regrowth in hippocampus
following deafferentation. Nature, 248: 71–73.
Pearlman, A.L. and Sheppard, A.M. (1996) Extracellular matrix
in early cortical development. Prog. Brain Res., 108: 117–134.
Pedersen, M.D., Minuzzi, L., Wirenfeldt, M., Meldgaard, M.,
Slidsborg, C., Cumming, P. and Finsen, B. (2006) Up-reg-
ulation of PK11195 binding in areas of axonal degeneration
coincides with early microglial activation in mouse brain.
Eur. J. Neurosci., 24: 991–1000.
Peterson, G.M. (1994) Sprouting of central noradrenergic
fibers in the dentate gyrus following combined lesions of its
entorhinal and septal afferents. Hippocampus, 4: 635–648.
Phinney, A.L., Calhoun, M.E., Woods, A.G., Deller, T. and
Jucker, M. (2004) Stereological analysis of the reorganization
of the dentate gyrus following entorhinal cortex lesion in
mice. Eur. J. Neurosci., 19: 1731–1740.
Poirier, J., May, P.C., Osterburg, H.H., Geddes, J., Cotman,
C.W. and Finch, C.E. (1990) Selective alterations of RNA in
rat hippocampus after entorhinal cortex lesioning. Proc.
Natl. Acad. Sci. U.S.A., 87: 303–307.
Poulsen, F.R., Lagord, C., Courty, J., Pedersen, E.B., Bar-
ritault, D. and Finsen, B. (2000) Increased synthesis of he-
parin affin regulatory peptide in the perforant path lesioned
mouse hippocampal formation. Exp. Brain Res., 135:
319–330.
Ramirez, J.J. (2001) The role of axonal sprouting in functional
reorganization after CNS injury: lessons from the hippocam-
pal formation. Restor. Neurol. Neurosci., 19: 237–262.
Ramirez, J.J. and Stein, D.G. (1984) Sparing and recovery of
spatial alternation performance after entorhinal cortex le-
sions in rats. Behav. Brain Res., 13: 53–61.
Rappert, A., Bechmann, I., Pivneva, T., Mahlo, J., Biber, K.,
Nolte, C., Kovac, A.D., Gerard, C., Boddeke, H.W., Nitsch,
R. and Kettenmann, H. (2004) CXCR3-dependent microglial
recruitment is essential for dendrite loss after brain lesion. J.
Neurosci., 24: 8500–8509.
Rappert, A., Biber, K., Nolte, C., Lipp, M., Schubel, A., Lu, B.,
Gerard, N.P., Gerard, C., Boddeke, H.W. and Kettenmann,
H. (2002) Secondary lymphoid tissue chemokine (CCL21)
activates CXCR3 to trigger a Cl current and chemotaxis in
murine microglia. J. Immunol., 168: 3221–3226.
Reeves, T.M. and Steward, O. (1986) Emergence of the capacity
for LTP during reinnervation of the dentate gyrus: evidence
that abnormally shaped spines can mediate LTP. Exp. Brain
Res., 65: 167–175.
Reeves, T.M. and Steward, O. (1988) Changes in the firing
properties of neurons in the dentate gyrus with denervation
and reinnervation: implications for behavioral recovery. Exp.
Neurol., 102: 37–49.
Rollins, B.J. (1997) Chemokines. Blood, 90: 909–928.
Rose, G., Lynch, G. and Cotman, C.W. (1976) Hypertrophy
and redistribution of astrocytes in the deafferented dentate
gyrus. Brain Res. Bull., 1: 87–92.
Sattler, R., Charlton, M.P., Hafner, M. and Tymianski, M.
(1998) Distinct influx pathways, not calcium load, determine
neuronal vulnerability to calcium neurotoxicity. J. Neuroc-
hem., 71: 2349–2364.
Savaskan, N.E. and Nitsch, R. (2001) Molecules involved in
reactive sprouting in the hippocampus. Rev. Neurosci., 12:
195–215.
Schafer, M., Brauer, A.U., Savaskan, N.E., Rathjen, F.G. and
Brummendorf, T. (2005) Neurotractin/kilon promotes neu-
rite outgrowth and is expressed on reactive astrocytes after
entorhinal cortex lesion. Mol. Cell Neurosci., 29: 580–590.
Schauwecker, P.E. and Steward, O. (1997) Genetic influences
on cellular reactions to brain injury: activation of microglia
in denervated neuropil in mice carrying a mutation (WLDs)
that causes delayed wallerian degeneration. J. Comp. Ne-
urol., 380: 82–94.
Scheff, S.W. and Cotman, C.W. (1977) Recovery of spontane-
ous alternation following lesions of the entorhinal cortex in
adult rats: possible correlation to axon sprouting. Behav.
Bio., 21: 286–293.
Schenk, F. and Morris, R.G.M. (1985) Dissociation between
components of spatial memory in rats after recovery from the
effects of retrohippocampal lesions. Exp. Brain Res., 58: 11–28.
Schwab, M.E., Kapfhammer, J.P. and Bandtlow, C.E. (1993)
Inhibitors of neurite growth. Annu. Rev. Neurosci., 16:
565–595.
Shi, B. and Stanfield, B.B. (1996) Differential sprouting re-
sponses in axonal fiber systems in the dentate gyrus following
lesions of the perforant path in WLDs mutant mice. Brain
Res., 740: 89–101.
Siebert, H., Sachse, A., Kuziel, W.A., Maeda, N. and Bruck,
W. (2000) The chemokine receptor CCR2 is involved in ma-
crophage recruitment to the injured peripheral nervous sys-
tem. J. Neuroimmunol., 110: 177–185.
Soriano, E. and Frotscher, M. (1994) Mossy cells of the rat
fascia dentata are glutamate-immunoreactive. Hippocampus,
4: 65–70.

Soriano, E., del Rio, J.A., Martinez, A. and Super, H. (1994)
Organization of the embryonic and early postnatal murine
hippocampus. I. Immunocytochemical characterization of
neuronal populations in the subplate and marginal zone. J.
Comp. Neurol., 342: 571–595.
Stanfield, B.B., Caviness, V.S.J. and Cowan, W.M. (1979) The
organization of certain afferents to the hippocampus and
dentate gyrus in normal and reeler mice. J. Comp. Neurol.,
185: 461–483.
Stanfield, B.B. and Cowan, W.M. (1979a) The development of
the hippocampus and dentate gyrus in normal and reeler
mice. J. Comp. Neurol., 185: 423–460.
Stanfield, B.B. and Cowan, W.M. (1979b) The morphology of
the hippocampus and dentate gyrus in normal and reeler
mice. J. Comp. Neurol., 185: 393–422.
Stanfield, B.B. and Cowan, W.M. (1982) The sprouting of
septal afferents to the dentate gyrus after lesions of the
entorhinal cortex in adult rats. Brain Res., 232: 162–170.
Stäubli, U., Gall, C. and Lynch, G. (1984) The distribution of
the commissural-associational afferents of the dentate gyrus
after perforant path lesions in one-day-old rats. Brain Res.,
292: 156–159.
Steward, O. (1976a) Reinnervation of the dentate gyrus by ho-
mologous afferents following entorhinal cortical lesion in
adult rats. Science, 194: 426–428.
Steward, O. (1976b) Topographic organization of the projec-
tions from the entorhinal area to the hippocampal formation
of the rat. J. Comp. Neurol., 167: 285–314.
Steward, O. (1980) Trajectory of contralateral entorhinal axons
which reinnervate the fascia dentata of the rat following
ipsilateral entorhinal lesions. Brain Res., 183: 277–289.
Steward, O. (1991) Synapse replacement on cortical neurons
following denervation. Cereb. Cortex, 9: 81–132.
Steward, O. (1992) Signals that induce sprouting in the central
nervous system: sprouting is delayed in a strain of mouse
exhibiting delayed axonal degeneration. Exp. Neurol., 118:
340–351.
Steward, O. (1994a) Cholinergic sprouting is blocked by re-
peated induction of electroconvulsive seizures, a manipula-
tion that induces a persistent reactive state in astrocytes. Exp.
Neurol., 129: 103–111.
Steward, O. (1994b) Reorganization of neuronal circuitry fol-
lowing central nervous system trauma: naturally occurring
processes and opportunities for therapeutic intervention. In:
Salzman S.K. and Faden A.I. (Eds.), The Neurobiology of
Central Nervous System Trauma. Oxford University Press,
New York, pp. 266–287.
Steward, O., Cotman, C.W. and Lynch, G. (1974) Growth of a
new fiber projection in the brain of adult rats: reinnervation
of the dentate gyrus by the contralateral entorhinal cortex
following ipsilateral entorhinal lesion. Exp. Brain Res., 20:
45–66.
Steward, O., Cotman, C.W. and Lynch, G.S. (1973) Re-estab-
lishment of electrophysiologically functional entorhinal cor-
tical input to the dentate gyrus deafferented by ipsilateral
entorhinal lesions: innervation by the contralateral cortex.
Exp. Brain Res., 18: 396–414.
527
Steward, O., Cotman, C.W. and Lynch, G.S. (1976) A quan-
titative autoradiographic and electrophysiological study of
the reinnervation of the dentate gyrus by the contralateral
entorhinal cortex following ipsilateral lesions. Brain Res.,
114: 181–200.
Steward, O., Kelley, M.S. and Schauwecker, P.E. (1997) Signals
that regulate astroglial gene expression: induction of GFAP
mRNA following seizures or injury is blocked by protein
synthesis inhibitors. Exp. Neurol., 148: 100–109.
Steward, O., Kelley, M.S. and Torre, E.R. (1993) The process
of reinnervation in the dentate gyrus of adult rats: temporal
relationship between changes in the level of glial fibrillary
acidic protein (GFAP ) and GFAP mRNA in reactive as-
trocytes. Exp. Neurol., 124: 167–183.
Steward, O. and Scoville, S.A. (1976) Cells of origin of
entorhinal cortical afferents to the hippocampus and fascia
dentata of the rat. J. Comp. Neurol., 169: 347–370.
Steward, O., Torre, E.R., Phillips, L. and Trimmer, P.A. (1990)
The process of reinnervation in the dentate gyrus of adult
rats: time course of increases in mRNA for glial fibrillary
acidic protein. J. Neurosci., 10: 2373–2384.
Steward, O. and Trimmer, P.A. (1997) Genetic influences on
cellular reactions to CNS injury: the reactive response of as-
trocytes in denervated neuropil regions in mice carrying a
mutation (Wld(S)) that causes delayed Wallerian degenera-
tion. J. Comp. Neurol., 380: 70–81.
Steward, O. and Vinsant, S.L. (1978) Identification of the cells
of origin of a central pathway which sprouts following lesions
in mature rats. Brain Res., 147: 223–243.
Steward, O. and Vinsant, S.L. (1983) The process of reinner-
vation in the dentate gyrus of the adult rat: a quantitative
electron microscopic analysis of terminal proliferation
and reactive synaptogenesis. J. Comp. Neurol., 214: 370–386.
Steward, O., Vinsant, S.L. and Davis, L. (1988) The process of
reinnervation in the dentate gyrus of adlut rats: an ultra-
structural study of changes in presynaptic terminals as a re-
sult of sprouting. J. Comp. Neurol., 267: 203–210.
Stone, D.J., Rozovsky, I., Morgan, T.E., Anderson, C.P. and
Finch, C.E. (1998) Increased synaptic sprouting in response to
estrogen via an apolipoprotein E-dependent mechanism: im-
plications for Alzheimer’s disease. J. Neurosci., 18: 3180–3185.
Storm-Mathisen, J. (1974) Choline acetyltransferase and ace-
tylcholinesterase in fascia dentata following lesion of the
entorhinal afferents. Brain Res., 80: 181–197.
Styren, S.D., Lagenaur, C.F., Miller, P.D. and DeKosky, S.T.
(1994) Rapid expression and transport of embryonic N-CAM
in dentate gyrus following entorhinal cortex lesion: ultra-
structural analysis. J. Comp. Neurol., 349: 486–492.
Styren, S.D., Miller, P.D., Lagenaur, C.F. and DeKosky, S.T.
(1995) Alternate strategies in lesion-induced reactive syna-
ptogenesis: differential expression of L1 in two populations
of sprouting axons. Exp. Neurol., 131: 165–173.
Supèr, H. and Soriano, E. (1994) The organization of the em-
bryonic and early postnatal murine hippocampus. II. Devel-
opment of entorhinal, commissural, and septal connections
studied with the lipophilic tracer dil. J. Comp. Neurol., 344:
101–120.
528
Thon, N., Haas, C.A., Rauch, U., Merten, T., Fässler, R.,
Frotscher, M. and Deller, T. (2000) The chondroitin sulphate
proteoglycan brevican is upregulated by astrocytes after
entorhinal cortex lesions in adult rats. Eur. J. Neurosci., 12:
2547–2558.
Ueki, A., Miwa, C., Oohara, K. and Miyoshi, K. (1996) His-
tological evidence for cholinergic alteration in the hippo-
campus following entorhinal cortex lesion. J. Neurol. Sci.,
142: 7–11.
White, F., Nicoll, J.A. and Horsburgh, K. (2001a) Alterations
in ApoE and ApoJ in relation to degeneration and regener-
ation in a mouse model of entorhinal cortex lesion. Exp.
Neurol., 169: 307–318.
White, F., Nicoll, J.A., Roses, A.D. and Horsburgh, K. (2001b)
Impaired neuronal plasticity in transgenic mice expressing
human apolipoprotein E4 compared to E3 in a model of
entorhinal cortex lesion. Neurobiol. Dis., 8: 611–625.
White, W.F., Goldowitz, D., Lynch, G. and Cotman, C.W.
(1976) Electrophysiological analysis of the projection from
the contralateral entorhinal cortex to the dentate gyrus in
normal rats. Brain Res., 114: 201–209.
Wiley, R.G., Oeltmann, T.N. and Lappi, D.A. (1991)
Immunolesioning: selective destruction of neurons using
immunotoxin to rat NGF receptor. Brain Res., 562: 149–153.
Wirenfeldt, M., Dalmau, I. and Finsen, B. (2003) Estimation of
absolute microglial cell numbers in mouse fascia dentata us-
ing unbiased and efficient stereological cell counting princi-
ples. Glia, 44: 129–139.
Woods, A.G., Guthrie, K.M., Kurlawalla, M.A. and Gall,
C.M. (1998) Deafferentation-induced increases in hippocam-
pal insulin-like growth factor-1 messenger RNA expression
are severely attenuated in middle aged and aged rats. Ne-
uroscience, 83: 663–668.
Woods, A.G., Poulsen, F.R. and Gall, C.M. (1999) Dexa-
methasone specifically suppresses microglial trophic
responses to hippocampal deafferentation. Neuroscience,
91: 1277–1289.
Ying, G.X., Huang, C., Jiang, Z.H., Liu, X., Jing, N.H. and
Zhou, C.F. (2002) Up-regulation of cystatin C expression in
the murine hippocampus following perforant path transec-
tions. Neuroscience, 112: 289–298.
Yoshida, K. and Gage, F.H. (1991) Fibroblast growth factors
stimulate nerve growth factor synthesis and secretion by
astrocytes. Brain Res., 538: 118–126.
Zappone, C.A. and Sloviter, R.S. (2001) Commissurally pro-
jecting inhibitory interneurons of the rat hippocampal dent-
ate gyrus: a colocalization study of neuronal markers and the
retrograde tracer Fluoro-Gold. J. Comp. Neurol., 441:
324–344.
Zimmer, J. and Hjorth-Simonsen, A. (1975) Crossed pathways
from the entorhinal area to the fascia dentata. II. Provokable
in rats. J. Comp. Neurol., 161: 71–102.
Zimmer, J., Laurberg, S. and Sunde, N. (1986) Non-cholinergic
afferents determine the distribution of the cholinergic septo-
hippocampal projection: a study of the AChE staining pat-
tern in the rat fascia dentata and hippocampus after lesions,
x-irradiation, and intracerebral grafting. Exp. Brain Res., 64:
158–168.
Zipp, F., Nitsch, R., Soriano, E. and Frotscher, M. (1989)
Entorhinal fibers form synaptic contacts on parvalbumin-
immunoreactive neurons in the rat fascia dentata. Brain Res.,
495: 161–166.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 28

Adult neurogenesis in the intact and epileptic

dentate gyrus

Jack M. Parent

Department of Neurology, University of Michigan Medical Center, 109 Zina Pitcher Place, 5021 BSRB,
Ann Arbor, MI 48109-2200, USA

Abstract: Neurogenesis persists throughout life in the adult mammalian dentate gyrus. Adult-born dentate
granule cells integrate into existing hippocampal circuitry and may provide network plasticity necessary for
certain forms of hippocampus-dependent learning and memory. Neural stem cells and neurogenesis in the
adult dentate gyrus are regulated by a variety of environmental, physiological, and molecular factors. These
include aging, stress, exercise, neurovascular components of the stem cell niche, growth factors, neuro-
transmitters, and hormones. Seizure activity also influences dentate granule cell neurogenesis. Production
of adult-born neurons increases in rodent models of temporal lobe epilepsy, and both newborn and
pre-existing granule neurons contribute to aberrant axonal reorganization in the epileptic hippocampus.
Prolonged seizures also disrupt the migration of dentate granule cell progenitors and lead to hilar-ectopic
granule cells. The ectopic granule neurons appear to integrate abnormally and contribute to network
hyperexcitability. Similar findings of granule cell layer dispersion and ectopic granule neurons in
human TLE suggest that aberrant neurogenesis contributes to epileptogenesis or learning and memory
disturbances in this epilepsy syndrome.

Keywords: neurogenesis; stem cell; epileptogenesis; temporal lobe epilepsy; migration; dentate granule cell

Dentate granule cell neurogenesis persists
throughout life
New neurons are generated throughout life in the
mammalian dentate gyrus (Altman and Das, 1965;
Kaplan and Hinds, 1977; Cameron et al., 1993;
Kuhn et al., 1996). This process, termed neuro-
genesis, begins with neural stem or progenitor cells
that give rise to transit amplifying precursors to
expand the numbers of progeny. In the dentate
gyrus, these progeny form dentate granule cells
Corresponding author. Tel.: +1 734 763 3776;
Fax: +1 734 763 7686; E-mail: parent@umich.edu
(DGCs). Persistent DGC neurogenesis occurs in
all mammalian species examined to date, including
human and nonhuman primates (Eriksson et al.,
1998; Gould et al., 1998; Kornack and Rakic,
1999). Although the majority of DGCs in the rat
are produced near the end of the first postnatal
week, new DGCs continue to be generated at a
lower rate throughout adulthood and into senes-
cence (Schlessinger et al., 1975; Kuhn et al., 1996)
(Fig. 1). The human dentate gyrus, for example,
generates neurons well into the seventh decade of
life (Eriksson et al., 1998; Kempermann et al.,
2004). Estimates suggest that approximately 9000
DGCs are added daily to the dentate gyrus of
young adult rats, and 6% of the granule cell

DOI: 10.1016/S0079-6123(07)63028-3 529

530

Fig. 1. Neonatal and adult dentate granule cell neurogenesis. (A, B) Bromodeoxyuridine (BrdU) immunostaining of coronal sections
through the dentate gyrus of 10 day (A) and 10 week (B) old rats shows proliferating granule cell progenitors concentrated at the inner
part of the granule cell layer at the hilar (h) border. Dentate granule cell neurogenesis peaks at the end of the 1st postnatal week (A) but
persists into adulthood (B). BrdU was injected 6 days earlier for both. (C) Retroviral nuclear localization signal-b-galactosidase
reporter labeling of proliferating cell clusters in the dentate subgranular zone (arrows) of an adult rat. Inset shows the right cluster at
higher magnification. Retrovirus was injected 3 days earlier. (D) Doublecortin immunolabeling of immature neurons (arrows) in the
adult rat, with cell bodies close to the SGZ but processes throughout the granule cell and molecular layer gcl, granule cell layers.

layer is thought to consist of recently born neurons
(Cameron and McKay, 2001).
Dentate gyrus neural stem cells reside in the
subgranular zone (SGZ) at the border of the hilus
and DGC layer. The stem cells are a subtype of
radial glia-like astrocyte that expresses glial fibril-
lary acidic protein (GFAP) and nestin (Seri et al.,
2001; Filippov et al., 2003). They proliferate and
give rise to clusters of transit amplifying precur-
sors that divide further to generate neuroblasts
(Fig. 1). The SGZ precursors and neuroblasts are
characterized by a sequence of marker expression
as they mature (reviewed in Kempermann et al.,
2004). The progenitors express doublecortin
(DCX) and polysialylated neural cell adhesion
molecule (PSA-NCAM) during cell division and as
early postmitotic neurons (Fig. 1D). With further
differentiation, postmitotic neurons continue to
expresss DCX and PSA-NCAM, but also begin to
express calretinin and Prox1; subsequently, they
differentiate further into mature granule neurons
that are immunoreactive for Prox1 and calbindin.
Integration and function of adult-born DGCs
Increasing attention has focused on the potential
for adult-generated DGCs to integrate into exist-
ing neural circuitry. Combined retrograde tracer
and mitotic labeling studies using tritiated thymi-
dine or its analogue, bromodeoxyuridine (BrdU),
in adult rodent indicate that mossy fibers of
newborn DGCs project to appropriate targets in
hippocampal area CA3 (Stanfield and Trice, 1988;
Hastings and Gould, 1999; Markakis and Gage,
1999). Acute hippocampal slice recordings from
rodents in which adult-generated DGCs were labe-
led with retroviral reporters or by transgenic
manipulations show that the cells integrate into
hippocampal circuitry and acquire some electro-
physiological characteristics of mature DGCs
(Song et al., 2002b; van Praag et al., 2002; Wadi-
che et al., 2005; Ge et al., 2006). The newborn
neurons appear to receive gamma-aminobutyric
acid (GABA)ergic synaptic inputs at 1 week after
birth and glutamatergic inputs within 2 weeks.

Adult-generated DGCs initially show depolarizing
responses to GABA (Tozuka et al., 2005; Wadiche
et al., 2005; Ge et al., 2006), similar to immature
neurons during early development. Unlike those in
the neonate, however, adult-generated DGCs take
much longer (perhaps 3–6 months) to obtain a full
dendritic arborization pattern (van Praag et al.,
2002). Compared to neighboring mature DGCs,
immature DGCs in the adult have a higher input
resistance and a subthreshold calcium conductance
that may enhance their excitability. In vitro elect-
rophysiological studies also show a decreased
threshold for long-term potentiation (LTP) or
long-term depression, as well as higher-amplitude
LTP, in newborn DGCs compared to their mature
counterparts (Wang et al., 2000; Schmidt-Hieber
et al., 2004).
The precise function of DGCs generated in
adulthood is unknown. Most of the experimental
evidence to date, however, supports a role for
DGC neurogenesis in learning and memory func-
tion (see Doetsch and Hen, 2005 for review). Stim-
ulation of adult DGC neurogenesis with
behavioral interventions such as exercise or envi-
ronmental enrichment is associated with better
performance on certain hippocampal learning
tasks (Bruel-Jungerman et al., 2005; van Praag
et al., 2005; Meshi et al., 2006), although the link
between increased neurogeneis and enrichment-
enhanced performance has recently been called
into question (Meshi et al., 2006). The degree of
adult neurogenesis in some inbred mouse strains
also correlates positively with learning ability
(Kempermann and Gage, 2002). More direct
evidence that neurogenesis supports hippocam-
pal-dependent learning and memory derives
from experiments in which the depletion of
adult-generated neurons disrupts certain forms of
learning (Shors et al., 2001, 2002; Winocur et al.,
2006). These data are supported by the findings
described above that suggest LTP is more easily
elicited and of higher amplitude in immature,
adult-born DGCs (Wang et al., 2000; Schmidt-
Hieber et al., 2004). Recent animal studies also
provide evidence that DGC neurogenesis in the
adult is necessary for the positive effects of anti-
depressant medication (Santarelli et al., 2003).
Despite these accumulating data, knowledge of the
531
entire spectrum of biological functions involving
adult DGC neurogenesis awaits more specific
means of manipulating DGC progenitors in the
adult.
Regulation of adult DGC neurogenesis
A host of physiological states and molecular
factors appear to regulate adult hippocampal
neurogenesis (see Table 1). Neuronal production
declines markedly with age (Seki and Arai, 1993;
Kuhn et al., 1996), probably due to reduced DGC
precursor proliferation and altered differentiation
or delayed maturation of their progeny (Kemper-
mann et al., 1998; Rao et al., 2005). The mecha-
nisms likely reflect active suppression of
neurogenesis with age because removal of cortico-
steroids by adrenalectomy or the infusion of spe-
cific growth factors significantly restores DGC
production in aged rats (Cameron and McKay,
1999; Lichtenwalner et al., 2001; Jin et al., 2003).
In addition to age, acute or chronic stress de-
creases DGC neurogenesis in young adult rodents
and primates (Mirescu and Gould, 2006).
Adult neurogenesis is stimulated by environ-
mental enrichment, which increases the survival
and neuronal differentiation of DGCs generated in
early adulthood through senescence (Kemper-
mann et al., 1997, 1998, 2002; Nilsson et al.,
1999). Exercise also enhances adult neurogenesis,
but this effect relates to increased precursor pro-
liferation rather than survival (van Praag et al.,
1999). Growth factors may play a significant role
in mediating these effects, as blocking brain up-
take of insulin-like growth factor-1 (IGF-1) or
vascular endothelial growth factor (VEGF) pre-
vents the increase in DGC production induced by
enrichment or exercise (Trejo et al., 2001; Fabel
et al., 2003; Cao et al., 2004). Administration of
these growth factors also increases adult neuro-
genesis in the rodent dentate gyrus (Aberg et al.,
2000; Jin et al., 2002; Cao et al., 2004; Schanzer
et al., 2004), but it is unclear whether this effect
results from increased progenitor cell proliferation
(Jin et al., 2002) or survival (Schanzer et al., 2004).
Other mediators of adult hippocampal neuro-
genesis include hormones and neurotransmitters.
532

Table 1. Modulation of adult hippocampal neurogenesis

Variable Change in Mechanism Change in Proliferation Neuronal Survival

neurogenesis neurogenesis differentiation

Physiologic state
Aging Decreased Proliferation/ Decreased m m
neuronal
differentiation
Stress Decreased Proliferation Decreased m
Exercise Increased Proliferation/ Increased m m? m?
?survival/
?differentiation
Environmental Increased Survival Increased m
enrichment
Growth factors
bFGF Increased (early Proliferation Increased m
postnatal) (postnatal)
BDNF Increased Differentiation/ Increased m m
survival
VEGF Increased Proliferation/ Increased m m
survival
IGF-1 Increased Proliferation/ Increased k m
survival/
?differentiation
Neurotransmitters/neuromodulators
Glutamate Decreased Proliferation Decreased m
GABA Increased Differentiation/ Increased m? m
?proliferation
Serotonin Increased Proliferation Increased m
Nitrous oxide Decreased Decreased Decreased k m
proliferation
(promotes
differentiation)
Transcription factors
Wnt Increased Proliferation/ Increased m m?
?differentiation
Sonic Increased Proliferation Increased m
hedgehog
Sox2 Increased ?Proliferation/ m? m?
?differentiation

Adrenalectomy and corticosteroid replacement receptor antagonists or lesion of DGC afferents

studies have shown that circulating stress hor-
mones suppress the rate of DGC production in
adulthood (reviewed in Mirescu and Gould, 2006).
In contrast, cell proliferation in the rat dentate
gyrus increases during proestrus, and the effect is
abolished by ovariectomy (Tanapat et al., 1999).
Glutamatergic inputs also appear to influence
dentate gyrus neurogenesis in the adult rodent.
Activation of N-methyl-D-aspartate (NMDA) re-
ceptors attenuates cell proliferation, and suppress-
ing glutamatergic neurotransmission with NMDA
exerts the opposite effect (Cameron et al., 1995).
Serotonin (5-hydroxytryptamine) also influences
adult neurogenesis. Chronic treatment with anti-
depressant drugs that act as serotonin reuptake
inhibitors increases cell proliferation in the rodent
dentate gyrus (Malberg et al., 2000), and blockade
of serotonin synthesis or administration of a
serotonin neurotoxin decreases new cell produc-
tion (Brezun and Daszuta, 1999). In terms of other
neuromodulators, the number of newly generated
DGCs is negatively regulated by nitric oxide,

which reduces DGC precursor proliferation but
promotes neuronal differentiation (Packer et al.,
2003). Neuronal addition thus continues in adult-
hood under the influence of a balance between
multiple positive and negative regulators of ne-
urogenic activity.
The local microenvironment in regions of on-
going neurogenesis in the adult mammalian brain
provides what is referred to as a ‘‘stem cell niche.’’
This niche governs neuronal production and in-
cludes progenitors, astrocytes, microvasculature,
and microglia (Wurmser et al., 2004). Astrocytes
appear to influence all phases of adult neurogen-
esis, including neural precursor cell proliferation,
migration, and differentiation (Lim and Alvarez-
Buylla, 1999; Song et al., 2002a). Studies of ne-
urogenesis in the adult rodent dentate gyrus and
subventricular zone (SVZ) implicate radial glia-
like astrocytes as the neural stem cells from which
neuroblasts derive (Doetsch et al., 1999; Seri et al.,
2001; Garcia et al., 2004; Merkle et al., 2004).
A vascular presence in the stem cell niche likely
modulates neurogenesis as precursors in the dent-
ate gyrus and SVZ reside in close proximity to the
microvasculature (Palmer et al., 2000; Shen et al.,
2004), and their proliferative activity appears to be
tightly linked with angiogenesis. In the adult, sup-
pression of hippocampal neurogenesis by irradia-
tion may reflect disruption of local angiogenesis
(Monje et al., 2002), and infusion of VEGF into
the adult brain increases the production both of
endothelial cells and DGCs (Jin et al., 2002). In
contrast to astrocytes and endothelial cells, recent
studies suggest that reactive microglia disrupt
neurogenesis in the adult dentate gyrus. Micro-
glial reaction and inflammation induced by brain
irradiation, seizures, or lipopolysaccharide treat-
ment decreases the survival of adult-generated
DGCs, and suppressing microglial activation with
minocycline or indomethacin treatment partially
restores neurogenesis (Ekdahl et al., 2003; Monje
et al., 2003).
Seizure-induced hippocampal neurogenesis
The persistence of neural stem cells in the adult
mammalian forebrain raises the possibility for
533
endogenous repair after brain injury or neurode-
generation. Various brain insults in the adult,
including mechanical lesions, hypoglycemia,
fluid percussion injury, and stroke, stimulate pro-
genitor cell proliferation in the hippocampal
dentate gyrus (Gould and Tanapat, 1997; Liu
et al., 1998; Jin et al., 2001; Suh et al., 2005).
As in the intact brain, many newly generated cells
die, but the vast majority of those that survive
differentiate into neurons, and typically less
than 20% of adult-born cells become glia or
remain undifferentiated. Similar findings are seen
after seizure-induced injury in temporal lobe
epilepsy (TLE) models, which were the first brain
injury models used to demonstrate altered hippo-
campal neurogenesis (Parent et al., 1997). Studies
of adult rodent models of limbic epileptogenesis
or acute seizures indicate that prolonged
seizures potently stimulate DGC neurogenesis
(Bengzon et al., 1997; Gray and Sundstrom,
1998; Parent et al., 1997, 1998; Scott et al.,
1998). In the adult rodent kainate and pilocar-
pine models of temporal lobe epilepsy,
chemoconvulsant-induced status epilepticus (SE)
increases dentate gyrus cell proliferation approx-
imately 5–10-fold after a latent period of at least
several days (Parent et al., 1997; Gray and
Sundstrom, 1998) (Fig. 2A and B). Approxi-
mately 80–90% of the newly generated cells
differentiate into DGCs.
DGC neurogenesis is also enhanced by kindling-
induced epileptogenesis. Electrical kindling of the
amygdala (Parent et al., 1998; Scott et al., 1998),
hippocampus (Bengzon et al., 1997) or perforant
path (Nakagawa et al., 2000) acutely increases
dentate gyrus cell proliferation and neurogenesis.
Similar neurogenic effects occur after acute
seizures induced by intermittent perforant path
or hippocampal stimulation in adult rats (Bengzon
et al., 1997; Parent et al., 1997). Even single,
discrete seizure-like afterdischarges induced by
hippocampal stimulation lead to increased num-
bers of newly differentiated DGCs (Bengzon et al.,
1997). Although more severe seizures enhance
neurogenesis to a greater extent, the survival of
the newborn DGCs may decrease with increased
seizure severity (Mohapel et al., 2004). This effect
may relate to the degree of microglial activation as
534

Fig. 2. Increased cell proliferation and altered dentate gyrus neuroblast migration after SE. (A, B) BrdU incorporation (arrows) in
adult rat dentate gyrus 7 days after saline (A) or pilocarpine (B) treatment followed by a 2 day post-BrdU survival. Subgranular zone
BrdU labeling increases markedly after SE (B) compared to the control (A). (C, D) Doublecortin (DCX) immunoreactivity in the
dentate gyrus of a control (Con; C) and an adult rat 14 days after pilocarpine-induced SE (14d; D). Note the increased hilar DCX
expression and chains of DCX-positive cells extending into the hilus (arrows in D) after SE. E, PSA-NCAM+ neuroblast chains (red,
arrows) alongside GFAP+ hilar astrocytes (green). Dashed lines in C–E denote the granule cell layer (gcl) — hilar (h) border.
(See Color Plate 28.2 in color plate section.)

it is reversed in part by minocycline treatment to project axons aberrantly in the epileptogenic

(Ekdahl et al., 2003).
Seizure-induced hippocampal neurogenesis:
functional significance
In the pilocarpine model of TLE, aberrant mossy
fiber sprouting occurs in parallel with the gener-
ation of increased numbers of DGCs (Parent et al.,
1997). Evidence suggests that developing axons
from newborn DGCs contribute to status epilep-
ticus (SE)-induced aberrant mossy fiber reorgan-
ization in both hippocampal area CA3 and the
dentate inner molecular layer (Parent et al., 1997).
Eliminating DGC progenitors by brain irradiation
before and after pilocarpine treatment, however,
does not suppress mossy fiber sprouting (Parent
et al., 1999). Thus, seizure-induced injury appears
to induce both pre-existing and adult-born DGCs
hippocampal formation.
An increasingly recognized abnormality of
DGC neurogenesis in experimental TLE is the
ectopic location of adult-born granule neurons.
The DGC layer in human TLE is often abnormal
due to dispersion and the presence of ectopic
granule-like neurons in the hilus and inner molec-
ular layer (Houser, 1990; Parent et al., 2006a).
An early report of seizure-induced neurogenesis
described newly differentiating neurons with gran-
ule cell morphology in the dentate hilus and inner
molecular layer after pilocarpine-induced SE in
adult rats (Parent et al., 1997). Hilar-ectopic
DGCs are found in several different experimental
TLE models and may persist for many months
(Parent et al., 1997, 2006a; Scharfman et al., 2000;
Dashtipour et al., 2001). The cells resemble the
granule-like neurons identified in surgical speci-
mens from humans with temporal lobe epilepsy,
both in terms of their morphology and marker
 535

Fig. 3. Schematic showing the effects of prolonged seizures on caudal subventricular zone (SVZ) and dentate gyrus progenitors. In the
intact adult rat (top), progenitors located in the infracallosal SVZ (purple) give rise to white matter oligodendrocytes (orange), while
those in the dentate gyrus (green) give rise to DGCs in the granule cell layer (gcl). After seizure-induced hilar and pyramidal cell layer
(pcl) injury (bottom; jagged arrows), progenitors migrate aberrantly from the caudal SVZ to form glia (a; purple) in the hippocampus
proper or from the dentate subgranular zone to form hilar-ectopic DGCs (b). (See Color Plate 28.3 in color plate section.)

expression (Scharfman et al., 2000; Dashtipour
et al., 2001; Parent et al., 2006a). Studies of ne-
urogenesis in the pilocarpine TLE model indicate
that the hilar-ectopic DGCs migrate aberrantly
from the dentate subgranular proliferative zone to
the hilus (Parent et al., 2006a) (Fig. 3).
Using intracellular recordings in hippocampal
slices from epileptic adult rats, Scharfman and
colleagues (Scharfman et al., 2000) have shown
that the hilar-ectopic granule-like cells are hyper-
excitable and fire in abnormal bursts synchro-
nously with CA3 pyramidal cells. Expression of
the activation-induced immediate early gene c-fos
by hilar-ectopic DGCs during spontaneous limbic
seizures also suggests that they participate in ep-
ileptic circuitry (Scharfman et al., 2002). In addi-
tion, many putatively newborn DGCs located in
the hilus or hilar aspect of the DGC layer after
seizures exhibit a much higher percentage of per-
sistent basal dendrites than is normally found
(Scharfman et al., 2000; Dashtipour et al., 2001).
Work by Dashtipour and colleagues suggests that
the basal dendrites of hilar-ectopic DGCs receive
increased excitatory input (Dashtipour et al.,
2001), suggesting that this structural plasticity
may be a mechanism for seizure generation or
propagation. Further evidence supporting the
epileptogenic nature of hilar-ectopic DGCs comes
from the work of Jung and colleagues (Jung et al.,
2004). Their group used the antimitotic agent
AraC to inhibit DGC neurogenesis after pilocar-
pine treatment, and found that these rats devel-
oped fewer and shorter spontaneous recurrent
seizures than controls. Taken together, these data
suggest that hilar-ectopic DGCs integrate abnor-
mally, are hyperexcitable, and thus may contribute
536
Brain insult in childhood (e.g.,
prolonged febrile seizure)
Altered developmental cues
Aberrant migration
of DGC progenitors
Fig. 4. Aberrant neurogenesis hypothesis of epileptogenesis and memory disturbance in mesial temporal lobe epilepsy. Note the
potential maladaptive positive feedback loop of recurrent seizures on aberrant neurogenesis.
to seizure generation or propagation. The presence
of hilar-ectopic DGCs in hippocampi surgically
resected to treat intractable epilepsy (Houser,
1990; Parent et al., 2006a) raises the possibility
that abnormalities in dentate gyrus persistent ne-
urogenesis contribute to epileptogenesis or cogni-
tive dysfunction in human TLE (Fig. 4).
Seizure-induced neurogenesis: mechanisms
The mechanisms by which seizure activity stimu-
lates neurogenesis or gliogenesis are unknown.
Experiments in which proliferating cells were labe-
led with BrdU prior to seizure induction have
shown that epileptic activity stimulates dentate
gyrus and caudal SVZ precursors that are actively
proliferating prior to injury (Parent et al., 1999,
2006b). Huttman and colleagues used a reporter
mouse in which green fluorescent protein selec-
tively labeled nestin-expressing stem-like cells to
show that kainate-induced seizures specifically
increased the proliferation of the radial glia-like
progenitors in dentate gyrus (Huttmann et al.,
2003). Seizures may act to increase neurogenesis
indirectly through activation of astrocytes, as as-
trocytes stimulate hippocampal neurogenesis via
wnt signaling and perhaps other mechanisms
(Song et al., 2002a; Lie et al., 2005). Alternatively,
the SE-induced death of some mature DGCs may
increase cell turnover in the dentate gyrus via other
mechanisms. Cell death is associated with subse-
quent cell birth in a number of postnatal
neurogenic systems, including the dentate gyrus
RECURRENT IMPAIRED
SEIZURES LEARNING/
MEMORY
Ectopically integrated DGCs
and olfactory bulb of adult rodents (Gould and
McEwen, 1993; Biebl et al., 2000). Seizures also
increase the expression of molecules with the
potential to increase neurogenesis or gliogenesis
such as growth factors (Humpel et al., 1993) and
neurotrophins (Isackson et al., 1991). Specific ne-
urotransmitters or neuromodulatory systems
that normally influence DGC neurogenesis (see
Table 1) also may be altered by seizure activity.
In terms of potential mechanisms underlying
hilar- or molecular-layer ectopic DGC formation,
molecular factors that influence neuronal migra-
tion during development are prime candidates.
The expression of one of these developmental fac-
tors, reelin, persists in the hippocampal dentate
gyrus of adult humans and rodents, and has been
implicated in DGC layer dispersion in human
TLE (Haas et al., 2002). The expression of reelin
decreases markedly in the dentate gyrus after
pilocarpine-induced SE (JMP, personal communi-
cation), suggesting that loss of this migration
guidance factor may be responsible for the aber-
rant migration of DGC progenitors during epile-
ptogenesis. Another potential mechanism is loss of
GABA caused by SE-induced depletion of dentate
interneurons, as this neurotransmitter influences
DGC differentiation (see Table 1) and GABA
decreases neuroblast migration in the other adult
neurogenic region, the SVZ-olfactory bulb path-
way (Liu et al., 2005). In an interesting study,
Scharfman and colleagues showed that hippocam-
pal brain-derived neurotrophic factor (BDNF)
infusion in adult rat increased DGC neurogenesis
and led to the appearance of ectopic DGCs
 537

(Scharfman et al., 2005). This result raises the pos- Altman, J. and Das, G.D. (1965) Autoradiographic and histo-
sibility that BDNF is involved in seizure-induced, logical evidence of postnatal hippocampal neurogenesis in
aberrant DGC neurogenesis as well. rats. J. Comp. Neurol., 124: 319–335.
Bengzon, J., Kokaia, Z., Elmer, E., Nanobashvili, A., Kokaia,
The multiple steps of DGC neurogenesis are in- M. and Lindvall, O. (1997) Apoptosis and proliferation of
fluenced by a delicate balance of factors affecting dentate gyrus neurons after single and intermittent limbic
DGC progenitors in the adult. Improved under- seizures. Proc. Natl. Acad. Sci. U.S.A., 94: 10432–10437.
standing of the biological functions of adult-born Biebl, M., Cooper, C.M., Winkler, J. and Kuhn, H.G. (2000)
DGCs, their regulation, and how they are influ- Analysis of neurogenesis and programmed cell death reveals
a self-renewing capacity in the adult rat brain. Neurosci.
enced by brain pathology may provide important Lett., 291: 17–20.
insights into the pathophysiology of TLE, its as- Brezun, J.M. and Daszuta, A. (1999) Depletion in serotonin
sociated cognitive impairments and other brain decreases neurogenesis in the dentate gyrus and the subven-
disorders. Manipulation of adult neurogenesis for tricular zone of adult rats. Neuroscience, 89: 999–1002.
therapeutic purposes therefore is likely to involve Bruel-Jungerman, E., Laroche, S. and Rampon, C. (2005) New
neurons in the dentate gyrus are involved in the expression
suppression of aberrant integration in certain
of enhanced long-term memory following environmental
instances rather than simply stimulation of neuro- enrichment. Eur. J. Neurosci., 21: 513–521.
genesis for neuronal replacement. Cameron, H.A., McEwen, B.S. and Gould, E. (1995) Regula-
tion of adult neurogenesis by excitatory input and NMDA
receptor activation in the dentate gyrus. J. Neurosci., 15:
4687–4692.
Cameron, H.A. and McKay, R.D. (1999) Restoring production
Abbreviations
of hippocampal neurons in old age. Nat. Neurosci., 2:
894–897.
BDNF brain-derived neurotrophic Cameron, H.A. and McKay, R.D. (2001) Adult neurogenesis
factor produces a large pool of new granule cells in the dentate
BrdU bromodeoxyuridine gyrus. J. Comp. Neurol., 435: 406–417.
DCX doublecortin Cameron, H.A., Woolley, C.S., McEwen, B.S. and Gould, E.
(1993) Differentiation of newly born neurons and glia in the
DGC dentate granule cell dentate gyrus of the adult rat. Neuroscience, 56: 337–344.
GABA gamma-aminobutyric acid Cao, L., Jiao, X., Zuzga, D.S., Liu, Y., Fong, D.M., Young, D.
GFAP glial fibrillary acidic protein and During, M.J. (2004) VEGF links hippocampal activity
IGF-1 insulin-like growth factor-1 with neurogenesis, learning and memory. Nat. Genet., 36:
LTP long-term potentiation 827–835.
Dashtipour, K., Tran, P.H., Okazaki, M.M., Nadler, J.V. and
mTLE medial temporal lobe epi- Ribak, C.E. (2001) Ultrastructural features and synaptic
lepsy connections of hilar ectopic granule cells in the rat dentate
NMDA N-methyl-D-aspartate gyrus are different from those of granule cells in the granule
PSA-NCAM polysialylated neural cell ad- cell layer. Brain Res., 890: 261–271.
hesion molecule Doetsch, F., Caille, I., Lim, D.A., Garcia-Verdugo, J.M. and
Alvarez-Buylla, A. (1999) Subventricular zone astrocytes are
SE status epilepticus
neural stem cells in the adult mammalian brain. Cell, 97:
SGZ subgranular zone 703–716.
SVZ subventricular zone Doetsch, F. and Hen, R. (2005) Young and excitable: the func-
TLE temporal lobe epilepsy tion of new neurons in the adult mammalian brain. Curr.
VEGF vascular endothelial growth Opin. Neurobiol., 15: 121–128.
Ekdahl, C.T., Claasen, J.H., Bonde, S., Kokaia, Z. and
factor
Lindvall, O. (2003) Inflammation is detrimental for neuro-
genesis in adult brain. Proc. Natl. Acad. Sci. U.S.A., 100:
13632–13637.
References Eriksson, P.S., Perfilieva, E., Bjork-Eriksson, T., Alborn, A.M.,
Nordborg, C., Peterson, D.A. and Gage, F.H. (1998) Ne-
Aberg, M.A., Aberg, N.D., Hedbacker, H., Oscarsson, J. and urogenesis in the adult human hippocampus. Nat. Med., 4:
Eriksson, P.S. (2000) Peripheral infusion of IGF-I selectively 1313–1317.
induces neurogenesis in the adult rat hippocampus. J. Ne- Fabel, K., Fabel, K., Tam, B., Kaufer, D., Baiker, A., Simm-
urosci., 20: 2896–2903. ons, N., Kuo, C.J. and Palmer, T.D. (2003) VEGF is
538
necessary for exercise-induced adult hippocampal neurogen-
esis. Eur. J. Neurosci., 18: 2803–2812.
Filippov, V., Kronenberg, G., Pivneva, T., Reuter, K., Steiner,
B., Wang, L.P., Yamaguchi, M., Kettenmann, H. and Kem-
permann, G. (2003) Subpopulation of nestin-expressing pro-
genitor cells in the adult murine hippocampus shows
electrophysiological and morphological characteristics of as-
trocytes. Mol. Cell Neurosci., 23: 373–382.
Garcia, A.D., Doan, N.B., Imura, T., Bush, T.G. and
Sofroniew, M.V. (2004) GFAP-expressing progenitors are
the principal source of constitutive neurogenesis in adult
mouse forebrain. Nat. Neurosci., 7: 1233–1241.
Ge, S., Goh, E.L., Sailor, K.A., Kitabatake, Y., Ming, G.L.
and Song, H. (2006) GABA regulates synaptic integration of
newly generated neurons in the adult brain. Nature, 439:
589–593.
Gould, E. and McEwen, B.S. (1993) Neuronal birth and death.
Curr. Opin. Neurobiol., 3: 676–682.
Gould, E. and Tanapat, P. (1997) Lesion-induced proliferation
of neuronal progenitors in the dentate gyrus of the adult rat.
Neuroscience, 80: 427–436.
Gould, E., Tanapat, P., McEwen, B.S., Flugge, G. and
Fuchs, E. (1998) Proliferation of granule cell precursors in
the dentate gyrus of adult monkeys is diminished by stress.
Proc. Natl. Acad. Sci. U.S.A., 95: 3168–3171.
Gray, W.P. and Sundstrom, L.E. (1998) Kainic acid increases
the proliferation of granule cell progenitors in the dentate
gyrus of the adult rat. Brain Res., 790: 52–59.
Haas, C.A., Dudeck, O., Kirsch, M., Huszka, C., Kann, G.,
Pollak, S., Zentner, J. and Frotscher, M. (2002) Role for
reelin in the development of granule cell dispersion in tem-
poral lobe epilepsy. J. Neurosci., 22: 5797–5802.
Hastings, N.B. and Gould, E. (1999) Rapid extension of axons
into the CA3 region by adult-generated granule cells. J.
Comp. Neurol., 413: 146–154.
Houser, C.R. (1990) Granule cell dispersion in the dentate
gyrus of humans with temporal lobe epilepsy. Brain Res.,
535: 195–204.
Humpel, C., Lippoldt, A., Chadi, G., Ganten, D., Olson, L. and
Fuxe, K. (1993) Fast and widespread increase of basic fib-
roblast growth factor messenger RNA and protein in the
forebrain after kainate-induced seizures. Neuroscience, 57:
913–922.
Huttmann, K., Sadgrove, M., Wallraff, A., Hinterkeuser, S.,
Kirchhoff, F., Steinhauser, C. and Gray, W.P. (2003) Seizures
preferentially stimulate proliferation of radial glia-like astro-
cytes in the adult dentate gyrus: functional and immunocyto-
chemical analysis. Eur. J. Neurosci., 18: 2769–2778.
Isackson, P.J., Huntsman, M.M., Murray, K.D. and Gall,
C.M. (1991) BDNF mRNA expression is increased in adult
rat forebrain after limbic seizures: temporal patterns of in-
duction distinct from NGF. Neuron, 6: 937–948.
Jin, K., Minami, M., Lan, J.Q., Mao, X.O., Batteur, S.,
Simon, R.P. and Greenberg, D.A. (2001) Neurogenesis
in dentate subgranular zone and rostral subventricular zone
after focal cerebral ischemia in the rat. Proc. Natl. Acad. Sci.
U.S.A., 98: 4710–4715.
Jin, K., Sun, Y., Xie, L., Batteur, S., Mao, X.O., Smelick, C.,
Logvinova, A. and Greenberg, D.A. (2003) Neurogenesis
and aging: FGF-2 and HB-EGF restore neurogenesis in
hippocampus and subventricular zone of aged mice. Aging
Cell, 2: 175–183.
Jin, K., Zhu, Y., Sun, Y., Mao, X.O., Xie, L. and Greenberg,
D.A. (2002) Vascular endothelial growth factor (VEGF)
stimulates neurogenesis in vitro and in vivo. Proc. Natl.
Acad. Sci. U.S.A., 99: 11946–11950.
Jung, K.H., Chu, K., Kim, M., Jeong, S.W., Song, Y.M.,
Lee, S.T., Kim, J.Y., Lee, S.K. and Roh, J.K. (2004)
Continuous cytosine-b-D-arabinofuranoside infusion re-
duces ectopic granule cells in adult rat hippocampus with
attenuation of spontaneous recurrent seizures following
pilocarpine-induced status epilepticus. Eur. J. Neurosci.,
19: 3219–3226.
Kaplan, M.S. and Hinds, J.W. (1977) Neurogenesis in the adult
rat: electron microscopic analysis of light radioautographs.
Science, 197: 1092–1094.
Kempermann, G. and Gage, F.H. (2002) Genetic determinants
of adult hippocampal neurogenesis correlate with acquisition,
but not probe trial performance, in the water maze task.
Eur. J. Neurosci., 16: 129–136.
Kempermann, G., Gast, D. and Gage, F.H. (2002) Neuroplas-
ticity in old age: sustained fivefold induction of hippocampal
neurogenesis by long-term environmental enrichment. Ann.
Neurol., 52: 135–143.
Kempermann, G., Jessberger, S., Steiner, B. and Kronenberg,
G. (2004) Milestones of neuronal development in the adult
hippocampus. Trends Neurosci., 27: 447–452.
Kempermann, G., Kuhn, H.G. and Gage, F.H. (1997) More
hippocampal neurons in adult mice living in an enriched en-
vironment. Nature, 386: 493–495.
Kempermann, G., Kuhn, H.G. and Gage, F.H. (1998) Expe-
rience-induced neurogenesis in the senescent dentate gyrus. J.
Neurosci., 18: 3206–3212.
Kornack, D.R. and Rakic, P. (1999) Continuation of neuro-
genesis in the hippocampus of the adult macaque monkey.
Proc. Natl. Acad. Sci. U.S.A., 96: 5768–5773.
Kuhn, H.G., Dickinson-Anson, H. and Gage, F.H. (1996)
Neurogenesis in the dentate gyrus of the adult rat: age-related
decrease of neuronal progenitor proliferation. J. Neurosci.,
16: 2027–2033.
Lichtenwalner, R.J., Forbes, M.E., Bennett, S.A., Lynch, C.D.,
Sonntag, W.E. and Riddle, D.R. (2001) Intracerebroven-
tricular infusion of insulin-like growth factor-I ameliorates
the age-related decline in hippocampal neurogenesis. Neuro-
science, 107: 603–613.
Lie, D.C., Colamarino, S.A., Song, H.J., Desire, L., Mira, H.,
Consiglio, A., Lein, E.S., Jessberger, S., Lansford, H.,
Dearie, A.R. and Gage, F.H. (2005) Wnt signalling regu-
lates adult hippocampal neurogenesis. Nature, 437:
1370–1375.
Lim, D.A. and Alvarez-Buylla, A. (1999) Interaction between
astrocytes and adult subventricular zone precursors stimu-
lates neurogenesis. Proc. Natl. Acad. Sci. U.S.A., 96:
7526–7531.

Liu, J., Solway, K., Messing, R.O. and Sharp, F.R. (1998) In-
creased neurogenesis in the dentate gyrus after transient glo-
bal ischemia in gerbils. J. Neurosci., 18: 7768–7778.
Liu, X., Wang, Q., Haydar, T.F. and Bordey, A. (2005) Nonsy-
naptic GABA signaling in postnatal subventricular zone
controls proliferation of GFAP-expressing progenitors. Nat.
Neurosci., 8: 1179–1187.
Malberg, J.E., Eisch, A.J., Nestler, E.J. and Duman, R.S.
(2000) Chronic antidepressant treatment increases neurogen-
esis in adult rat hippocampus. J. Neurosci., 20: 9104–9110.
Markakis, E.A. and Gage, F.H. (1999) Adult-generated neu-
rons in the dentate gyrus send axonal projections to field CA3
and are surrounded by synaptic vesicles. J. Comp. Neurol.,
406: 449–460.
Merkle, F.T., Tramontin, A.D., Garcia-Verdugo, J.M. and
Alvarez-Buylla, A. (2004) Radial glia give rise to adult neural
stem cells in the subventricular zone. Proc. Natl. Acad. Sci.
U.S.A., 101: 17528–17532.
Meshi, D., Drew, M.R., Saxe, M., Ansorge, M.S., David, D.,
Santarelli, L., Malapani, C., Moore, H. and Hen, R. (2006)
Hippocampal neurogenesis is not required for behavioral
effects of environmental enrichment. Nat. Neurosci., 9:
729–731.
Mirescu, C. and Gould, E. (2006) Stress and adult neurogenesis.
Hippocampus, 16: 233–238.
Mohapel, P., Ekdahl, C.T. and Lindvall, O. (2004) Status ep-
ilepticus severity influences the long-term outcome of neuro-
genesis in the adult dentate gyrus. Neurobiol. Dis., 15:
196–205.
Monje, M.L., Mizumatsu, S., Fike, J.R. and Palmer, T.D.
(2002) Irradiation induces neural precursor-cell dysfunction.
Nat. Med., 8: 955–962.
Monje, M.L., Toda, H. and Palmer, T.D. (2003) Inflammatory
blockade restores adult hippocampal neurogenesis. Science,
302: 1760–1765.
Nakagawa, E., Aimi, Y., Yasuhara, O., Tooyama, I., Shimada,
M., McGeer, P.L. and Kimura, H. (2000) Enhancement of
progenitor cell division in the dentate gyrus triggered by in-
itial limbic seizures in rat models of epilepsy. Epilepsia, 41:
10–18.
Nilsson, M., Perfilieva, E., Johansson, U., Orwar, O. and Er-
iksson, P.S. (1999) Enriched environment increases neuro-
genesis in the adult rat dentate gyrus and improves spatial
memory. J. Neurobiol., 39: 569–578.
Packer, M.A., Stasiv, Y., Benraiss, A., Chmielnicki, E., Grinb-
erg, A., Westphal, H., Goldman, S.A. and Enikolopov, G.
(2003) Nitric oxide negatively regulates mammalian adult
neurogenesis. Proc. Natl. Acad. Sci. U.S.A., 100: 9566–9571.
Palmer, T.D., Willhoite, A.R. and Gage, F.H. (2000) Vascular
niche for adult hippocampal neurogenesis. J. Comp. Neurol.,
425: 479–494.
Parent, J.M., von dem Bussche, N. and Lowenstein, D.H.
(2006b) Prolonged seizures recruit caudal subventricular zone
glial progenitors into the injured hippocampus. Hippocam-
pus, 16: 321–328.
Parent, J.M., Elliott, R.C., Pleasure, S.J., Barbaro, N.M. and
Lowenstein, D.H. (2006a) Aberrant seizure-induced
539
neurogenesis in experimental temporal lobe epilepsy. Ann.
Neurol., 59: 81–91.
Parent, J.M., Janumpalli, S., McNamara, J.O. and Lowenstein,
D.H. (1998) Increased dentate granule cell neurogenesis fol-
lowing amygdala kindling in the adult rat. Neurosci. Lett.,
247: 9–12.
Parent, J.M., Tada, E., Fike, J.R. and Lowenstein, D.H. (1999)
Inhibition of dentate granule cell neurogenesis with brain ir-
radiation does not prevent seizure-induced mossy fiber
synaptic reorganization in the rat. J. Neurosci., 19:
4508–4519.
Parent, J.M., Yu, T.W., Leibowitz, R.T., Geschwind, D.H.,
Sloviter, R.S. and Lowenstein, D.H. (1997) Dentate granule
cell neurogenesis is increased by seizures and contributes to
aberrant network reorganization in the adult rat hippocam-
pus. J. Neurosci., 17: 3727–3738.
van Praag, H., Kempermann, G. and Gage, F.H. (1999) Run-
ning increases cell proliferation and neurogenesis in the adult
mouse dentate gyrus. Nat. Neurosci., 2: 266–270.
van Praag, H., Schinder, A.F., Christie, B.R., Toni, N., Palmer,
T.D. and Gage, F.H. (2002) Functional neurogenesis in the
adult hippocampus. Nature, 415: 1030–1034.
van Praag, H., Shubert, T., Zhao, C. and Gage, F.H. (2005)
Exercise enhances learning and hippocampal neurogenesis in
aged mice. J. Neurosci., 25: 8680–8685.
Rao, M.S., Hattiangady, B., Abdel-Rahman, A., Stanley, D.P.
and Shetty, A.K. (2005) Newly born cells in the ageing dent-
ate gyrus display normal migration, survival and neuronal
fate choice but endure retarded early maturation. Eur. J.
Neurosci., 21: 464–476.
Santarelli, L., Saxe, M., Gross, C., Surget, A., Battaglia, F.,
Dulawa, S., Weisstaub, N., Lee, J., Duman, R., Arancio, O.,
Belzung, C. and Hen, R. (2003) Requirement of hippocampal
neurogenesis for the behavioral effects of antidepressants.
Science, 301: 805–809.
Schanzer, A., Wachs, F.P., Wilhelm, D., Acker, T., Cooper-
Kuhn, C., Beck, H., Winkler, J., Aigner, L., Plate, K.H. and
Kuhn, H.G. (2004) Direct stimulation of adult neural stem
cells in vitro and neurogenesis in vivo by vascular endothelial
growth factor. Brain Pathol., 14: 237–248.
Scharfman, H., Goodman, J., Macleod, A., Phani, S., Anton-
elli, C. and Croll, S. (2005) Increased neurogenesis and the
ectopic granule cells after intrahippocampal BDNF infusion
in adult rats. Exp. Neurol., 192: 348–356.
Scharfman, H.E., Goodman, J.H. and Sollas, A.L. (2000)
Granule-like neurons at the hilar/CA3 border after status
epilepticus and their synchrony with area CA3 pyramidal
cells: functional implications of seizure-induced neurogenesis.
J. Neurosci., 20: 6144–6158.
Scharfman, H.E., Sollas, A.L. and Goodman, J.H. (2002)
Spontaneous recurrent seizures after pilocarpine-induced sta-
tus epilepticus activate calbindin-immunoreactive hilar cells
of the rat dentate gyrus. Neuroscience, 111: 71–81.
Schlessinger, A.R., Cowan, W.M. and Gottlieb, D.I. (1975) An
autoradiographic study of the time of origin and the pattern
of granule cell migration in the dentate gyrus of the rat. J.
Comp. Neurol., 159: 149–175.
540
Schmidt-Hieber, C., Jonas, P. and Bischofberger, J. (2004)
Enhanced synaptic plasticity in newly generated granule cells
of the adult hippocampus. Nature, 429: 184–187.
Scott, B.W., Wang, S., Burnham, W.M., De Boni, U. and
Wojtowicz, J.M. (1998) Kindling-induced neurogenesis in the
dentate gyrus of the rat. Neurosci. Lett., 248: 73–76.
Seki, T. and Arai, Y. (1993) Highly polysialylated neural cell
adhesion molecule (NCAM-H) is expressed by newly gener-
ated granule cells in the dentate gyrus of the adult rat.
J. Neurosci., 13: 2351–2358.
Seri, B., Garcia-Verdugo, J.M., McEwen, B.S. and Alvarez-
Buylla, A. (2001) Astrocytes give rise to new neurons in
the adult mammalian hippocampus. J. Neurosci., 21:
7153–7160.
Shen, Q., Goderie, S.K., Jin, L., Karanth, N., Sun, Y.,
Abramova, N., Vincent, P., Pumiglia, K. and Temple, S.
(2004) Endothelial cells stimulate self-renewal and expand
neurogenesis of neural stem cells. Science, 304:
1338–1340.
Shors, T.J., Miesegaes, G., Beylin, A., Zhao, M., Rydel, T. and
Gould, E. (2001) Neurogenesis in the adult is involved in the
formation of trace memories. Nature, 410: 372–376.
Shors, T.J., Townsend, D.A., Zhao, M., Kozorovitskiy, Y. and
Gould, E. (2002) Neurogenesis may relate to some but not all
types of hippocampal-dependent learning. Hippocampus, 12:
578–584.
Song, H., Stevens, C.F. and Gage, F.H. (2002a) Astroglia in-
duce neurogenesis from adult neural stem cells. Nature, 417:
39–44.
Song, H.J., Stevens, C.F. and Gage, F.H. (2002b) Neural stem
cells from adult hippocampus develop essential properties of
functional CNS neurons. Nat. Neurosci., 5: 438–445.
Stanfield, B.B. and Trice, J.E. (1988) Evidence that granule cells
generated in the dentate gyrus of adult rats extend axonal
projections. Exp. Brain Res., 72: 399–406.
Suh, S.W., Fan, Y., Hong, S.M., Liu, Z., Matsumori, Y., Wein-
stein, P.R., Swanson, R.A. and Liu, J. (2005) Hypoglycemia
induces transient neurogenesis and subsequent progenitor cell
loss in the rat hippocampus. Diabetes, 54: 500–509.
Tanapat, P., Hastings, N.B., Reeves, A.J. and Gould, E. (1999)
Estrogen stimulates a transient increase in the number of new
neurons in the dentate gyrus of the adult female rat. J. Ne-
urosci., 19: 5792–5801.
Tozuka, Y., Fukuda, S., Namba, T., Seki, T. and Hisatsune, T.
(2005) GABAergic excitation promotes neuronal differenti-
ation in adult hippocampal progenitor cells. Neuron, 47:
803–815.
Trejo, J.L., Carro, E. and Torres-Aleman, I. (2001) Circulating
insulin-like growth factor I mediates exercise-induced in-
creases in the number of new neurons in the adult hippo-
campus. J. Neurosci., 21: 1628–1634.
Wadiche, L.O., Bromberg, D.A., Bensen, A.L. and Westbrook,
G.L. (2005) GABAergic signaling to newborn neurons in
dentate gyrus. J. Neurophysiol., 94: 4528–4532.
Wang, S., Scott, B.W. and Wojtowicz, J.M. (2000) Heteroge-
nous properties of dentate granule neurons in the adult rat. J.
Neurobiol., 42: 248–257.
Winocur, G., Wojtowicz, J.M., Sekeres, M., Snyder, J.S. and
Wang, S. (2006) Inhibition of neurogenesis interferes with
hippocampus-dependent memory function. Hippocampus,
16: 296–304.
Wurmser, A.E., Palmer, T.D. and Gage, F.H. (2004) Neuro-
science: cellular interactions in the stem cell niche. Science,
304: 1253–1255.
Plate 28.2. Increased cell proliferation and altered dentate gyrus neuroblast migration after SE. (A, B) BrdU incorporation (arrows) in
adult rat dentate gyrus 7 days after saline (A) or pilocarpine (B) treatment followed by a 2 day post-BrdU survival. Subgranular zone
BrdU labeling increases markedly after SE (B) compared to the control (A). (C, D) Doublecortin (DCX) immunoreactivity in the
dentate gyrus of a control (Con; C) and an adult rat 14 days after pilocarpine-induced SE (14d; D). Note the increased hilar DCX
expression and chains of DCX-positive cells extending into the hilus (arrows in D) after SE. E, PSA-NCAM+ neuroblast chains (red,
arrows) alongside GFAP+ hilar astrocytes (green). Dashed lines in C–E denote the granule cell layer (gcl) — hilar (h) border. (For
B/W version, see page 534 in the volume.)

Plate 28.3. Schematic showing the effects of prolonged seizures on caudal subventricular zone (SVZ) and dentate gyrus progenitors. In
the intact adult rat (top), progenitors located in the infracallosal SVZ (purple) give rise to white matter oligodendrocytes (orange), while
those in the dentate gyrus (green) give rise to DGCs in the granule cell layer (gcl). After seizure-induced hilar and pyramidal cell layer
(pcl) injury (bottom; jagged arrows), progenitors migrate aberrantly from the caudal SVZ to form glia (a; purple) in the hippocampus
proper or from the dentate subgranular zone to form hilar-ectopic DGCs (b). (For B/W version, see page 535 in the volume.)
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 29

Unmasking recurrent excitation generated by mossy

fiber sprouting in the epileptic dentate gyrus: an
emergent property of a complex system

Thomas P. Sutula1, and F. Edward Dudek2

1
Department of Neurology H6/570 CSC, University of Wisconsin, 600 Highland Avenue, Madison, WI 53792, USA
2
Department of Physiology, University of Utah, Salt Lake City, UT, USA

Abstract: Seizure-induced sprouting of the mossy fiber pathway in the dentate gyrus has been observed
nearly universally in experimental models of limbic epilepsy and in the epileptic human hippocampus.
The observation of progressive mossy fiber sprouting induced by kindling demonstrated that even a few
repeated seizures are sufficient to alter synaptic connectivity and circuit organization. As it is now
recognized that seizures induce synaptic reorganization in hippocampal and cortical pathways, the im-
plications of seizure-induced synaptic reorganization for circuit properties and function have been subjects
of intense interest. Detailed anatomical characterization of the sprouted mossy fiber pathway has revealed
that the overwhelming majority of sprouted synapses in the inner molecular layer of the dentate gyrus form
recurrent excitatory connections, and are thus likely to contribute to recurrent excitation and potentially to
enhanced susceptibility to seizures. Nevertheless, difficulties in detecting functional abnormalities in circuits
reorganized by mossy fiber sprouting and the fact that some sprouted axons appear to form synapses with
inhibitory interneurons have been cited as evidence that sprouting may not contribute to seizure suscep-
tibility, but could form recurrent inhibitory circuits and be a compensatory response to prevent seizures.
Quantitative analysis of the synaptic connections of the sprouted mossy fiber pathway, assessment of the
functional features of sprouted circuitry using reliable physiological measures, and the perspective
of complex systems analysis of neural circuits strongly support the view that the functional effects of the
recurrent excitatory circuits formed by mossy fiber sprouting after seizures or injury emerge only con-
ditionally and intermittently, as observed with spontaneous seizures in human epilepsy. The recognition
that mossy fiber sprouting is induced after hippocampal injury and seizures and contributes conditionally
to emergence of recurrent excitation has provided a conceptual framework for understanding how injury
and seizure-induced circuit reorganization may contribute to paroxysmal network synchronization, epile-
ptogenesis, and the consequences of repeated seizures, and thus has had a major influence on understanding
of fundamental aspects of the epilepsies.

Keywords: hippocampus; seizures; epilepsy; synaptic reorganization; granule cells; plasticity

Corresponding author. Tel.: +1 608 263 5448; Fax:

+1 608 263 9405; E-mail: sutula@neurology.wisc.edu

DOI: 10.1016/S0079-6123(07)63029-5 541

542
Introduction
Axon sprouting and growth are important cellular
processes in the coordinated sequence of cell birth,
differentiation, migration, and neurite outgrowth
that culminate in the formation of synapses,
circuits, and networks in the developing nervous
system. The earliest events in formation of neural
circuits are genetically determined, but as devel-
opment proceeds, activity-dependent sprouting
and axon growth play an important role in or-
ganizing and refining neural connections (Sur and
Rubenstein, 2005). Electrical activity in immature
circuits promotes synapse formation and organizes
synaptic connectivity into networks supporting
adult function, behavior, and adaptation. The ob-
servation, nearly 20 years ago, that repeated brief
seizures evoked by kindling induced sprouting
of the mossy fiber pathway in the dentate gyrus of
adult rats demonstrated that the activity-depend-
ent processes of sprouting and axon growth re-
quired for circuit formation in the developing
nervous system also remodel and reorganize adult
pathways after brief seizures (Sutula et al., 1988).
The association of axon sprouting with injury in
neural circuits
Sprouting in the adult nervous system was first
recognized in response to injury and damage. In
addition to the mossy fiber pathway, the septo-
hippocampal, associational, and crossed ent-
orhinal pathways within the dentate gyrus and
hippocampal formation undergo sprouting and
rearrangement of synaptic connectivity in response
to a variety of lesions and injuries (Steward et al.,
1973, 1974, 1976; Steward, 1976; Steward and
Messenheimer, 1978; Nadler et al., 1980a, 1981;
Staubli et al., 1984; Steward, 1992). The first sug-
gestion of a link between seizures, sprouting, and
circuit reorganization was the observation that
sprouted axons of the crossed entorhinal pathway
to the dentate gyrus gained access to hippocampal
networks modified by kindling (Messenheimer et
al., 1979). Evidence of relationships between in-
jury, sprouting, and seizures was provided by
studies of kainic acid-induced status epilepticus.
Status epilepticus induced by chemoconvulsants or
electrical stimulation induces macroscopic damage
in the hilus of the dentate gyrus, subfields of the
hippocampus, and a variety of extrahippocampal
regions, and is accompanied by reactive sprout-
ing of the mossy fiber axons in the dentate gyrus
(Ben-Ari et al., 1979, 1980; Nadler et al., 1980b;
Ben-Ari, 1985). Mossy fiber sprouting in these
models is a consequence of both direct neurotoxic
damage and excitotoxicity induced by status epi-
lepticus (Ben-Ari et al., 1980). The massive, mac-
roscopic damage associated with status epilepticus
in these models precluded definitely distinguishing
whether axon sprouting and reorganization was a
consequence of direct injury, lesion-induced deaff-
erentation, seizures, or a combination of these
processes.
Progressive mossy fiber sprouting induced by
repeated brief seizures
Axon sprouting was definitively identified as a
consequence of the recurring brief seizures that
define epilepsy by the observation in kindled rats
that mossy fiber axons expanded their terminal
field to the supragranular region of the dentate
gyrus in the absence of overt damage or injury
(Sutula et al., 1988). Mossy fiber sprouting, as
demonstrated by Timm histochemistry, develops
after only a few brief seizures, progresses with re-
peated seizures, one is permanent (Sutula et al.,
1988; Represa et al., 1989a; Cavazos et al., 1991,
1992). Further evidence that mossy fiber sprouting
is routinely and reliably induced by seizures was
provided by studies indicating that sprouting is
induced by seizures propagating into the hippo-
campus from remote regions (Golarai et al., 1992).
Mossy fiber sprouting has been observed in genetic
mouse models of epilepsy (Stanfield, 1989; Qiao
and Noebels, 1993), and after seizures evoked by
maximal electroshock, flurothyl, pentylenetetrazol,
intrahippocampal tetanus toxin, and by hype-
rthermia in immature rats (Golarai et al., 1992;
Holmes et al., 1998, 1999; Anderson et al., 1999;
Gombos et al., 1999; Vaidya et al., 1999; Bender
et al., 2003). Mossy fiber sprouting has been ob-
served in primates after complex partial seizures

induced by alumina gel injection (Ribak et al.,
1998). Sprouting has been observed in CA3, CA1,
and neocortex (Represa and Ben-Ari, 1992; Salin
et al., 1995; Esclapez et al., 1999; Lehmann et al.,
2001; Smith and Dudek, 2001; Cavazos et al.,
2004; Cross and Cavazos, 2007), and in the human
epileptic temporal lobe (Represa et al., 1989b;
Sutula et al., 1989; Houser et al., 1990; El Bahh et
al., 1999). Axonal remodeling may be a continuing
process in chronic poorly controlled epilepsy as
expression of growth associated proteins B-50 and
polysialyated neural cell adhesion molecule (PSA-
NCAM) are observed in resected human epileptic
hippocampus (Mikkonen et al., 1998). Because
mossy fiber sprouting is progressively induced by
repeated brief seizures, poorly controlled seizures
in humans may induce progressive sprouting and
synaptic reorganization.
Seizure-induced mossy fiber sprouting is
accompanied by progressive neuronal loss
As sprouting is induced by status epilepticus with
extensive injury and by direct damage to hippo-
campal pathways, and also by kindling which is
not accompanied by obvious overt damage, there
has been uncertainty about whether seizures can
induce sprouting in the absence of neuronal dam-
age. Mossy fiber sprouting has been induced by
repetitive stimulation trains that evoke long-term
potentiation (Adams et al., 1997), and in associ-
ation with spatial learning in behavioral tasks such
as the Morris water maze, which would not be
expected to produce neuronal loss (Ramirez-
Amaya et al., 1999; Holahan et al., 2006). In the
case of mossy fiber sprouting induced by seizures,
however, it is difficult to conclusively eliminate the
possibility that mild neuronal loss may be occur-
ring even with the most sensitive immunohisto-
chemical and stereological techniques for detection
of apoptosis and neuronal loss. Studies from mul-
tiple laboratories using techniques including unbi-
ased estimates of neuronal numbers, TUNEL
staining for apoptotic neurons, and silver impreg-
nation for degenerating cells have demonstrated
that single and repeated brief seizures induce neu-
ronal death which contributes to reorganization of
543
neural circuits (Cavazos and Sutula, 1990;
Cavazos et al., 1991, 1994; Bengzon et al., 1997;
Pretel et al., 1997; Zhang et al., 1998; Dalby
et al., 1998; Haas et al., 2001; Kotloski et al.,
2002). A single kindled seizure doubles the rate of
apoptosis in the hilar neurons of the dentate gyrus
(Bengzon, 1997), and kindling induces apoptosis in
CA1, subiculum, and neocortex (Pretel, 1997).
Cumulative neuronal loss has been detected after
repeated evoked secondary generalized seizures
evoked by kindling using stereological counting
methods in the CA1, CA3 subfields of the hippo-
campus and the hilus of the dentate gyrus resem-
bling classical human hippocampal sclerosis
(Cavazos and Sutula, 1990; Cavazos et al., 1994;
Kotloski et al., 2002). As neuronal loss is progres-
sively induced by kindling and occurs in topo-
graphically specific patterns in hippocampal and
limbic areas as a function of the site of seizure
initiation, mossy fiber sprouting after seizures in
the adult nervous system is closely associated with
seizure-induced neuronal loss, implying that even a
relatively small number of repeated seizures which
induce sprouting are potentially accompanied not
only by neuronal loss, but by a variety of cellular
alterations.
Mossy fiber sprouting is associated with a variety of
seizure-induced cellular alterations
Repeated seizures not only induce mossy fiber
sprouting and neuronal loss, but also a diverse
range of cellular and molecular alterations that
may potentially modify functional properties
of neural circuits. The range of alterations in-
duced in neurons and circuits by seizures spans
molecular to network levels. Neurons are born and
added to neural circuits in the dentate gyrus after
status epilepticus and only a few partial seizures
induced by kindling stimulation (Parent et al.,
1998), integrate into local circuits, and become
activated during recurring seizures (Scharfman
et al., 2002). Seizures alter dendritic branching
and promote spine loss (Jiang M et al., 1998), and
induce growth of basal dendrites by granule cells
of the dentate gyrus (Buckmaster and Dudek,
1999; Ribak et al., 2000). Repeated brief seizures
544
evoked by kindling upregulate GFAP (glial fibril-
lary acidic protein) mRNA and protein levels in a
time-dependent manner (Hansen et al., 1990;
Bonthius et al., 1995), cause glial cell hypertrophy
and proliferation (Khurgel and Ivy, 1996; Stringer,
1996). Kindled seizures modify vesicle release pro-
teins (Matveeva et al., 2006), alter extracellular
matrix proteins (Niquet et al., 1995), and modify
motor cortex maps (van Rooyen et al., 2006). The
association of mossy fiber sprouting with a diverse
variety of activity-dependent seizure-induced al-
terations in hippocampal pathways in both exper-
imental and human epilepsy has implications for
assessment of the functional effects of sprouting
and these other associated alterations, and poses
significant experimental challenges for interpreta-
tion of how and to what degree any one of these
seizure-induced alterations contribute to func-
tional properties of reorganized neural circuits.
Assessing the functional effects of mossy fiber
sprouting
The functional effects of sprouting are unavoida-
bly confounded by numerous neuronal and glial
alterations. Despite this potential complexity, the
pattern of alterations of the terminal field of the
reorganized sprouted pathway and the synaptic
targets of sprouted axons are important determi-
nants of the functional effects of mossy fiber
sprouting. Detailed anatomic characterization of
the terminal field of the sprouted mossy pathway
and quantitative morphological analysis of the
synaptic targets of sprouted mossy fiber terminals
have generated clear and unambiguous predictions
about the possible functional effects of the
sprouted pathway in hippocampal circuits.
Functional implications of seizure-induced
alterations in the terminal field of the mossy fiber
pathway
Reorganization of terminal fields by sprouting
would be expected to have potentially significant
functional consequences in neural networks. While
mossy fiber sprouting into the supragranular
region of the dentate gyrus has received a great
deal of attention because it is easily detected by a
variety of methods and is a prominent histological
alteration, seizures also induce axon growth in the
hilus, promote infrapyramidal to suprapryamidal
(interblade) connections not observed in normal
rats, and expand the terminal field of the mossy
fiber pathway in the inner molecular layer of the
dentate gyrus along the septotemporal axis of the
hippocampus over distances as long as 700 mm
(Sutula et al., 1998; Buckmaster and Dudek, 1999)
(Fig. 1). As an example of the potential functional
effects of sprouting, reorganization along the
septotemporal axis may have specific effects on
hippocampal dependent spatial memory, as right-
and left-specific place cells are organized in lamel-
lar patterns along this axis (Hampson et al., 1999).
Afferent projections to the dentate gyrus are topo-
graphically organized in the rat along the septo-
temporal axis (Dolorfo and Amaral, 1998), and as
sprouting in the inner molecular layer varies along
this axis as a function of site of seizure induction,
might disrupt the normal topographic organiza-
tion of afferent inputs undergoing processing
in hippocampal circuitry, which would be expected
to result in functional abnormalities including
paradigm specific behavioral and cognitive
dysfunction.
Functional implications of seizure-induced
alterations in synaptic targets of mossy fibers
Formation of synaptic contacts by sprouted axons
on principal cells or inhibitory interneurons deter-
mines whether seizure-induced reorganization
results in new excitatory or inhibitory neural
circuits. There is considerable information about
the targets of mossy fiber axons in both normal
and reorganized circuitry after seizures. The
number of synapses formed on principal cells
or interneurons determines whether the overall
impact of the new circuits is likely to be net ex-
citation or inhibition. While both types of synaptic
contacts have been observed in both normal and
epileptic hippocampus, the new synapses formed
by sprouted axons in the inner molecular layer
are predominantly on principal cells and form re-
current excitatory circuits (Franck et al., 1995;
 545

Fig. 1. Anatomical features and alterations in terminal fields of sprouted mossy fibers in the reorganized dentate gyrus. (A) Sprouting
by a biocytin-filled mossy fiber axon of a granule cell from a rat that received kainic acid in the supragranular region and inner
molecular layer and growth of the axon plexus in the hilus with increased numbers of synaptic varicosities. In addition, growth of basal
dendrites is well documented, but not shown. (B) Sprouted axons not only extend into the supragranular and inner molecular layer, but
cross the hilus from the infrapyramidal blade and thereby increase connectivity between blades of the dentate gyrus. (C) Sprouted
mossy fiber axons establish connectivity in the supragranular layer along the septotemporal axis as demonstrated by a biocytin filled
granule cell in a saggital section extending from the septal to temporal pole of the hippocampus and dentate gyrus. (D) The sprouted
axons of the granule cell in (C) shown at higher magnification establish connectivity in the supragranular layer extending nearly 700 mm
along the septotemporal axis. Scale bars (in mm): A, B, C ¼ 50, D ¼ 500. Adapted with permission from Sutula et al., 1998.

Okazaki et al., 1995; Wenzel et al., 2000; Buck-
master et al., 2002; Cavazos et al., 2003) (Fig. 2).
Analysis of the synaptic targets of sprouted mossy
fibers at the ultrastructural EM level has been
facilitated by the ability to identify mossy fiber
terminals based on size, morphological features,
immunoreactivity to dynophin, and labeling by
Timm histochemistry (Zhang and Houser, 1999;
Buckmaster et al., 2002; Cavazos et al., 2003). The
typical mossy fiber in the normal dentate gyrus
forms about 8–10 giant terminals that have asym-
metric (excitatory) synaptic contacts with neurons
in the hilus and with pyramidal neurons in CA3
(Amaral, 1979; Amaral and Dent, 1981; Claiborne
et al., 1986; Lim et al., 1997; Zhang and Houser,
1999; Gonzales et al., 2001). Although the ‘‘giant
mossy fiber terminal’’ is a characteristic distin-
guishing anatomical feature of mossy fiber axons
in the normal dentate gyrus and hippocampus, the
more numerous terminals of mossy fibers of the
normal dentate gyrus are small and form asym-
metric excitatory contacts with mossy cells and
with inhibitory interneurons in the hilus (Acsady
et al., 1998). About 98% of the small terminals in
the hilus are on inhibitory interneurons (Acsady
et al., 1998). These observations imply that while
the net contribution of the mossy fiber pathway
to CA3 is excitatory, collaterals in the hilus may
546

Fig. 2. Examples of synaptic targets of sprouted mossy fibers in the reorganized dentate gyrus. Presynaptic terminals from sprouted
mossy fibers form axo-spinous, axo-dendritic, and axo-somatic synapses, and occasionally form synapses with interneurons. (A) A
presynaptic terminal of a sprouted mossy fiber axon labeled with silver grains from pre-embedding Timm methods in the inner
molecular layer of the dentate gyrus of a kindled rat that experienced 50 class V seizures forms a synapse with a spine head (s)
connected to a dendrite (d) in the inner molecular layer. The same terminal also makes contact with another dendrite in cross section.
(B) A presynaptic terminal labeled by dynorphin-A immunoreactivity from a sprouted mossy fiber axon forms an asymmetrical
synapse with a dendritic spine in the inner molecular layer of a rat that experienced 50 class V kindled seizures. The spine appears
perforated and contains a spine apparatus. (C) A dendrite (d) in the inner molecular layer of the dentate gyrus is extensively contacted
by synaptic terminals from sprouted mossy fiber axons labeled by preembedding Timm technique in a kindled rat that experienced 50
class V seizures. (D) A dentate basket cell with prominent nuclear infolding receives numerous abnormally large presynaptic terminals
(shown in inset) exhibiting the morphological characteristics of mossy fiber boutons (cytoplasm with tightly packed vesicles, mi-
tochondria, and dense core vesicles), which form asymmetrical synapses with the dendrite. (E) A sprouted mossy fiber presynaptic
terminal (arrow) labeled by dynorphin-A is apposed to the plasma membrane of the cell bodies of granule cells (gc) in the granule cell
layer of the dentate gyrus of a rat that experienced 50 class V kindled seizures. (F) The sprouted synaptic terminal in (E) shown at
higher magnification forms an asymmetric postsynaptic density on the granule cell. Scale bars (in mm): A, B, C, E, F ¼ 1; D ¼ 5,
inset ¼ 1.25. Adapted with permission from Cavazos et al. (2003, panels A, B, C, E, F), and Sloviter et al. (2006; panel D).

predominantly contribute to local recurrent inhib-
itory circuits. In addition, mossy fibers in the nor-
mal dentate gyrus can form dense innervation
patterns along interneurons extending dendrites
into the molecular layer (Ribak and Peterson,
1991). These regionally specific anatomical
features of the normal mossy fiber pathway indi-
cate that functional contributions of asymmetric
excitatory terminals of mossy fibers to net excita-
tion and inhibition may vary systematically in
different regions of hippocampal circuitry, even
from the hilus to CA3.

It should not be surprising that similar local
circuit complexity is encountered in the reorgan-
ized dentate gyrus. Except for the occasional in-
terneuron dendrite that receives dense innervation
by mossy fibers in the normal dentate gyrus (Ribak
and Peterson, 1991), the terminal field of mossy
fibers does not project to the inner molecular layer
and this region is therefore devoid of recurrent
circuits. Seizures alter the terminal field and the
synaptic targets of mossy fibers in the inner mo-
lecular layer by sprouting and growth of new
mossy fiber axons which predominantly contact
excitatory principal cells and to a much lesser ex-
tent form synapses with interneurons (Frotscher
and Zimmer, 1983; Franck et al., 1995; Okazaki
et al., 1995; Kotti et al., 1997; Wenzel et al., 2000;
Pierce and Milner, 2001; Buckmaster et al., 2002;
Cavazos et al., 2003; Sloviter et al., 2006) (Figs. 2
and 3). Quantitative analyses of the relative fre-
quency of synapses have provided unequivocal
evidence that the overwhelming number of new
synapses are on granule cells and form excitatory
circuits. In the most comprehensive study of
synaptic targets of sprouted mossy fibers, 96% of
sprouted terminals contacted granule cell dendrites
in the inner molecular layer and thus formed
recurrent excitatory circuits (Buckmaster et al.,
2002) (Fig. 3). In another study, the number
of asymmetric excitatory axo-spinous synapses
characteristic of granule cell synapses exceeded
asymmetric axo-dendritic shaft synapses putatively
on inhibitory dendrites by nearly 2 to 1 in the
reorganized inner molecular layer of kainic acid-
treated rats (Cavazos et al., 2003). Perforated axo-
spinous synapses on granule cells are observed
in kindled rats (Geinisman et al., 1992; Cavazos
et al., 2003), as well as axo-somatic synapses which
are not typically observed in normal rats (Cavazos
et al., 2003). While some authors have emphasized
the very occasional to rare synapses formed by
sprouted mossy fibers on interneurons (Kotti
et al., 1997; Sloviter et al., 2006), the overwhelm-
ing anatomical evidence from the most compre-
hensive and quantitative analyses clearly indicates
that the predominant circuit formed by seizure-
induced sprouting in the inner molecular layer is
a recurrent excitatory circuit (Coulter, 2002;
Sutula, 2002; Dudek and Shao, 2004).
547
The quantitative anatomical studies indicating
that the overwhelming number of new synapses
formed by sprouted mossy fibers in the inner mo-
lecular layer form new recurrent excitatory circuits
are not incompatible with the fact that many
mossy fiber terminals in the hilus of normal rats
form recurrent inhibitory circuits (Acsady et al.,
1998), and that these circuits and potentially new
mossy fiber connections to inhibitory interneurons
formed in the hilus by sprouted axons may con-
tribute to inhibition in the reorganized dentate
gyrus. The important point, however, is that new
recurrent excitatory circuits are formed in the
inner molecular layer, a region that normally lacks
any recurrent excitatory connections among gran-
ule cells. It is noteworthy that some studies have
reported expression of the 67 kDa isoform of glut-
amic acid decarboxylase (GAD67, the synthetic
enzyme for GABA) in mossy fiber terminals after
status epilepticus (Sloviter et al., 1996), suggesting
that sprouted asymmetric MF synapses may
contribute to inhibition in some circumstances
(Gutierrez and Heinemann, 2001), although the
overall physiological significance of this alteration
remains unclear. So while mossy fibers in the nor-
mal and reorganized dentate gyrus form regionally
specific and heterogeneous synaptic connections
with both CA3 pyramidal neurons and interneu-
rons in the hilus, seizures importantly induce
a new and predominantly excitatory circuit among
granule cells that does not exist in the normal
dentate gyrus.
The paradox of epileptic circuits: episodic
dysfunction and synchronization in association with
permanent structural reorganization
Despite the abundant evidence for mossy fiber
sprouting and a variety of chronic structural and
molecular alterations in the epileptic hippocam-
pus, it has been surprisingly difficult to detect
evidence for functional abnormalities in epileptic
circuitry in normal physiological conditions both
in vitro and in vivo, including in the resected
human epileptic hippocampus (Cronin and
Dudek, 1988; Cronin et al., 1992; Golarai and
Sutula, 1996; Wuarin and Dudek, 1996, 2001;
548

Fig. 3. Sprouted mossy fibers synapse with GABA-negative and GABA-positive targets identified by EM immunohistochemistry. The
predominant circuit formed by sprouted mossy fibers in the supragranular and inner molecular layer is a recurrent excitatory circuit.
(A) A biocytin-labeled axon (black) forms synaptic contacts (arrowheads) with two GABA-negative spines. Nearby GABA-positive
structures are labeled with 10 nm-diameter colloidal gold particles, and a GABA-positive axon terminal forms a symmetric synapse
(arrow) with a granule cell body. (B) A biocytin-labeled axon (black) in the molecular layer forms a synaptic contact (arrowhead) with a
GABA-positive dendritic shaft. (C) Sprouted mossy fibers synapse preferentially with GABA-negative dendritic spines. Example of a
reconstructed sprouted mossy fiber with synaptic contacts indicated by markers. Squares indicate that the postsynaptic target was a
dendritic spine; circles indicate a dendritic shaft. Open markers indicate that the postsynaptic target was GABA-negative; filled markers
indicate GABA-positive. Borders between strata (h, hilus; gcl, granule cell layer; ml, molecular layer) are indicated by lines. The vast
majority of synapses (93–96% in the granule cell layer and inner molecular layer respectively) form synapses with GABA-negative
structures demonstrating that sprouted mossy fibers in the supragranular and inner molecular predominantly form recurrent excitatory
circuits. Adapted with permission from Buckmaster et al., 2002.

Buckmaster and Dudek, 1997a, b, 1999; Lynch
and Sutula, 2000; Sutula, 2002). Membrane prop-
erties and evoked responses in resected human
epileptic hippocampal slices in standard physio-
logical conditions appear surprisingly normal
(Kim et al., 1993; Dudek et al., 1995). Initial anal-
yses of evoked field potentials in the dentate gyrus
of hippocampal slices from kainic acid-treated rats
and resected hippocampal slices from patients with
temporal lobe epilepsy demonstrated an associa-
tion between the duration and complexity of
antidromically evoked extracellular field potentials
and the extent of mossy fiber sprouting examined
by the Timm method (Tauck and Nadler, 1985;
Masukawa et al., 1992), but other studies revealed
only minimal evidence for spontaneous epileptic
burst discharges in hippocampal slices from kainic
acid-treated epileptic rats (Cronin and Dudek,
1988; Cronin et al., 1992; Nissinen et al., 2001).
Attempts to correlate the extent of mossy fiber
sprouting with spontaneous seizures in chronic in
vivo models and human epileptic hippocampus
also failed to reveal a direct relationship between
sprouting and seizure frequency, and were inter-
preted as evidence that mossy fiber sprouting does
not contribute to abnormal excitation and epilepsy
(Longo and Mello, 1998, 1999; Pitkanen et al.,
2000; Nissinen et al., 2001). Studies indicating that
prevention of mossy fiber sprouting by co-admin-
istration of the protein synthesis inhibitor cyclo-
heximide did not prevent development of seizures
after pilocarpine (Longo and Mello, 1998, 1999),
which were subsequently not replicated (Williams
et al., 2002; Toyoda and Buckmaster, 2005), also
were cited as evidence that sprouting is not related
to epileptogenesis. Other studies noting that
seizures could be induced in the absence of sprout-
ing similarly were interpreted as evidence that
sprouting is not necessary for seizure induction by
kindling (Armitage et al., 1998; Mohapel et al.,
2000), which should not be surprising given that
previous studies demonstrated that kindled
seizures can be evoked from limbic pathways
when granule cells in the dentate gyrus have been
selectively lesioned by colchicine (Dasheiff and
McNamara, 1982; Sutula et al., 1986). Skepticism
about the potential importance of mossy fiber
sprouting thus developed as multiple studies did
549
not detect simple direct relationships between
sprouting and outcome variables such as sponta-
neous seizure frequency, and led some to conclude
that sprouting and other cellular alterations in
the reorganized epileptic hippocampus have neg-
ligible functional importance for epileptogenesis
(Benardo, 2002).
Contrary to this viewpoint, synaptic transmis-
sion in the terminal field of the sprouted mossy
fiber pathway has been directly demonstrated in
vivo by current source density analysis in kindled
rats (Golarai and Sutula, 1996) (Fig. 4A–E).
An inward current (sink) corresponding spatially
to the terminal field of the sprouted mossy fiber
terminals in the inner molecular layer of
the dentate gyrus develops in kindled rats at a
latency consistent with disynaptic transmission
in response to perforant path stimulation. The
difficulty of detecting functional alterations in
chronically reorganized epileptic circuitry despite
multiple neuronal and circuit alterations and
functional synaptic transmission in the terminal
field of the sprouted mossy fiber pathway might
appear to be a paradox. This paradox should not
be entirely surprising, however, given the clinical
fact that patients with epilepsy manifest seizures
only sporadically and briefly, so continuous
evidence for network synchronization and un-
derlying imbalance of excitation and inhibition
should not be expected. Although chronic sus-
ceptibility to recurring network synchronization
and seizures is the defining feature of epilepsy,
seizures are an infrequent emergent event even in
intractable patients, and may not be associated
with continuously detectable functional abnor-
malities (Litt et al., 2001; Litt and Echauz, 2002;
Worrell et al., 2004). The paradox of infrequent
expression of spontaneous functional abnormal-
ity despite the presence of permanent structural
alterations has implications for design and in-
terpretation of experiments seeking to identify
functional abnormalities associated with epilep-
tic circuits reorganized by sprouting and a vari-
ety of cellular alterations. The paradox has
produced confusion that is in part based on lack
of understanding about emergent functional
properties in complex systems such as neural
circuits.
Fig. 4. Synaptic transmission in the terminal field of the sprouted mossy fiber pathway and unmasking of functional abnormalities in
the reorganized dentate gyrus with mossy fiber sprouting. (A) Surface plot of current source density (CSD) as function of depth and
time in the dentate gyrus of a rat that experienced three afterdischarges but prior to development of mossy fiber sprouting. Inward
currents (sinks) are upward and outward currents (sources) are downward. (B) Corresponding CSD contour plot demonstrates that the
inward current sink (solid lines) develops in the middle and outer molecular layer at short latency, and is followed by a lower amplitude
sink developing in the inner molecular layer at 16 ms. (C) Surface plot of CSD in the dentate gyrus of a rat with mossy fiber sprouting
after 105 class V seizures demonstrates a prominent inward current sink developing in the inner molecular layer at a latency of 9–10 ms
(arrow) corresponding to the terminal field of the sprouted mossy fiber pathway. (D) Corresponding CSD contour plot also dem-
onstrates the prominent inward current sink developing in the inner molecular layer at a latency of 9–10 ms (arrow). (E) Schematic
summary of the spatial and temporal features of inward current sinks corresponding to the terminal field of sprouted mossy fiber
pathway compared to normal control rats without sprouting. Black areas in the schematic dendrites are inward currents of highest
amplitude and gray areas are inward currents of lower amplitude. The initial inward current at 3 ms in both the normal and sprouted
dentate gyrus is in the distal dendrites (black area), corresponding to monosynaptic transmission in the perforant path. At 9–12 ms in
the normal DG, inward current are not yet developed in the proximal dendrites of the inner molecular layer. In the reorganized dentate
gyrus with sprouting, a large inward current develops in the proximal dendrites at 9–12 ms, which colocalizes with the laminar
distribution of sprouted mossy fiber terminals. This current is consistent with disynaptic transmission in the sprouted pathway in
response to perforant path stimulation. (F) Spontaneous epileptic bursts are unmasked in the dentate gyrus of hippocampal slices from
rats with mossy fiber sprouting in 10–30 mM bicuculline. The spontaneous bursts consist of prolonged negative shifts in the extra-
cellular field potential and synchronous action potentials. Slices without sprouting had small intracellular depolarizations followed by
hyperpolarizations and positive-going extracellular fields; no action potentials or extracellular population spikes were observed. In
slices with sprouting, large spontaneous depolarizations that evoked intracellular action potentials could occur synchronously with
slow extracellular negative shifts and population spikes. Arrows show expansion of traces. (G) Unmasking of functional abnormalities
in the reorganized dentate gyrus with mossy fiber sprouting by altering the extracellular ionic environment. In standard bathing
solution (3.5 mM K+), antidromic stimulation evoked epileptiform activity only in the presence of robust recurrent mossy fiber
growth. Increasing Timm scores indicate increased density of mossy fiber sprouting. For each value of the Timm score, raising [K+]o
increased the percentage of slices that responded with epileptiform activity. In slices with a Timm score of 1, epileptiform activity
appeared only when [K+]o was increased. Adapted with permission from Golarai and Sutula (1996; panels A–E), Cronin et al. (1992;
panel F), and Hardison et al. (2002; panel G).

Episodic paroxysmal abnormalities in epileptic
circuitry: the phenomena of unmasking and
emergent functional properties in reorganized
circuitry
Evidence for functional abnormalities in brain
slices from experimental models of epilepsy exam-
ined by in vitro physiological methods generally
emerges only when the reorganized circuitry
is perturbed by alterations in the extracellular
environment such as elevated [K+]o, or reduced
[Ca2+]o, or by disinhibition by GABAA antago-
nists (Cronin et al., 1992; Wuarin and Dudek,
1996; Patrylo and Dudek, 1998; Molnar and
Nadler, 1999; Patrylo et al., 1999; Hardison
et al., 2000; Lynch and Sutula, 2000; Wuarin and
Dudek 2001; Sutula, 2002) (Fig. 4F and G). Sur-
gically resected human epileptic hippocampus
similarly appears normal unless the extracellular
environment is altered directly or in response to
tetanic stimulation (Masukawa et al., 1989, 1992;
Wuarin et al., 1990; Kim et al., 1993; Dudek et al.,
1995). These observations suggest that the reor-
ganized circuits are a substrate for dysfunction,
but functional abnormality becomes unmasked
and emerges only in the context of some other
perturbation or transient alterations. These obser-
vations in hippocampal slices and the fact that
spontaneous seizures occur sporadically and
unpredictably in both experimental and human
epilepsy indicate that seizures and underlying
epileptic functional abnormalities are emergent
properties of epileptic neural circuitry altered
by primary pathologies or by seizure-induced re-
organization.
Unmasking emergent functional abnormalities in
recurrent sprouted circuits by modifying
GABAergic neurotransmission
Functional abnormalities in reorganized epileptic
circuits including spontaneous synchronous net-
work discharge become unmasked by reducing
GABAergic inhibition. The fact that epileptic net-
work synchronization can emerge even in normal
neural circuits when inhibition is reduced indicates
the narrow dynamic balance of excitatory and
551
inhibitory processes, and complicates efforts to
directly identify the contributions of new recurrent
excitatory circuits in the epileptic dentate gyrus.
How and when functional effects emerge at the
network and behavioral levels as a consequence
of the abnormal excitatory disynaptic activation of
granule cell dendrites associated with sprouted
terminal field in the inner molecular layer is a crit-
ical question. The phenomena of unmasking and
dynamic emergent properties must be considered
in the design and interpretation of experiments
seeking to identify the functional consequences of
sprouting and other circuit alterations associated
with epilepsy. Emergent functional effects of new
recurrent excitatory circuits formed by mossy fiber
sprouting will be influenced by the state of inhi-
bition in the dentate gyrus, which must remain
sufficiently robust even in intractable epilepsy to
prevent runaway excitation and continuous syn-
chronization. Dynamic alterations in inhibition
are therefore a key factor in assessing func-
tional effects of sprouting in reorganized epileptic
circuity of the dentate gyrus.
The state of inhibition in the reorganized dentate
gyrus: a critical variable for emergence of recurrent
excitation
The state of GABAergic inhibition thus plays an
important and indeed critical role in regulating
emergence of network synchronization and other
functional abnormalities in both the normal and
reorganized dentate gyrus. GABAergic neuro-
transmission is not a unitary or simple process
but rather represents a diverse range of potentially
complex and dynamic cellular and physiological
mechanisms which undergo activity-dependent
modification (McCarren and Alger, 1985;
Brooks-Kayal et al., 1998; Coulter, 2001; Cossart
et al., 2005). In some circumstances, such as in
early postnatal development and during acute
conditions when intracellular Cl
increases, GAB-
Aergic neurotransmission can contribute to exci-
tation as well as cellular inhibition (Staley et al.,
1995; Khalilov et al., 1999; Leinekugel et al., 1999;
Dzhala and Staley, 2003). Assessing the functional
effects of sprouting therefore requires analysis of
552
alterations in both inhibitory and excitatory
synaptic transmission and their dynamic interac-
tions in circuits that have undergone reorganiza-
tion as a result of seizures and primary injury or
pathology. Dynamic alterations in excitatory and
inhibitory synaptic transmission and seizure-in-
duced formation of new circuits vary significantly
in different regions of the hippocampus and inde-
pendently as a function of time after seizures. The
time course of alterations in synaptic transmission
evolve over periods of as long as months after the
last episode of network synchronization in circuits
undergoing remodeling and reorganization, so ex-
periments seeking to understand the emergence of
functional abnormalities associated with sprouting
also need to evaluate how inhibition and altera-
tions in connectivity systematically vary across
these prolonged periods in order to understand
functional effects of sprouting and phenomena
such as latent periods after initial injuries.
The normal dentate gyrus has powerful systems
of inhibition as demonstrated by the fact that
granule cells are resistant to repetitive discharge in
response to synaptic activation, infrequently gen-
erate spontaneous activity, and typically do not
generate repetitive discharges even when GABAA
inhibition is blocked (Fricke and Prince, 1984;
Lynch and Sutula, 2000; Lynch et al., 2000). The
state of inhibition in the dentate gyrus is predict-
ably modified by episodic network synchroniza-
tion, specifically as a function of both the timing
and duration of seizures. A variety of physiolog-
ical and pharmacological methods both in vivo
and in vitro have demonstrated that repeated brief
seizures evoked by kindling increase inhibition in
the dentate gyrus (Tuff et al., 1983a, b; de Jonge
and Racine, 1987; Stringer and Lothman, 1989;
Otis et al., 1994; Buhl et al., 1996; Nusser et al.,
1998; Sayin et al., 2003), while sustained seizures
as during status epilepticus initially have opposite
effects and produce rundown or acute reduction of
inhibition (Kapur et al., 1989c; Kapur and Mac-
donald, 1997), which is later followed by increases
in inhibition in granule cells (Cohen et al., 2003).
After a few repeated brief seizures evoked during
the early stages of kindling, paired pulse inhibition
in the dentate gyrus is acutely increased at inter-
pulse intervals of 15–40 ms, which are associated
with the GABA-dependent Cl
mediated conduct-
ance change and provide an indirect measure
of GABA-dependent inhibition of evoked granule
cell discharge under appropriately controlled stim-
ulation and timing parameters (Tuff et al., 1983a,
b; de Jonge and Racine, 1987; Stringer and
Lothman, 1989; Sayin et al., 2003). The increase
in inhibition implied by the indirect measure of
paired pulse inhibition has been directly confirmed
by demonstration of increased frequency and am-
plitude of IPSCs in granule cells (Otis et al., 1994;
Buhl et al., 1996; Nusser et al., 1998). It is also
clear that inhibition after status epilepticus is
increased in the dentate gyrus after an initial
period of depression (Hellier et al., 1999; Gorter
et al., 2002; Harvey and Sloviter, 2005; Sloviter
et al., 2006).
Although inhibition in the dentate gyrus is
initially increased by kindling, inhibition is even-
tually lost after 100 evoked Class V seizures
in association with emergence of spontaneous sei-
zures, as directly confirmed by reduction in
amplitude and alterations in kinetics of the
monosynaptic IPSC measured by single electrode
voltage clamp techniques (Sayin et al., 2003). The
loss of inhibition and emergence of spontaneous
seizures is associated with reduction of CCK and
GAT-1 subpopulations of interneurons, which
provide axo-somatic and axo-axonic inhibitory
terminals on granule cells. These observations are
consistent with the view that recurrent excitatory
circuits which are progressively formed by sprout-
ing in kindled rats gradually increase capacity to
generate recurrent excitation in the dentate gyrus
but are insufficient at early stages of seizure-
induced reorganization to generate spontaneous
seizures unless inhibition is reduced. With seizure-
induced loss of interneurons and critical axo-
somatic and axo-axonic inhibitory terminals at
advanced stages of kindling, recurrent excitation
generated by sprouted circuits among granule
cells, while still for the most part checked by
inhibition, periodically becomes sufficient to drive
spontaneous emergent network synchronization.
These studies in rats experiencing brief seizures
evoked by kindling and status epilepticus indicate
that the state of inhibition in the dentate gyrus
varies both acutely and chronically in response to

ongoing episodes of network synchronization as a
function of the duration, frequency, and number
of seizures. Efforts to assess the potential func-
tional properties of recurrent circuits formed by
sprouting must be considered with recognition of
these dynamic alterations in inhibition as a con-
sequence of timing and duration of recent seizures,
and pursued with a variety of electrophysiological
techniques.
In addition to the effects of timing and duration
of seizures, it is also clear from in vivo and in vitro
studies that the effects of seizures on network
inhibition vary significantly as function of location
in the hippocampus, and that results in other
regions of hippocampal and neocortical circuitry
do not necessarily apply to the dentate gyrus.
The effects of seizures in the dentate gyrus specifi-
cally contrast with CA1, where in vivo methods
have demonstrated reduction in inhibition after
brief evoked seizures and status epilepticus (Kapur
et al., 1989a–c; Michelson et al., 1989; Gorter
et al., 2002). In further contrast to the dentate
gyrus, GABAA receptor dependent currents
in granule cells are increased after status epilep-
ticus induced by pilocarpine but are reduced
in pyramidal neurons of CA1 (Gibbs et al.,
1997), and dendritic GABAergic inhibition
in CA1 as measured by analysis of frequency
and amplitude of IPSCs is reduced while somatic
inhibition is increased (Cossart et al., 2001).
Audiogenic kindling reduces GABAA depend-
ent IPCSs in the inferior colliculus (Evans et al.,
2006), and increases amplitude and duration of
mIPSCs in piriform cortex (Gavrilovici et al.,
2006). These contrasting observations in the
dentate gyrus, CA1, and other regions clearly
demonstrate that assessment of inhibitory and ex-
citatory processes in both normal and reorgan-
ized circuitry cannot be simply characterized
by a single technique or in a single model or in
a single region of hippocampal circuitry,
and that multiple techniques are required to
systematically characterize the effects of seizures
on these synaptic and circuit properties. This is
an important perspective for efforts to assess
the effects of sprouting or any other putative
cellular causes of seizures and chronic
epilepsy.
553
Direct evidence for recurrent excitation unmasked
by disinhibition in the reorganized dentate gyrus
The functional abnormalities that become
unmasked in the reorganized dentate gyrus by
reducing inhibition or altering the extracellular
ionic environment include direct evidence for
development of recurrent excitation in association
with mossy fiber sprouting. Evidence for recurrent
excitation in association with mossy fiber sprout-
ing has been demonstrated by focal stimulation by
glutamate application using microdrop or flash
photolysis techniques (Wuarin and Dudek, 1996,
2001; Molnar and Nadler, 1999; Lynch and
Sutula, 2000). Focal application of glutamate
microdrops to dendrites and cell bodies of gran-
ule cells remote from the recorded granule cell
in hippocampal slices from normal rats evokes
no responses when inhibition is blocked, but mi-
crodrop application in disinhibited slices with
sprouting evokes EPSPs at long and variable
latency (Wuarin and Dudek, 1996; Lynch and
Sutula, 2000). In kindled rats with mossy fiber
sprouting, trains of EPSPs and population dis-
charges can be evoked by glutamate microstimu-
lation remote from the recording site at one week
after induction of kindled seizures, when sprouting
is first detectable by Timm histochemistry, but are
not evoked in hippocampal slices from kindled rats
examined at 24 h after a single afterdischarge prior
to the development of sprouting (Lynch and
Sutula, 2000). Stimulation by flash photolysis of
caged glutamate at sites remote from the recording
site in granule cells also evokes EPSCs in granule
cells (Molnar and Nadler, 1999; Wuarin and
Dudek, 2001) which increase in correlation with
the development of sprouting (Fig. 5A and B). The
long and variable latency of responses evoked by
focal glutamate stimulation in these studies sug-
gested that recurrent excitation was generated by
multisynaptic rather than monosynaptic circuits,
but monosynaptic EPSPs can be evoked at short
latency (2.670.36 ms) between blades of the dent-
ate gyrus under conditions in which recurrent
inhibitory circuits are blocked by bicuculline,
polysynaptic activity is suppressed by 10 mM
Ca2+ in the bathing medium, and perforant path
activation is prevented by knife cuts (Lynch and
554

Fig. 5. Formation of recurrent excitatory circuits in the reorganized dentate gyrus with mossy fiber sprouting. (A) Evidence for
development of excitatory connectivity in the reorganized dentate gyrus with mossy fiber sprouting. Photostimulation of caged
glutamate in the granule cell layer in conditions of reduced inhibition (bicuculline, 30 mM) and elevated [K+]o (6 mM) which evokes no
responses in the normal dentate gyrus evokes epileptiform bursts of action potentials in a granule cell from a rat 33 weeks after kainate
treatment. The patch pipette in the diagram indicates the position of the recorded granule cell in the outer blade. The numbers show the
locations of the photostimulations in the diagram and the corresponding evoked bursts of action potentials. The granule cell was
recorded in the whole cell current-clamp configuration at resting membrane potential. Arrowheads show stimulation artifact produced
by the flash. (B) Excitatory connectivity as revealed by flash photolysis increases with time after treatment with kainic acid in
association with increasing mossy fiber sprouting. Plot of the percentage of granule cells from saline- and kainate-injected animals
responding to photostimulation with an increase in EPSCs as a function of time after treatment. The number of cells tested is indicated
for each age group. (C) Monosynaptic excitatory connections between blades of the dentate gyrus in a hippocampal slice from a rat
treated with kainic acid. The transected transverse hippocampal slice contained only the infrapyramidal and suprapyramidal blades of
the dentate gyrus, the intervening hilus, a small sector of CA3c, and CA1. The transection removed the crest of the dentate gyrus,
which severed perforant path connections between the blades. In bathing solution containing 10 mM bicuculline to suppress inhibitory
postsynaptic potentials and 10 mM [Ca2+]o to suppress polysynaptic activity, stimulation of the molecular layer of the infrapyramidal
blade (site indicated by arrows) with a 100 ms constant-voltage pulse of 7 V evoked an EPSP in a granule cell in the suprapyramidal
blade (location indicated by the arrowhead). The latency of the EPSP was 2.7 ms. Focal electrical microstimulation evoked EPSPs in 5
of 18 suprapyramidal granule cells in kainic acid-treated rats with an average latency of 2.5970.36 ms. In contrast, stimulation of the
infrapyramidal blade in hippocampal slices from normal rats failed to evoke EPSPs in 15 suprapyramidal granule cells. Calibration
bars: 2 mV, 8 ms. (D) Monosynaptic connections directly demonstrated between granule cells in slices from rats with mossy fiber
sprouting. Recordings from a pair of simultaneously recorded granule cells are shown in (5D1): presynaptic neuron (top), postsynaptic
neuron (bottom). Intracellular current (a 150 ms rectangular current pulse; start and end of the pulse are marked by the dots) triggered
an action potential (AP) in the presynaptic cell. Immediately thereafter, a small depolarization occurred in the postsynaptic cell. An
arrow marks the capacitative artifact of the presynaptic cell’s AP. Calibration: presynaptic cell, 20 mV, 30 ms; postsynaptic cell, 4 mV,
30 ms. (5D2): recordings from the same pair of neurons with higher gain. Several postsynaptic responses are overlapped to show the
variability in the response to the presynaptic AP. Calibration: presynaptic cell, 20 mV, 4 ms; postsynaptic cell, 3 mV, 4 ms. (5D3): in a
different pair of granule cells, tonic intracellular current was used to depolarize both the putative presynaptic (top) and postsynaptic
(bottom) cells. A spontaneous AP in the presynaptic cell triggered an AP in the second cell. Membrane potentials: top,
55 mV;
bottom,
54 mV. Calibration: 15 mV, 25 ms. Adapted with permission from Wuarin and Dudek (2001; panel A), Hardison et al. (2000;
panel B), Lynch et al. (2000; panel C), and Scharfman et al. (2003; panel D).

Sutula, 2000) (Fig. 5C). These studies from mul-
tiple laboratories are evidence that seizures induce
recurrent excitatory connectivity in the dentate
gyrus which emerges only under conditions of re-
duced inhibition or alterations in the extracellular
ionic environment.
This physiological evidence for formation of
recurrent excitatory connections has been con-
firmed by dual recordings from pairs of granule
cells in the dentate gyrus reorganized by mossy
fiber sprouting (Fig. 5D1–3). Examination of 903
granule cell pairs in hippocampal slices from
epileptic pilocarpine-treated rats revealed monosy-
naptic connections in 1/150 granule cell pairs
which were not observed in any of 285 pairs from
normal animals (Scharfman et al., 2003). While the
number of pairs was small, the results directly
confirm the presence of recurrent excitatory
circuits in epileptic dentate gyrus when inhibition
was intact, and leave open the possibility that
reduction in inhibition might reveal additional
evidence for recurrent excitatory circuits masked
by powerful inhibition present in both the normal
and reorganized dentate gyrus.
Dynamic and evolving plasticity in reorganized
excitatory and inhibitory circuits of the epileptic
dentate gyrus: implications for assessment of the
functional effects of sprouting
The molecular and cellular processes underlying
network inhibition and excitation may be altered
by primary pathologies that cause epilepsy. In
addition, network inhibition and excitation are
systematically altered as a consequence of repeated
network synchronization and seizures, i.e. by sei-
zure-induced plasticity or kindling. Inhibition and
excitation undergo independent acute and long-
term alterations that extend for as long as months
after seizures. For example, in addition to the
rewiring of synaptic connections by mossy fiber
sprouting, enhancement of kainate-receptor and
NMDA-receptor dependent glutamatergic trans-
mission also results in alterations in excitatory
synaptic properties that may contribute to net-
work synchronization (Sayin et al., 1999; Behr
et al., 2001; Epsztein et al., 2005). If the primary
555
pathological process and the subsequent processes
of seizure-induced plasticity following an initial
episode of network synchronization are sustained
or permanent, recurring seizures that define epi-
lepsy can emerge. The strength and dynamic bal-
ance of excitatory and inhibitory transmission,
which undergo activity-dependent alterations after
seizures in reorganizing circuitry with sprouting,
will evolve and vary as a function of time and
duration of recent seizures. These potentially com-
plex interactions and the temporal relationships
of acute and chronic seizure-induced plasticity
support the view that recurrent excitation and net-
work synchronization be considered as emergent
circuit properties in different experimental models.
As the neurobiological phenomena of plasticity
associated with repeated network synchronization
and seizures in neural circuits, the gradually evolv-
ing time course of circuit alteration induced
by kindling has been informative for assessment
of how mossy fiber sprouting contributes to
network dysfunction and epileptogenesis. Unlike
models of status epilepticus which produce
massive initial damage followed by mossy fiber
sprouting, kindling induces gradually progressive
sprouting accompanied by incremental but cumu-
lative neuronal loss. These features of gradually
evolving cumulative circuit alterations after brief
seizures evoked by kindling have enabled physio-
logical analysis at different time points after
a range of induced seizures, and thereby provide
an experimental opportunity to examine both
acute and chronic effects of repeated network syn-
chronization in normal and reorganized circuits
with sprouting. Exploitation of these features
of kindling has enabled recognition of the distinct
contributions of increased NMDA dependent
synaptic currents, fading of inhibition in associa-
tion with seizure-induced loss of interneuron sub-
populations, and identification of recurrent
excitatory circuits at both early and advanced
stages of circuit reorganization (Golarai and
Sutula, 1996; Sayin et al., 1999, 2003; Lynch
et al., 2000; Lynch and Sutula, 2000). Similar
experimental designs have been employed for
analysis of structural and functional alterations
in models of status epilepticus (Houser and
Esclapez, 1996; Buckmaster and Dudek, 1997a, b,
556
1999; Cossart et al., 2001; Austin and Buck-
master, 2004; Peng et al., 2004; Peng and Houser,
2005; Ang et al., 2006). Attempts to define the
functional effects of sprouting in both kindling
that gradually evolves into spontaneous seizures
and after status epilepticus that more quickly
induces spontaneous seizures after a latent period
have provided some straightforward lessons about
the pitfalls of simple approaches in complex sys-
tems. For example, efforts to characterize the
functional effects of sprouting anticipating simple
associations with the presence or absence of
recurrent excitatory or inhibitory circuits formed
by sprouted mossy fibers which ignore independ-
ently evolving activity-dependent plasticity in
excitatory and inhibitory synaptic transmission
are subject to significant misinterpretation (Harvey
and Sloviter, 2005; Sloviter et al., 2006). At both
early and advanced stages of reorganization with
minimal or extensive sprouting, recurrent excita-
tion and seizures are emergent events in a complex
system. How recurrent excitatory circuits formed
by mossy fiber sprouting contribute to recurrent
excitation needs to be considered from the
perspective of complex systems.
Epilepsy as a ‘‘complex systems’’ disorder
There is increasing awareness that biological sys-
tems often behave as ‘‘complex systems’’, and that
neural circuitry of the brain and dentate gyrus
fulfill criteria of a ‘‘complex system’’ (Koch and
Laurent, 1999). As the circuits formed by sprout-
ing are but one component of a complex system of
molecular and cellular pathways in the reorganized
dentate gyrus, when and how these new circuits are
activated and contribute to network function can
be anticipated to be governed by principles of
‘‘complex systems’’. The phenomena of complex
systems are well-known in engineering design, but
have received attention in biology only relatively
recently (Gallagher and Appenzeller, 1999). Com-
plex systems are difficult to fully explain through
reductionistic understanding of their component
parts, and typically demonstrate functional prop-
erties that are context dependent and cannot
be readily predicted from isolated analysis of
component parts. The functional properties of
complex systems typically include emergent events
which are context dependent.
The heterogeneous etiologies of epilepsy and the
diverse variety of underlying molecular and cellu-
lar mechanisms are consistent with the view that
epilepsy is a ‘‘complex systems’’ disorder which
cannot be fully explained by simple understanding
of one or perhaps even a few processes or com-
ponents. Multiple molecular and physiological
mechanisms work alone or together to promote
epileptogenesis (Dudek et al., 2002; Sutula, 2002;
Dudek and Shao, 2004), as demonstrated by the
rapidly growing list of transgenic mice demon-
strating epilepsy. An example is the interaction
between cellular mechanisms studied in an isolated
neuron, and the profound functional alterations
that may emerge when neurons with subtle abnor-
malities are components of a neural network.
Alterations that appear subtle at one level may
produce cumulative and profound alterations
manifesting as emergent properties at another
level, such as recurrent excitation and behavioral
seizures.
Emergent properties in complex systems may
not be associated with continuously detectable
abnormalities (Gallagher and Appenzeller, 1999;
Sole and Goodwin, 2000). Application of reduc-
tionistic approaches to ‘‘complex systems’’ often
result in causal ‘‘gaps’’ between one level of
understanding and the next (Sole and Goodwin,
2000). Phenomena at one level cannot be viewed in
isolation, but need to be considered together with
other processes at the same and at other levels.
These features apply to the interactions of inhibi-
tion and excitation generated at both synaptic and
network levels in the circuitry of the normal and
reorganized dentate gyrus and hippocampus.
Observations that spontaneous seizures are
infrequent even in intractable epilepsy and func-
tional abnormalities that become unmasked by
disinhibition or extracellular ionic alterations in
hippocampal slices from epileptic rats and human
epileptic temporal lobe fulfill the defining criteria
for emergent properties of a complex system. Un-
derstanding of constituent parts of a complex sys-
tem, for example the functional properties
generated by the sprouted mossy fiber pathway,

requires not only precise knowledge of the parts,
but also the context in which the part operates.
These principles are highly relevant to the inter-
play of activity-dependent alterations in inhibition,
unmasking of recurrent excitation, generation
of episodic network synchronization, and recur-
ring behavioral seizures which are the defining
features of epilepsy and are prototypical emergent
events in the complex system of neural circuitry in
the dentate gyrus.
Recurrent excitation in the dentate gyrus
reorganized by sprouting: an emergent property of a
complex system
While the heterogeneity and complexity of epilepsy
at the systems biology level is obvious, studies of
potential underlying mechanisms for seizures and
chronic epilepsies frequently employ experimental
designs which anticipate linear relationships be-
tween a given alteration such as an ion channel
mutation with altered biophysical properties and
emergent properties such as network synchroniza-
tion and behavioral seizures. At the circuit level,
numerous past studies of specific molecular or cel-
lular alterations associated with epilepsy have been
pursued as possible mechanistic ‘‘causes’’ of the
disorder, and then dismissed when it is recognized
that sporadically expressed emergent phenomena,
such as network synchronization or behavioral
seizures, are not universally or linearly related to
the specific defect (Dudek, 2002). This interpretive
flaw is common in epilepsy research, and has been
contributed to skepticism about the function and
importance of mossy fiber sprouting (Armitage
et al., 1998; Longo and Mello, 1998, 1999;
Timofeeva and Peterson, 1999; Mohapel et al.,
2000; Nissinen et al., 2001; Romcy-Pereira and
Garcia-Cairasco, 2003; Sloviter et al., 2006).
Experiments seeking to define the functional
effects of mossy sprouting in the complex system
of the reorganized dentate gyrus must include
design and interpretive perspectives recognizing
that sprouting is but one alteration among the
many molecular and cellular alterations in epilep-
tic neural circuitry. Characterization of its effects
will require more than correlational assessment of
557
time dependent dynamic changes in neural circuits,
but in addition manipulation of multiple variables
to isolate context dependent processes. This level
of complexity demands detailed quantitative char-
acterization of the structural features of sprouting
and reorganized circuitry, physiological assess-
ment using multiple techniques, and experimental
designs which recognize evolving and time
dependent alterations in neurons and circuits sub-
ject to activity dependent modification across
intervals spanning milliseconds to months or
more. Analysis of relationships between structural
alterations such as mossy fiber sprouting in the
inner molecular layer of the dentate gyrus and
context dependent emergent events such as recur-
rent excitation will be flawed if analysis is pursued
without appreciation of regional variations and
critical quantitative features such as numbers
of excitatory and inhibitory synapses and targets.
Attempting to define functional effects of sprout-
ing based on simple association, correlation,
or with anticipation of linear relationships will
invariably overlook important context depend-
ent emergent properties and will be subject to
misinterpretation and flawed perspectives.
The preceding sections have reviewed some of
the anatomical and physiological considerations
that need to be addressed in order to characterize
the functional effects of sprouting in the reorgan-
ized dentate gyrus. Recurrent excitatory circuits
between granule cells are the predominant recur-
rent connection formed by sprouted mossy fiber
axons in the inner molecular layer of the reorgan-
ized dentate gyrus. The available quantitative
characterizations of the synaptic connections
of the sprouted mossy fiber pathway, assessment
of the functional features of sprouted circuitry
using multiple physiological measures and exper-
imental designs, and the perspective of complex
systems analysis of neural circuits strongly support
the conclusion that mossy fiber sprouting induced
in the inner molecular layer by seizures or injury
forms predominantly recurrent excitatory circuits
in this region whose functional effects emerge only
conditionally and intermittently. The episodic
emergence of recurrent excitation in this region
of circuitry that is commonly involved in human
focal and limbic epilepsy is a potentially important
558
functional property of circuitry reorganized by
primary pathologies and recurring seizures,
but should not be anticipated to be a necessary
condition or absolute requirement for network
synchronization underlying the heterogeneous
variety of epileptic syndromes.
Recurrent excitation as a functional effect of mossy
fiber sprouting: experimental challenges and
therapeutic opportunities
The dentate gyrus and hippocampus are promi-
nently involved in the most common form of
intractable human epilepsy. The recognition
of mossy fiber sprouting in experimental models
and in human epileptic temporal lobe and appre-
ciation of its capacity to episodically generate
recurrent excitation has been a milestone for
epilepsy research and a major influence on under-
standing of fundamental aspects of the epilepsies.
The perspective that mossy fiber sprouting con-
tributes conditionally to emergence of recurrent
excitation provides a conceptual framework
for understanding how injury and seizure-induced
circuit reorganization may contribute to paroxys-
mal network synchronization, epileptogenesis, and
the consequences of repeated seizures. The accom-
plishments of nearly two decades of investigation
on mossy fiber sprouting set the stage for efforts
to modify seizure-induced sprouting and processes
of plasticity in the reorganizing dentate gyrus as
a major translational and therapeutic opportunity
for epilepsy research.
References
Acsady, L., Kamondi, A., Sik, A., Freund, T. and Buzsaki, G.
(1998) GABAergic cells are the major postsynaptic targets of
mossy fibers in the rat hippocampus. J. Neurosci., 18:
3386–3403.
Adams, B., Lee, M., Fahnestock, M. and Racine, R.J. (1997)
Long-term potentiation trains induce mossy fiber sprouting.
Brain Res., 775: 193–197.
Amaral, D.G. (1979) Synaptic extensions from the mossy fibers
of the fascia dentata. Anat. Embryol., 155: 241–251.
Amaral, D.G. and Dent, J.A. (1981) Development of the mossy
fibers of the dentate gyrus: I. A light and electron microscopic
study of the mossy fibers and their expansions. J. Comp.
Neurol., 195: 51–86.
Anderson, A.E., Hrachovy, R.A., Antalffy, B.A., Armstrong,
D.L. and Swann, J.W. (1999) A chronic focal epilepsy with
mossy fiber sprouting follows recurrent seizures induced by
intrahippocampal tetanus toxin injection in infant rats. Ne-
uroscience, 92: 73–82.
Ang, C.W., Carlson, G.C. and Coulter, D.A. (2006) Massive
and specific dysregulation of direct cortical input to the hip-
pocampus in temporal lobe epilepsy. J. Neurosci., 26:
11850–11856.
Armitage, L.L., Mohapel, P., Jenkins, E.M., Hannesson, D.K.
and Corcoran, M.E. (1998) Dissociation between mossy fiber
sprouting and rapid kindling with low-frequency stimulation
of the amygdala. Brain Res., 781: 37–44.
Austin, J.E. and Buckmaster, P.S. (2004) Recurrent excitation
of granule cells with basal dendrites and low interneuron
density and inhibitory postsynaptic current frequency in the
dentate gyrus of macaque monkeys. J. Comp. Neurol., 476:
205–218.
Behr, J., Heinemann, U. and Mody, I. (2001) Kindling induces
transient NMDA receptor-mediated facilitation of high-fre-
quency input in the rat dentate gyrus. J. Neurophysiol., 85:
2195–2202.
Benardo, L.S. (2002) Evidence against a pathogenic role for
mossy fiber sprouting. Epilepsy Curr., 2: 96–97.
Ben-Ari, Y. (1985) Limbic seizure and brain damage produced
by kainic acid: mechanisms and relevance to human temporal
lobe epilepsy. Neuroscience, 14(2): 375–403.
Ben-Ari, Y., Tremblay, E., Ottersen, O.P. and Meldrum, B.S.
(1980) The role of epileptic activity in hippocampal and ‘‘re-
mote’’ cerebral lesions induced by kainic acid. Brain Res.,
191: 79–97.
Ben-Ari, Y., Tremblay, E., Ottersen, O.P. and Naquet, R.
(1979) Evidence suggesting secondary epileptogenic lesions
after kainic acid: pre-treatment with diazepam reduces dis-
tant but not local damage. Brain Res., 165: 362–365.
Bender, R.A., Dube, C., Gonzalez-Vega, R., Mina, E.W. and
Baram, T.Z. (2003) Mossy fiber plasticity and enhanced hip-
pocampal excitability, without hippocampal cell loss or al-
tered neurogenesis, in an animal model of prolonged febrile
seizures. Hippocampus, 13: 399–412.
Bengzon, J., Kokaia, Z., Elmer, E., Nanobashvili, A., Ko-
kaia, M. and Lindvall, O. (1997) Apoptosis and prolifer-
ation of dentate gyrus neurons after single and intermittent
limbic seizures. Proc. Natl. Acad. Sci. U.S.A., 94:
10432–10437.
Bonthius, D.J., Lothman, E.W. and Steward, O. (1995) The
role of extracellular ionic changes in upregulating the mRNA
for glial fibrillary acidic protein following spreading depres-
sion. Brain Res., 674: 314–328.
Brooks-Kayal, A.R., Shumate, M.D., Jin, H., Rikhter, T.Y.
and Coulter, D.A. (1998) Selective changes in single cell
GABA(A) receptor subunit expression and function in tem-
poral lobe epilepsy. Nat. Med., 4: 1166–1172.
Buckmaster, P.S. and Dudek, F.E. (1997a) Network properties
of the dentate gyrus in epileptic rats with hilar neuron loss
and granule cell axon reorganization. J. Neurophysiol., 77:
2685–2696.

Buckmaster, P.S. and Dudek, F.E. (1997b) Neuron loss, gran-
ule cell axon reorganization, and functional changes in the
dentate gyrus of epileptic kainate-treated rats. J. Comp.
Neurol., 385: 385–404.
Buckmaster, P.S. and Dudek, F.E. (1999) In vivo intracellular
analysis of granule cell axon reorganization in epileptic rats.
J. Neurophysiol., 81: 712–721.
Buckmaster, P.S., Zhang, G.F. and Yamawaki, R. (2002) Axon
sprouting in a model of temporal lobe epilepsy creates a pre-
dominantly excitatory feedback circuit. J. Neurosci.,
22: 6650–6658.
Buhl, E.H., Otis, T.S. and Mody, I. (1996) Zinc-induced col-
lapse of augmented inhibition by GABA in a temporal lobe
epilepsy model. Science, 271: 369–373.
Cavazos, J.E., Das, I. and Sutula, T.P. (1994) Neuronal loss
induced in limbic pathways by kindling: evidence for induc-
tion of hippocampal sclerosis by repeated brief seizures. J.
Neurosci., 14: 3106–3121.
Cavazos, J.E., Golarai, G. and Sutula, T.P. (1991) Mossy fiber
synaptic reorganization induced by kindling: time course
of development, progression, and permanence. J. Neurosci.,
11: 2795–2803.
Cavazos, J.E., Golarai, G. and Sutula, T.P. (1992) Septotem-
poral variation of the supragranular projection of the mossy
fiber pathway in the dentate gyrus of normal and kindled
rats. Hippocampus, 2: 363–372.
Cavazos, J.E., Jones, S.M. and Cross, D.J. (2004) Sprouting
and synaptic reorganization in the subiculum and CA1 region
of the hippocampus in acute and chronic models of partial-
onset epilepsy. Neuroscience, 126: 677–688.
Cavazos, J.E. and Sutula, T.P. (1990) Progressive neuronal loss
induced by kindling: a possible mechanism for mossy fiber
synaptic reorganization and hippocampal sclerosis. Brain
Res., 527: 1–6.
Cavazos, J.E., Zhang, P., Qazi, R. and Sutula, T.P. (2003)
Ultrastructural features of sprouted mossy fiber synapses
in kindled and kainic acid-treated rats. J. Comp. Neurol.,
458: 272–292.
Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1986) A light
and electron microscopic analysis of the mossy fibers of the
rat dentate gyrus. J. Comp. Neurol., 246: 435–458.
Cohen, A.S., Lin, D.D., Quirk, G.L. and Coulter, D.A. (2003)
Dentate granule cell GABA(A) receptors in epileptic hippo-
campus: enhanced synaptic efficacy and altered pharmacol-
ogy. Eur. J. Neurosci., 17: 1607–1616.
Cossart, R., Bernard, C. and Ben-Ari, Y. (2005) Multiple facets
of GABAergic neurons and synapses: multiple fates
of GABA signalling in epilepsies. Trends Neurosci.,
28: 108–115.
Cossart, R., Dinocourt, C., Hirsch, J.C., Merchan-Perez, A.,
De Felipe, J., Ben-Ari, Y., Esclapez, M. and Bernard, C.
(2001) Dendritic but not somatic GABAergic inhibition
is decreased in experimental epilepsy. Nat. Neurosci.,
4: 52–62.
Coulter, D.A. (2001) Epilepsy-associated plasticity in gamma-
aminobutyric acid receptor expression, function, and inhib-
itory synaptic properties. Int. Rev. Neurobiol., 45: 237–252.
559
Coulter, D.A. (2002) Sprouted mossy fibers form primarily
excitatory connections. Epilepsy Curr., 2: 194–195.
Cronin, J. and Dudek, F.E. (1988) Chronic seizures and col-
lateral sprouting of dentate mossy fibers after kainic acid
treatment in rats. Brain Res., 474: 181–184.
Cronin, J., Obenaus, A., Houser, C.R. and Dudek, F.E. (1992)
Electrophysiology of dentate granule cells after kainate-
induced synaptic reorganization of the mossy fibers. Brain
Res., 573: 305–310.
Cross, D.J. and Cavazos, J.E. (2007) Synaptic reorganization
in subiculum and CA3 after early-life status epilepticus
in the kainic acid rat model. Epilepsy Res.,
73(2): 156–165.
Dalby, N.O., West, M. and Finsen, B. (1998) Hilar somatosta-
tin-mRNA containing neurons are preserved after perforant
path kindling in the rat. Neurosci. Lett., 255: 45–48.
Dasheiff, R.M. and McNamara, J.O. (1982) Intradentate col-
chicine retards the development of amygdala kindling. Ann.
Neurol., 11: 347–352.
Dolorfo, C.L. and Amaral, D.G. (1998) Entorhinal cortex of
the rat: topographic organization of the cells of origin of the
perforant path projection to the dentate gyrus. J. Comp.
Neurol., 398: 25–48.
Dudek, F.E. and Shao, L.R. (2004) Mossy fiber sprouting and
recurrent excitation: direct electrophysiologic evidence and
potential implications. Epilepsy Curr., 4: 184–187.
Dudek, F.E., Staley, K.J. and Sutula, T.P. (2002) The search
for animal models of epileptogenesis and pharmacoresist-
ance: are there biologic barriers to simple validation strate-
gies? Epilepsia, 43: 1275–1277.
Dudek, F.E., Wuarin, J.P., Tasker, J.G., Kim, Y.I. and Pea-
cock, W.J. (1995) Neurophysiology of neocortical slices
resected from children undergoing surgical treatment for
epilepsy. J. Neurosci. Methods, 59: 49–58.
Dzhala, V.I. and Staley, K.J. (2003) Excitatory actions of
endogenously released GABA contribute to initiation of ictal
epileptiform activity in the developing hippocampus. J.
Neurosci., 23: 1840–1846.
El Bahh, B., Lespinet, V., Lurton, D., Coussemacq, M., Le Gal
La Salle, G. and Rougier, A. (1999) Correlations between
granule cell dispersion, mossy fiber sprouting, and hippo-
campal cell loss in temporal lobe epilepsy. Epilepsia, 40:
1393–1401.
Epsztein, J., Represa, A., Jorquera, I., Ben-Ari, Y. and Crepel,
V. (2005) Recurrent mossy fibers establish aberrant kainate
receptor-operated synapses on granule cells from epileptic
rats. J. Neurosci., 25: 8229–8239.
Esclapez, M., Hirsch, J.C., Ben-Ari, Y. and Bernard, C. (1999)
Newly formed excitatory pathways provide a substrate for
hyperexcitability in experimental temporal lobe epilepsy. J.
Comp. Neurol., 408: 449–460.
Evans, M.S., Cady, C.J., Disney, K.E., Yang, L. and Lag-
uardia, J.J. (2006) Three brief epileptic seizures reduce in-
hibitory synaptic currents, GABA(A) currents, and
GABA(A)-receptor subunits. Epilepsia, 47: 1655–1664.
Franck, J.E., Pokorny, J., Kunkel, D.D. and Schwartzkroin,
P.A. (1995) Physiologic and morphologic characteristics of
560
granule cell circuitry in human epileptic hippocampus. Epile-
psia, 36: 543–558.
Fricke, R.A. and Prince, D.A. (1984) Electrophysiology of
dentate gyrus granule cells. J. Neurophysiol., 51: 195–209.
Frotscher, M. and Zimmer, J. (1983) Lesion-induced mossy
fibers to the molecular layer of the rat fascia dentata: iden-
tification of postsynaptic granule cells by the Golgi-EM
technique. J. Comp. Neurol., 215: 299–311.
Gallagher, R. and Appenzeller, T. (1999) Beyond reductionism.
Science, 284: 79.
Gavrilovici, C., D’Alfonso, S., Dann, M. and Poulter, M.O.
(2006) Kindling-induced alterations in GABA(A) receptor-
mediated inhibition and neurosteroid activity in the rat
piriform cortex. Eur. J. Neurosci., 24: 1373–1384.
Geinisman, Y., Morrell, F. and deToledo-Morrell, L. (1992)
Increase in the number of axospinous synapses with
segmented postsynaptic densities following hippocampal
kindling. Brain Res., 569: 341–347.
Gibbs III, J.W., Shumate, M.D. and Coulter, D.A. (1997)
Differential epilepsy-associated alterations in postsynaptic
GABA(A) receptor function in dentate granule and CA1
neurons. J. Neurophysiol., 77: 1924–1938.
Golarai, G., Cavazos, J.E. and Sutula, T.P. (1992) Activation of
the dentate gyrus by pentylenetetrazol evoked seizures
induces mossy fiber synaptic reorganization. Brain Res.,
593: 257–264.
Golarai, G. and Sutula, T.P. (1996) Functional alterations in
the dentate gyrus after induction of long-term potentiation,
kindling, and mossy fiber sprouting. J. Neurophysiol.,
75: 343–353.
Gombos, Z., Spiller, A., Cottrell, G.A., Racine, R.J. and
McIntyre Burnham, W. (1999) Mossy fiber sprouting induced
by repeated electroconvulsive shock seizures. Brain Res.,
844: 28–33.
Gonzales, R.B., DeLeon Galvan, C.J., Rangel, Y.M. and Clai-
borne, B.J. (2001) Distribution of thorny excrescences on
CA3 pyramidal neurons in the rat hippocampus. J. Comp.
Neurol., 430: 357–368.
Gorter, J.A., van Vliet, E.A., Aronica, E. and Lopes da Silva,
F.H. (2002) Long-lasting increased excitability differs in
dentate gyrus vs. CA1 in freely moving chronic epileptic rats
after electrically induced status epilepticus. Hippocampus,
12: 311–324.
Gutierrez, R. and Heinemann, U. (2001) Kindling induces
transient fast inhibition in the dentate gyrus — CA3 projec-
tion. Eur. J. Neurosci., 13: 1371–1379.
Haas, K.Z., Sperber, E.F., Opanashuk, L.A., Stanton, P.K. and
Moshe, S.L. (2001) Resistance of immature hippocampus
to morphologic and physiologic alterations following status
epilepticus or kindling. Hippocampus, 11: 615–625.
Hampson, R.E., Simeral, J.D. and Deadwyler, S.A. (1999)
Distribution of spatial and nonspatial information in dorsal
hippocampus. Nature, 402: 610–614.
Hansen, A., Jorgensen, O.S., Bolwig, T.G. and Barry, D.I.
(1990) Hippocampal kindling alters the concentration of glial
fibrillary acidic protein and other marker proteins in rat
brain. Brain Res., 531: 307–311.
Hardison, J.L., Okazaki, M.M. and Nadler, J.V. (2000) Mod-
est increase in extracellular potassium unmasks effect
of recurrent mossy fiber growth. J. Neurophysiol.,
84: 2380–2389.
Harvey, B.D. and Sloviter, R.S. (2005) Hippocampal granule
cell activity and c-Fos expression during spontaneous sei-
zures in awake, chronically epileptic, pilocarpine-treated rats:
implications for hippocampal epileptogenesis. J. Comp.
Neurol., 488: 442–463.
Hellier, J.L., Patrylo, P.R., Dou, P., Nett, M., Rose, G.M. and
Dudek, F.E. (1999) Assessment of inhibition and epilepti-
form activity in the septal dentate gyrus of freely behaving
rats during the first week after kainate treatment. J. Neuro-
sci., 19: 10053–10064.
Holahan, M.R., Rekart, J.L., Sandoval, J. and Routtenberg, A.
(2006) Spatial learning induces presynaptic structural
remodeling in the hippocampal mossy fiber system of two
rat strains. Hippocampus, 16: 560–570.
Holmes, G.L., Gairsa, J.L., Chevassus-Au-Louis, N. and Ben-
Ari, Y. (1998) Consequences of neonatal seizures in the rat:
morphological and behavioral effects. Ann. Neurol.,
44: 845–857.
Holmes, G.L., Sarkisian, M., Ben-Ari, Y. and Chevassus-Au-
Louis, N. (1999) Mossy fiber sprouting after recurrent sei-
zures during early development in rats. J. Comp. Neurol.,
404: 537–553.
Houser, C.R. and Esclapez, M. (1996) Vulnerability and plas-
ticity of the GABA system in the pilocarpine model of spon-
taneous recurrent seizures. Epilepsy Res., 26: 207–218.
Houser, C.R., Miyashiro, J.E., Swartz, B.E., Walsh, G.O.,
Rich, J.R. and Delgado-Escueta, A.V. (1990) Altered
patterns of dynorphin immunoreactivity suggest mossy fiber
reorganization in human hippocampal epilepsy. J. Neurosci.,
10: 267–282.
Jiang, M., Lee, C.L., Smith, K.L. and Swann, J.W. (1998) Spine
loss and other persistent alterations of hippocampal pyram-
idal cell dendrites in a model of early-onset epilepsy. J.
Neurosci., 18: 8356–8368.
de Jonge, M. and Racine, R.J. (1987) The development and
decay of kindling-induced increases in paired-pulse depres-
sion in the dentate gyrus. Brain Res., 412: 318–328.
Kapur, J., Bennett Jr., J.P., Wooten, G.F. and Lothman, E.W.
(1989a) Evidence for a chronic loss of inhibition in the hip-
pocampus after kindling: biochemical studies. Epilepsy Res.,
4: 100–108.
Kapur, J. and Macdonald, R.L. (1997) Rapid seizure-induced
reduction of benzodiazepine and Zn2+ sensitivity of hippo-
campal dentate granule cell GABA(A) receptors. J. Neuro-
sci., 17: 7532–7540.
Kapur, J., Michelson, H.B., Buterbaugh, G.G. and Lothman,
E.W. (1989b) Evidence for a chronic loss of inhibition in
the hippocampus after kindling: electrophysiological studies.
Epilepsy Res., 4: 90–99.
Kapur, J., Stringer, J.L. and Lothman, E.W. (1989c) Evidence
that repetitive seizures in the hippocampus cause a lasting
reduction of GABAergic inhibition. J. Neurophysiol.,
61: 417–426.

Khalilov, I., Dzhala, V., Ben-Ari, Y. and Khazipov, R. (1999)
Dual role of GABA in the neonatal rat hippocampus. Dev.
Neurosci., 21: 310–319.
Khurgel, M. and Ivy, G.O. (1996) Astrocytes in kindling:
relevance to epileptogenesis. Epilepsy Res., 26: 163–175.
Kim, Y.I., Peacock, W.J. and Dudek, F.E. (1993) Properties
and synaptic mechanisms of bicuculline-induced epileptiform
bursts in neocortical slices from children with intractable
epilepsy. J. Neurophysiol., 70: 1759–1766.
Koch, C. and Laurent, G. (1999) Complexity and the nervous
system. Science, 284: 96–98.
Kotloski, R., Lynch, M., Lauersdorf, S. and Sutula, T. (2002)
Repeated brief seizures induce progressive hippocampal neu-
ron loss and memory deficits. Prog. Brain Res., 135: 95–110.
Kotti, T., Riekkinen Sr., P.J. and Miettinen, R. (1997) Char-
acterization of target cells for aberrant mossy fiber collaterals
in the dentate gyrus of epileptic rat. Exp. Neurol.,
146: 323–330.
Lehmann, T.N., Gabriel, S., Eilers, A., Njunting, M., Kovacs,
R., Schulze, K., Lanksch, W.R. and Heinemann, U. (2001)
Fluorescent tracer in pilocarpine-treated rats shows wide-
spread aberrant hippocampal neuronal connectivity. Eur. J.
Neurosci., 14: 83–95.
Leinekugel, X., Khalilov, I., McLean, H., Caillard, O., Gaiarsa,
J.L., Ben-Ari, Y. and Khazipov, R. (1999) GABA is the
principal fast-acting excitatory transmitter in the neonatal
brain. Adv. Neurol., 79: 189–201.
Lim, C., Blume, H.W., Madsen, J.R. and Saper, C.B. (1997)
Connections of the hippocampal formation in humans: I. The
mossy fiber pathway. J. Comp. Neurol., 385: 325–351.
Litt, B. and Echauz, J. (2002) Prediction of epileptic seizures.
Lancet Neurol., 1: 22–30.
Litt, B., Esteller, R., Echauz, J., D’Alessandro, M., Shor, R.,
Henry, T., Pennell, P., Epstein, C., Bakay, R., Dichter, M.
and Vachtsevanos, G. (2001) Epileptic seizures may begin
hours in advance of clinical onset: a report of five patients.
Neuron, 30: 51–64.
Longo, B.M. and Mello, L.E. (1998) Supragranular mossy fiber
sprouting is not necessary for spontaneous seizures in the
intrahippocampal kainate model of epilepsy in the rat.
Epilepsy Res., 32: 172–182.
Longo, B.M. and Mello, L.E. (1999) Effect of long-term spon-
taneous recurrent seizures or reinduction of status epilepticus
on the development of supragranular mossy fiber sprouting.
Epilepsy Res., 36: 233–241.
Lynch, M., Sayin, U., Golarai, G. and Sutula, T. (2000)
NMDA receptor-dependent plasticity of granule cell spiking
in the dentate gyrus of normal and epileptic rats. J.
Neurophysiol., 84: 2868–2879.
Lynch, M. and Sutula, T. (2000) Recurrent excitatory connec-
tivity in the dentate gyrus of kindled and kainic acid-treated
rats. J. Neurophysiol., 83: 693–704.
Masukawa, L.M., Higashima, M., Kim, J.H. and Spencer,
D.D. (1989) Epileptiform discharges evoked in hippocampal
brain slices from epileptic patients. Brain Res., 493: 168–174.
Masukawa, L.M., Uruno, K., Sperling, M., O’Connor, M.J.
and Burdette, L.J. (1992) The functional relationship between
561
antidromically evoked field responses of the dentate gyrus
and mossy fiber reorganization in temporal lobe epileptic
patients. Brain Res., 579: 119–127.
Matveeva, E.A., Vanaman, T.C., Whiteheart, S.W. and Slevin,
J.T. (2006) Asymmetric accumulation of hippocampal 7S
SNARE complexes occurs regardless of kindling paradigm.
Epilepsy Res., 73(3): 266–274.
McCarren, M. and Alger, B.E. (1985) Use-dependent depres-
sion of IPSPs in rat hippocampal pyramidal cells in vitro. J.
Neurophysiol., 53: 557–571.
Messenheimer, J.A., Harris, E.W. and Steward, O. (1979)
Sprouting fibers gain access to circuitry transsynaptically al-
tered by kindling. Exp. Neurol., 64: 469–481.
Michelson, H.B., Kapur, J. and Lothman, E.W. (1989) Reduc-
tion of paired pulse inhibition in the CA1 region of the hip-
pocampus by pilocarpine in naive and in amygdala-kindled
rats. Exp. Neurol., 104: 264–271.
Mikkonen, M., Soininen, H., Kalvianen, R., Tapiola, T., Ylin-
en, A., Vapalahti, M., Paljarvi, L. and Pitkanen, A. (1998)
Remodeling of neuronal circuitries in human temporal lobe
epilepsy: increased expression of highly polysialylated neural
cell adhesion molecule in the hippocampus and the ent-
orhinal cortex. Ann. Neurol., 44: 923–934.
Mohapel, P., Armitage, L.L., Gilbert, T.H., Hannesson, D.K.,
Teskey, G.C. and Corcoran, M.E. (2000) Mossy fiber sprout-
ing is dissociated from kindling of generalized seizures in the
guinea-pig. Neuroreport, 11: 2897–2901.
Molnar, P. and Nadler, J.V. (1999) Mossy fiber-granule cell
synapses in the normal and epileptic rat dentate gyrus studied
with minimal laser photostimulation. J. Neurophysiol., 82:
1883–1894.
Nadler, J.V., Perry, B.W. and Cotman, C.W. (1980a) Selective
reinnervation of hippocampal area CA1 and the fascia den-
tata after destruction of CA3-CA4 afferents with kainic acid.
Brain Res., 182: 1–9.
Nadler, J.V., Perry, B.W., Gentry, C. and Cotman, C.W.
(1980b) Degeneration of hippocampal CA3 pyramidal cells
induced by intraventricular kainic acid. J. Comp. Neurol.,
192: 333–359.
Nadler, J.V., Perry, B.W., Gentry, C. and Cotman, C.W. (1981)
Fate of the hippocampal mossy fiber projection after de-
struction of its postsynaptic targets with intraventricular
kainic acid. J. Comp. Neurol., 196: 549–569.
Niquet, J., Jorquera, I., Faissner, A., Ben-Ari, Y. and Represa,
A. (1995) Gliosis and axonal sprouting in the hippocampus
of epileptic rats are associated with an increase of tenascin-C
immunoreactivity. J. Neurocytol., 24: 611–624.
Nissinen, J., Lukasiuk, K. and Pitkanen, A. (2001) Is mossy
fiber sprouting present at the time of the first spontaneous
seizures in rat experimental temporal lobe epilepsy? Hippo-
campus, 11: 299–310.
Nusser, Z., Hajos, N., Somogyi, P. and Mody, I. (1998)
Increased number of synaptic GABA(A) receptors underlies
potentiation at hippocampal inhibitory synapses. Nature,
395: 172–177.
Okazaki, M.M., Evenson, D.A. and Nadler, J.V. (1995) Hip-
pocampal mossy fiber sprouting and synapse formation after
562
status epilepticus in rats: visualization after retrograde trans-
port of biocytin. J. Comp. Neurol., 352: 515–534.
Otis, T.S., De Koninck, Y. and Mody, I. (1994) Lasting
potentiation of inhibition is associated with an increased
number of gamma-aminobutyric acid type A receptors acti-
vated during miniature inhibitory postsynaptic currents.
Proc. Natl. Acad. Sci. U.S.A., 91: 7698–7702.
Parent, J.M., Janumpalli, S., McNamara, J.O. and Lowenstein,
D.H. (1998) Increased dentate granule cell neurogenesis fol-
lowing amygdala kindling in the adult rat. Neurosci. Lett.,
247: 9–12.
Patrylo, P.R. and Dudek, F.E. (1998) Physiological unmasking
of new glutamatergic pathways in the dentate gyrus of hip-
pocampal slices from kainate-induced epileptic rats. J.
Neurophysiol., 79: 418–429.
Patrylo, P.R., Schweitzer, J.S. and Dudek, F.E. (1999) Abnor-
mal responses to perforant path stimulation in the dentate
gyrus of slices from rats with kainate-induced epilepsy and
mossy fiber reorganization. Epilepsy Res., 36: 31–42.
Peng, Z. and Houser, C.R. (2005) Temporal patterns of fos
expression in the dentate gyrus after spontaneous seizures
in a mouse model of temporal lobe epilepsy. J. Neurosci.,
25: 7210–7220.
Peng, Z., Huang, C.S., Stell, B.M., Mody, I. and Houser, C.R.
(2004) Altered expression of the delta subunit of the
GABA(A) receptor in a mouse model of temporal lobe
epilepsy. J. Neurosci., 24: 8629–8639.
Pierce, J.P. and Milner, T.A. (2001) Parallel increases in the
synaptic and surface areas of mossy fiber terminals following
seizure induction. Synapse, 39: 249–256.
Pitkanen, A., Nissinen, J., Lukasiuk, K., Jutila, L., Paljarvi, L.,
Salmenpera, T., Karkola, K., Vapalahti, M. and Ylinen, A.
(2000) Association between the density of mossy fiber sprout-
ing and seizure frequency in experimental and human tem-
poral lobe epilepsy. Epilepsia, 41: S24–S29.
Pretel, S., Applegate, C.D. and Piekut, D. (1997) Apoptotic and
necrotic cell death following kindling induced seizures. Acta
Histochem., 99: 71–79.
Qiao, X. and Noebels, J.L. (1993) Developmental analysis of
hippocampal mossy fiber outgrowth in a mutant mouse with
inherited spike-wave seizures. J. Neurosci., 13: 4622–4635.
Ramirez-Amaya, V., Escobar, M.L., Chao, V. and Bermudez-
Rattoni, F. (1999) Synaptogenesis of mossy fibers induced by
spatial water maze overtraining. Hippocampus, 9: 631–636.
Represa, A. and Ben-Ari, Y. (1992) Kindling is associated with
the formation of novel mossy fibre synapses in the CA3
region. Exp. Brain Res., 92: 69–78.
Represa, A., Le Gall La Salle, G. and Ben-Ari, Y. (1989a)
Hippocampal plasticity in the kindling model of epilepsy in
rats. Neurosci. Lett., 99: 345–350.
Represa, A., Robain, O., Tremblay, E. and Ben-Ari, Y. (1989b)
Hippocampal plasticity in childhood epilepsy. Neurosci.
Lett., 99: 351–355.
Ribak, C.E. and Peterson, G.M. (1991) Intragranular mossy
fibers in rats and gerbils form synapses with the somata
and proximal dendrites of basket cells in the dentate gyrus.
Hippocampus, 1: 355–364.
Ribak, C.E., Seress, L., Weber, P., Epstein, C.M., Henry, T.R.
and Bakay, R.A. (1998) Alumina gel injections into the tem-
poral lobe of rhesus monkeys cause complex partial seizures
and morphological changes found in human temporal lobe
epilepsy. J. Comp. Neurol., 401: 266–290.
Ribak, C.E., Tran, P.H., Spigelman, I., Okazaki, M.M. and
Nadler, J.V. (2000) Status epilepticus-induced hilar basal
dendrites on rodent granule cells contribute to recurrent
excitatory circuitry. J. Comp. Neurol., 428: 240–253.
Romcy-Pereira, R.N. and Garcia-Cairasco, N. (2003) Hippo-
campal cell proliferation and epileptogenesis after audiogenic
kindling are not accompanied by mossy fiber sprouting
or Fluoro-Jade staining. Neuroscience, 119: 533–546.
van Rooyen, F., Young, N.A., Larson, S.E. and Teskey, G.C.
(2006) Hippocampal kindling leads to motor map expansion.
Epilepsia, 47: 1383–1391.
Salin, P., Tseng, G.F., Hoffman, S., Parada, I. and Prince, D.A.
(1995) Axonal sprouting in layer V pyramidal neurons
of chronically injured cerebral cortex. J. Neurosci.,
15: 8234–8245.
Sayin, U., Osting, S., Hagen, J., Rutecki, P. and Sutula, T.
(2003) Spontaneous seizures and loss of axo-axonic and axo-
somatic inhibition induced by repeated brief seizures in kin-
dled rats. J. Neurosci., 23: 2759–2768.
Sayin, U., Rutecki, P. and Sutula, T. (1999) NMDA-dependent
currents in granule cells of the dentate gyrus contribute to
induction but not permanence of kindling. J. Neurophysiol.,
81: 564–574.
Scharfman, H.E., Sollas, A.L., Berger, R.E. and Goodman, J.H.
(2003) Electrophysiological evidence of monosynaptic excita-
tory transmission between granule cells after seizure-induced
mossy fiber sprouting. J. Neurophysiol., 90: 2536–2547.
Scharfman, H.E., Sollas, A.L. and Goodman, J.H. (2002)
Spontaneous recurrent seizures after pilocarpine-induced sta-
tus epilepticus activate calbindin-immunoreactive hilar cells
of the rat dentate gyrus. Neuroscience, 111: 71–81.
Sloviter, R.S., Dichter, M.A., Rachinsky, T.L., Dean, E.,
Goodman, J.H., Sollas, A.L. and Martin, D.L. (1996) Basal
expression and induction of glutamate decarboxylase and
GABA in excitatory granule cells of the rat and monkey
hippocampal dentate gyrus. J. Comp. Neurol., 373: 593–618.
Sloviter, R.S., Zappone, C.A., Harvey, B.D. and Frotscher, M.
(2006) Kainic acid-induced recurrent mossy fiber innervation
of dentate gyrus inhibitory interneurons: possible anatomical
substrate of granule cell hyper-inhibition in chronically
epileptic rats. J. Comp. Neurol., 494: 944–960.
Smith, B.N. and Dudek, F.E. (2001) Short- and long-term
changes in CA1 network excitability after kainate treatment
in rats. J. Neurophysiol., 85: 1–9.
Sole, R. and Goodwin, B. (2000) Signs of Life — How Com-
plexity Invades Biology. Basic Books, New York.
Staley, K.J., Soldo, B.L. and Proctor, W.R. (1995) Ionic mech-
anisms of neuronal excitation by inhibitory GABA(A)
receptors. Science, 269: 977–981.
Stanfield, B.B. (1989) Excessive intra- and supragranular mossy
fibers in the dentate gyrus of tottering (tg/tg) mice. Brain
Res., 480: 294–299.

Staubli, U., Gall, C. and Lynch, G. (1984) The distribution of
the commissural-associational afferents of the dentate gyrus
after perforant path lesions in one-day-old rats. Brain Res.,
292: 156–159.
Steward, O. (1976) Reinnervation of dentate gyrus by homol-
ogous afferents following entorhinal cortical lesions in adult
rats. Science, 194: 426–428.
Steward, O. (1992) Lesion-induced synapse reorganization in the
hippocampus of cats: sprouting of entorhinal, commissural/
associational, and mossy fiber projections after unilateral
entorhinal cortex lesions, with comments on the normal
organization of these pathways. Hippocampus, 2: 247–268.
Steward, O., Cotman, C. and Lynch, G. (1976) A quantitative
autoradiographic and electrophysiological study of the rein-
nervation of the dentate gyrus by the contralateral entorhinal
cortex following ipsilateral entorhinal lesions. Brain Res.,
114: 181–200.
Steward, O., Cotman, C.W. and Lynch, G.S. (1973) Re-estab-
lishment of electrophysiologically functional entorhinal cor-
tical input to the dentate gyrus deafferented by ipsilateral
entorhinal lesions: innervation by the contralateral ent-
orhinal cortex. Exp. Brain Res., 18: 396–414.
Steward, O., Cotman, C.W. and Lynch, G.S. (1974) Growth
of a new fiber projection in the brain of adult rats: re-inner-
vation of the dentate gyrus by the contralateral entorhinal
cortex following ipsilateral entorhinal lesions. Exp. Brain
Res., 20: 45–66.
Steward, O. and Messenheimer, J.A. (1978) Histochemical
evidence for a post-lesion reorganization of cholinergic affer-
ents in the hippocampal formation of the mature cat. J.
Comp. Neurol., 178: 697–709.
Stringer, J.L. (1996) Repeated seizures increase GFAP and
vimentin in the hippocampus. Brain Res., 717: 147–153.
Stringer, J.L. and Lothman, E.W. (1989) Repetitive seizures
cause an increase in paired-pulse inhibition in the dentate
gyrus. Neurosci. Lett., 105: 91–95.
Sur, M. and Rubenstein, J.L. (2005) Patterning and plasticity of
the cerebral cortex. Science, 310: 805–810.
Sutula, T. (2002) Seizure-induced axonal sprouting: assessing
connections between injury, local circuits, and epileptogen-
esis. Epilepsy Curr., 2: 86–91.
Sutula, T., Cascino, G., Cavazos, J., Parada, I. and Ramirez, L.
(1989) Mossy fiber synaptic reorganization in the epileptic
human temporal lobe. Ann. Neurol., 26: 321–330.
Sutula, T., Harrison, C. and Steward, O. (1986) Chronic epile-
ptogenesis induced by kindling of the entorhinal cortex: the
role of the dentate gyrus. Brain Res., 385: 291–299.
Sutula, T., He, X.X., Cavazos, J. and Scott, G. (1988) Synaptic
reorganization in the hippocampus induced by abnormal
functional activity. Science, 239: 1147–1150.
Sutula, T., Zhang, P., Lynch, M., Sayin, U., Golarai, G. and
Rod, R. (1998) Synaptic and axonal remodeling of mossy
fibers in the hilus and supragranular region of the dentate
563
gyrus in kainate-treated rats. J. Comp. Neurol., 390:
578–594.
Tauck, D.L. and Nadler, J.V. (1985) Evidence of functional
mossy fiber sprouting in hippocampal formation of kainic
acid-treated rats. J. Neurosci., 5: 1016–1022.
Timofeeva, O.A. and Peterson, G.M. (1999) Dissociation of
mossy fiber sprouting and electrically-induced seizure sensi-
tivity: rapid kindling versus adaptation. Epilepsy Res., 33:
99–115.
Toyoda, I. and Buckmaster, P.S. (2005) Prolonged infusion
of cycloheximide does not block mossy fiber sprouting in
a model of temporal lobe epilepsy. Epilepsia, 46: 1017–1020.
Tuff, L.P., Racine, R.J. and Adamec, R. (1983a) The effects
of kindling on GABA-mediated inhibition in the dentate
gyrus of the rat. I. Paired-pulse depression. Brain Res.,
277: 79–90.
Tuff, L.P., Racine, R.J. and Mishra, R.K. (1983b) The effects
of kindling on GABA-mediated inhibition in the dentate
gyrus of the rat. II. Receptor binding. Brain Res., 277: 91–98.
Vaidya, V.A., Siuciak, J.A., Du, F. and Duman, R.S. (1999)
Hippocampal mossy fiber sprouting induced by chronic elec-
troconvulsive seizures. Neuroscience, 89: 157–166.
Wenzel, H.J., Woolley, C.S., Robbins, C.A. and Schwartzkroin,
P.A. (2000) Kainic acid-induced mossy fiber sprouting and
synapse formation in the dentate gyrus of rats. Hippocam-
pus, 10: 244–260.
Williams, P.A., Wuarin, J.P., Dou, P., Ferraro, D.J. and
Dudek, F.E. (2002) Reassessment of the effects of cyclohexi-
mide on mossy fiber sprouting and epileptogenesis in the
pilocarpine model of temporal lobe epilepsy. J. Ne-
urophysiol., 88: 2075–2087.
Worrell, G.A., Parish, L., Cranstoun, S.D., Jonas, R., Baltuch,
G. and Litt, B. (2004) High-frequency oscillations and seizure
generation in neocortical epilepsy. Brain, 127: 1496–1506.
Wuarin, J.P. and Dudek, F.E. (1996) Electrographic seizures
and new recurrent excitatory circuits in the dentate gyrus
of hippocampal slices from kainate-treated epileptic rats. J.
Neurosci., 16: 4438–4448.
Wuarin, J.P. and Dudek, F.E. (2001) Excitatory synaptic input
to granule cells increases with time after kainate treatment. J.
Neurophysiol., 85: 1067–1077.
Wuarin, J.P., Kim, Y.I., Cepeda, C., Tasker, J.G., Walsh, J.P.,
Peacock, W.J., Buchwald, N.A. and Dudek, F.E. (1990)
Synaptic transmission in human neocortex removed for
treatment of intractable epilepsy in children. Ann. Neurol.,
28: 503–511.
Zhang, L.X., Smith, M.A., Li, X.L., Weiss, S.R.B. and Post,
R.M. (1998) Apoptosis of hippocampal neurons after
amygdala kindled seizures. Mol. Brain Res., 55: 198–208.
Zhang, N. and Houser, C.R. (1999) Ultrastructural localization
of dynorphin in the dentate gyrus in human temporal lobe
epilepsy: a study of reorganized mossy fiber synapses. J.
Comp. Neurol., 405: 472–490.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 30

A behavioral analysis of dentate gyrus function

Raymond P. Kesner

University of Utah, Department of Psychology, 380 S. 1530 E., Room 502, Salt Lake City, UT 84121, USA

Abstract: Computational models of the dentate gyrus (DG) have suggested based on anatomical, electro-
physiological, and computer simulation data that the DG plays an important role in learning and memory
by processing and representing spatial information on the basis of conjunctive encoding, pattern sepa-
ration, and encoding of spatial information in conjunction with the CA3. Behavioral evidence supports a
role for the DG in mnemonic processing of spatial information based on the operation of conjunctive
encoding of multiple sensory inputs, pattern separation of spatial (especially metric) information, and
subsequent encoding in cooperation with CA3. A potential role of the DG in mediating processes, such as
recall of sequential information and short-term memory as well as temporal order for remote memory, are
also discussed.

Keywords: conjunctive encoding; spatial pattern separation; dentate gyrus; encoding; retrieval

Introduction Conjunctive encoding

The anatomy and neural circuitry of the dentate
gyrus (DG) has been described in detail in previ-
ous chapters in this volume (Chapter 1, Amaral,
Scharfman and Lavenex). Based on the anatomy
of the DG, its input and output pathways, and the
development of a computational model, Rolls
(1996) and Rolls and Kesner (2006) have sug-
gested that the DG has three major functions in-
cluding conjunctive encoding of multiple sensory
inputs, spatial pattern separation, and facilitation
of encoding of spatial information based on its
outputs to CA3. This is accomplished by a com-
petitive learning network with Hebb-like modifi-
ability to remove redundancy from the inputs and
produce a more orthogonal, sparse, and catego-
rized set of outputs.
Corresponding author. Tel.: +801 581 7430;
Fax: +801 581 5841; E-mail: rpkesner@behsci.utah.edu
It can be shown that the DG receives multiple
sensory inputs including vestibular, olfactory, vis-
ual, auditory, and somatosensory from the peri-
rhinal cortex and lateral entorhinal cortex in
conjunction with spatially organized grid cells
from the medial entorhinal cortex (Hafting et al.,
2005) to represent metric spatial representations.
The perforant path input in the DG can be divided
into a medial and lateral component. The medial
component processes spatial information and the
lateral component processes non-spatial (e.g. ob-
jects, odors) information (Witter et al., 1989a;
Hargreaves et al., 2005). Based on the idea that the
medial perforant path input into the DG mediates
spatial information via activation of NMDA re-
ceptors and the lateral perforant path input into
the DG mediates visual object information via
activation of opioid receptors, the following ex-
periment was conducted. Using a paradigm deve-
loped by Poucet (1989), rats were tested for the

DOI: 10.1016/S0079-6123(07)63030-1 567

568
detection of a novel spatial change and the detec-
tion of a novel visual object change under the in-
fluence of direct infusions of AP5 (an NMDA
antagonist) or naloxone (a m opiate antagonist)
into the DG. The results indicated that naloxone
infusions into the DG disrupted both novelty de-
tection of a spatial location and a visual object,
whereas AP5 infusions into the DG disrupted only
detection of a novel spatial location, but had no
effect on detection of a novel object. In contrast,
infusions of AP5 into the CA1 region disrupts only
the detection of a spatial location change, but not
a visual object change, whereas naloxone into the
CA1 region disrupts only the detection of a visual
object change, but not a spatial location change
(Hunsaker et al., 2007). These data suggest that
the DG uses conjunctive encoding of visual object
and spatial information to provide for a spatial
representation, which I will show below might be
based on metric information.
Spatial pattern separation
It can clearly be demonstrated that single cells
within the hippocampus are activated by most
sensory inputs, including vestibular, olfactory, vis-
ual, auditory, and somatosensory as well as higher
order integration of sensory stimuli (Cohen and
Eichenbaum, 1993). The question of importance is
whether these sensory inputs via conjunctive en-
coding have a memory representation within the
hippocampus. One possible role for the hippo-
campus in processing all sensory information
might be to provide sensory markers to demar-
cate a spatial location, so that the hippocampus
can more efficiently mediate spatial information.
Thus, it is possible that one of the main processing
functions of the hippocampus is to encode and
separate spatial events from each other. This
would ensure that new highly processed sensory
information is organized within the hippocampus
and enhances the possibility of encoding and tem-
porarily remembering one place as separate from
another place. This may be accomplished via pat-
tern separation of event information, so that sim-
ilar spatial events can be separated from each
other and spatial interference is reduced. To the
extent that DG acts to produce separate represen-
tations of different but similar places, it is pre-
dicted that the DG will be especially important
when memories must be formed about similar
places.
Rolls’ model proposes that pattern separation is
facilitated by sparse connections in the mossy-fiber
system, which connects DG granular cells to CA3
pyramidal neurons. The separation of patterns is
accomplished based on the low probability that
any two CA3 neurons will receive mossy fiber in-
put synapses from a similar subset of DG cells.
The mossy fiber inputs to CA3 from DG are sug-
gested to be essential during learning and may in-
fluence which CA3 neurons will fire based on the
distributed activity in the DG. The cells of the DG
are suggested to act as a competitive learning
network with Hebb-like modifiability to reduce
redundancy and produce sparse, orthogonal out-
puts. O’Reilly and McClelland (1994) and Shapiro
and Olton (1994) also suggest that the mossy fiber
connections between the DG and CA3 may sup-
port pattern separation. Rolls (1996) notes addi-
tional characteristics of the mossy fiber projection
system that may promote pattern separation in the
DG-CA3 system and hence, efficient information
storage in CA3. First, mossy fiber synapses are
very large and terminate close to the soma of the
CA3 pyramidal neurons in the pyramidal layer.
Therefore, the mossy fiber synapses will be rela-
tively powerful in activating the CA3 cell. The
projections from layer II of entorhinal cortex to
CA3 terminate in the lacunosum moleculare layer
of the CA3 pyramidal cells, which is much further
from the soma. Second, the firing activity of
granule cells within the DG is sparse (Jung and
McNaughton, 1992) and coupled with the small
number of connections to CA3 cells, which would
produce a sparse signal that may be transformed
into an even sparser signal in CA3. It also has been
demonstrated that the place fields of DG cells
(Mizumori et al., 1990) and specifically granular
cells (Jung and McNaughton, 1992) are small and
highly reliable, which may support the role of DG
in pattern separation. In addition, the mossy fibers
demonstrate non-associative plasticity (Brown
et al., 1989), which may enhance the signal-
to-noise ratio such that the mossy fiber cell would

produce a non-linearly amplified current in the
CA3 cell. Due to this type of plasticity in the DG,
a particular stimulus such as spatial location
would be likely to activate the same population
of CA3 neurons across subsequent presentations,
which may result in economical storage, because
any given CA3 cell could be used in different
memories.
If disruption of DG function results in ineffi-
cient pattern separation, then deficits on spatial
tasks may occur when there is increased overlap or
similarity among distal cues and presumably in-
creased similarity among representations within
the DG. Remembering a specific location in an
8-arm maze, a water maze, or a spatial context in
fear conditioning may be influenced by the degree
of overlap among critical distal spatial cues. Rats
with lesions in DG have been tested on a working
memory version of the radial 8-arm maze. The re-
sults demonstrated that a lesion of the DG resulted
in deficits similar to complete hippocampal lesions
(Walsh et al., 1986; Tilson et al., 1987; McLamb
et al., 1988; Emerich and Walsh, 1989). In addi-
tion, rats with DG lesions were tested on the
Morris water maze task and showed deficits com-
parable to rats with complete hippocampal lesions
when the start location varied on each trial
(Sutherland et al., 1983; Nanry et al., 1989;
Xavier et al., 1999; Jeltsch et al., 2001). Lee and
Kesner (2004a) tested rats with DG lesions on ac-
quisition and retrieval of contextual fear condi-
tioning. Rats with DG lesions showed initial
impairments in freezing behavior during acquisi-
tion of the task but eventually reached the level of
freezing of controls with subsequent testing. When
retrieval of contextual fear was examined in rats
with DG lesions 24 h after acquisition, the animals
showed a significant deficit in freezing compared
to controls. Based on these studies it is clear that
DG lesions impair spatial memory. The DG le-
sioned animals may be impaired as a result of
impaired pattern separation; however, it is difficult
to determine if a particular memory process is
impaired in these animals using these three
paradigms.
To examine the contribution of the DG to spa-
tial pattern separation, Gilbert et al. (2001) tested
rats with DG lesions using a paradigm that
569
measured short-term memory for spatial location
information as a function of spatial similarity bet-
ween two spatial locations (Gilbert et al., 1998).
Specifically, the study was designed to examine the
role of the DG subregion in discriminating spatial
locations when rats were required to remember a
spatial location based on distal environmental cues
and differentiate between the to-be-remembered
location and a distractor location with different
degrees of similarity or overlap among the distal
cues. Animals were tested using a cheeseboard
maze apparatus on a delayed-match-to-sample for
a spatial location task. Animals were trained to
displace an object that was randomly positioned to
cover a baited food well in 1 of 15 locations along
a row of food wells. Following a short delay, the
animals were required to choose between two ob-
jects identical to the sample phase object. One ob-
ject was in the same location as the sample phase
object and the second object was in a different
location along the row of food wells. An animal
was rewarded for displacing the object in the same
spatial location as the sample phase object (correct
choice) but received no reward for displacing the
foil object (incorrect choice). Five spatial separa-
tions, from 15 cm to 105 cm, were used to separate
the correct object and the foil object during the
choice phase. The results showed that rats with
DG lesions were significantly impaired at short
spatial separations; however, the performance of
the DG lesioned animals increased as a function of
increased spatial separation between the correct
object and the foil on the choice phases. The per-
formance of rats with DG lesioned matched con-
trols at the largest spatial separation. The graded
nature of the impairment and the significant linear
improvement in performance as a function of in-
creased separation illustrate the deficit in pattern
separation. Based on these results, it was con-
cluded that lesions of the DG decrease the effi-
ciency of spatial pattern separation, which resulted
in impairments on trials with increased spatial
proximity and increased spatial similarity among
working memory representations.
Based on the evidence that the hippocampus,
and especially the DG, receives inputs from all
sensory modalities, there is a possibility that the
DG uses sensory markers to demarcate a spatial
570
location, allowing the DG to more efficiently rep-
resent spatial information. Thus, one function of
the DG may be to encode and separate events in
space resulting in spatial pattern separation. Spa-
tial pattern separation would ensure that new
highly processed sensory information is organized
within the hippocampus, which in turn enhances
the possibility of encoding and temporarily re-
membering one spatial location as separate from
another location. Based on the observations that
cells in the CA3 and CA1 region respond to
changes in the metric and topological aspects of
the environment (O’Keefe and Burgess, 1996;
Jeffery and Anderson, 2003), one can ask the
question whether these different features of the
spatial environment are processed via the DG and
subsequently transferred to the CA3 subregion or
are these features communicated via the direct
perforant path projection to the CA3 subregion?
In both cases the information may then be trans-
ferred to the CA1 subregion. To answer this ques-
tion, Goodrich-Hunsaker et al. (2005) examined
the contribution of the DG to memory for metric
and topological spatial relationships. Using a
modified version of an exploratory paradigm de-
veloped by Poucet (1989), rats with DG, CA3, and
CA1 lesions and controls were tested on tasks in-
volving a metric spatial manipulation. In this task,
a rat was allowed to explore two different visual
objects that were separated by a specific distance
on a cheeseboard maze. On the initial presentation
of the objects, the rat explored each object. How-
ever, across subsequent presentations of the ob-
jects in the same spatial locations, the rat
habituated and eventually spent less time explor-
ing the objects. Once the rat had habituated to the
objects in their locations, the metric spatial dis-
tance between the two objects was manipulated so
the two objects were either closer together or fur-
ther apart. The time the rat spent exploring each
moved object was recorded. The results showed
that DG lesions impaired detection of the metric
distance change, in that rats with DG lesions spent
significantly less time exploring the two objects
that were displaced relative to controls. Rats with
CA3 or CA1 lesions displayed smaller reductions
in re-exploration and the DG was reliably im-
paired relative to controls and CA1. It is also
possible to examine the detection of a topological
spatial change. In the topological manipulation
condition, rats were allowed to explore four differ-
ent visual objects that were positioned in a square
on the cheeseboard maze. The rats again were al-
lowed to explore the objects and eventually habitu-
ated to the objects with subsequent exposure.
However, following habituation, the locations of
two of the objects were switched and the time the
rat spent exploring each object was recorded. The
results showed that rats with CA1 lesions impair,
whereas DG or CA3, lesions do not impair the
detection of the topological manipulation. The re-
sults suggest that neurons in the DG may be crit-
ically involved in processing spatial information
on a metric scale, but may not be necessary for
representing topological space. The results of both
experiments provide for empirical validation of the
role of DG in spatial pattern separation and sup-
port the predictions of computational models
(Rolls, 1996; Rolls and Kesner, 2006).
There are some studies in the literature that have
demonstrated that hippocampal lesions, including
the DG, can result in deficits for spatial tasks that
can be interpreted to be a function of increased
interference and an impairment in the utilization
of a spatial pattern separation process. A few ex-
amples will suffice. Because rats are started in
different locations in the standard water maze
task, there is potential interference among similar
and overlapping spatial patterns. Thus, the obser-
vation that hippocampal lesioned rats are im-
paired in learning and subsequent consolidation of
important spatial information in this task could be
due to difficulty to separate spatial patterns re-
sulting in enhanced spatial interference. Support
for this idea comes from the observation of
Eichenbaum et al. (1990) who demonstrated that
when fimbria-fornix lesioned rats are trained on
the water maze task from only a single starting
position (less spatial interference) there are hardly
any learning deficits, whereas training from many
different starting points resulted in learning diffi-
culties. In a somewhat similar study, it was shown
that total hippocampal lesioned rats learned or
consolidated rather readily that only one spatial
location was correct on an 8-arm maze (Hunt
et al., 1994). In a different study, McDonald and

White (1995) used a place preference procedure in
an 8-arm maze. In this procedure food is placed at
the end of one arm and no food is placed at
the end of another arm. In a subsequent preference
test normal rats prefer the arm that contained the
food. In this study fornix lesioned rats acquired
the place preference task as quickly as controls if
the arm locations were opposite each other,
but the fornix lesioned rats were markedly im-
paired if the locations were adjacent to each other.
Clearly, there would be greater spatial interference
when the spatial locations are adjacent to each
other rather than far apart. In a different task, rats
were trained on a pair-wise visual spatial discrim-
ination of vertical lines that were not very far
apart. Compared to controls, rats with hippocam-
pal lesions were not able to learn the task (Lazaris
et al., 2006). Thus, spatial pattern separation may
play an important role in the acquisition of new
spatial information and there is a good possibility
that the DG may have been the subregion respon-
sible for the impairments in the various tasks de-
scribed above.
Do other subregions of the hippocampus en-
gage in spatial pattern separation? It appears that
the CA3 region may also engage spatial pattern
separation processes. For example, Tanila (1999)
showed that CA3 place cells were able to maintain
distinct representations of two visually identical
environments and selectively reactivate either one
of the representation patterns depending on the
experience of the rat. Also, Leutgeb et al. (2004)
recently showed that when rats experienced a
completely different environment, CA3 place cells
developed orthogonal representations of those
different environments by changing their firing
rates between the two environments, whereas CA1
place cells maintained similar responses. In a
different study Vazdarjanova and Guzowski
(2004) placed rats in two different environments
separated by 30 min. The two environments
differed greatly in that different objects were lo-
cated in each room. The authors were able to
monitor the time course of activations of ensem-
bles of neurons in both CA3 and CA1, using a
new immediate-early gene-based brain-imaging
method (Arc/H1a catFISH). When the two envi-
ronments were significantly different, CA3
571
neurons exhibited lower overlap in their activity
between the two environments compared to CA1
neurons. This could be consistent with the com-
putational point that if CA3 is an autoassociator,
the pattern representations within it should be as
orthogonal as possible to maximize memory ca-
pacity and minimize interference. The actual
pattern separation may be performed, the com-
putational model holds, as a result of the
operation of the dentate granule cells as a
competitive net and the nature of the mossy
fiber connections to CA3 cells. Thus, CA3 may
represent different environments relatively ortho-
gonally. In the computational account, each
environment would be a separate chart, and the
number of charts that could be stored in CA3
would be high if the representations in each chart
were relatively orthogonal to those in other charts
and further, charts could operate independently
(Stringer et al., 2004). Any one chart of a given
spatial environment can be understood as a con-
tinuous attractor network with place cells with
Gaussian-shaped place fields that overlap contin-
uously with each other. In different charts (differ-
ent spatial environments) the same neurons may
represent very different regions of space, and neu-
rons representing close places in one environment
may represent distant locations in another envi-
ronment (chart). It appears that the DG mediated
pattern separation is based on reducing interfer-
ence between discrete spatial locations based on
metric representations of distance between the
spatial locations, whereas the CA3 mediated pat-
tern separation is based on reducing interference
across global environments based on representa-
tion of unique charts. Whether the actual pattern
separation in CA3 is performed as a result of the
operation of the dentate granule cells as a com-
petitive net and the nature of the mossy fiber
connections to CA3 cells remains to be investi-
gated. An alternative possibility is that pattern
separation observed for the global environment is
separate from pattern separation observed for
discrete elements in the DG. The dissociation bet-
ween these two forms of pattern separation could
be achieved by assuming that the perforant path
input into CA3 is critical for representing a global
environment or context. The perforant path into
572
CA3 can also be divided into a medial and lateral
component. Because there is associative LTP in
CA3 between the medial or lateral perforant path
and the intrinsic commissural/associational-CA3
synapses (Martinez et al., 2002), either place, ob-
ject or contextual cues can be processed by the
associative medial and lateral perforant path con-
nections to CA3 cells. This is consistent with the
finding that disruption of the perforant path input
impairs the multiple trial acquisition of an object-
place paired associate task [perhaps because re-
trieval of what has been previously learned is im-
paired (unpublished observations)].
In addition, Martinez et al. (2002) demonstrated
associative (cooperative) LTP between the medial
and lateral perforant path inputs to the CA3 neu-
rons. This could provide a mechanism for object
(lateral perforant path) — place (medial perforant
path) associative learning in conjunction with a
global context. Further support for this idea is
based on testing rats for the detection of a novel
spatial change and the detection of a novel visual
object change under the influence of direct infu-
sions of AP5 or naloxone into the CA3. The
results indicated that naloxone or AP5 infusions
into CA3 disrupted both novelty detection of a
spatial location and a visual object, suggesting
that the perforant path input into CA3 could pro-
vide the context for integration of both spatial and
visual object information (Hunsaker et al., 2007).
These studies thus indicate that both the DG
and CA3 play an important role in spatial pattern
separation for spatial memory tasks. Other forms
of pattern separation do not involve the DG sub-
region of the hippocampus. For example, temporal
pattern separation is mediated by the CA1, but not
the DG (Gilbert et al., 2001). Furthermore, hippo-
campal lesions including DG do not produce a
deficit for pattern separation of reward values,
visual objects, or motor responses (Gilbert and
Kesner, 2002, 2003b, 2006; Saksida et al., 2006).
Instead, the perirhinal cortex subserves pattern
separation for visual objects (Bussey et al., 2002;
Gilbert and Kesner, 2003), the amygdala subserves
pattern separation for reward value (Gilbert and
Kesner, 2002) and the caudate nucleus subserves
pattern separation for motor responses (Kesner
and Gilbert, 2006).
Encoding vs. retrieval of spatial information
Rolls’ (1996) computational model postulates that
the dentate/mossy fiber system is necessary for
setting up the appropriate conditions for optimal
storage of new information in the CA3 system
(which could be called encoding). In order to test
the model rats with DG or CA3 lesions were ad-
ministered 10 learning trials per day in a Hebb-
Williams maze. The results based on a within-day
analysis indicated that DG and CA3 lesions im-
paired the acquisition of this task, consistent with
an encoding or learning impairment (Lee and
Kesner, 2004b; Jerman et al., 2006). However,
when tested using a between days analysis, re-
trieval of what had been learned previously was
not impaired by DG lesions (Lee and Kesner,
2004b; Jerman et al., 2006). In the Hebb-Williams
learning task, one can measure improvement in
performance within a day, reflecting the operation
of encoding of new information based in part on
short-term memory representations and one can
also measure improvement in performance bet-
ween days, reflecting the operation of retrieval of
information either based on intermediate-term
memory representations mediated by synaptic con-
solidation and/or access to stored information. It
is recognized that the separation of encoding from
retrieval processes is extremely difficult. Therefore,
it is assumed that during acquisition within a day,
there will be a greater involvement of encoding
compared to retrieval processes, and that during
retention across days, there will be a greater in-
volvement of retrieval compared to encoding
processes. It is also assumed that encoding en-
compasses spatial pattern separation processes in
conjunction with associative processes and repre-
sentations within short-term memory. Even
though there is likely to be some retrieval from
short-term memory that may also occur during
acquisition, it is assumed not to be the dominant
factor governing performance within the first 10
trials on Day 1. Furthermore, it is assumed that
retrieval 24 h later encompasses associative proc-
esses as well as representations within intermedi-
ate-term memory. Even though there is likely to be
some encoding that may also occur during re-
trieval, it is assumed not to be the most critical

determinant of performance within the first five
trials on Day 2.
It is therefore of interest that the results indicate
that both DG and CA3 lesions disrupted encod-
ing, but not retrieval, even though CA3 lesions did
produce more errors during acquisition than DG
lesions. The DG lesion data are consistent with the
results of (Lassalle et al., 2000), who showed that
in a water maze learning task, lesions of the mossy
fibers disrupted encoding, but not retrieval. The
CA3 lesion data are consistent with the idea that
short-term encoding of new information is medi-
ated by CA3. Other studies involving alteration of
CA3 function support this idea (Lee and Kesner,
2002, 2003, 2004a; Nakazawa et al., 2003). Thus,
the data suggest a possible cooperation between
the DG and CA3 in encoding of information in a
Hebb-Williams maze, in that there is a deficit in
encoding of information following either a DG or
CA3 lesion.
The present data appear to be in conflict with
the results of Lee and Kesner (2004b), wherein a
lesion to the perforant path into CA3 caused a
deficit in retrieval but no effect for encoding. There
is the possibility that the lesion to the perforant
path may have disrupted the Schaffer collateral
output from CA3 into CA1. Since the Lee and
Kesner paper was published (2004b), Hebb-
Williams maze data for CA1 lesioned animals
from our laboratory has shown that animals with
lesions to dorsal CA1 show no deficit in encoding
but a significant deficit in retrieval relative to a
control group (Vago et al., 2007). If the Schaffer
collateral input into CA1 were disrupted by the
perforant path lesion performed by Lee and
Kesner (2004b), then the effects seen in that study
would be more indicative of a CA1 effect rather
than a CA3 effect. This would be more consistent
with the present results indicating that a lesion to
CA3 does not result in a deficit for retrieval of
information — only encoding.
Given that there was an encoding, but not a re-
trieval deficit for the DG and CA3, could there be
an interaction in terms of the encoding processes
between the DG and the CA3? This cooperation
between DG and CA3 could theoretically derive
from the observation that the DG granule cells
project to CA3 pyramidal neurons via mossy fiber
573
projections forming the primary output of the DG.
Based on characteristics of the mossy fiber system,
Rolls (1996) suggests that pattern separation may
be a function of the DG and its mossy fiber pro-
jections to CA3 and thus may facilitate the encod-
ing of spatial information via an interaction
between the DG and CA3. However, recent stud-
ies have shown that the functions of these two
hippocampal subregions can be dissociated using
behavioral tasks. For example, Gilbert and Kesner
(2003a) demonstrated that DG lesioned animals
were able to learn object-place and odor-place
paired-associate tasks as quickly as controls. How-
ever, rats with CA3 lesions showed significant
learning impairments on both tasks. Furthermore,
a lesion of the perforant path input into the CA3
also disrupted object-place paired associate learning
(unpublished observations). These results could in-
dicate that the CA3, but not the DG, subregion is
involved in associative learning. Furthermore, DG,
but not CA3, lesioned rats produce a deficit in the
acquisition of the standard version water maze
(Sutherland et al., 1983; Nanry et al., 1989; Xavier
et al., 1999; Lassalle et al., 2000; Jeltsch et al., 2001).
The dissociations that arise between DG and
CA3 are primarily due to the CA3 subregion of the
hippocampus having two major inputs with a di-
rect connection from the DG via the mossy fibers
and a direct input from the perforant path that
bypasses the DG. An alternate explanation is pos-
sible since a lesion of the perforant path disrupts
retrieval, but not encoding (Lee and Kesner,
2004b). The differences between DG and CA3 in
the above mentioned tasks could be simply a func-
tion of differential intrinsic processing of similar
spatial information. As an another example
Gilbert et al. (2001) and Gilbert and Kesner
(2006) tested rats with DG or CA3 lesions using
a paradigm that measured one-trial short-term
memory for spatial location information as a
function of spatial similarity between two spatial
locations. The results showed that rats with DG
lesions were significantly impaired at short spatial
separations; however, the performance of the DG
lesioned rats improved as a function of increased
spatial separation between the correct object and
the foil on the choice phases. In contrast, CA3
lesioned rats were impaired for all spatial
574
separations, suggesting a disruption of a short-
term memory process subserved by CA3 as an
intrinsic contribution to the spatial pattern sepa-
ration task. Therefore, there is no guarantee that
any potential cooperation between DG and CA3
in encoding information in the Hebb-Williams
maze is due to the direct connection between these
two regions. To examine this issue, a disconnection
study was carried out using an ipsilateral lesion
(DG and CA3 lesion on one side only) group vs. a
contralateral lesion (DG on one side and CA3 on
the other side) group. It is assumed the right and
left hemispheres operate in parallel. Crossed le-
sions (i.e. unilateral lesions in contralateral hemi-
spheres), therefore, would disrupt communication
within each of the two hemispheres, thus func-
tionally disconnecting the two brain regions. If the
DG and CA3 subregions of the hippocampus in-
teract, then crossed lesioned rats should be mark-
edly impaired relative to animals with lesions on
the same side. If these two regions do not interact,
then crossed lesions would not produce a deficit,
suggesting that each subregion produces a deficit
in encoding for different reasons such as spatial
pattern separation problems vs. associative or
short-term memory problems.
The results indicate that rats with an ipsilateral
lesion of both DG and CA3 do not produce a
deficit in either encoding or retrieval in the Hebb-
Williams maze. In contrast, rats with a DG lesion
on one side of the brain and a CA3 lesion on the
other side of the brain disrupted encoding, but not
retrieval, of information on the Hebb-Williams
maze. Thus, the DG and CA3 interact to support
encoding of spatial information in the present task.
There is always the possibility that following a DG
lesion on one side and a CA3 lesion on the other
side, there would be a significant amount of degen-
eration within CA3 following the DG lesion, so that
there would be a lesion of the CA3 lesion on the
other side in effect creating a bilateral lesion of CA3
and similarly a significant amount of degeneration
within DG following the CA3 lesion in effect cre-
ating a bilateral DG lesion. Against this argument
is that upon histological analysis, there is no clear
loss of the contralateral CA3 following the DG le-
sion and no loss of the contralateral DG following
the CA3 lesion (see Jerman et al., 2005 for data
concerning subregional specificity of CA3 and DG
lesions). Thus, the data suggest that the DG and
CA3 cooperate and interact with each other in the
encoding of new spatial information in the Hebb-
Williams maze without affecting retrieval.
Additional potential functions of the dentate gyrus
Based on the observation that neurogenesis occurs
in the DG and that new DG granule cells can be
formed across time, it has been proposed that the
DG mediates a temporal pattern separation mech-
anism that can generate patterns of episodic mem-
ories within remote memory (Aimone et al., 2006).
There are currently no behavioral data available to
assess the proposal. It should be noted that the
CA1, but not DG, is involved in temporal ordering
of spatial information within an intermediate
memory system, but remote memory has not been
investigated (Kesner et al., 2004).
There are some data that suggest that disruption
of neurogenesis using MAM (an anti-mitotic
agent) impairs the learning of trace eye-blink con-
ditioning, but has no effect on delay eye-blink
conditioning, water maze spatial navigation, or
contextual fear conditioning (Shors, 2002). Fur-
thermore, trace eye-lid conditioning prevents the
loss of newly developed granule cells (Shors, 2004).
These data suggest that the DG may play a role in
temporal processing, but it appears to be based on
a different mechanism than the one proposed by
Aimone et al. (2006).
Based on the observation that the DG has a re-
current network system in which DG granule cells
excite mossy cells which feed back modifiable ex-
citatory connections to the granule cells and that
CA3 pyramidal cells have a feedback connection to
the dentate network (see Scharfman’s Chapter 34,
this volume), Lisman (1999) has suggested that this
recurrent circuit can provide for short-term mem-
ory across seconds and thus can play an important
role in providing accurate recall of sequences.
From a behavioral perspective, there is evidence
that the DG contributes to short-term or working
memory (Xavier et al., 1999; Lee and Kesner,
2003), which could be a function of both the CA3
and DG recurrent circuits. However, there is

currently no behavioral evidence to demonstrate
that the DG contributes to sequence recall.
In summary, based on the development of com-
putational models of DG and behavioral evidence
based on dysfunction of DG, it appears that the
DG mediates mnemonic processing of spatial in-
formation. The processes subserved by DG include
(a) the operation of conjunctive encoding of mul-
tiple sensory inputs, implying an integration of
sensory inputs to determine a spatial representa-
tion, (b) pattern separation of spatial (especially
metric) information, involving the reduction of in-
terference between similar spatial locations, and
(c) subsequent encoding in cooperation with CA3,
suggesting an important role in the acquisition of
spatial information. A potential role of the DG in
recall of sequential information requires more em-
pirical evidence.
References
Aimone, J.B., Wiles, J. and Gage, F.H. (2006) Potential role for
adult neurogenesis in the encoding of time in new memories.
Nat. Neurosci., 9: 723–727.
Brown, T.H., Ganong, A.H., Kairiss, E.W., Keenan, C.L. and
Kelso, S.R. (1989) Long-term potentiation in two synaptic
systems of the hippocampal brain slice. In: Byrne J.H. and
Berry W.O. (Eds.), Neural Models of Plasticity. Academic
Press, San Diego, CA, pp. 266–306.
Bussey, T.J., Saksida, L.M. and Murray, E.A. (2002) Perirhinal
cortex resolves and feature ambiguity in complex discrimi-
nations. Eur. J. Neurosci., 15: 365–374.
Cohen, N.J. and Eichenbaum, H.B. (1993) Memory, Amnesia,
and Hippocampal Function. MIT Press, Cambridge.
Eichenbaum, H., Stewart, C. and Morris, R.G.M. (1990) Hip-
pocampal representation in spatial learning. J. Neurosci., 10:
331–339.
Emerich, D.F. and Walsh, T.J. (1989) Selective working mem-
ory impairments following intradentate injection of col-
chicine: attenuation of the behavioral not the
neuropathological effects by gangliosides GM1 and AGF2.
Physiol. Behav., 45: 93–101.
Gilbert, P.E. and Kesner, R.P. (2002) The amygdala but not the
hippocampus is involved in pattern separation based on re-
ward value. Neurobiol. Learn. Mem., 77: 338–353.
Gilbert, P.E. and Kesner, R.P. (2003a) Localization of function
within the dorsal hippocampus: the role of the dorsal CA3
subregion in paired-associate learning. Behav. Neurosci., 117:
1385–1394.
Gilbert, P.E. and Kesner, R.P. (2003b) Recognition memory
for complex visual discrimination is influenced by stimulus
interference in rodents with perirhinal cortex damage. Learn.
Mem., 10: 525–530.
575
Gilbert, P.E. and Kesner, R.P. (2006) The role of dorsal CA3
hippocampal subregion in spatial working memory and pat-
tern separation. Behav. Brain Res., 169: 142–149.
Gilbert, P.E., Kesner, R.P. and DeCoteau, W.E. (1998) The
role of the hippocampus in mediating spatial pattern sepa-
ration. J. Neurosci., 18: 804–810.
Gilbert, P.E., Kesner, R.P. and Lee, I. (2001) Dissociating
hippocampal subregions: a double dissociation between the
dentate gyrus and CA1. Hippocampus, 11: 626–636.
Goodrich-Hunsaker, N.J., Hunsaker, M.R. and Kesner, R.P.
(2005) Effects of hippocampus sub-regional lesions for metric
and topological spatial information processing. Society for
Neuroscience Abstracts, Number 647.1.
Hafting, T., Fyhn, M., Molden, S., Moser, M.B. and Moser,
E.I. (2005) Microstructure of a spatial map in the entorhinal
cortex. Nature, 436: 801–806.
Hargreaves, E.L., Rao, G., Lee, I. and Knierim, J.J. (2005)
Major dissociation between medial and lateral entorhinal in-
put to dorsal hippocampus. Science, 308: 1792–1794.
Hunsaker, M.R., Mooy, G.G., Swift, J. and Kesner, R.P.
(2007) Dissociations of the role of the medial and lateral
perforant path projections into dorsal DG, CA3 and CA1 for
spatial and nonspatial (visual object) information processing.
Behav. Neuroscience (in press).
Hunt, M.E., Kesner, R.P. and Evans, R.B. (1994) Memory for
spatial location: functional dissociation of entorhinal cortex
and hippocampus. Psychobiology, 22: 186–194.
Jeffery, K.J. and Anderson, M.I. (2003) Dissociation of the
geometric and contextual influences on place cells. Hippo-
campus, 13: 868–872.
Jeltsch, H., Bertrand, F., Lazarus, C. and Cassel, J.-C. (2001)
Cognitive performances and locomotor activity following
dentate granule cell damage in rats: role of lesion extent
and type of memory tested. Neurobiol. Learn. Mem., 76:
81–105.
Jerman, T., Kesner, R.P. and Hunsaker, M.R. (2006) Discon-
nection analysis of CA3 and DG in mediating encoding but
not retrieval in a spatial maze learning task. Learn. Mem., 13:
458–464.
Jerman, T.S., Kesner, R.P., Lee, I. and Berman, R.F. (2005)
Patterns of hippocampal cell loss based on subregional le-
sions of the hippocampus. Brain Res., 1065: 1–7.
Jung, M.W. and McNaughton, B.L. (1992) Spatial selectivity of
unit activity in the hippocampal granular layer. Hippocam-
pus, 3: 165–182.
O’Keefe, J. and Burgess, N. (1996) Geometric determinants of
the place field of hippocampal neurons. Nature, 381:
425–428.
Kesner, R.P. and Gilbert, P.E. (2006) The role of the medial
caudate nucleus, but not the hippocampus, in a matching-to
sample task for a motor response. Eur. J. Neurosci., 23:
1888–1894.
Kesner, R.P., Lee, I. and Gilbert, P. (2004) A behavioral as-
sessment of hippocampal function based on a subregional
analysis. Rev. Neurosci., 15: 333–351.
Lassalle, J.M., Bataille, T. and Halley, H. (2000) Reversible
inactivation of the hippocampal mossy fiber synapses in mice
576
impairs spatial learning, but neither consolidation nor mem-
ory retrieval, in the Morris navigation task. Neurobiol.
Learn. Mem., 73: 243–257.
Lazaris, A., Talpos, J.C., Dias, R., Saksida, L.M. and Bussey, T.J.
(2006) Lesions of the hippocampus impair a pair-wise spatial
discrimination task in a touchscreen-equipped operant box. 5th
Forum of European Neuroscience, Program No. A160.9 Poster
board 390. FENS Forum Abstracts, Vienna, Austria.
Lee, I. and Kesner, R.P. (2002) Differential contribution of
NMDA receptors in hippocampal subregions to spatial
working memory. Nat. Neurosci., 5: 162–168.
Lee, I. and Kesner, R.P. (2003) Differential role of dorsal hip-
pocampus subregions in spatial working memory with short
versus intermediate delay. Behav. Neurosci., 117: 1044–1053.
Lee, I. and Kesner, R.P. (2004a) Differential contributions of
dorsal hippocampal subregions to memory acquisition and
retrieval in contextual fear-conditioning. Hippocampus, 14:
301–310.
Lee, I. and Kesner, R.P. (2004b) Encoding versus retrieval of
spatial memory: double dissociation between the dentate
gyrus and the perforant path inputs into CA3 in the dorsal
hippocampus. Hippocampus, 14: 66–76.
Leutgeb, S., Leutgeb, J.K., Treves, A., Moser, M.B. and Moser,
E.L. (2004) Distinct ensemble codes in hippocampal areas
CA3 and CA1. Science, 305: 1295–1298.
Lisman, J.E. (1999) Relating hippocampal circuitry to function:
recall of memory sequences by reciprocal dentate-CA3 inter-
actions. Neuron, 22: 233–242.
Martinez, C.O., Do, V.H., Martinez Jr., J.L. and Derrick, B.E.
(2002) Associative long-term potentiation (LTP) among ex-
trinsic afferents of the hippocampal CA3 region in vivo.
Brain Res., 940: 86–94.
McDonald, R.J. and White, N.M. (1995) Hippocampal and
nonhippocampal contributions to place learning in rats. Be-
hav. Neurosci., 109: 579–593.
McLamb, R.L., Mundy, W.R. and Tilson, H.A. (1988) Intra-
dentate colchicine disrupts the acquisition and performance
of a working memory task in the radial arm maze. Neuro-
toxicology, 9: 521–528.
Mizumori, S.J.Y., Perez, G.M., Alvarado, M.C., Barnes, C.A.
and McNaughton, B.L. (1990) Reversible inactivation of the
medial septum differentially affects two forms of learning in
rats. Brain Res., 528: 12–20.
Nakazawa, K., Sun, L.D., Quirk, M.C., Rondi-Reig, L., Wil-
son, M.A. and Tonegawa, S. (2003) Hippocampal CA3
NMDA receptors are crucial for memory acquisition of one-
time experience. Neuron, 38: 305–315.
Nanry, K.P., Mundy, W.R. and Tilson, H.A. (1989) Colchicine-
induced alternations of reference memory in rats: role of
spatial versus non-spatial task components. Behav. Brain
Res., 35: 45–53.
Poucet, B. (1989) Object exploration, habituation, and response
to a spatial change in rats following septal or medial frontal
cortical damage. Behav. Neurosci., 103: 1009–1016.
O’Reilly, R.C. and McClelland, J.L. (1994) Hippocampal con-
junctive encoding, storage, and recall: avoiding a trade- off.
Hippocampus, 4: 661–682.
Rolls, E.T. (1996) A theory of hippocampal function in mem-
ory. Hippocampus, 6: 601–620.
Rolls, E.T. and Kesner, R.P. (2006) A computational theory of
hippocampal function, and empirical tests of the theory.
Prog. Neurobiol., 79: 1–48.
Saksida, L.M., Bussey, T.J., Buckmaster, C.A. and Murray,
E.A. (2006) No effect of hippocampal lesions on perirhinal
cortex-dependent feature-ambiguous visual discriminations.
Hippocampus, 16: 421–430.
Shapiro, M.L. and Olton, D.S. (1994) Hippocampal func-
tion and interference. In: Schacter D.L. and Tulving E.
(Eds.), Memory Systems 1994. MIT Press, London,
pp. 141–146.
Shors, T.J. (2002) Neurogenesis may relate to some but not all
types of hippocampal-dependent learning. Hippocampus, 12:
578–584.
Shors, T.J. (2004) Memory traces of trace memories: neuro-
genesis, synaptogenesis and awareness. Trends Neurosci., 27:
250–256.
Stringer, S.M., Rolls, E.T. and Trappenberg, T.P. (2004) Self-
organising continuous attractor networks with multiple ac-
tivity packets, and the representation of space. Neurl. Net-
wk., 17: 5–27.
Sutherland, R.J., Whitshaw, I.Q. and Kolb, B. (1983) A be-
havioural analysis of spatial localization following electro-
lytic, kainate- or colchicine-induced damage to the
hippocampal formation in the rat. Behav. Brain Res., 7:
133–153.
Tanila, H. (1999) Hippocampal place cells can develop distinct
representations of two visually identical environments.
Hippocampus, 9: 235–246.
Tilson, H.A., Rogers, B.S., Crimes, L., Harry, J.G., Peterson,
N.J., Hong, J.S. and Dryer, R.S. (1987) Time-dependent ne-
urobiological effects of colchicine administered directly into
the hippocampus of rats. Brain Res., 408: 163–172.
Vago D.R., Bevan, A. and Kesner, R.P. (2007) The role of the
direct perforant path input to the CA1 subregion of the dor-
sal hippocampus in memory retention and retrieval. Hippo-
campus (in press).
Vazdarjanova, A. and Guzowski, J.F. (2004) Differences in
hippocampal neuronal population responses to modifications
of an environmental context: evidence for distinct, yet com-
plementary, functions of CA3 and CA1 ensembles. J. Neu-
rosci., 24: 6489–6496.
Walsh, T.J., Schulz, D., Tilson, H.A. and Schmechel, D.E.
(1986) Colchicine-induced granule cell loss in rat hippocam-
pus: selective behavioral and histological alterations. Brain
Res., 389: 23–36.
Witter, M.P., Groenewegen, H.J., Lopes da Silva, F.H. and
Lohman, A.H. (1989) Functional organization of the extrin-
sic and intrinsic circuitry of the parahippocampal region.
Prog. Neurobiol., 33: 161–253.
Xavier, G.F., Oliveira-Filho, F.J.B. and Santos, A.M.G.
(1999) Dentate gyrus-selective colchicine lesion and
disruption of performance in spatial tasks: difficulties in
‘‘place strategy’’ because of a lack of flexibility in the use of
environmental cues? Hippocampus, 9: 668–681.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 31

Models, structure, function: the transformation

of cortical signals in the dentate gyrus

László Acsády1, and Szabolcs Káli1,2

1
Institute of Experimental Medicine, Hungarian Academy of Sciences, PO Box 67, 1450 Budapest, Hungary
2
HAS-PPCU-SU Neurobionics Research Group, Budapest, Hungary

Abstract: Our central question is why the hippocampal CA3 region is the only cortical area capable of
forming interference-free representations of complex environmental events (episodes), given that appar-
ently all cortical regions have recurrent excitatory circuits with modifiable synapses, the basic substrate for
autoassociative memory networks. We review evidence for the radical (but classic) view that a unique
transformation of incoming cortical signals by the dentate gyrus and the subsequent faithful transfer of the
resulting code by the mossy fibers are absolutely critical for the appropriate association of memory items by
CA3 and, in general, for hippocampal function.
In particular, at the gate of the hippocampal formation, the dentate gyrus possesses a set of unusual
properties, which selectively evolved for the task of code transformation between cortical afferents and the
hippocampus. These evolutionarily conserved anatomical features enable the dentate gyrus to translate the
noisy signal of the upstream cortical areas into the sparse and specific code of hippocampal formation,
which is indispensable for the efficient storage and recall of multiple, multidimensional memory items.
To achieve this goal the mossy fiber pathway maximally utilizes the opportunity to differentially regulate
its postsynaptic partners. Selective innervation of CA3 pyramidal cells and interneurons by distinct ter-
minal types creates a favorable condition to differentially regulate the short-term and long-term plasticity
and the motility of various mossy terminal types. The utility of this highly dynamic system appears to be the
frequency-dependent fine-tuning of excitation and inhibition evoked by the large and the small mossy
terminals respectively. This will determine exactly which CA3 cell population is active and induces per-
manent modification in the autoassociational network of the CA3 region.

Introduction The learning of arbitrary associations of complex,

The hippocampus, a peculiar cortical structure,
has long been known to be involved in higher or-
der cognitive functions, most notably, memory
formation and spatial navigation (Scoville and
Milner, 1957; O’Keefe and Dostrovsky, 1971). Not
all kinds of memory depend on the hippocampus.
Corresponding author. Tel.: +36-1-210-9413;
Fax: +36-1-210-9412; E-mail: acsady@koki.hu
multimodal items during a single exposure (i.e.,
episodic memory) is irreversibly compromised fol-
lowing hippocampal lesions, but the acquisition of
simple associative pairings (e.g., classical condi-
tioning) and motor learning remain intact (Squire,
1992). Hippocampal-dependent memory traces
can be used flexibly, i.e., they can be activated in
a context different from the one where they were
learned. Clinical studies in humans, as well as a
large number of behavioral and physiological

DOI: 10.1016/S0079-6123(07)63031-3 577

578
experiments in other mammalian species (mainly
rodents) have also implicated the hippocampus in
the formation and flexible use of world-centered
(allocentric) spatial representations (O’Keefe and
Nadel, 1978). It has been argued repeatedly that
these two domains share several important char-
acteristics, and may have fairly similar computa-
tional requirements. In particular, both episodic
memory and spatial navigation may require the
fast storage and subsequent recall of specific con-
junctions of environmental stimuli. It has been
suggested that the bulk of neocortex, which is as-
sumed to be involved in continuously creating and
refining an internal representation of the general
structure of the observed world, may not be well
suited for the rapid, interference-free acquisition of
such specific memory traces (McClelland et al.,
1995). On the other hand, the hippocampus might
be optimized for exactly this operation, and could
thus complement the generic learning capabilities
of the neocortex.
However, the hippocampus constitutes only a
tiny fraction of the cortical areas, and it has a rel-
atively simple structure. It receives input from and
sends information back to multimodal associatio-
nal cortical areas. The question is why neocortex
needs the hippocampal loop to implement rapid
learning of arbitrary complex associations? Why is
it that other cortical areas with multiple cellular
layers cannot do the same job? What is so special
about hippocampus that enables it to establish the
most complex memory traces? In short, what is the
‘‘trick’’ of the hippocampus?
For a long time, the central role for memory
formation was assigned to the hippocampal CA3
region (Marr, 1971; McNaughton and Morris,
1987; Rolls, 1989). This hippocampal subfield was
favored by most computational neuroscientists
and electrophysiologists since its principal cells,
the CA3 pyramidal cells, form a so-called ‘‘auto-
associative memory network’’ with their abundant,
local recurrent collaterals (Li et al., 1994). In com-
putational models these types of network were
found to be optimal for the efficient storage of a
large number of memory items and for reactiva-
tion of complete memory traces if only part of the
trace was provided, which are key components of
episodic memory. Moreover, the synapses among
CA3 pyramidal cells, as well as those between py-
ramidal cells in CA3 and their primary down-
stream target CA1, are subject to associative and
cooperative long-term potentiation (LTP), a pos-
sible molecular mechanism underlying memory
formation (Debanne et al., 1998). Interestingly,
however, many features of CA3 pyramidal cells
are shared by pyramidal cells of other cortical ar-
eas. Neocortical pyramidal cells have just as pro-
fuse recurrent local collaterals and their synapses
are subject to long-term plastic changes. CA3 and
layer II–III cells also have many intrinsic electro-
physiological characteristics in common. Thus,
autoassociative networks with plastic synapses are
abundant in the cortex. Therefore, it is not imme-
diately obvious why the formation of complex
memory traces is restricted to the hippocampal
formation and cannot be performed by other mul-
timodal associational cortical areas.
Two features of the CA3 area, however, clearly
distinguish it from other cortical regions. CA3 py-
ramidal cells receive a prominent excitatory input
to their proximal apical dendrites (Ramon y Cajal,
1911; Claiborne et al., 1986), which enables an
unusually strong spike coupling between an up-
stream region — the dentate gyrus (DG) — and
CA3 (Henze et al., 2002). The faithfulness of this
synaptic transmission is unparalleled in excitatory
cortical circuits. The second feature (shared by the
principal cells of the other hippocampal subre-
gions) is the way in which the firing pattern of
pyramidal cell codes the environmentally relevant
stimuli. Most pyramidal cells in other cortical
regions are characterized by higher spontaneous
firing rates, and a lower rate of modulation by the
appropriate stimuli. In sharp contrast, back-
ground activity in hippocampal principal cells,
and granule cells in particular, is very low (Muller
et al., 1987; Barnes et al., 1990; Quirk et al., 1992;
Jung and McNaughton, 1993). However, when
hippocampal principal cells participate in infor-
mation transfer (e.g., place cells), their activity in-
creases enormously. In computational terms, the
hippocampus uses a sparse code, whereas coding
in other cortical areas is denser. Apparently, the
hippocampus and the other cortical regions
‘‘speak’’ different neuronal languages. Since the
environmentally specific information reaches the

hippocampus via other cortical areas, the imme-
diate consequence of the difference in coding strat-
egies is that densely encoded cortical information
reaching the hippocampus needs to be translated
into a sparser hippocampal code. In other words,
an interface is needed between the hippocampus
and neocortex. The main task of this interface
would be the translation of the neocortical code
into a hippocampal one, and to transfer the new
code as efficiently as possible to the next station of
information processing. According to the classic
view, this interface is the DG, the first step of the
trisynaptic hippocampal loop.
We argue that the DG is the ‘‘odd-man-out’’
among cortical regions. In particular, at the gate of
the hippocampal formation, the DG possesses a
set of unusual properties that selectively evolved
for the task of code transformation between cor-
tical afferents and the hippocampus. These evolu-
tionarily conserved anatomical features enable the
DG to translate the noisy signal of the upstream
cortical areas to the sparse and specific code of
hippocampal formation, which is indispensable for
the formation of multiple, multidimensional mem-
ory items.
Computational requirements for the formation
of episodic memories
There is a sort of general consensus about how the
hippocampus could contribute to cortical memory
functions (Alvarez and Squire, 1994; Treves and
Rolls, 1994; McClelland et al., 1995; Kali and
Dayan, 2004; Rolls and Kesner, 2006). First, the
hippocampus is assumed to be capable of rapidly
creating and storing a memory trace, which is dis-
tinct from all existing traces. The memory trace is
associated with a snapshot of activity in medial
temporal neocortex (particularly entorhinal cor-
tex), which, in turn, is thought to represent a com-
pressed version of activity in the rest of neocortex.
Second, the hippocampus is thought to be capable
of retrieving particular stored traces if the ent-
orhinal activity pattern provides only a partial or
noisy version of the corresponding original pat-
tern. Using this cue, the hippocampus reinstates
the original pattern in the entorhinal cortex and,
579
subsequently, the rest of neocortex. This operation
is referred to as pattern completion, or autoasso-
ciative memory function. Finally, the hippocam-
pus may also be capable of autonomously
reactivating stored memory traces in the absence
of specific retrieval cues, thereby reactivating com-
plete cortical memory representations during ‘‘off-
line’’ behavioral states (e.g., slow-wave sleep).
Such replay may contribute in various ways to
the consolidation (transfer to a final repository)
and maintenance of episodic and semantic mem-
ory (Kali and Dayan, 2004).
But do we have any reason to believe that the
hippocampus is even capable of carrying out these
operations, and if so, that it is in some sense opt-
imized for exactly these tasks? The most widely
cited evidence is the existence of the extensive re-
current collateral network of pyramidal neurons in
area CA3. Such recurrent networks are the classic
examples of autoassociative memory devices,
whose properties have been extensively investi-
gated by theoretical means and computer simula-
tions (Marr, 1971; McNaughton and Morris, 1987;
Willshaw and Buckingham, 1990; Treves and
Rolls, 1992; Samsonovich and McNaughton,
1997; Kali and Dayan, 2000). In particular, the
storage capacity of such networks, i.e., the number
of patterns they can store and retrieve reliably, has
been determined (Treves and Rolls, 1992), and was
found to depend substantially on the properties of
the set of input patterns that we attempt to store.
Capacity is roughly inversely proportional to the
sparsity of individual patterns (i.e., the proportion
of active units in a pattern), and generally in-
creases as the overlap between different stored
patterns decreases. Therefore, it is reasonable to
assume that the need to maximize storage capacity
is an important reason for the conspicuously low
activity levels of principal cell populations in the
hippocampus (the proportion of active principal
cells in any hippocampal subfield at any given
moment is thought to be on the order of a few
percent (Barnes et al., 1990; Quirk et al., 1992;
Jung and McNaughton, 1993; Treves and Rolls,
1994 #5013).
However, activity patterns in most areas of neo-
cortex, including entorhinal cortex, appear to be
much denser (involving a larger proportion of
580
neurons at any given time), which is thought to be
beneficial for generalization (McClelland et al.,
1995), an important characteristic of the kind of
representational learning that the neocortex may
be engaged in. On the other hand, if the optimal
type of activity pattern (dense vs. sparse) is differ-
ent in the neocortex and the hippocampus, then
the information contained in the entorhinal dense
code needs to be ‘‘translated’’ into a sparse code
when hippocampal representations are created
(and vice versa). In principle, such translation
could be implemented directly by a single set of
projections from the input area to the autoassoci-
ative network (i.e., by the direct perforant path
input from entorhinal cortex to the CA3 region),
without an intervening specific interface for spars-
ification. However, as argued on theoretical
grounds by Treves and Rolls (1992), a projection
with the characteristics of the perforant path — a
large number of relatively weak, associatively
modifiable synapses on each target cell — may
be optimal during retrieval, but a different type of
input to the CA3 recurrent network — one with
a small number of individually strong synapses
per cell — is probably required for the storage of
new memory traces. The mossy fiber pathway, the
projection to CA3 from granule cells, meets the
requirements for this second type of input. Addi-
tional theoretical and computer simulation studies
by O’Reilly and McClelland (1994) indicated that
the two-stage pathway from EC to DG to CA3
could perform very effective pattern separation,
provided that individual mossy fiber connections
were sufficiently strong to transfer the benefits of
pattern separation in the DG to CA3. They further
argued that the coexistence of this indirect path-
way with the direct EC–CA3 connection enables
the hippocampus to avoid an inherent conflict be-
tween pattern separation and pattern completion,
both of which are important for the efficient
operation of the hippocampal autoassociator. In
summary, these studies suggest that a possible role
of the DG is to form sparse, pattern-separated
representations of entorhinal activity patterns, and
transmit this sparse representation reliably for
subsequent storage in the CA3 recurrent network.
But how does the DG create sparse representa-
tions from the relatively dense entorhinal activity
patterns, which constitute its only major cortical
input? Since no direct experimental investigation
of this issue has been undertaken, we need to rely
on indirect evidence and the results of computa-
tional studies to try to answer this question. At an
abstract level, well-known computational algo-
rithms exist which create sparse, pattern-separated
representations from distributed input patterns. A
simple example is the competitive-learning pattern
classification device described by Rumelhart and
Zipser (1986), which has been shown to be capable
of generating hippocampal place field-like activity
patterns from input patterns resembling neocorti-
cal sensory representations (Sharp, 1991). This
algorithm operates on binary input patterns and
generates binary output patterns. For each pres-
entation of an input pattern, the current values of
the synaptic weights (which are initially set to ran-
dom values) are used to determine the feedforward
activation of units in the output layer. Then the
output unit with the highest level of feedforward
input is allowed to become active (this step is
assumed to reflect the action of feedforward and
feedback inhibitory circuits), and the incoming
weights of this unit are allowed to change.
This weight change is assumed to be Hebbian in
nature: weights from active input units are in-
creased, while weights from inactive units are de-
creased in such a way that the sum of all incoming
weights remains constant (and identical to the sum
of incoming weights to all other output units). This
way, the ‘‘winning’’ unit will have an even higher
level of feedforward activation the next time the
same input pattern appears. It will also show an
increased response to other similar input patterns;
on the other hand, its response to patterns that are
very different (with a low degree of overlap) will
diminish. As a result, different output units will
eventually respond to different kinds of patterns,
and end up partitioning the input space among
themselves into non-overlapping groups of similar
patterns. The algorithm performs both sparsificat-
ion — since only a single output unit becomes ac-
tive for any given (distributed) input pattern —
and orthogonalization (pattern separation) —
since relatively dissimilar, but still overlapping
input patterns end up activating different output
units (zero overlap).

However, is there any experimental evidence
that sparsification requires a separate relay station,
and that the DG has unique features — besides its
well-known ‘‘detonator’’ type of terminals — for
code conversion and reliable transmission? In the
following pages we review behavioral, morpholog-
ical, and physiological data relevant to these
questions.
Lesion studies
Let us first examine whether the putative role of
the DG as described above is consistent with be-
havioral data on the effects of specific lesions. In
general, we expect that some, but not all tasks that
are sensitive to global hippocampal lesions will
also be sensitive to more specific lesions of the DG,
and might hope that the pattern of impairments
that occur after selective DG lesions sheds light on
the role of the DG in hippocampal processing.
Most lesion studies have taken advantage of the
fact that intrahippocampal injections of colchicine
cause a fairly selective destruction of the granule
cells of the DG, but other techniques, such as
neonatal X-ray irradiation and adrenalectomy,
which also cause a similar pattern of damage, have
also been used. Perhaps the most consistent find-
ing after DG lesions in rodents has been a severe
impairment in the acquisition of the reference
memory task in the Morris water maze (Suther-
land et al., 1983; McNaughton et al., 1989; Conrad
and Roy, 1993; Xavier et al., 1999). Working
memory versions of the task are also affected,
although perhaps to a lesser extent (Xavier et al.,
1999; Jeltsch et al., 2001). Reference and working
memory performance in the radial arm maze are
also compromised (McNaughton et al., 1989;
Jeltsch et al., 2001). In some more recent studies,
DG lesions were also found to cause impairment
in a delayed-matching-to-place task, in a temporal
task (Costa et al., 2005), and in a task which
required the detection of metric distance change
between objects (Goodrich-Hunsaker et al., 2005).
Recently, there have been some attempts to test
more directly the proposed contributions of the
DG to hippocampal function. In particular,
Gilbert et al. (2001) designed a short-term spatial
581
memory task where the proximity of relevant
locations could be varied systematically. They
found that DG lesions impaired performance at
small, but not at large separations, consistent with
the role of the DG in spatial pattern separation.
Lassalle et al. (2000) examined the consequences of
selective and reversible inactivation of mossy fiber
synapses in CA3 in mice during various stages of a
reference memory task in the Morris water maze.
They found that the mossy fiber input from DG to
CA3 was essential during learning, but not during
the retrieval phase of the task, or in the period
directly following learning (the early stages of
consolidation). Similarly, Lee and Kesner (2004)
attempted to distinguish encoding and retrieval
deficits during the acquisition of a navigation task
in the Hebb-Williams maze following lesions of
various type. They found that lesions of the DG
impaired learning, but not retrieval, in this task;
conversely, lesions of the perforant path input to
CA3 affected retrieval, but not learning.
In summary, the available behavioral data from
rodents with lesions to dentate granule cells or their
mossy fiber output are generally consistent with a
crucial role of the DG in providing input to area
CA3 during the acquisition of allocentric spatial
information, and provide some support for a more
specific function in (spatial) pattern separation. So
what are the morphological peculiarities of the DG
that support its role in orthogonalization and make
it indispensable for proper CA3 function?
Morphological arguments — heterogeneous
terminal types of the mossy fibers
Let’s see first how Cajal described of the unusual
terminals types of the mossy fibers:
‘‘ythere arise either short and thick
divergent appendages or quite long
fine threads that end in a swelling. Thus,
we reproduced here the arrangement
(although less distinctly) that we
described in certain branched fibers of
the cerebellum the mossy fibers. There-
fore without further ado let us apply the
same name to the axons of the granules
of the fascia dentata.’’
582
It is clear from the above description that the
investigator who named the mossy fibers by ex-
amining Golgi-stained material clearly identified
that this peculiar fiber system has more than one
terminal type. The well-recognizable giant endings
(the large mossy terminals) gave rise to thin fila-
mentous structures that ended in terminal-like
swellings. Apparently the name ‘‘mossy fiber’’ is
based on the presence of these filopodial exten-
sions. Later electron microscopic work identified
that filopodial terminals indeed establish asym-
metrical synapses (Amaral, 1979; Claiborne et al.,
1986) (Fig. 1). In addition, a third terminal type
has been described, small ‘‘en passant’’ boutons,
which resemble the most the conventional axonal
varicosities of cortical pyramidal cells (Claiborne
et al., 1986) (Fig. 1). Together the two smaller
mossy terminal types far outnumber their larger
counterpart, but still, the total number of termi-
nals is actually very low (Claiborne et al., 1986;
Acsady et al., 1998). A single granule cell has no
more than
200 terminals, which is at least two
orders of magnitude less than the number of vari-
cosities along the axonal arbor of a cortical py-
ramidal cell. How can these few but variable
terminals account for the code conversion and
efficient transmission as outlined above?
Unique features for code conversion
The computational studies summarized above sug-
gested that memory formation in the hippocampus
may be a two-step process: the sparsification of the
entorhinal signal by granule cells, followed by an
association of the now sparse code in the CA3
network. Modeling studies suggest that the ‘‘win-
ner-take-all’’ method of sparsification requires the
recruitment of strong feedback inhibition. Accord-
ing to this scheme, the output of a small popula-
tion of granule cells that become active in a given
environmental context (e.g., during a couple of
theta cycles as the rat passes through their place
fields) activate GABAergic interneurons which
exert fast and strong feedback inhibition on the
somata and dendrites of the non-coding granule
cells, shunting their entorhinal inputs and preclud-
ing their firing. As a result, synaptic plasticity only
takes place at the perforant path input of the cod-
ing (granule) cells. However, feedback inhibition is
well-known in all cortical regions. Is there any
reason to suppose that feedback inhibition is more
powerful in the DG than in other cortical regions,
which makes it especially useful for pattern sepa-
ration?
In cortical regions, interneurons constitute
10–20% of all the neurons. Cortical pyramidal
neurons innervate their postsynaptic principal and
interneuron targets in a quasi-random manner,
i.e., the incidence of the targets is determined by
the relative distribution of the neuron types
(Gulyas et al., 1993; Sik et al., 1993). Thus, the
estimated ratio of interneurons among the postsy-
naptic targets of cortical pyramidal cells is around
10%.
In the DG granule cells have axon collaterals
only in the hilus below the granule cell layer not
in stratum granulosum or stratum moleculare
(Claiborne et al., 1986; Acsady et al., 1998). As a
consequence only inhibitory neurons having so-
mata and/or dendrites in the hilus can participate
in feedback inhibition. Close to 50% of the hilar
neurons are GABAergic (Houser and Esclapez,
1994) which suggests that inhibitory cells may be
abundant among the postsynaptic targets of gran-
ule cells. The 5–8 hilar collaterals of the granule
cells possess around 7–12 large mossy boutons and
102–147 small terminals (filopodial and en passant
boutons) (Acsady et al., 1998). The postsynaptic
targets of the small terminal types are almost ex-
clusively interneurons, whereas targets of the large
mossy terminals are mainly excitatory mossy cells
(Acsady et al., 1998) (Fig. 2). Thus, due to the
surprising target selectivity of granule cell terminal
types, interneurons may constitute up to 90% of
the postsynaptic targets of the mossy fibers in the
hilus, in contrast to the 10–20% GABAergic tar-
gets in other cortical regions. Many of these hilar
neurons provide feedback inhibitory control of the
granule cells, suggesting that proportionally stron-
ger feedback inhibition is recruited here than in
other cortical regions.
A characteristic cell type of the hilus, the
somatostatin-immunoreactive interneuron, pro-
vides a good example of the strong recurrent
inhibition operating in this system. The
 583

Fig. 1. Electron micrographs of different terminal types along the mossy fibers in the CA3 region (A–C, E) and of a CA3 pyramidal
cell terminal (D) for comparison. All electron micrographs have the same magnification. A, B, A small en passant terminal establishes a
single asymmetrical synapse on a dendritic shaft, showing the characteristic long perforated postsynaptic density of small terminal
types on interneurons (arrows). C, A filopodial extension of a mossy fiber terminal forms a synapse (arrow) with a Substance P
receptor-immunoreactive interneuron. D, A CA3 pyramidal cell establishes asymmetrical synapse on a simple spine of a CA1 py-
ramidal neuron. E, A large, double-headed mossy fiber terminal forms multiple contacts (arrows) with thorny excrescences of a CA3
pyramidal cell. All active zones converged on the same pyramidal cell. The individual release sites are short. Scale bars: A–D, 0.5 mm;
E, 1 mm. Reprinted with permission from Acsady et al., 1998, Society for Neuroscience.

somatostatin-containing hilar neurons restrict they are also called HIPP cells, i.e., HIlar
their entire dendritic arbor to the hilus and inner- Interneurons with Perforant Path associated
vate the dendritic segment of granule cells in the axon terminal) (Han et al., 1993; Sik et al.,
zone where entorhinal afferents terminate (hence 1997). Unlike many other interneuron types that
584
Fig. 2. Filopodial extensions of mossy fiber terminals are spe-
cialized to innervate GABAergic cells. Artistic rendition of two
large mossy terminals, each equipped with four filopodial ex-
tensions (large arrowheads). The mossy fibers were labeled by
intracellular injection of biocytin into two neighboring granule
cells. All filopodial terminals were examined in the electron
microscope (not shown) and all contacted the dendrites or
spines of altogether six GABAergic neurons. Four of the GAB-
Aergic neurons were identified by their Substance P receptor-
content and two of them by ultrastructural characteristics. Five
of the six postsynaptic interneurons were spiny cells. Arrows
point to the main axons. Reprinted with permission from Ac-
sady et al., 1998, Society for Neuroscience.
are characterized by a smooth dendritic surface,
the dendrites of HIPP cells are densely covered
with thousands of long thin, elaborated spines
(Baude et al., 1993). In contrast to the compart-
mentalized spines of the pyramidal cells, which
usually receive a single terminal, these spines are
contacted by multiple asymmetrical synapses (up
to 8–10) (Acsady et al., 1998). Thus, these spines
increase the total synaptic input of the HIPP cells
enormously. Small terminals of granule cells have
been described to contact these spines (at least
30% of the total synaptic output of granule cells
contacts HIPP cells) (Acsady et al., 1998), whereas
the axons of CA3 pyramidal cells that project back
to the hilus selectively avoided HIPP cells (Wittner
et al., 2006), suggesting that the majority of the
excitatory input of these cells originates from
granule cells. Since the number of contact between
a granule cell and a HIPP cell is only 1 or 2 these
morphological data indicate the convergence of
several thousand granule cells on any one of these
interneurons. The axons of HIPP cells terminate in
the same layers as the perforant path, and there-
fore are in a critical position to modulate the
information transfer from the entorhinal cortex to
the DG (Han et al., 1993). A single HIPP cell may
form an unusually large number of axon terminals
in the molecular layer (up to 80,000 compared to
5000–10,000 in the case of a basket cell), the vast
majority of which innervate granule cells (Sik
et al., 1997). The little data available about the
activity of HIPP cells suggest that these neurons
are not fast-firing cells; rather, their activity is
principally driven by the firing pattern of granule
cells, a key feature for appropriately timed feed-
back inhibition (Buckmaster and Schwartzkroin,
1995).
Another unique morphological feature of the
connectivity in DG is the lack of interaction
among the major inhibitory basket cell classes
and among basket cells and other interneurons
(Acsady et al., 2000). In other hippocampal re-
gions as well as in other cortical regions, basket
cells densely innervate other basket- and non-
basket-type interneurons (Sik et al., 1995; Cobb et
al., 1997; Tamas et al., 1998) forming an interact-
ing local GABAergic network. For example, in the
CA1 region the somatic region of parvalbumin-
positive basket cells is contacted by more GAB-
Aergic terminals than the somatic region of py-
ramidal cells (Gulyas et al., 1999; Megias et al.,
2001). In sharp contrast, GABAergic cells in the
hilus of the DG receive, on average, 15–40 times
less input from local basket cells than mossy cells,
the principal excitatory cell type of this region
(Acsady et al., 2000). Since GABAergic cells in-
hibit each other, this connectivity pattern suggests
minimal disinhibitory influence in the hilus and
different GABAergic network dynamics.
But what are the physiological properties of the
granule cell–interneuron synapses? Examination of
synaptic transmission at the granule cell–basket
cell synapses demonstrated very fast kinetics: the
postsynaptic conductance of the unitary current
demonstrated submillisecond rise and decay
(Geiger et al., 1997). This effect was largely at-
tributed to the high synchrony of transmitter re-
lease and the rapid time course of AMPA receptor

deactivation. The fast postsynaptic response
allows rapid activation of feedback inhibition,
which supports a role in pattern separation.
These data suggest that the basic principles of
connectivity between excitatory and inhibitory
neurons in the DG are significantly different from
those in any other cortical region. The peculiar
arrangement of excitatory and inhibitory connec-
tions in the DG suggests an unusually strong re-
cruitment of inhibition that can be used to
suppress the activity of granule cells in a compet-
itive manner. As a result, only granule cells with
the strongest entorhinal excitatory drive will par-
ticipate in the information transfer to CA3. This
small population of active granule cells could
effectively prevent large granule cell populations
from reaching firing threshold via the strong feed-
back inhibitory system outlined above. Only those
entorhinal inputs will be potentiated which contact
active granule cells. This would further strengthen
the competitive process, which could lead to the
sparsification of the entorhinal signal resulting,
e.g., in the sharp and focused place fields of the
granule cells which are in sharp contrast to the
grid-like entorhinal signal (see below). The next
step is to faithfully transmit this recoded cortical
signal to the associational station, the CA3 region,
and to ensure that only the activated subset of
CA3 pyramidal cells would be included in the
autoassociation network responsible for memory
storage.
Unique features for efficient and sparse transmission
The giant mossy terminals display all the morpho-
logical features of a classic ‘‘detonator’’ or
‘‘driver’’ type terminal. No other cortical synapse
is comparable to them, but several subcortical
structures (e.g., thalamus, cerebellum) utilize sim-
ilar synaptic arrangements to secure faithful
synaptic transmission (Sherman and Guillery,
1998). A single large mossy terminal may estab-
lish up to 30–40 release sites, all converging on the
proximal dendrite of a single postsynaptic pyram-
idal cell (Chicurel and Harris, 1992; Acsady et al.,
1998). One would predict that the short electro-
tonic distance from the soma would maximize the
585
efficacy of the input in driving the postsynaptic
cell to threshold. Recent in vitro and in vivo data
confirm this assumption. Monosynaptic AMPA/
kainate receptor-mediated EPSCs from granule
cell–pyramidal cell pairs had a mean peak ampli-
tude of 163.0723 pA at 70 mV in organotypic
slice cultures which displayed morphological prop-
erties similar to the in vivo condition (Mori et al.,
2004). Strong excitatory action has been described
earlier between granule cells and their excitatory
targets in the hilus, the mossy cells (Scharfman
et al., 1990). In the in vivo anesthetized preparation,
repetitive firing in a single granule cell reliably
induced action potentials in monosynaptically
connected CA3 pyramidal cells (Henze et al.,
2002). It has to be emphasized that such strong
coupling is extremely rare among excitatory cells
in cortical circuits. The general rule is that a large
number of excitatory inputs have to be simultane-
ously active to reach postsynaptic spike threshold.
But how sparse is the detonator signal in mor-
phological terms?
The number of large mossy terminals along the
single unbranching axon of granule cells within
area CA3 is very low (average: 12.3; range: 10–18)
(Acsady et al., 1998). If one considers that a single
CA3 pyramidal cell may have up to 60,000 termi-
nals, it is straightforward to conclude that granule
cells are a specific cortical cell type designed to
transfer the sparse code generated in DG very
effectively to only a restricted set of postsynaptic
pyramidal cells.
Two recent studies described additional surpris-
ing features of the mossy fibers. These features not
only support faithful transmission through this
pathway, but also demonstrate the computational
power of the axon terminals in unexpected ways.
The first study (Engel and Jonas, 2005) describes
the active properties of the large mossy terminals,
which facilitates reliable transmission of high
frequency trains of action potentials. Apparently,
mossy terminals have a very high density of spe-
cialized Na+ channels with faster activation and
inactivation kinetics than somatic Na+ channels.
These Na+ channels enable reliable action poten-
tial invasion into large mossy terminals and in-
crease presynaptic Ca2+ influx, resulting in up to
16-fold increase of transmitter release. In addition,
586
modeling studies suggest that this active property
of the terminals is absolutely necessary to induce
Ca2+ influx into the filopodial extensions to trig-
ger glutamate release at the mossy fiber–interneu-
ron synapse. This mechanism may underlie the
high release probability of the filopodial connec-
tion compared to the mossy fiber–CA3 pyramidal
cell synapse (Jonas et al., 1993; Lawrence et al.,
2004).
The second study (Alle and Geiger, 2006) dem-
onstrates that mossy fibers transmit not only ac-
tion potentials, but also postsynaptic potentials
originating in the soma-dendritic compartment.
These ‘‘excitatory presynaptic potentials’’ alter the
transmitter release of the subsequent action po-
tential (within a 10–20 ms delay), thus enabling the
axon to integrate subthreshold and suprathreshold
signals provided they are temporally proximal.
In sum, mossy terminals fulfill all criteria for a
classical, sparse ‘‘detonator’’ synapse. The strong
granule cell–pyramidal cell connection, however,
poses two major problems for the autoassociative
network of the CA3 region.
Problem 1. Spontaneously active granule cells,
which represent only background noise, may in-
duce irrelevant spiking of CA3 pyramidal cells,
resulting in non-coding CA3 networks which may
overlap with the coding networks.
Problem 2. Spontaneously active CA3 pyramidal
cells may generate action potentials coincident
with the CA3 spikes evoked by granule cell firing,
which represent the given environment. These non-
coding EPSPs will be also strengthened in the
autoassociative CA3 network, which would ruin
the orthogonalized, sparse code.
A first solution to Problem 1 is that granule cells
have very low spontaneous firing rates
(0.1–0.01 spike/s) (Jung and McNaughton, 1993),
probably due to their hyperpolarized resting mem-
brane potential (80 mV in vivo) (Penttonen et al.,
1997) and/or strong inhibitory control. Still, if we
consider that there are one million granule cells per
hippocampus in rodents (Seress, 1988) and we take
the low end of their spontaneous discharge fre-
quency (0.01 Hz), we can calculate that at least 10
granule cells will fire in any one millisecond,
activating
120 CA3 pyramidal cells, each of
which has around 30,000 terminals within the CA3
(Li et al., 1994). Thus, other solutions are needed
to solve Problem 1.
Similar to the hilus, the number of small mossy
fiber terminals in the CA3 region exceeds the
number of large mossy fiber terminals by a factor
of at least four (Acsady et al., 1998). As in the
hilus, the small terminals selectively innervate in-
hibitory cells. All examined inhibitory cell classes
(perisomatic, dendritic, and interneuron-selective)
were among the postsynaptic elements of granule
cells. Since a single mossy fiber rarely innervates a
postsynaptic interneuron via multiple contacts
(Acsady et al., 1998), granule cell firing activates
at least four times as many inhibitory as excitatory
cells. Physiological data confirm strong activation
of this feedforward inhibitory circuit. In paired
recordings, a single action potential in the granule
cell induced a biphasic response in the postsynap-
tic CA3 pyramidal cell: a brief EPSC followed by a
pronounced IPSC (Mori et al., 2004). The direc-
tion of the summated charge transfer was outward,
indicating a net inhibitory synaptic response. The
authors calculated that approximately four inter-
neurons fired together to evoke the measured in-
hibitory responses, which corresponds well to the
anatomical data. Reliable activation of interneu-
rons was also observed in vivo. Multiple granule
cell spikes induced firing in monosynaptically cou-
pled interneurons with quite high probability
(Henze et al., 2002) (Fig. 3D), suggesting that al-
though the granule cell–interneuron contact rarely
contains multiple release sites, the single active
zone of the small terminals is still efficient enough
to induce postsynaptic firing of the GABAergic
cell. Recruitment of more inhibitory than excita-
tory cells suggests that the net effect of the exci-
tatory mossy fiber system on the CA3 pyramidal
cell population is, counter-intuitively, inhibitory.
A peculiar EEG transient, the dentate spike, can
be utilized to demonstrate the impact of granule
cell activation on the postsynaptic cell populations
along the dentate–CA3 axis during a ‘‘natural’’
stimulus. Dentate spikes are short-duration, large-
amplitude field potentials caused by synchronous
activation of the entorhinal input, which occur
during behavioral immobility and slow-wave sleep
 587

C D 0.12
0.03

0.08
Probability

0.02

0.01 0.04

0 0.00
0 2 4 6 8 10 0 2 4 6 8 10

E 1.0 100 Hz F 0.3

0.8 60 Hz
40 Hz
Probability

0.2
Probability

0.6
20 Hz
0.4
10 Hz
0.1
0.2

0.0
0.0
0 0.2 0.4 0.6 0.8 1 1 3 5 7 9
Time (s) Spike number

Fig. 3. Spike transmission dynamics between a granule cell and its interneuron and pyramidal cell targets in CA3c in vivo. (A) Camera
lucida reconstruction of the extracellular electrode track and biocytin-labeled granule cell. Inset, a higher-power view of the mossy fiber
axon near the probe track. Arrowheads, mossy fiber boutons. (B) Superimposed (n ¼ 60) intracellularly evoked action potentials in a
granule cell (bottom traces) and simultaneously recorded extracellular units (filtered 0.8–8 kHz). Note the time-locked response of a
putative pyramidal cell to the granule cell action potentials. (C–D) Cross-correlograms between the evoked granule cell action
potentials and the activity of a putative CA3c pyramidal cell (C) or interneuron (D). The values are shuffle corrected and expressed as
probability (number of unit spikes per bin/total number of granule cell spikes). Arrowhead, peak time of the granule cell action
potential. (E) Representative results of the effect of intratrain frequency on spike transmission probability for a putative pyramidal cell.
(F) Spike transmission probability (shuffle-corrected probability of spike in 6 ms following granule cell spike) as a function of spike
number in evoked 100 Hz train (7s.e.m). Solid line, putative interneurons (n ¼ 24); dotted line, putative pyramidal cells (n ¼ 21). Scale
bars, (A) 50 mm; inset 20 mm; (B) 1 ms, 25 mV, 75 mV. m, molecular layer; g, granule cell layer; h, hilus; IC, intracellular electrode track;
EC, extracellular electrode track.

(Bragin et al., 1995). Extracellular recordings dur- 1997). It is worth mentioning that this pattern of
ing dentate spikes demonstrated increased unit activity is in sharp contrast to the neuronal be-
activity in hilar neurons (many of which are GAB- havior that can be observed during the other major
Aergic), but suppressed multiunit activity in the excitatory field transient in the hippocampus, the
CA3 region (Bragin et al., 1995). Intracellular sharp wave, which originates in CA3. Since in this
studies confirmed depolarization of granule cells case there is no inhibitory ‘‘barrier’’ comparable
and hilar interneurons but hyperpolarization in to the dentate GABAergic network which could
CA3 and CA1 pyramidal cells (Penttonen et al., block the spread of excitation, sharp waves
588
propagate not only to the CA1 region but also to
parahippocampal cortical regions via polysynaptic
activation (Chrobak and Buzsaki, 1994, 1996).
Most recently the efficacy of the dentate inhib-
itory ‘‘barrier’’ was demonstrated by comparing
the correlation of intracellular activity of neurons
in various cortical fields with the UP and DOWN
states in the EEG during slow cortical oscillation
(Isomura et al., 2006). Surprisingly, all cortical
regions (including entorhinal cortex and DG)
changed intracellular activity coherently with the
EEG states with the exception of CA3 region. In
contrast to basically all cortical cells CA3 pyram-
idal cells were not active during the UP states.
Apparently the inhibitory network launched by
the dentate during the UP states is able to shield
CA3 even from the most synchronous excitatory
events of the cortical mantle.
In summary, anatomical and physiological data
provide convergent evidence for the conclusion
that the output of the granule cells activates an
unusually strong feedforward inhibition to area
CA3. Since even a single basket cell is able to block
action potential generation in a large number of
pyramidal cells (Miles et al., 1996), feedforward
inhibition will effectively reduce the number of
active CA3 pyramidal cells during dentate activa-
tion. In addition, in the CA3 region, specialized
interneuron types exist which restrict their dendri-
tic (Gulyas et al., 1991) or axonal (Vida and
Frotscher, 2000) arbor to stratum lucidum of CA3,
suggesting selective control of this pathway. Thus,
the likely solution of Problems 1 and 2 is that
‘‘background’’ or ‘‘non-coding’’ EPSPs and action
potentials in CA3 are actively inhibited during
dentate–CA3 information transfer by the strong
feedforward inhibition. In this way only the small
population of CA3 cells receiving the decorrelated
sparse dentate signal will be active, and following
the Hebbian rule, only the recurrent synapses be-
tween the activated CA3 cells will be potentiated.
According to the modeling studies (see above) the
matrix of potentiated synapses will represent the
memory trace.
However, what happens to the EPSPs generated
by the spontaneous activity of CA3 pyramidal cells
immediately before the dentate input arrives?
The EPSPs among pyramidal cells are quite slow
(half duration, 27 ms) (Miles and Wong, 1986) and
accidentally the mossy fiber input can arrive
together with their peak, thus these early back-
ground (unwanted) EPSPs can be potentiated be-
fore the disynaptic feed forward inhibition arrives.
A recent report provides a possible solution for
this problem (Kobayashi and Poo, 2004). This
study describes LTP at the recurrent synapses of
the CA3 network induced by paired stimulation of
mossy fiber input and the associational/commis-
sural input. First, this study provides direct and
elegant evidence for the role of mossy fibers in
changing the synaptic strength of the autoassoci-
ative CA3 matrix. Second, an interesting observa-
tion from this study helps resolve the problem of
early EPSPs. The results demonstrate that the po-
tentiation of associational input depends on the
relative timing of mossy fiber and associational
spike trains. The potentiation was smaller if addi-
tional associational spikes were added before the
paired stimulation, compared to the protocol that
added spikes after the paired (associational/mossy
fiber) pulses. The effect depended on mGluR1 ac-
tivation. Thus, apparently the system favors the
potentiation of CA3 EPSPs arriving coincidentally
or after the dentate signal. Since the mossy fiber
input is able to induce CA3 spiking in vivo, these
EPSPs will mostly represent the spiking of CA3
pyramidal cells evoked by the sparse, orthogonali-
zed dentate input. In this way only the associat-
ional EPSPs representing dentate activity will be
potentiated, but EPSPs arriving earlier, represent-
ing spontaneous CA3 activity, will not.
Target-dependent plasticity — the meaning of
various terminals types
If we consider the strong feedforward inhibition
operating along the dentate–CA3 axis the follow-
ing, third problem arises:
Problem 3. How to overcome the strong feedfor-
ward inhibition when the specific dentate pattern
has to activate the pyramidal cell?
Apparently the solution to Problem 3 is that
short- and long-term plasticity at large and
small types of mossy fiber ending are different.

The physiological data clearly demonstrate that
the small terminals are not only distinct morpho-
logical units, evolved to contact a large number of
interneurons, but discrete compartments which
harbor distinct molecular machinery for plasticity
different from their giant cousins.
Small and large mossy terminals display differ-
ent types of short-term plasticity. The giant mossy
terminal–pyramidal cell connection express very
strong short-term facilitation (Salin et al., 1996;
Toth et al., 2000), reaching up to threefold ampli-
tude increase on average at the fifth EPSC in case
of 20 Hz stimulation. (This is highly unusual in the
case of giant ‘‘driver’’-like terminals; e.g., in the
thalamus excitatory terminals with a similar
synaptic arrangement show strong depression in
relay cells; Reichova and Sherman, 2004). In con-
trast the small mossy terminal–interneuron con-
nection shows short-term depression after
repetitive mossy fiber stimulation in about 50%
of the interneurons. Others show modest facilita-
tion at 20 Hz (Toth et al., 2000), but at 40 Hz, all
responses are depressed after the 8th action
potential (Mori et al., 2004). The third synapse
in the feedforward inhibitory circuit, the interneu-
ron–CA3 pyramidal cell contact, expresses pro-
nounced short-term depression (Mori et al., 2004)
at all frequencies tested.
This variability in short-term plasticity at differ-
ent mossy fiber synapses favors conditions for
a single granule cell action potential to induce
weaker excitation, but relatively strong feedfor-
ward inhibition. Repetitive firing, however, rapidly
increases the effect of excitation and at the same
time decreases the efficacy of inhibition. Thus, the
net effect of mossy fiber activation on CA3 py-
ramidal cells depends heavily on the frequency of
granule cell firing. The system appears to act as a
high-pass filter, where spike transmission from
granule cell to pyramidal cell is blocked at low
frequencies but favored when the frequency of
granule cell discharge increases. Recently this as-
sumption was tested directly in in vitro paired
recordings of granule cells and pyramidal cells
(Mori et al., 2004). The postsynaptic potentials
evoked in pyramidal cells by granule cell stimula-
tion were measured at increasing frequencies
(10–40 Hz). The inhibitory dominant PSPs
589
observed during low frequency trains switched to
excitatory dominant PSPs at high frequencies.
At 40 Hz EPSPs dominated the response already
after the third granule cell action potentials,
whereas at 10 Hz the response remained inhibi-
tory even at the 15th action potential.
Frequency-dependent facilitation of mossy fiber
transmission was demonstrated in monosynaptic-
ally coupled granule cell — pyramidal cell pairs
in vivo as well (Henze et al., 2002) (Fig. 3). The
probability of postsynaptic pyramidal cell spikes
rapidly increased with increasing granule cell firing
and reached 0.8 at 100 Hz (Fig. 3E), which is an
extremely high value in cortical circuits and results
in an almost one-to-one relay of the granule cell
activity. Within the spike train the probability of
CA3 pyramidal spikes increased with the number
of presynaptic granule cell spikes. The maximum
spike transmission probability was reached after
the 4–5th spike (Fig. 3F). In contrast, in the case of
interneurons, the probability of transmission did
not increase with an increasing number of presy-
naptic spikes at 100 Hz and remained lower than
that of the pyramidal cell (Fig. 3F). These in vitro
and in vivo data indicate that differential short-
term plasticity in this feed-forward circuit result
in a frequency-dependent shift of the polarity of
postsynaptic response.
In the freely moving condition, granule cells
have very low spontaneous firing rates, which can
rapidly increase to 40 Hz (Jung and McNaughton,
1993) as the animal enters the place field of the
neuron. As a consequence of the frequency-
dependent switch from excitation to inhibition de-
scribed above, the probability of spike transmis-
sion between the DG and CA3 is very low at low
firing rates, despite the ‘‘detonator’’ nature of the
synapse. When granule cells code the specific
information of the environment, the probability
of spike transmission to CA3 becomes very high.
Thus, granule cells can act as a ‘‘conditional
detonator’’ as suggested by Henze et al. (2002).
This mechanism solves both the problem of
filtering out non-coding spontaneous activity
(Problem 1), and resolves the issue that strong
feedforward inhibition must be overcome when
specific information must be transmitted from
dentate to CA3 (Problem 3).
590
The delicate frequency-dependent balance be-
tween excitation and inhibition substantially in-
creases the computational power of the mossy
fibers. It raises the possibility that they not only
participate in the faithful transmission of an or-
thogonalized dentate signal but they themselves
participate in creating the non-overlapping repre-
sentations in the CA3 region. Recent data suggest
(Leutgeb et al., 2007, see below) that the repre-
sentation of the environment is more orthogonali-
zed in CA3 than in the DG, which questions the
role of dentate as the sole contributor to this
process. Due to its high-pass filter nature, the
dentate–CA3 circuit may refine the dentate code,
and, e.g., participate in creating the single recep-
tive field observed in CA3 place cells as opposed to
the multiple place fields of dentate place cells. The
critical variable in this process will be the exact
firing pattern of the dentate granule cell. For
instance, assume that a granule cell fires at 40 Hz
in one of its receptive fields but only 10 Hz in the
other. Based on the data discussed above, only
the higher firing rate will induce firing in the CA3
pyramidal cell, the lower activity will be filtered
out, and the CA3 pyramidal cell will display a
single receptive field as a result.
What about long-term changes? One presenta-
tion of the dentate code may not be sufficient to
induce long-term changes in the CA3 recurrent
network. Multiple presentations (e.g., crossing the
place field several times) or autonomous replay of
the memory trace in absence of the original con-
dition, may be necessary to induce long-term plas-
ticity. Replay of a given firing pattern may occur
during different EEG states, as has been shown for
theta and sharp wave activity (Nadasdy et al.,
1999), or during the subsequent sleep episodes
(Wilson and Mcnaughton, 1993).
Tetanic stimulation of mossy fibers induces LTP
in pyramidal neurons, but is either without effect,
or it induces depression at synapses onto inter-
neurons (Maccaferri et al., 1998). Since the mossy
fiber–LTP onto pyramidal cells critically depends
on cAMP, the effect can be explained by the
absence of the adenylyl cyclase-cAMP cascade
from the filopodial and ‘‘en passant’’ small termi-
nals. As a result of this differential LTP, the crit-
ical frequency at which inhibition switches to
excitation may change and/or the steepness of the
high-pass filter cutoff may increase. In this way,
presynaptic mossy fiber–LTP may help the
orthogonalization process performed by the dent-
ate–CA3 circuit and creates a good opportunity
for the faithful activation of the same CA3 circuit
in case of repetitive presentation.
Finally, morphological plasticity of the small
terminals provides yet another way to fine-tune the
balance of excitation and inhibition in the mossy
fiber pathway. The structure of the filopodial ter-
minals strongly resembles rapidly advancing and
retracting axonal filopodia observed during axonal
development and synapse formation. Recent stud-
ies of slice cultures indeed demonstrated that over
one-third of the filopodia are highly active
(De Paola et al., 2003; Tashiro et al., 2003). In
addition, in mature slices, approximately 9% of
the small en passant boutons were also labile (half-
life, approximately 1 day), in contrast to the sta-
bility of large terminals. Interestingly, in mature
cultures, the total number of synapses remained
stable in the presence of substantial turnover of
individual terminal structures. Motility of the
filopodial extensions was observed not only in
slice cultures but also in the acute whole-mount
hippocampal preparation and acute hippocampal
slices. This actin-based motility can be regulated
by brain-derived neurotrophic factor (BDNF),
AMPA and/or kainate receptors, in a cAMP de-
pendent manner (De Paola et al., 2003; Tashiro
et al., 2003). These data strongly suggest that tar-
get selectivity of the small terminal types of the
mossy fiber is based on the dynamic morpholog-
ical properties of these axonal elements. In addi-
tion, they suggest that even in mature animals
activity change, traumatic injury or cell loss may
induce rapid remodeling of the mossy fiber-to-
interneuron connections from granule cells. In-
deed, ischemic damage, which results in the loss of
hilar and stratum lucidum interneurons, dramat-
ically reduces the number of filopodia (Arabadzisz
and Freund, 1999).
In summary, the mossy fiber pathway maxim-
ally utilizes the opportunity to differentially regu-
late its postsynaptic partners. Selective innervation
of pyramidal cells and interneurons by distinct
terminal types creates a favorable condition to

differentially regulate short-term and long-term
plasticity and the motility of various mossy termi-
nal types. The ‘‘bottom-line’’ of this highly
dynamic system appears to be the fine-tuning of
the balance between the excitation evoked by the
large ‘‘detonator’’ terminals and the feedforward
inhibition activated by the small terminals. This
will determine exactly which dentate firing patterns
will induce permanent modification of the auto-
associational network of the CA3 region. The final
test for all predictions regarding dentate function
is the examination of firing activity in the freely
moving animal.
Neuronal activity patterns in vivo and the processing
of entorhinal input by the dentate gyrus
Compared to the vast amount of data available on
spatial (and non-spatial) representations in area
CA1 of the hippocampus, relatively little is known
about the in vivo firing properties of neurons in the
DG under various behavioral conditions. During
exploration, granule cells display spatially selective
activity, and their firing behavior is, at least under
most conditions studied so far, qualitatively quite
similar to that of pyramidal cells (place cells) in
areas CA1 and CA3. In particular, in the radial
arm maze, granule cells have directional place
fields, although the average number of subfields is
somewhat higher and the average size of these
subfields is slightly smaller than in CA3 place cells
(Jung and McNaughton, 1993). The firing activity
of dentate granule cells is modulated by local field
potential oscillations in the theta frequency range,
and the timing of individual action potentials
changes during traversals of the place field similar
to phase precession observed in CA1 pyramidal
cells (Skaggs et al., 1996). When different spatial
cues were put in conflict by manipulating the en-
vironment, sudden coherent transitions (known as
reference frame shifts) could be observed in the
activity of the set of simultaneously recorded
granule cells, also analogous to the behavior of
the CA1 cell population under these conditions
(Gothard et al., 2001). In an experiment where an
explicit attempt was made to identify non-spatial
as well as spatial responses, a subpopulation of
591
granule cells showed position-selective or position-
independent reward site responses, whereas an-
other population showed pure place responses
(Tabuchi et al., 2003).
When rats are allowed to explore a novel envi-
ronment for the first time, both dentate granule
cells and CA1 pyramidal cells acquire distinct
spatial preferences within the first few minutes
(Nitz and McNaughton, 2004). However, concur-
rently recorded interneurons showed very different
behavior in the two regions: while most CA1
interneurons transiently decreased their activity
while the animal was exploring a novel environ-
ment, the majority of interneurons in the DG sig-
nificantly increased their activity during the same
epoch. These data are congruent with the role of
strong feedback inhibition in establishing sparse
and decorrelated output in the DG.
In order to understand the nature of the com-
putations performed by the DG, it is essential to
have an accurate description of the cortical input it
receives. The last few years have witnessed a major
revolution in our understanding of spatial repre-
sentations in entorhinal cortex. Until a few years
ago, existing data on EC representations indicated
that, although neurons in EC were spatially selec-
tive, their place fields were much larger and less
clearly defined than those in hippocampal areas
(Quirk et al., 1992; Frank et al., 2000). Indeed, this
was perhaps the most direct experimental evidence
for the assumption that sparse, orthogonal, ‘‘hip-
pocampal-type’’ representations are created first in
the DG. However, when spatial firing patterns
were measured in the part of medial entorhinal
cortex (mEC) which projects to the dorsal HC
(where place fields are normally recorded), much
smaller place fields, similar in size to correspond-
ing hippocampal place fields, were found (Fyhn
et al., 2004), while areas of mEC with large place
fields projected to more ventral parts of the HC,
which itself was found to have large place fields
(Maurer et al., 2005). From these data, it appeared
that there was in fact no major transformation of
spatial representations from EC to the DG (and
the rest of the hippocampus), although subtler
differences (especially in response to environmen-
tal manipulations) could not be ruled out. How-
ever, it soon emerged that if spatial firing patterns
592
are recorded over a larger spatial scale, entorhinal
and hippocampal representations are again fun-
damentally different. In particular, EC place fields
were found to repeat periodically at the vertices of
a regular hexagonal lattice, and different EC ‘‘grid
cells’’ differ in the center location (phase), orien-
tation, and scale of the grid (Hafting et al., 2005).
In contrast, hippocampal place cells have at most a
few discrete place fields even in these larger envi-
ronments, and their fields typically do not have
any special geometrical relationship. Thus, it cur-
rently appears that hippocampal spatial represen-
tations are in fact different in nature, and in
particular, much sparser over a large environment
than representations in the (medial) entorhinal
cortex, and a transformation (in fact, a pattern
separation operation) by the DG is still required.
Indeed, if we consider two locations in the envi-
ronment which are separated by the grid period
(assumed to be relatively invariable) in the area of
entorhinal cortex which projects to a given part of
the HC, the activity of the grid cell population will
be quite similar, while the hippocampal activity
pattern will be distinct, reflecting the outcome of
some kind of pattern separation process. In fact,
the need for pattern separation becomes even more
obvious if we consider two environments. The
root of the problem is the observation that the
relative grid parameters (phase and orientation) of
different grid cells appear to be fixed in all envi-
ronments, so, in principle, there must be corre-
sponding locations (and orientations) in two
distinct environments where the activity of the en-
tire grid cell population is identical. This raises an
obvious question: how distinct spatial representa-
tions of the two environments could be formed in
the HC based on a single entorhinal grid cell rep-
resentation. One possible answer to this question is
based on the fact that the entorhinal grids are not
perfectly regular; for instance, peaks in the grid
vary in amplitude, so that, in principle, there could
be sufficient spatial information present in the
amplitudes to distinguish different environments.
Another, perhaps more plausible explanation is
that the hippocampus (and, in particular, the DG)
receives, in addition to input from medial EC,
where grid cells are located, input from lateral EC,
where neuronal firing patterns have a lower spatial
information content, but instead, carry more
information about relevant objects (Hargreaves
et al., 2005). Such object-based information is
probably sufficient to disambiguate different envi-
ronments, and create distinct codes in the HC.
Hippocampal pattern separation between envi-
ronments of varying degrees of similarity has re-
cently been investigated in a series of experiments
(Leutgeb et al., 2004, 2005a,b), and the initial
analysis in areas CA3 and CA1 has now been par-
tially extended to mEC and the DG (Leutgeb
et al., 2007; Hafting et al., 2006). By analyzing the
spatial firing fields of neurons within environments
of varying degrees of similarity (manipulations in-
cluded switching between different recording loca-
tions, as well as changes in the shape and/or the
color of the enclosure), Moser and colleagues
found that firing patterns of neurons in all hippo-
campal areas distinguished between different en-
vironments much better than those of grid cells in
mEC. Essentially no pattern separation was
detected in EC, as population firing patterns in
enclosures of different shapes, colors, or even in
different rooms were not significantly different
(any observed changes occurred coherently in all
recorded neurons; Hafting et al., 2006). However,
they also found that the basic properties of pattern
separation were different between different sub-
fields of the hippocampus. In area CA3, the phe-
nomenon termed ‘‘rate remapping’’ was observed
when the shape of the enclosure was varied con-
tinuously: the firing rate pattern of the active cell
population changed continuously as the environ-
ment was gradually transformed, while place field
locations remained constant (Leutgeb et al.,
2005a). In the end, environments of clearly dis-
tinct shapes (e.g., circle vs. square) activated CA3
pyramidal cell populations with relatively little
overlap — indeed, when recordings were made in
two different rooms, the two populations of active
CA3 neurons appeared to be chosen independ-
ently, a phenomenon referred to as ‘‘global
remapping’’ (Leutgeb et al., 2005b).
Interestingly, preliminary data from recent ex-
periments have revealed a radically different type
of pattern separation in the DG. First, unlike in
CA3, the same dentate cells were found to be ac-
tive in different environments (Leutgeb et al.,

2007). Second, DG neurons typically had multiple
place fields even in a single environment, consist-
ent with earlier data (Jung and McNaughton,
1993). Finally, the peak firing rate within any one
place field of a single cell varied with even small
changes in the shape of the enclosure, independ-
ently of rate changes in other fields of the same cell
(Leutgeb et al., 2007). These results show that the
DG performs a pattern separation operation on its
entorhinal inputs, but pattern separation appears
to work differently from what was previously as-
sumed. The observation that the proportion of si-
multaneously active cells is much lower in the DG
than in EC (a fact that was confirmed by the data
described above, and independently by another
recent study, which measured immediate early
gene expression in the DG following spatial expe-
rience; Chawla et al., 2005) suggested that different
environments might activate different sets of neu-
rons in the DG, thereby implementing a particu-
larly efficient type of pattern separation. The
results of Moser and colleagues now suggest that
this might not be the case, and different environ-
ments (as well as different locations within these
environments) might be encoded by different firing
rate patterns in essentially the same population of
active DG neurons. This latter scheme potentially
also allows a fine discrimination of different envi-
ronments (and locations) based on the DG pop-
ulation activity pattern (especially since the firing
rates of DG neurons appear to be rather sensitive
to changes in environmental features). However,
since the CA3 code appears to be more efficiently
orthogonalized than the DG code, an additional
processing step may be needed. The physiological
data reviewed above indicate that the mossy fiber
projection may be utilized to arrive at the sparser,
more completely pattern-separated representation
recorded in area CA3, but further contributions
from other sources (temporo-ammonic pathway
and local network connections) cannot be ex-
cluded.
The different behavior of spatial representations
at different stages of hippocampal processing
clearly argues against the simple view that all
properties of hippocampal place cells originate in
the DG, and downstream areas simply inherit
these properties. A more complex view of the
593
formation of hippocampal representations is also
indicated by recordings of neuronal activity pat-
terns following selective lesions. In particular, fol-
lowing colchicine lesions of dentate granule cells as
described above, place cell representations could
still be observed in area CA1 (McNaughton et al.,
1989). Similarly, the spatial representation in area
CA1 was largely intact after surgical separation
from all other hippocampal areas, which left it
with the direct projection from entorhinal cortex
as its sole cortical input (Brun et al., 2002). There-
fore, hippocampal areas other than the DG must
be capable of creating sparse, distributed, place-
field-like representations on their own under some
circumstances, probably based on their direct
entorhinal inputs. However, it has to be kept in
mind that despite the presence of a proper place
cell representation in CA1, navigation memory
was compromised following both types of lesion,
suggesting that utilization of the spatial code at
the behavioral level requires intact dentate–CA3
interaction.
In summary, these data suggest that the DG
does perform a pattern separation of the ent-
orhinal signal, which, however, needs further
processing to achieve the sparse and decorrelated
activity pattern observed in the CA3 region.
The dentate gyrus: an evolutionary-developmental
perspective
Apparently the basic organization of the dent-
ate–CA3 network has the deepest phylogenetic
root among cortical regions. Before the divergence
of reptilian-mammalian lineages which apparently
preceded the mass extinction of the Permian pe-
riod (
250 million years ago) neurons in all cor-
tical areas were most probably packed in a single
cellular layer. During the ontogenesis they likely
followed an outside-in pattern of histogenesis,
where newly generated neurons settle below the
older ones, like in extant reptiles (Goffinet et al.,
1986). Their main excitatory afferents entered and
terminated in the embryonic marginal zone, above
the cortical plate, i.e., in the same zone where the
apical dendrites of their main targets (excitatory
principal cells) were present (Ten Donkelaar,
594
1998). As suggested recently, this arrangement
does not support the evolution of a cortical struc-
ture with multiple cellular layers and distinct areas
(Super and Uylings, 2001).
As the mammalian nervous system evolved, the
dorsal and lateral cortex of the ancient reptilian
cortex underwent a significant modification that
opened up tremendous opportunities for areal and
cellular diversification (Super and Uylings, 2001).
This included changing the pattern of histogenesis
to the inside-out pattern, where newly generated
neurons settle above the older ones, and changing
the pattern of axonal ingrowth (for a review, see
Super and Uylings, 2001). In the mammalian neo-
cortex, most of the afferents fibers enter not above
but below the cortical plate, via the subplate
(Allendoerfer and Shatz, 1994; Molnar, 2000).
This new developmental scheme allowed the de-
velopment of multilayered cortical structures with
highly variable cell types, rich reciprocal interac-
tion with the thalamus and the establishment of
new cortical regions.
Little developmental change occurred, however,
in the medial and dorsomedial cortex, which be-
came the mammalian allocortex. In the dorsome-
dial cortex that is homologous with the Cornu
Ammonis, the pattern of histogenesis changed to
the inside-out pattern, but the pattern of axonal
ingrowth did not (Super et al., 1998). However, in
the most ‘‘stubborn’’ structure, the DG, the basic
reptilian developmental pattern was retained
(Bayer, 1980) histogenesis follows the outside-in
pattern and the main excitatory afferents enter
above the cortical plate.
Neocortex has evolved rapidly to become a
highly complex multilayered structure, with exten-
sive intracortical and thalamo-cortical reciprocal
connections (Butler, 1994; Nieuwenhuys, 1994;
Rakic, 1995; Northcutt and Kaas, 1995). Allocor-
tex remained a unilayered structure with a basi-
cally unidirectional information flow, DG being
the only cortical structure without thalamic input
(Amaral and Witter, 1989). Neocortex has been
significantly expanded laterally and was parceled
into numerous functionally segregated areas, but
the hippocampus retained the original two major
subfields, the DG and Cornu Ammonis. For these
two regions, only the Cornu Ammonis showed
some areal segregation, and the original reptilian
dorsomedial cortex became CA3, CA2, CA1, and
subiculum (Amaral et al., 1990). Again, the DG
showed no areal segregation.
But why did the DG–CA3 connection remain
essentially unchanged during the course of evolu-
tion, when the rest of the cortical mantle under-
went a dramatic reorganization? Here we would
like to propose that the reason behind the pro-
tracted evolutionary pattern of the hippocampus,
and especially the DG, is the structural constraints
of hippocampal function. Apparently, the forma-
tion of freely accessible multidimensional memory
traces can only be performed in a two-step process
that includes a segregation of input followed by an
associative step. The reciprocally coupled, multi-
layered cortical structures evolved for a different
role (for a more complex interaction with the
environment). Apparently, the basic plan of the
two-step information processing through the hip-
pocampal formation remained essentially the same
from lizard to human. Similarly to mammals, in
reptiles, the dentate-equivalent medial cortex
receives the cortical input. This cortical region
lacks an autoassociative network and projects
the recoded information unidirectionally, to the
reptilian analogue of the Cornu Ammonis, the
dorsomedial cortex (Lopez-Garcia and Martinez-
Guijarro, 1988; Martinez-Guijarro et al., 1991; de
la Iglesia et al., 1994). Association functions in the
reptile may take place in the next step here in the
dorsomedial cortex, where an extensive recurrent
collateral system exists, similar to the CA3 region
in mammals (Martinez-Guijarro et al., 1984).
Interestingly, damage to the reptilian homologue
of hippocampus causes similar learning problems
as hippocampal lesions do in mammals (Rodriguez
et al., 2002), suggesting an analogous structural-
functional relationship.
Has anything changed in DG during the mam-
malian evolution? In primates, the volumetric ratio
of the DG and CA3 has changed in favor of the
first. Indeed, DG is ‘‘dentate’’ sensu stricto only in
primates, where it includes numerous infoldings.
Among these areas the hilus showed the largest
relative increase in volume and cell number
(Seress, 1988) underlying the importance of the
region, where most of the peculiarities in

microcircuits have been noticed. Apparently, the
feed back regulation of granule cells became
more elaborated with the increasing complexity
of information to be categorized by the system.
Conclusions, unresolved questions
The DG appears to be an ancient cortical structure
from the phylogenetic perspective, yet it displays a
number of unique features not found in other cor-
tical regions. Its morphological and physiological
properties are highly specialized, and these can
explain, at least in part, the role assigned to the
DG and CA3 by computational theories. In par-
ticular, mossy fibers are characterized by distinct
mechanisms of signal transfer at excitatory vs.
inhibitory targets, and unusually strong activation
of GABAergic circuits. This arrangement allows
a delicate balance of excitation and inhibition,
which is utilized for code conversion and sparsifi-
cation at the entorhinal–dentate connection and
for frequency-dependent spike transfer at the dent-
ate–CA3 connection.
Several issues, however, remain unresolved.
More comparable data are needed from freely
moving animals to understand the precise compu-
tation that takes place in the DG and in CA3.
Since DG–CA3 transmission appears to be de-
pendent on the frequency of granule cell discharge,
DG firing patterns should be carefully analyzed,
and particularly with respect to multiple receptive
fields. The role of two major excitatory inputs of
the granule cells, not discussed here, the mossy cell
input and the supramammillary afferents, is nec-
essary to clarify DG information processing com-
prehensively. Both mossy cells and supramam-
millary afferents contact the proximal dendrites of
granule cells, and therefore are likely to exert a
powerful influence. Mossy cells of the hilus have
highly divergent axons, and thus may link distant
DG populations involved in coding similar envi-
ronmental events, whereas the supramammillary
input may mediate the modulation of granule cells
by the theta rhythm. The role of rhythmic EEG
activities (theta, gamma) in mediating signal trans-
fer, code conversion, and the short- and long-term
plasticity is unclear at the present time. Similarly,
595
the role of dentate spikes is not explored. It is
tempting to speculate that they may participate in
memory replay, like the sharp wave in the
CA3–CA1 network (Buzsaki, 1989), but definitive
proof is currently unavailable. From the compu-
tational perspective, it is not clear how the lack of
connectivity among granule cells and the relative
paucity of the direct backprojection from the CA3
to the DG (Li et al., 1994) helps the categorization
function in DG. In conclusion, this peculiar neo-
cortex–archicortex interface will likely keep us
busy for a long time.
Acknowledgment
This work was supported by the Wellcome Trust
(A.L. is the recipient of a Wellcome Trust Inter-
national Senior Fellowship), the Institut de Cerv-
eau et de la Moelle épiniere, the Hungarian
Scientific Research Fund (OTKA T 049100) and
the EU Framework 6.
References
Acsady, L., Kamondi, A., Sik, A., Freund, T. and Buzsaki, G.
(1998) GABAergic cells are the major postsynaptic targets of
mossy fibers in the rat hippocampus. J. Neurosci., 18:
3386–3403.
Acsady, L., Katona, I., Martinez-Guijarro, F.J., Buzsaki, G.
and Freund, T.F. (2000) Unusual target selectivity of peri-
somatic inhibitory cells in the hilar region of the rat hippo-
campus. J. Neurosci., 20: 6907–6919.
Alle, H. and Geiger, J.R. (2006) Combined analog and action
potential coding in hippocampal mossy fibers. Science, 311:
1290–1293.
Allendoerfer, K.L. and Shatz, C.J. (1994) The subplate, a tran-
sient neocortical structure—its role in the development of
connections between thalamus and cortex. Annu. Rev.
Neurosci., 17: 185–218.
Alvarez, P. and Squire, L.R. (1994) Memory consolidation and
the medial temporal lobe: a simple network model. Proc.
Natl. Acad. Sci. U.S.A., 91: 7041–7045.
Amaral, D.G. (1979) Synaptic extensions from the mossy fibers
of the fascia dentata. Anat. Embryol. (Berl), 155: 241–251.
Amaral, D.G., Ishizuka, N. and Claiborne, B. (1990) Neurons,
numbers and the hippocampal network. Understanding the
Brain Through the Hippocampus, 83.
Amaral, D.G. and Witter, M.P. (1989) The three-dimensional
organization of the hippocampal formation: a review of an-
atomical data. Neuroscience, 31: 571–591.
596
Arabadzisz, D. and Freund, T.F. (1999) Changes in excitatory
and inhibitory circuits of the rat hippocampus 12–14 months
after complete forebrain ischemia [In Process Citation].
Neuroscience, 92: 27–45.
Barnes, C.A., McNaughton, B.L., Mizumori, S.J., Leonard,
B.W. and Lin, L.H. (1990) Comparison of spatial and tem-
poral characteristics of neuronal activity in sequential stages
of hippocampal processing. Prog. Brain Res., 83: 287–300.
Baude, A., Nusser, Z., David, J., Roberts, B., Mulvihill, E.,
McIlhinney, R.A.J. and Somogyi, P. (1993) The metabo-
tropic glutamate receptor (mGluR1Ó) is concentrated at
perisynaptic membrane of neuronal subpopulations as
detected by immunogold reaction. Neuron, 11: 771–787.
Bayer, S.A. (1980) Development of the hippocampal region in
the rat. I. Neurogenesis examined with 3H-thymidine auto-
radiography. J. Comp. Neurol., 190: 87–114.
Bragin, A., Jando, G., Nadasdy, Z., van Landeghem, M. and
Buzsaki, G. (1995) Dentate EEG spikes and associated
interneuronal population bursts in the hippocampal hilar
region of the rat. J. Neurophysiol., 73: 1691–1705.
Brun, V.H., Otnass, M.K., Molden, S., Steffenach, H.A.,
Witter, M.P., Moser, M.B. and Moser, E.I. (2002) Place cells
and place recognition maintained by direct entorhinal-
hippocampal circuitry. Science, 296: 2243–2246.
Buckmaster, P.S. and Schwartzkroin, P.A. (1995) Interneurons
and inhibition in the dentate gyrus of the rat in vivo. J.
Neurosci., 15: 774–789.
Butler, A.B. (1994) The evolution of the dorsal thalamus of
jawed vertebrates, including mammals: cladistic analysis and
a new hypothesis. Brain Res. Brain Res. Rev., 19: 29–65.
Buzsaki, G. (1989) Two-stage model of memory trace forma-
tion: a role for ‘noisy’ brain states. Neuroscience, 310:
551–570.
Chawla, M.K., Guzowski, J.F., Ramirez-Amaya, V., Lipa, P.,
Hoffman, K.L., Marriott, L.K., Worley, P.F., McNaughton,
B.L. and Barnes, C.A. (2005) Sparse, environmentally selec-
tive expression of Arc RNA in the upper blade of the rodent
fascia dentata by brief spatial experience. Hippocampus, 15:
579–586.
Chicurel, M.E. and Harris, K.M. (1992) 3-dimensional analysis
of the structure and composition of CA3 branched dendritic
spines and their synaptic relationships with mossy fiber
boutons in the rat hippocampus. J. Comp. Neurol., 325:
169–182.
Chrobak, J.J. and Buzsaki, G. (1994) Selective activation of
deep layer (V–VI) retrohippocampal cortical neurons during
hippocampal sharp waves in the behaving rat. J. Neurosci.,
14: 6160–6170.
Chrobak, J.J. and Buzsaki, G. (1996) High-frequency oscilla-
tions in the output networks of the hippocampal-entorhinal
axis of the freely behaving rat. J. Neurosci., 16: 3056–3066.
Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1986) A light
and electron microscopic analysis of the mossy fibers of the
rat dentate gyrus. J. Comp. Neurol., 246: 435–458.
Cobb, S.R., Halasy, K., Vida, I., Nyiri, G., Tamas, G., Buhl,
E.H. and Somogyi, P. (1997) Synaptic effects of identified
interneurons innervating both interneurons and pyramidal
cells in the rat hippocampus [published erratum appears in
Neuroscience 1997 Oct; 80(3): 971]. Neuroscience, 79:
629–648.
Conrad, C.D. and Roy, E.J. (1993) Selective loss of hippocam-
pal granule cells following adrenalectomy: implications for
spatial memory. J. Neurosci., 13: 2582–2590.
Costa, V.C., Bueno, J.L. and Xavier, G.F. (2005) Dentate
gyrus-selective colchicine lesion and performance in temporal
and spatial tasks. Behav. Brain Res., 160: 286–303.
Debanne, D., Gahwiler, B.H. and Thompson, S.M. (1998)
Long-term synaptic plasticity between pairs of individual
CA3 pyramidal cells in rat hippocampal slice cultures. J.
Physiol., 507(Pt 1): 237–247.
De Paola, V., Arber, S. and Caroni, P. (2003) AMPA receptors
regulate dynamic equilibrium of presynaptic terminals in
mature hippocampal networks. Nat. Neurosci., 6: 491–500.
Engel, D. and Jonas, P. (2005) Presynaptic action potential
amplification by voltage-gated Na+ channels in hippocam-
pal mossy fiber boutons. Neuron, 45: 405–417.
Frank, L.M., Brown, E.N. and Wilson, M. (2000) Trajectory
encoding in the hippocampus and entorhinal cortex. Neuron,
27: 169–178.
Fyhn, M., Molden, S., Witter, M.P., Moser, E.I. and Moser,
M.B. (2004) Spatial representation in the entorhinal cortex.
Science, 305: 1258–1264.
Geiger, J.R., Lubke, J., Roth, A., Frotscher, M. and Jonas, P.
(1997) Submillisecond AMPA receptor-mediated signaling
at a principal neuron-interneuron synapse. Neuron, 18:
1009–1023.
Gilbert, P.E., Kesner, R.P. and Lee, I. (2001) Dissociating hip-
pocampal subregions: double dissociation between dentate
gyrus and CA1. Hippocampus, 11: 626–636.
Goffinet, A.M., Daumerie, C., Langerwerf, B. and Pieau, C.
(1986) Neurogenesis in reptilian cortical structures: 3H-
thymidine autoradiographic analysis. J. Comp. Neurol., 243:
106–116.
Goodrich-Hunsaker, N.J., Hunsaker, M.R. and Kesner, R.P.
(2005) Dissociating the role of the parietal cortex and dorsal
hippocampus for spatial information processing. Behav.
Neurosci., 119: 1307–1315.
Gothard, K.M., Hoffman, K.L., Battaglia, F.P. and McN-
aughton, B.L. (2001) Dentate gyrus and ca1 ensemble activity
during spatial reference frame shifts in the presence and
absence of visual input. J. Neurosci., 21: 7284–7292.
Gulyas, A.I., Megias, M., Emri, Z. and Freund, T.F. (1999) Total
number and ratio of excitatory and inhibitory synapses con-
verging onto single interneurons of different types in the CA1
area of the rat hippocampus. J. Neurosci., 19: 10082–10097.
Gulyas, A.I., Miettinen, R., Jacobowitz, D.M. and Freund,
T.F. (1991) Calretinin-immunoreacti cells in the rat hippo-
campus. I. A new type of neuron — specifically associated
with the mossy fibre system — revealed. Eur. J. Neurosci.
Suppl., 4: 158.
Gulyas, A.I., Miles, R., Hájos, N. and Freund, T.F. (1993)
Precision and variability in postsynaptic target selection
of inhibitory cells in the hippocampal CA3 region. Eur. J.
Neurosci., 5: 1729–1751.

Hafting, T., Fyhn, M., Molden, S., Moser, M.B. and Moser,
E.I. (2005) Microstructure of a spatial map in the entorhinal
cortex. Nature, 436: 801–806.
Hafting, T., Fyhn, M., Treves, A., Moser, E.I. and Moser, M.B.
(2006) Coherent reallignment of entorhinal grid cells
coincides global remapping in the hippocampus. FENS,
Vienna.
Han, Z.S., Buhl, E.H., Lorinczi, Z. and Somogyi, P. (1993) A
high degree of spatial selectivity in the axonal and dendritic
domains of physiologically identified local-circuit neurons in
the dentate gyrus of the rat hippocampus. Eur. J. Neurosci.,
5: 395–410.
Hargreaves, E.L., Rao, G., Lee, I. and Knierim, J.J. (2005)
Major dissociation between medial and lateral entorhinal in-
put to dorsal hippocampus. Science, 308: 1792–1794.
Henze, D.A., Wittner, L. and Buzsaki, G. (2002) Single granule
cells reliably discharge targets in the hippocampal CA3 net-
work in vivo. Nat. Neurosci., 5: 790–795.
Houser, C.R. and Esclapez, M. (1994) Localization of mRNAs
encoding two forms of glutamic acid decarboxylase in the rat
hippocampal formation. Hippocampus, 4: 530–545.
de la Iglesia, J.A., Martinez-Guijarro, F.I. and Lopez-Garcia,
C. (1994) Neurons of the medial cortex outer plexiform layer
of the lizard Podarcis hispanica: Golgi and immunocyto-
chemical studies. J. Comp. Neurol., 341: 184–203.
Isomura, Y., Sirota, A., Ozen, S., Montgomery, S., Mizuseki,
K., Henze, D.A. and Buzsaki, G. (2006) Integration and
segregation of activity in entorhinal-hippocampal subregions
by neocortical slow oscillations. Neuron, 52: 871–882.
Jeltsch, H., Bertrand, F., Lazarus, C. and Cassel, J.C. (2001)
Cognitive performances and locomotor activity following
dentate granule cell damage in rats: role of lesion extent and
type of memory tested. Neurobiol. Learn. Mem., 76: 81–105.
Jonas, P., Major, G. and Sakmann, B. (1993) Quantal compo-
nents of unitary EPSCs at the mossy fibre synapse on CA3
pyramidal cells of rat hippocampus. J. Physiol. Paris Lon-
don, 472: 615–663.
Jung, M.W. and McNaughton, B.L. (1993) Spatial selectivity of
unit activity in the hippocampal granular layer. Hippocam-
pus, 3: 165–182.
Kali, S. and Dayan, P. (2000) The involvement of recurrent
connections in area CA3 in establishing the properties of
place fields: a model. J. Neurosci., 20: 7463–7477.
Kali, S. and Dayan, P. (2004) Off-line replay maintains declar-
ative memories in a model of hippocampal-neocortical inter-
actions. Nat. Neurosci., 7: 286–294.
O’Keefe, J. and Dostrovsky, J. (1971) The hippocampus as a
spatial map. Preliminary evidence from unit activity in the
freely-moving rat. Brain Res., 34: 171–175.
O’Keefe, J. and Nadel, L. (1978) The hippocampus as a cog-
nitive map. Clarendon, Oxford.
Kobayashi, K. and Poo, M.M. (2004) Spike train timing-de-
pendent associative modification of hippocampal CA3 recur-
rent synapses by mossy fibers. Neuron, 41: 445–454.
Lassalle, J.M., Bataille, T. and Halley, H. (2000) Reversible
inactivation of the hippocampal mossy fiber synapses in mice
impairs spatial learning, but neither consolidation nor
597
memory retrieval, in the Morris navigation task. Neurobiol.
Learn. Mem., 73: 243–257.
Lawrence, J.J., Grinspan, Z.M. and McBain, C.J. (2004) Quan-
tal transmission at mossy fibre targets in the CA3 region of
the rat hippocampus. J. Physiol., 554: 175–193.
Lee, I. and Kesner, R.P. (2004) Encoding versus retrieval of
spatial memory: double dissociation between the dentate
gyrus and the perforant path inputs into CA3 in the dorsal
hippocampus. Hippocampus, 14: 66–76.
Leutgeb, J.K., Leutgeb, S., Moser, M.B. and Moser, E.I. (2007)
Pattern separation in the dentate gyrus and CA3 of the hip-
pocampus. Science, 315: 961–966.
Leutgeb, J.K., Leutgeb, S., Treves, A., Meyer, R., Barnes, C.A.,
McNaughton, B.L., Moser, M.B. and Moser, E.I. (2005a)
Progressive transformation of hippocampal neuronal repre-
sentations in ‘‘morphed’’ environments. Neuron, 48:
345–358.
Leutgeb, S., Leutgeb, J.K., Barnes, C.A., Moser, E.I., McN-
aughton, B.L. and Moser, M.B. (2005b) Independent codes
for spatial and episodic memory in hippocampal neuronal
ensembles. Science, 309: 619–623.
Leutgeb, S., Leutgeb, J.K., Treves, A., Moser, M.B. and Moser,
E.I. (2004) Distinct ensemble codes in hippocampal areas
CA3 and CA1. Science, 305: 1295–1298.
Li, X.G., Somogyi, P., Ylinen, A. and Buzsaki, G. (1994) The
hippocampal ca3 network — an in vivo intracellular labeling
study. J. Comp. Neurol., 339: 181–208.
Lopez-Garcia, C. and Martinez-Guijarro, F.J. (1988) Neurons
in the medial cortex give rise to Timm-positive boutons in the
cerebral cortex of lizards. Brain Res., 463: 205–217.
Maccaferri, G., Toth, K. and McBain, C.J. (1998) Target-spe-
cific expression of presynaptic mossy fiber plasticity. Science,
279: 1368–1370.
Marr, D. (1971) Simple memory: a theory for archicortex. Phi-
los. Trans. R. Soc. Lond. B Biol. Sci., 262: 23–81.
Martinez Guijarro, F.J., Berbel, P.J., Molowny, A. and Lopez
Garcia, C. (1984) Apical dendritic spines and axonic termi-
nals in the bipyramidal neurons of the dorsomedial cortex of
lizards (Lacerta). Anat. Embryol., 170: 321–326.
Martinez-Guijarro, F.J., Soriano, E., Del Rio, J.A. and Lopez-
Garcia, C. (1991) Zinc-positive boutons in the cerebral cortex
of lizards show glutamate immunoreactivity. J. Neurocytol.,
20: 834–843.
Maurer, A.P., Vanrhoads, S.R., Sutherland, G.R., Lipa, P. and
McNaughton, B.L. (2005) Self-motion and the origin of
differential spatial scaling along the septo-temporal axis of
the hippocampus. Hippocampus, 15: 841–852.
McClelland, J.L., McNaughton, B.L. and O’Reilly, R.C. (1995)
Why there are complementary learning systems in the hip-
pocampus and neocortex: insights from the successes and
failures of connectionist models of learning and memory.
Psychol. Rev., 102: 419–457.
McNaughton, B.L., Barnes, C.A., Meltzer, J. and Sutherland,
R.J. (1989) Hippocampal granule cells are necessary for nor-
mal spatial learning but not for spatially-selective pyramidal
cell discharge. Experimental brain research. Experimentelle
Hirnforschung, 76: 485–496.
598
McNaughton, B.L. and Morris, N.G. (1987) Hippocampal
synaptic enhancement and information storage within a dis-
tributed memory system. TINS, 10: 408–415.
Megias, M., Emri, Z., Freund, T.F. and Gulyas, A.I. (2001)
Total number and distribution of inhibitory and excitatory
synapses on hippocampal CA1 pyramidal cells. Neurosci-
ence, 102: 527–540.
Miles, R., Toth, K., Gulyas, A.I., Hájos, N. and Freund, T.F.
(1996) Differences between somatic and dendritic inhibition
in the hippocampus. Neuron, 16: 815–823.
Miles, R. and Wong, R.K. (1986) Excitatory synaptic interac-
tions between CA3 neurones in the guinea-pig hippocampus.
J. Physiol., 373: 397–418.
Molnar, Z. (2000) Development and evolution of thalamocor-
tical interactions. Eur. J. Morphol., 38: 313–320.
Mori, M., Abegg, M.H., Gahwiler, B.H. and Gerber, U. (2004)
A frequency-dependent switch from inhibition to excitation
in a hippocampal unitary circuit. Nature, 431: 453–456.
Muller, R.U., Kubie, J.L. and Ranck Jr., J.B. (1987) Spatial
firing patterns of hippocampal complex-spike cells in a fixed
environment. J. Neurosci., 7: 1935–1950.
Nadasdy, Z., Hirase, H., Czurko, A., Csicsvari, J. and Buzsaki,
G. (1999) Replay and time compression of recurring spike
sequences in the hippocampus [in process citation]. J. Ne-
urosci., 19: 9497–9507.
Nieuwenhuys, R. (1994) The neocortex. An overview of its ev-
olutionary development, structural organization and syna-
ptology. Anat. Embryol., 190: 307–337.
Nitz, D. and McNaughton, B. (2004) Differential modulation
of CA1 and dentate gyrus interneurons during exploration of
novel environments. J. Neurophysiol., 91: 863–872.
Northcutt, R.G. and Kaas, J.H. (1995) The emergence and ev-
olution of mammalian neocortex. Trends Neurosci., 18:
373–379.
Penttonen, M., Kamondi, A., Sik, A., Acsady, L. and Buzsaki,
G. (1997) Feed-forward and feed-back activation of the
dentate gyrus in vivo during dentate spikes and sharp wave
bursts. Hippocampus, 7: 437–450.
Quirk, G.J., Muller, R.U., Kubie, J.L. and Ranck Jr., J.B.
(1992) The positional firing properties of medial entorhinal
neurons: description and comparison with hippocampal place
cells. J. Neurosci., 12: 1945–1963.
Rakic, P. (1995) A small step for the cell, a giant leap for man-
kind: a hypothesis of neocortical expansion during evolution.
Trends Neurosci., 18: 383–388.
Ramon y Cajal, S. (1911) Histologie de systeme nerveux de
l’homme et des vertebres tomme II. Maloine, Paris.
Reichova, I. and Sherman, S.M. (2004) Somatosensory co-
rticothalamic projections: distinguishing drivers from modu-
lators. J. Neurophysiol., 92: 2185–2197.
O’Reilly, R.C. and McClelland, J.L. (1994) Hippocampal con-
junctive encoding, storage, and recall: avoiding a trade-off.
Hippocampus, 4: 661–682.
Rodriguez, F., Lopez, J.C., Vargas, J.P., Gomez, Y., Broglio,
C. and Salas, C. (2002) Conservation of spatial memory
function in the pallial forebrain of reptiles and ray-finned
fishes. J. Neurosci., 22: 2894–2903.
Rolls, E.T. (1989) Functions of neuronal networks in the hip-
pocampus and cerebral cortex in memory. In: Cotterill
R.M.J. (Ed.), Models of Brain Function. Cambridge
University Press, Cambridge, pp. 15–33.
Rolls, E.T. and Kesner, R.P. (2006) A computational theory of
hippocampal function, and empirical tests of the theory.
Prog. Neurobiol., 79: 1–48.
Rumelhart, D.E. and Zipser, D. (1986) Feature discovery by
competitive learning. In: Rumelhart D.E., McClelland J. and
Group. P.R. (Eds.), Parallel Distributed Processing: Explo-
rations in the Microstructure of Cognition. Vol. 1. Founda-
tions. MIT Press, Cambridge, MA, pp. 151–193.
Salin, P.A., Scanziani, M., Malenka, R.C. and Nicoll, R.A.
(1996) Distinct short-term plasticity at two excitatory
synapses in the hippocampus. Proc. Natl. Acad. Sci.
U.S.A., 93: 13304–13309.
Samsonovich, A. and McNaughton, B.L. (1997) Path integra-
tion and cognitive mapping in a continuous attractor neural
network model. J. Neurosci., 17: 5900–5920.
Scharfman, H.E., Kunkel, D.D. and Schwartzkroin, P.A.
(1990) Synaptic Connections of Dentate Granule Cells and
Hilar Neurons — Results of Paired Intracellular Recordings
and Intracellular Horseradish Peroxidase Injections. Neuro-
science, 37: 693–707.
Scoville, W.B. and Milner, B. (1957) Loss of recent memory
after bilateral hippocampal lesions. J. Neurol. Neurosurg.
Psychiatry, 20: 11–21.
Seress, L. (1988) Interspecies comparison of the hippocampal
formation shows increased emphasis on the regio superior in
the Ammon’s horn of the human brain. Journal fur Hirn-
forschung, 29: 335–340.
Sharp, P.E. (1991) Computer simulation of hippocampal place
cells. Psychobiology, 19: 103–115.
Sherman, S.M. and Guillery, R.W. (1998) On the actions that
one nerve cell can have on another: distinguishing ‘‘drivers’’
from ‘‘modulators.’’ Proc. Natl. Acad. Sci. U.S.A., 95:
7121–7126.
Sik, A., Penttonen, M. and Buzsaki, G. (1997) Interneurons in
the hippocampal dentate gyrus: an in vivo intracellular study.
Eur. J. Neurosci., 9: 573–588.
Sik, A., Penttonen, M., Ylinen, A. and Buzsaki, G. (1995)
Hippocampal CA1 interneurons: an in vivo intracellular
labeling study. J. Neurosci., 15(10): 6651–6665.
Sik, A., Tamamaki, N. and Freund, T.F. (1993) Complete axon
arborization of a single CA3 pyramidal cell in the rat hip-
pocampus, and its relationship with postsynaptic parvalbu-
min-containing interneurons. Eur. J. Neurosci., 5:
1719–1728.
Skaggs, W.E., McNaughton, B.L., Wilson, M.A. and Barnes,
C.A. (1996) Theta phase precession in hippocampal neuronal
populations and the compression of temporal sequences.
Hippocampus, 6: 149–172.
Squire, L.R. (1992) Memory and the hippocampus: a synthesis
from findings with rats, monkeys, and humans. Psychol.
Rev., 99: 195–231.
Super, H., Martinez, A., Del Rio, J.A. and Soriano, E. (1998)
Involvement of distinct pioneer neurons in the formation of

layer-specific connections in the hippocampus. J. Neurosci.,
18: 4616–4626.
Super, H. and Uylings, H.B. (2001) The early differentiation of
the neocortex: a hypothesis on neocortical evolution. Cere-
bral cortex New York, N.Y., 11: 1101–1109.
Sutherland, R.J., Whishaw, I.Q. and Kolb, B. (1983) A behav-
ioural analysis of spatial localization following electrolytic,
kainate- or colchicine-induced damage to the hippocampal
formation in the rat. Behav. Brain Res., 7: 133–153.
Tabuchi, E., Mulder, A.B. and Wiener, S.I. (2003) Reward
value invariant place responses and reward site associated
activity in hippocampal neurons of behaving rats. Hippo-
campus, 13: 117–132.
Tamas, G., Somogyi, P. and Buhl, E.H. (1998) Differentially
interconnected networks of GABAergic interneurons in the
visual cortex of the cat. J. Neurosci., 18: 4255–4270.
Tashiro, A., Dunaevsky, A., Blazeski, R., Mason, C.A. and
Yuste, R. (2003) Bidirectional regulation of hippocampal
mossy fiber filopodial motility by kainate receptors: a two-
step model of synaptogenesis. Neuron, 38: 773–784.
Ten Donkelaar, H. (1998) Reptiles. In: Nieuwenhuys R., Ten
Donkelaar H.J. and Nicholson C. (Eds.), The Central Nerv-
ous System of Vertebrates. Springer, Berlin.
Toth, K., Suares, G., Lawrence, J.J., Philips-Tansey, E. and
McBain, C.J. (2000) Differential mechanisms of transmission
599
at three types of mossy fiber synapse. J. Neurosci., 20:
8279–8289.
Treves, A. and Rolls, E.T. (1992) Computational constraints
suggest the need for two distinct input systems to the hip-
pocampal CA3 network. Hippocampus, 2: 189–199.
Treves, A. and Rolls, E.T. (1994) Computational analysis of the
role of the hippocampus in memory. Hippocampus, 4:
374–391.
Vida, I. and Frotscher, M. (2000) A hippocampal interneuron
associated with the mossy fiber system. Proc. Natl. Acad. Sci.
U.S.A., 97: 1275–1280.
Willshaw, D.J. and Buckingham, J.T. (1990) An assessment of
Marr theory of the hippocampus as a temporary memory
store. Phil. Trans. R. Soc. London, 329: 205–215.
Wilson, M.A. and Mcnaughton, B.L. (1993) Dynamics of the
hippocampal ensemble code for space. Science, 261:
1055–1058.
Wittner, L., Henze, D.A., Zaborszky, L. and Buzsaki, G. (2006)
Hippocampal CA3 pyramidal cells selectively innervate
aspiny interneurons. Eur. J. Neurosci., 24: 1286–1298.
Xavier, G.F., Oliveira-Filho, F.J. and Santos, A.M. (1999)
Dentate gyrus-selective colchicine lesion and disruption of
performance in spatial tasks: difficulties in ‘‘place strategy’’
because of a lack of flexibility in the use of environmental
cues? Hippocampus, 9, 668–681.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 32

The dentate gyrus as a filter or gate: a look back and

a look ahead

David Hsu

Department of Neurology, University of Wisconsin, 600 Highland Avenue, H6/526, Madison, WI 53792, USA

Abstract: The idea of the dentate gyrus as a gate or filter at the entrance to the hippocampus, blocking or
filtering incoming excitation from the entorhinal cortex, has been an intriguing one. Here we review the
historical development of the idea, and discuss whether it may be possible to be more specific in defining
this gate. We propose that dentate function can be understood within a context of Hebbian association and
competition: hilar mossy cells help the dentate granule cells to recognize incoming entorhinal patterns of
activity (Hebbian association), after which patterns that are consistently and repetitively presented to the
dentate gyrus are passed through, while random, more transient patterns are blocked (non-associative
Hebbian competition). Translamellar inhibition as well as translamellar potentiation can be understood in
this context. The dentate-hilar complex thus plays the role of a ‘‘pattern excluder’’, not a pattern completer.
The unique role of pattern exclusion may explain the peculiar qualities of dentate granule cells and hilar
mossy cells.

Keywords: dentate gate; dentate filter; pattern excluder

Introduction 1992; Staley et al., 1992; Williamson et al., 1993),

The dentate gyrus, sitting between the entorhinal
cortex and area CA3, is both anatomically well
positioned and physiologically predisposed to play
the role of a gate, blocking or filtering excitatory
activity from the entorhinal cortex and controlling
the amount of excitation that gets through to the
hippocampus. Normal adult granule cells rarely
generate action potentials. In part this is because
there is little direct interconnectivity between
dentate granule cells under normal conditions
(reviewed in Chapter 1 of this volume). In addi-
tion, granule cells have a high resting membrane
potential (Fricke and Prince, 1984; Scharfman,
Corresponding author. Tel.: +1 608 263 8551;
Fax: +1 608 263 0412; E-mail: hsu@neurology.wisc.edu
and strong GABA receptor-mediated inhibition
(Mody et al., 1992; Coulter, 1999; Nusser and
Mody, 2002; Stell and Mody, 2002; Cohen et al.,
2003; Mody, 2005).
History of the idea
Data that the dentate gyrus may serve as a gate
appeared in at least as early as 1966 in work by
Andersen et al. (1966). In these experiments, the
hippocampal formation of adult rabbits was ex-
posed by removal of the overlying neocortex and
corpus callosum. Stimulating electrodes were
placed into entorhinal cortex or in the perforant
pathway. Extracellular as well as intracellular re-
cordings were made from electrodes placed in the

DOI: 10.1016/S0079-6123(07)63032-5 601

602
transverse plane of the hippocampus, piercing
CA1 and both blades of the dentate gyrus. Perfo-
rant pathway stimulation resulted in a negative
wave reflecting granule cell excitatory postsynaptic
potentials (EPSPs) generated in the middle third of
the dentate gyrus molecular layer. The EPSP was
followed by a large and slow inhibitory postsy-
naptic potential (IPSP), persisting for 100–150 ms.
That EPSPs were followed by IPSPs persisting
for 100–150 ms suggested that perforant pathway
stimulation at frequencies higher than 10 Hz
should result in a decremental granule cell popu-
lation spike response (habituation). For pulse
stimulus durations of less than 1 s, this decrement-
al response was indeed observed. However, for
longer pulse durations of 6 and 9 s, there was a
gradual incremental granule cell response (facili-
tation), beginning after a few seconds of repetitive
stimulation. The degree of facilitation increased
with increasing stimulation frequency up to a
maximum of 10 Hz, and was abolished by a stim-
ulation frequency higher than 20 Hz (Andersen et
al., 1966). Although not stated explicitly by the
authors, these experiments established one way
that the dentate gyrus appears to act as a gate.
Short-duration stimuli carried at a certain fre-
quency are blocked, but longer duration stimuli
carried at the same frequency are facilitated. Later
studies demonstrated additional ways the dentate
gyrus acts as a gate.
In 1976, Alger and Teyler applied repetitive per-
forant pathway stimulation to rat hippocampal
slices at 1 Hz for a total duration of 10 s (Alger and
Teyler, 1976; Teyler and Alger, 1976). They found
an incremental EPSP and population spike re-
sponse with each succeeding stimulus (facilitation)
in CA3 and CA1 but a decremental response
(habituation) in the dentate gyrus. In contrast,
stimulation of the perforant pathway at 15 Hz for
15 s produced potentiation of EPSP and population
spike responses of dentate gyrus, CA3 and CA1.
These results are similar to those of Andersen et al.
(1966), showing that the dentate gyrus suppresses
flow of excitation from perforant pathway input but
that this suppression can be reversed with longer
duration stimuli in an appropriate frequency range.
Winson and Abzug (1977, 1978) placed stimu-
lating electrodes into the angular bundle of the
perforant pathway in behaving rats, and compared
dentate granular layer population spike ampli-
tudes and molecular layer EPSPs in slow wave and
rapid eye movement (REM) sleep vs. the alert and
still state. The authors found that slow wave and
REM sleep states are associated with larger pop-
ulation spikes but smaller EPSPs compared to the
still, alert state. To explain these findings, the au-
thors hypothesized that there was relative hyper-
polarization of the dentate granule cell membrane
potential during the still, alert state compared to
slow-wave sleep. The relative hyperpolarization
was interpreted as a gating mechanism.
Collins et al. (1983) studied the role of the dent-
ate gyrus in seizure propagation in behaving rats.
Stimulation with either focal chemoconvulsant in-
jection into or electrical stimulation of entorhinal
cortex produced a graded response. If focal
chemoconvulsant injection induced fewer than 10
spikes per minute in entorhinal cortex, no or min-
imal behavioral changes were noted, and no met-
abolic changes were noted on later sectioning and
deoxyglucose autoradiography. If 10–30 spikes per
minute are induced in entorhinal cortex, then there
is increased deoxyglucose uptake in the entorhinal
cortex and in the dentate gyrus, but restricted in
dentate gyrus to the molecular layer. Behaviors
associated with weak seizure activity were ob-
served. However, if greater than 40 spikes per
minute were induced, moderate seizures devel-
oped. Metabolic changes then spread to the entire
dentate gyrus, areas CA3 and CA1, the septal nu-
cleus, and occasionally to the amygdala, nucleus
accumbens, and ventral pallidum-lateral preoptic
area. Spread to the contralateral side also oc-
curred. Because metabolic changes were initially
restricted to the molecular layer of the dentate
gyrus, and only with further increases of chemical
or electrical stimulation were these changes able to
spread to other parts of the hippocampus and to
extrahippocampal structures, the authors con-
cluded that ‘‘the sequential changes in [deoxyglu-
cose] metabolism suggest that the dentate gyrus
acts as a restrictive gateway for seizure spread
from entorhinal cortex to the rest of the limbic
systemy.’’ (Collins et al., 1983).
Further work by Lothman and coworkers was
reviewed in 1992 (Lothman et al., 1992). Working

in vivo with unanesthetized rats, Stringer et al.
(1989) and Stringer and Lothman (1992) devel-
oped the concept of maximal dentate activation
(MDA). Their experiments involved stimulation of
the perforant pathway or area CA3 until a max-
imal, apparently saturating level of activity was
recorded in the dentate gyrus. This state was
defined as MDA. MDA was most easily obtained
with a stimulating frequency between 10–40 Hz.
The onset of MDA occurred with a pronounced
negative shift of the DC potential, reflecting a de-
polarization of the granule cell layer. In addition,
there was an abrupt rise in extracellular K+ con-
centration, and the appearance of bursts of large
amplitude population spikes. If stimulation was
triggered above the threshold for MDA, afterdis-
charges were observed, which could persist for a
short time after stimulation ceased. MDA in the
intact rat was always bilateral and associated with
synchronous epileptiform discharges in bilateral
CA3, CA1, subiculum, and entorhinal cortex.
Lesioning the entorhinal cortex on one side can
block MDA on that side, but the contralateral
side retained the ability to reach MDA. If stimu-
lation were applied to the perforant pathway,
MDA occurred in the dentate gyrus before activity
was recorded in CA1. Thus, it appeared that
transmission flowed from entorhinal cortex to
dentate gyrus to CA3 and CA1, and not by the
direct entorhinal to CA1 (temporoammonic)
pathway.
Lothman et al. (1992) and Stringer and Loth-
man (1992) interpreted their data in the following
way: ‘‘(1) MDA serves to initiate and sustain re-
verberatory seizure activity in [the] hippocam-
pal–parahippocampal loop; (2) this reverberatory
seizure activity bombards extrahippocampal struc-
tures; (3) MDA can directly (within the hippo-
campal–parahippocampal loop) and indirectly
(by influencing sites outside the hippocam-
pal–parahippocampal loop) modulate the length
of electrographic seizures and, in turn, their prop-
agation, thereby affecting the expression of vari-
ous types of behavioral seizures; (4) MDA can be
accessed from any point within the hippocam-
pal–parahippocampal loop; (5) MDA can also be
accessed from points outside this loop....’’. Thus
the dentate gyrus, when its function as a control
603
point is breached, actually acts as a ‘‘promoter’’ or
‘‘amplifier’’ of seizural discharges.
Walther et al. (1986) studied rat brain slices in
superfusate containing a low concentration of
magnesium, which was shown to induce repetitive
burst discharges in their slices. The entorhinal cor-
tex demonstrated prolonged epilepiform dis-
charges, lasting minutes at a time. The subiculum
was also capable of spontaneous discharges, last-
ing up to 9 s. In contrast, isolated minislices of area
CA3 were only capable of brief spontaneous tran-
sients, and the dentate gyrus demonstrated no ac-
tivity at all when connections to the entorhinal
cortex were disrupted. Because even prolonged
epileptiform discharges in the entorhinal cortex
elicited only brief transient activity in the dentate
gyrus, the authors suggested that ‘‘the dentate
gyrus y may serve as a filter which reduces the
excitatory load into CA3 and hence into CA1’’
(Walther et al., 1986). Further details of the phar-
macology and electrophysiology of these slices
were reviewed in Heinemann et al. (1992).
A striking visual demonstration of dentate
gating was presented by Iijima et al. (1996) using
optical imaging with fluorescence voltage-sensitive
dye in rat brain slices. After superfusing their slices
with an antagonist of GABA-A receptors, electri-
cal stimulation in the superficial layers of the
entorhinal cortex led to signals reflecting robust
excitation of the entorhinal cortex. This first
spread throughout the superficial layers of the
entorhinal cortex, then involved the deep layers.
At 33.6 ms after stimulation, excitation invaded
the hippocampus, lasting until 151.2 ms after stim-
ulation. The entorhinal cortex remained active
through this time, and activity reverberated within
the entorhinal cortex for the next 200 ms. In this
period, the hippocampus showed only weak and
partial activation. A second stimulus, delivered
352.8 ms after the first, caused further reverbera-
tory activity in the entorhinal cortex, which then
again penetrated to the hippocampus and led to
hippocampal excitation lasting about 70 ms. A
second experiment was also performed in normal
solution (without the GABA-A receptor antago-
nist), using 1 Hz repetitive stimulation instead
of single stimuli. Each stimulation resulted in
increased activity in the entorhinal cortex, but it
604
was not until the seventh stimulus that activity
penetrated to the hippocampus. These results
demonstrated that entorhinal cortex activation,
even when robust, does not easily penetrate into
the hippocampus, presumably due to gating at the
level of the entrance to the hippocampus, i.e., at
the dentate gyrus.
Behr et al. (1998) used kindled animals to dem-
onstrate the dentate gate and its ‘‘breakdown’’.
The authors first superfused entorhinal-hippocam-
pal brain slices in low-magnesium solution. Spon-
taneous seizure-like activity in both entorhinal
cortex and in area CA3 was then recorded, but
activity was significantly larger in amplitude and
longer in duration in area CA3 of kindled slices
compared to control slices. Transecting the perfo-
rant pathway greatly diminished epileptiform ac-
tivity in area CA3, but transecting the subiculum-
to-entorhinal pathway did not. These results dem-
onstrated that epileptiform activity can be passed
from entorhinal cortex through the dentate gyrus
into area CA3, and that passage is made more
likely after kindling. The authors then devised an
experiment so that only the entorhinal cortex was
locally perfused with both a GABA-A receptor
antagonist and elevated K+ to induce epileptiform
activity in a spatially-specific manner. Electrical
stimulation of the entorhinal cortex resulted in a
much stronger response in the dentate gyrus of
kindled slices. Furthermore, in a separate experi-
ment, spontaneous entorhinal interictal activity
failed to trigger epileptiform discharges in dentate
gyrus in 8 out of 8 control slices, but did trigger
them in 7 out of 9 kindled slices. Taken together,
these results show that, in normal brain, the dent-
ate gyrus appears to prevent epileptiform activity
in the entorhinal cortex from reaching the hippo-
campus, but after epileptogenesis (exemplified by
kindling), the dentate gyrus no longer functions as
a gate.
Evidence for dentate gating of seizures was also
found by monitoring the level of c-fos protein ex-
pression after the development of spontaneous
seizures in a pilocarpine model of epilepsy in mice
(Peng and Houser, 2005). The expression of c-fos
protein is a marker for neuronal activity. Increased
c-fos protein levels are evident 20–40 min after
c-fos activation, at least in many neuronal types
where it has been studied. At 15 min after a
1–2 min spontaneous behavioral seizure in epilep-
tic mice, c-fos labeling appeared in dentate granule
cells, spread throughout the entire extent of the
dentate gyrus but not involving the interneurons of
the dentate-hilar border or the dentate molecular
layer. At 30 min, c-fos staining was intense in the
dentate gyrus, involving both granule cells and
dentate-hilar border interneurons, and increased
c-fos staining also spread to the rest of the hippo-
campus. At 1–2 h, c-fos staining began to fade in
the dentate granule cells but was intense in inter-
neurons of the dentate-hilar border and the dent-
ate molecular layer. At 4 h, c-fos staining was
lighter throughout the hippocampus, including the
dentate gyrus, compared to controls. These data
suggested that dentate granule cell activation was
likely to have been an early event in spontaneous
seizures. The authors commented that the rate of
c-fos expression varies between cell types, so that
c-fos expression occurred first in dentate granule
cells and later in dentate interneurons is suggestive,
but not proof that activity in granule cells pre-
ceded that in dentate interneurons.
The breakdown of dentate gating and its pre-
sumed relationship to epileptogenesis motivated
much of the research described above, at least as
early as the work by Collins et al. (1983). It was
recognized that the dentate gyrus is normally re-
sistant to the propagation of discharges from the
entorhinal cortex. In the setting of limbic epilepsy
(i.e., temporal lobe epilepsy), the gate is thought to
be compromised, so that seizure activity from ent-
orhinal cortex is allowed into the hippocampus,
and propagated in a reverberatory cycle back to
entorhinal cortex again (Stringer and Lothman,
1992). This suggestion has led to many studies,
which have focused on reasons why the dentate
gyrus gate may ‘‘breakdown’’ in temporal lobe
epilepsy. Based primarily on animal models of
temporal lobe epilepsy, the results have suggested
that the dentate gyrus gate may breakdown be-
cause of a change in the balance of excitation
and inhibition of dentate gyrus granule cells. De-
creased inhibition of granule cells could develop
because of seizure-induced loss of GABAergic
neurons or altered expression of GABAA recep-
tors, among many other reasons (Mody et al.,

1992; Mody, 2005). Increased excitation may de-
velop because of mossy fiber sprouting, as well as
other factors (Stringer and Lothman, 1992; Sutula
et al., 1992; Jackson and Scharfman, 1996; Scharf-
man, 2004). A more complex interplay between
initial hyperexcitability followed by chronic hyper-
inhibition has also been suggested (Sloviter et al.,
2006).
The idea of dentate gate vs. filter
In summary, there is now a series of studies, which
suggest that activity in the entorhinal cortex is of-
ten halted, delayed, or diminished at the dentate
gyrus. The decreased excitability appears to be re-
lated at least in part to the strong, prolonged
dentate granule cell IPSP first described by And-
ersen et al. (1966). Further, as also found in that
study, repetitive perforant pathway stimulation for
a prolonged period of time (a few seconds or
longer) results in facilitation of succeeding stimu-
lations. Thus if the dentate gyrus is a gate, it is a
gate that can be opened if one is persistent, i.e., if
one keeps ‘‘knocking’’ on it.
Why does the dentate gate open with repeated
stimulation? Meticulous simultaneous intracellular
recordings seem to show that dentate and hilar
interneurons respond strongly and faithfully to
dentate granule cell discharges (Scharfman et al.,
1990). However, with repeated granule cell dis-
charge, the interneuronal response switch from
action potentials to EPSPs — the interneurons still
hear the command to fire but stop firing. Con-
versely, dentate and hilar interneuronal discharges
produce IPSPs in dentate granule cells, but with
many failures. Interestingly, failures are more
likely after a large IPSP. These results suggest
that at least part of the gating function resides in
the local granule cell and interneuron circuitry,
and that this inhibitory circuit is tuned down in
efficacy with repeated activation (Scharfman et al.,
1990).
Such activity-dependent disinhibition, involving
principal neurons with their local interneuronal
circuitry, appears to be an important recurring
theme in other brain areas as well (Ben-Ari et al.,
1980; Wong and Watkins, 1982; McCarren and
605
Alger, 1985; Deisz and Prince, 1989; Thompson
and Gahwiler, 1989a, b, c; Scanziani et al., 1991;
Mott and Lewis, 1992; Thomson et al., 1993). For
instance, in area CA3, repeated stimulation of py-
ramidal neurons leads to decreased IPSCs via two
mechanisms: (a) prolonged activation of GABA-A
receptors, which leads to a chloride influx into the
principal neuron, which leads to a decrease in
driving force for chloride-mediated GABAergic
inhibition, and (b) presynaptic negative feedback
of GABA onto GABA-B receptors, which leads to
decreased presynaptic GABA release (Thompson
and Gahwiler, 1989a, b, c). Further details on
complex GABAergic responses have been studied
and reviewed (Kaila, 1994; Kaila et al., 1997;
Staley, 2004).
Given that the dentate gyrus can function as a
gate, is there also a way in which the dentate gyrus
might act as a filter? The prolonged IPSP in effect
acts as a high-frequency filter, but is there a more
specific way in which the dentate gyrus might act
as a filter of information? We would like to suggest
that the dentate gyrus does indeed filter informa-
tion in a specific way. We prepare for discussion of
dentate filtering function with the following com-
ments on hippocampal anatomy, mossy cell func-
tion, the relation of translamellar inhibition and
potentiation to associative Hebbian learning, and
the role of synaptic scaling in non-associative
Hebbian competition.
A detailed review of hippocampal anatomy ap-
pears in other chapters of this volume. We note
here only that the hippocampus has a striking
lamellar structure, with the projections of the
perforant pathway, the mossy fibers, the Schaffer
collaterals and the alvear pathway, all appearing
to be on nearly a plane (‘‘lamella’’) perpendicular
to the longitudinal axis of the hippocampus
(Andersen et al., 1969, 1971; Amaral and Witter,
1989). A point to be emphasized here for the dis-
cussion below is that hilar mossy cells project
maximally onto granule cells that are septally and
temporally displaced from the mossy cells of
origin, not onto the granule cells from which the
mossy cells receive input. That is, mossy cells pro-
jections are unique in being preferentially perpen-
dicular to the plane of the lamellae (Amaral and
Witter, 1989).
606
Why do mossy cells project in this way? A de-
tailed discussion of mossy cell function appears in
a separate chapter of this volume. We summarize
the principal features and suggest its role in dent-
ate-hilar function as follows: (1) mossy cells can
mediate both translamellar inhibition as well as
potentiation; (2) translamellar potentiation as me-
diated by mossy cells is weak, and an individual
mossy cell, by itself, cannot cause a dentate gran-
ule cell to discharge; (3) optimal potentiation of
dentate granule cells requires near-simultaneous
stimulation of both the perforant and association
pathways (which we refer to as ‘‘double input’’), to
within a time interval on the order of about 5 ms.
A caveat is that the functional width of a lamella
at the level of the dentate gyrus may be as thick as
2.5 mm (Zappone and Sloviter, 2004).
We also propose that granule cells scale their
activity to pass only the most favored or most po-
tentiated dentate patterns. That is, individual
granule cells evaluate not only whether individual
synapses are favorable or not (the traditional, as-
sociative Hebbian LTP or LTD), but also whether
the number of action potentials averaged over
some period of time is too low or too high. If the
average activity of one particular neuron is too
low, all synaptic strengths of this neuron are scaled
up; and if too high, all synaptic strengths are
scaled down, so as to preserve the average activity
within some characteristic range (LeMasson et al.,
1993; Turrigiano and Nelson, 2000). Experimental
evidence for this kind of activity-dependent
homeostatic synaptic scaling in other brain areas
exists (Royer and Pare, 2003; Wierenga et al.,
2005). Hebbian systems that are not capable of
similar homeostasis of activity evolve inevitably
into a state of tonic hyperactivity or global silence
(Miller, 1996, Marder and Prinz, 2002).
As a consequence of activity-dependent synaptic
scaling, the establishment of potentiated input
patterns causes the response of a neural system to
non-potentiated patterns to be scaled down, even
in the absence of specific LTD mechanisms for the
non-potentiated patterns. That is, for synapses to
survive in competition with other synapses, it is
not enough that they not be specifically identified
as being unfavorable; synapses will nonetheless be
scaled down in strength if there are other synapses
that are systematically scaled up or potentiated.
We refer to synaptic scaling as being representative
of a type of non-associative Hebbian competition.
Thus we suggest that translamellar potentiation
be viewed in terms of Hebbian associative learn-
ing, with the additional twist that double input
from both the perforant and associative pathways
results in more effective potentiation. Indeed, in-
put from only one source, e.g., the association
pathway only, may actually result in depotentiat-
ion through Hebbian competition. To see this,
consider repetitive perforant pathway input that
arrives at dentate granule cells in a certain number
of lamellae (Fig. 1). The dentate granule cells in
these lamellae fire multiple action potentials. These
granule cells cause mossy cells downstream in the
same lamellae to fire multiple action potentials as
well. These mossy cells then send signals to many
other lamellae. Some of these other lamellae re-
ceive near-simultaneous perforant and association
pathway input, and some do not. For lamellae
that do receive near-simultaneous perforant and
association pathway input, one expects the mossy
EC
DG
MC
Fig. 1. Dentate-hilar potentiation is mediated by double input
from both entorhinal cortex neurons and from hilar mossy cells.
EC ¼ entorhinal cortex; DG ¼ dentate gyrus granule cells;
MC ¼ mossy cells of hilus. Repetitive stimulation of dentate
granule cells by entorhinal cortex neurons causes transmission
of excitation to mossy cells in the hilus. The mossy cells then
stimulate extralamellar granule cells (plus interneurons near
those granule cells). Those granule cells that receive input from
both entorhinal cortex and from mossy cells become potenti-
ated, while those that do not, become depotentiated. Potenti-
ated connections are represented by thick arrows.
Depotentiated connections are represented by dashed arrows.

cell-to-granule cell connection to be potentiated.
However, Hebbian competition then requires that
non-potentiated connections be scaled down in
strength. Thus lamellae that do not receive simul-
taneous perforant and association pathway input
will find their mossy cell input weakened.
What is the timescale for translamellar inhibi-
tion? A scaling mechanism should take place on
a timescale that is much longer than the baseline
dentate granule cell firing interval, because a scal-
ing mechanism requires monitoring and averaging
the firing rate over some period of time. One set of
experiments (Zappone and Sloviter, 2004) found
translamellar inhibition to appear on a timescale
of 200 s. A timescale this long is consistent with
a scaling mechanism, and is not consistent with
direct connectivity-related effects (i.e., disynaptic
mossy cell to basket cell to granule cell transmis-
sions).
Translamellar potentiation and inhibition can
now be put together in a consistent scheme for
system learning. Translamellar potentiation allows
associative learning, while translamellar inhibi-
tion helps maintain dynamical system stability.
Consistently successful double input from both
perforant and association pathways results in
translamellar potentiation (Steward et al., 1977;
Buzsaki and Eidelberg, 1982; Strowbridge et al.,
1992; Hetherington et al., 1994; Strowbridge and
Schwartzkroin, 1996; Kleschevnikov and Rout-
tenberg, 2003), while multiple inputs from mossy
cells without concomitant perforant pathway in-
put result in translamellar inhibition (Zappone
and Sloviter, 2004). Translamellar inhibition and
potentiation are thus complementary mechanisms,
both necessary for a stable system capable of con-
tinual learning.
Dentate-hilar filtering function: a hypothesis
Various ways in which mossy cells can help the
dentate gyrus function as a filter have been pro-
posed. Buckmaster and Schwartzkroin (1994) have
suggested a granule cell association hypothesis,
wherein the mossy cells help to link subpopulat-
ions of granule cells. They suggested that the dent-
ate-hilar role is one of pattern recognition, where
607
the role of the mossy cells is to fill in missing
components of perforant pathway input. For ex-
ample, if the dentate-hilar complex learns to rec-
ognize a pattern involving co-activation of granule
cells in lamellae A, B, and C, but later receives
perforant pathway input only at lamellae A and B,
then the mossy cells via association pathway po-
tentiation will nonetheless stimulate granule cells
in lamella C to fire. This type of pattern recogni-
tion is often referred to as ‘‘pattern completion’’.
It is tuned to be sensitive but not specific. The
mossy cells will cause granule cell co-activation in
a remembered pattern, if the input pattern is
‘‘close enough’’ to the remembered pattern. That
mossy cell projections are perpendicular to the
perforant pathway and project preferentially to
distant granule cells, sets up an ideal geometry for
the mossy cells to play an associative role. It was
not clear to these authors why associations be-
tween distant granule cells should be mediated by
a separate cell population (the mossy cells), but it
was conjectured that this arrangement allowed for
independent influences to act on granule cells and
mossy cells separately, and that nearest-neighbor
granule cell co-activations may be discouraged,
thus preventing a dangerous accretion of co-local-
ized excitation (Buckmaster and Schwartzkroin,
1994).
Alternatively, considering that granule cells are
difficult to activate while mossy cells are easily ac-
tivated, Jackson and Scharfman (1996) proposed
that mossy cells act as a switch: ‘‘By keeping
activity in the granule cells either above or below a
threshold for potentiation of synapses on pyram-
idal cells, mossy cells could create a bistable
system, and thus form a gate to control whether
or not information will be stored in the down-
stream elements of the trisynaptic circuit’’.
Other influences on the mossy cells could presum-
ably determine whether the switch is turned on or
off.
We offer yet another explanation of the dentate
filter, closer to that of Buckmaster and Schwa-
rtzkroin (1994) but differing in an important way.
Mossy cells can help bring granule cells closer
to threshold but rarely trigger granule cell action
potentials by themselves (Hetherington et al.,
1994; Scharfman, 1995; Kleschevnikov and
608
Routtenberg, 2003). Such weak mossy cell associ-
ation does not lend itself to high-sensitivity pattern
recognition, as missing components of a perforant
pathway pattern are not likely to be filled in by
mossy cell collateral input. The finding that tem-
poral to septal association is absent or very weak
in rats (Hetherington et al., 1994) would also lead
to very poor pattern completion capabilities, as
essentially there is no pattern completion capabil-
ity in the temporal to septal direction. Further-
more, association pathway potentiation, as
discussed in the previous section, is best triggered
with double input from both the perforant and
association pathways (Fig. 1). Inconsistent pattern
input may not simply be ignored, but may lead to
loss of potentiation and possibly even inhibition of
granule cell response. Thus, we agree that the
dentate-hilar complex is a pattern recognition
complex, but we propose that it represents an
unforgiving pattern recognizer, one that is specific
but not necessarily sensitive. We propose that the
dentate-hilar complex is not a pattern completer,
but a pattern excluder. Its job is to exclude input
patterns that are not exactly right.
Dentate-hilar function, in our conjecture, thus
consists of the following steps: (1) patterns are
presented to the dentate-hilar complex via repet-
itive perforant pathway input, (2) mossy cells
strengthen granule cell responses to patterns that
are repeated in a consistent and persistent way
(translamellar potentiation), and weaken random
or erratically presented patterns (translamellar
inhibition), (3) Hebbian competition, through
non-associative mechanisms, scales granule cell
firing thresholds to fire only with the most highly
potentiated pattern or patterns, (4) with future
repetitions, the most highly potentiated pattern or
patterns are allowed to pass through, while more
random patterns are blocked.
In this model, the dentate gyrus functions as
both gate and filter. It is a gate that can be opened
by persistent, repetitive stimulation, and it is a fil-
ter in that it prefers that the stimulation be con-
sistent, in terms of the pattern of the stimulation as
distributed along the longitudinal axis. That is,
one must ‘‘knock’’ on the gate not only many
times, but in nearly exactly the same way each
time. Random knocks are ignored.
What is the optimal frequency for opening the
dentate gate? Comparing repetitive stimulation at
0.5, 4, 7, 10, and 20 Hz, Andersen et al. (1966)
found that 10 Hz was optimal. Comparing repet-
itive stimulation at 0.5, 5, and 100 Hz, Mott and
Lewis (1992) found that 5 Hz was optimal. The
frequencies 5–10 Hz fall in the theta–alpha band,
which is known to be prominent in limbic struc-
tures including the hippocampus (Bland, 1986;
Freund and Buzsaki, 1996; Buzsaki, 2002; Buzsaki
et al., 2003). We therefore conjecture that theta
oscillations indicate activity requiring opening of
the dentate gate.
We comment that the dentate-hilar complex
combines a sluggish but powerful excitatory
source (the dentate granule cells) with a highly la-
bile but weaker component (the mossy cells). The
sluggishness of the granule cells and the weakness
of mossy cell output allow the granule cells to
maintain high specificity, but the lability of the
mossy cells nonetheless allows highly responsive
associative learning. This dual need for specificity
and association may explain why dentate-hilar as-
sociation is mediated by a specialized cell popula-
tion (the mossy cells), while in CA3 and in the
neocortex, association occurs directly between
principal cells. The dentate-hilar complex is unique
in being tuned for specificity.
A look ahead
Our conjecture for dentate-hilar function is spec-
ulative, but testable. The simplest type of exper-
iment would be to monitor all EPSPs from the
dentate granular layer in one lamella, and to cal-
culate the initial slope of each EPSP, denoted E in
units of volts per second. One would also keep
track of which E’s result in population spikes.
Then a distribution function can be constructed,
G(E), giving the probability of observing each
value of E (Fig. 2, filled squares). One can also
determine the threshold E0, defined as the smallest
E above which a population spike is likely (e.g.,
with a likelihood greater than 95%). Alternatively,
one can define various threshold functions with
parameters that can be extracted from experiment.
For instance, one can hypothesize a threshold

0.030
0.025
0.020
G(E) and Eo
0.015
0.010
0.005
0.000
0 2
Fig. 2. Hypothetical distributions of G(E) and E0 with E in arbitrary units. Filled squares: baseline G(E). The single solid square at E
¼ 5.28 represents the threshold E0 such that 95% of the distribution has EoE0. Open triangles: hypothetical G(E) after potentiation.
The threshold is at E ¼ 6.72. Open circles: hypothetical G(E) in chronic epilepsy. The threshold is at E ¼ 6.00.
function of the form
PðEÞ ¼ A expðBðE
E 0 ÞÞ,
where P(E) is the probability that a given E results
in a population spike. Here A is a normalization
constant, E0 is the firing threshold for E, and B
controls how sharp is the transition to firing. The
parameters B and E0 can be extracted from exper-
iment, by plotting log P(E) vs. E. The idea of an
adjustable firing threshold E0 is key to our model
for dentate-hilar function.
What happens to G(E) and E0 after potentiat-
ion? If one stimulates the perforant pathway re-
petitively at a frequency fS with a certain number
of repetitions NR, one expects potentiation of the
dentate response to future stimulations. The ideal
frequency for potentiation may be 5–10 Hz for
6–9 s (Andersen et al., 1966; Mott and Lewis,
1992), as discussed above. The relevant E to mon-
itor may be the median or maximal value over the
NR repetitions. After potentiation, one may repeat
the procedure for constructing G(E), and see how
this distribution function has changed. One may
hypothesize that G(E) develops a new peak at
609
4 6 8 10
E
higher E, representing potentiated EPSPs, with
threshold E0 between this new peak and the old
peak (Fig. 2, open triangles). EPSPs from the new
high-E peak are passed by the dentate gyrus, while
those from the old peak are blocked. The more
distinct is this new peak, the easier it becomes to
exclude incorrect patterns. Small fluctuations in
the value of E0 would not greatly affect specificity.
What happens to G(E) and E0 in chronic epi-
lepsy? With loss of mossy cells, one might hypoth-
esize that it becomes more difficult to create a
distinct high-E peak. One might hypothesize, for
instance, that only a high-E shoulder is created
(Fig. 2, open circles). The function of the dentate-
hilar complex as a pattern excluder would thus be
degraded. The threshold E0, furthermore, would
have to be placed on a steeper part of the curve for
G(E). Any slight fluctuation of E0 would cause
dentate activity either to be overly inhibited
(E0 too high) or overly excitable (E0 too low).
The effect of recurrent excitatory mossy fiber
collaterals (reviewed in Chapter 29 by Sutula and
Dudek in this volume) on G(E) and E0 should
also be very interesting. One might expect either a
610
high-E shoulder or a separate higher-E peak, sim-
ilar to that seen due to potentiation in the normal
state discussed above. However, unlike the high-E
contribution mediated by mossy cells, the high-E
contribution from recurrent mossy fiber collaterals
carries no useful information, because the associ-
ated EPSPs are the result of local recurrent exci-
tations. Furthermore, if the high-E contributions
from recurrent mossy fiber collaterals and from
mossy cells overlap, then the filtering function of the
dentate-hilar complex may be severely degraded.
A more ambitious experimental goal would be
to determine the functional width of a lamella, and
to develop techniques to stimulate and to record
from individual lamellae reliably. This goal is
likely to be technically challenging. The simplest
alternative would be to place a single stimulating
electrode into the angular bundle, one in each
hemisphere, and a single recording electrode into
the dentate granular layer, again one in each hem-
isphere. Presumably, one is guaranteed in this ar-
rangement to have one pair of stimulating and
recording electrodes in each of two distinct, non-
overlapping lamellae (one in each hemisphere).
However many distinct lamellae are accessible
to experiment, one may then stimulate a subset of
them simultaneously and repetitively, at a certain
frequency of repetition for a certain number of
repetitions, NR. This frequency may again be
taken in the range of 5–10 Hz (Andersen et al.,
1966; Mott and Lewis, 1992). The spatial pattern
of the stimuli represents the pattern to be learned.
A distribution function G(E) can then be con-
structed for this system, with one G(E) for each
lamella.
Of interest would be the number of repetitions,
NR, needed to teach a given target pattern, and
whether the train of NR repetitions need to be re-
peated a certain number of times. After potentiat-
ion of the target pattern is achieved, a test of
sensitivity would be to present, simultaneously, the
target spatial pattern plus a random pattern of
variable amplitude, and then see if the random
component is blocked while the target pattern is
allowed through. A test of specificity would be to
present only a part of the target spatial pattern,
and see how close the presented pattern has to be
to the target pattern to be passed through.
The stimulation strength necessary to produce
EPSPs and population spikes is also of interest.
One expects a threshold effect for the stimulation
strength S (in units of volts), wherein a minimal
value of this necessary before population spikes
are seen. A distribution function can be defined,
D(E,S), giving the probability of observing a given
E and S. It would be of interest to know what
happens to D(E,S), E0 and S0 after potentiation,
and in the context of chronic epilepsy.
Finally, if it turns out to be true that the dent-
ate-hilar complex is a pattern excluder, one may
then employ similar arguments as developed in this
chapter to speculate on the function of the auxil-
iary pathways, e.g., the direct pathways from ent-
orhinal cortex to area CA3 (Hjorth-Simonsen and
Jeune, 1972; Steward and Scoville, 1976; Witter
and Amaral, 1991). One possible function for the
auxiliary pathways may be to help validate signal
passed into the hippocampus. We discuss this
point at a little greater length below.
By the activity-dependent homeostasis hypoth-
esis (LeMasson et al., 1993; Turrigiano and Nel-
son, 2000), all principal neurons have a preferred
target firing rate, with homeostatic mechanisms to
return to this rate over some period of time if per-
turbed away from it. If we assume that this hy-
pothesis applies to granule cells, then there must
be some rate at which granule cells fire action
potentials spontaneously, even in the absence of
perforant pathway stimulation. This rate is low
but cannot be zero. How can CA3 principal neu-
rons downstream from granule cells know if the
signals they receive from granule cells are due to
spontaneous granule cell activity or due to perfo-
rant pathway stimulated activity? There should
be some way to ignore the inevitable (if rare)
spontaneous granule cell discharge, while not com-
promising a faithful response to legitimate, stim-
ulated granule cell discharge.
The answer may be that CA3 principal neurons
have their own G(E) distribution function, and
they scale their firing thresholds to fire only with
the most highly potentiated inputs. Thus if a set of
CA3 principal neurons have been trained to expect
‘‘double input’’ from granule cells and from ent-
orhinal neurons via the auxiliary entorhinal-to-
CA3 pathway, then these CA3 principal neurons

are less likely to fire if they receive input only from
granule cells. Thus, auxiliary pathways may play a
crosschecking role, validating information arriving
via the main pathways. If a confirmatory signal
does not arrive via an auxiliary pathway, then in-
formation arriving via the main pathway may be
ignored. The additional input from the auxiliary
pathway may be needed to push a principal neuron
above the firing threshold.
In summary, the current wealth of experimental
data on the dentate gyrus shows that the dentate
gyrus does function as a gate. We further conjec-
ture that it also functions as a highly specific pat-
tern recognizer, or filter. The data to date do not
directly address this conjecture. We suggest future
experiments that may help to prove or disprove the
filtering conjecture. Even if the specifics of our
conjecture are wrong, we hope these experiments
will deepen our insight into the structure and
function of the dentate gyrus and hippocampus.
Acknowledgments
I am grateful to Drs. Helen Scharfman and
Thomas Sutula for helpful advice and for the op-
portunity to write this chapter. I also thank the
American Epilepsy Society for support, and the
National Institutes of Health and National Center
for Research Resources K12 Roadmap, Project
number 8K12RR023268-02.
References
Alger, B.E. and Teyler, T.J. (1976) Long-term and short-term
plasticity in the CA1, CA3, and dentate regions of the rat
hippocampal slice. Brain Res., 110(3): 463–480.
Amaral, D.G. and Witter, M.P. (1989) The three-dimensional
organization of the hippocampal formation: a review of an-
atomical data. Neuroscience, 31(3): 571–591.
Andersen, P., Bliss, T.V., et al. (1969) Lamellar organization of
hippocampal excitatory pathways. Acta Physiol. Scand.,
76(1): 4A–5A.
Andersen, P., Bliss, T.V., et al. (1971) Lamellar organization of
hippocampal pathways. Exp. Brain Res., 13(2): 222–238.
Andersen, P., Holmqvist, B., et al. (1966) Entorhinal activation
of dentate granule cells. Acta Physiol. Scand., 66(4): 448–460.
Behr, J., Lyson, K.J., et al. (1998) Enhanced propagation of
epileptiform activity through the kindled dentate gyrus. J.
Neurophysiol., 79(4): 1726–1732.
611
Ben-Ari, Y., Krnjevic, K., et al. (1980) Lability of synaptic
inhibition of hippocampal pyramidal cells. J. Physiol., 298:
36P–37P [proceedings].
Bland, B.H. (1986) The physiology and pharmacology of hip-
pocampal formation theta rhythms. Prog. Neurobiol., 26(1):
1–54.
Buckmaster, P.S. and Schwartzkroin, P.A. (1994) Hippocampal
mossy cell function: a speculative view. Hippocampus, 4(4):
393–402.
Buzsaki, G. (2002) Theta oscillations in the hippocampus.
Neuron, 33(3): 325–340.
Buzsaki, G., Buhl, D.L., et al. (2003) Hippocampal network
patterns of activity in the mouse. Neuroscience, 116(1):
201–211.
Buzsaki, G. and Eidelberg, E. (1982) Direct afferent excitation
and long-term potentiation of hippocampal interneurons. J.
Neurophysiol., 48(3): 597–607.
Cohen, A.S., Lin, D.D., et al. (2003) Dentate granule cell
GABA(A) receptors in epileptic hippocampus: enhanced
synaptic efficacy and altered pharmacology. Eur. J. Neuro-
sci., 17(8): 1607–1616.
Collins, R.C., Tearse, R.G., et al. (1983) Functional anatomy of
limbic seizures: focal discharges from medial entorhinal cor-
tex in rat. Brain Res., 280(1): 25–40.
Coulter, D.A. (1999) Chronic epileptogenic cellular alterations
in the limbic system after status epilepticus. Epilepsia,
40(Suppl 1): S23–S33 discussion S40–S41.
Deisz, R.A. and Prince, D.A. (1989) Frequency-dependent de-
pression of inhibition in guinea-pig neocortex in vitro by
GABAB receptor feed-back on GABA release. J. Physiol.,
412: 513–541.
Freund, T.F. and Buzsaki, G. (1996) Interneurons of the hip-
pocampus. Hippocampus, 6(4): 347–470.
Fricke, R.A. and Prince, D.A. (1984) Electrophysiology of
dentate gyrus granule cells. J. Neurophysiol., 51(2): 195–209.
Heinemann, U., Beck, H., et al. (1992) The dentate gyrus as a
regulated gate for the propagation of epileptiform activity.
Epilepsy Res. Suppl., 7: 273–280.
Hetherington, P.A., Austin, K.B., et al. (1994) Ipsilateral asso-
ciational pathway in the dentate gyrus: an excitatory feed-
back system that supports N-methyl-D-aspartate-dependent
long-term potentiation. Hippocampus, 4(4): 422–438.
Hjorth-Simonsen, A. and Jeune, B. (1972) Origin and termina-
tion of the hippocampal perforant path in the rat studied by
silver impregnation. J. Comp. Neurol., 144(2): 215–232.
Iijima, T., Witter, M.P., et al. (1996) Entorhinal-hippocampal
interactions revealed by real-time imaging. Science,
272(5265): 1176–1179.
Jackson, M.B. and Scharfman, H.E. (1996) Positive feedback
from hilar mossy cells to granule cells in the dentate gyrus
revealed by voltage-sensitive dye and microelectrode record-
ing. J. Neurophysiol., 76(1): 601–616.
Kaila, K. (1994) Ionic basis of GABAA receptor channel func-
tion in the nervous system. Prog. Neurobiol., 42(4): 489–537.
Kaila, K., Lamsa, K., et al. (1997) Long-lasting GABA-medi-
ated depolarization evoked by high-frequency stimulation in
pyramidal neurons of rat hippocampal slice is attributable to
612
a network-driven, bicarbonate-dependent K+ transient. J.
Neurosci., 17(20): 7662–7672.
Kleschevnikov, A.M. and Routtenberg, A. (2003) Long-term
potentiation recruits a trisynaptic excitatory associative net-
work within the mouse dentate gyrus. Eur. J. Neurosci.,
17(12): 2690–2702.
LeMasson, G., Marder, E., et al. (1993) Activity-dependent
regulation of conductances in model neurons. Science,
259(5103): 1915–1917.
Lothman, E.W., Stringer, J.L., et al. (1992) The dentate gyrus
as a control point for seizures in the hippocampus and
beyond. Epilepsy Res. Suppl., 7: 301–313.
Marder, E. and Prinz, A.A. (2002) Modeling stability in neuron
and network function: the role of activity in homeostasis.
Bioessays, 24: 1145–1154.
McCarren, M. and Alger, B.E. (1985) Use-dependent depres-
sion of IPSPs in rat hippocampal pyramidal cells in vitro.
J. Neurophysiol., 53(2): 557–571.
Miller, K.D. (1996) Synaptic economics: competition and
cooperation in synaptic plasticity. Neuron, 17: 371–374.
Mody, I. (2005) Aspects of the homeostaic plasticity of GAB-
AA receptor-mediated inhibition. J. Physiol., 562(Pt 1):
37–46.
Mody, I., Otis, T.S., et al. (1992) The balance between excita-
tion and inhibition in dentate granule cells and its role in
epilepsy. Epilepsy Res. Suppl., 9: 331–339.
Mott, D.D. and Lewis, D.V. (1992) GABAB receptors mediate
disinhibition and facilitate long-term potentiation in the
dentate gyrus. Epilepsy Res. Suppl., 7: 119–134.
Nusser, Z. and Mody, I. (2002) Selective modulation of tonic
and phasic inhibitions in dentate gyrus granule cells. J.
Neurophysiol., 87(5): 2624–2628.
Peng, Z. and Houser, C.R. (2005) Temporal patterns of fos
expression in the dentate gyrus after spontaneous seizures in
a mouse model of temporal lobe epilepsy. J. Neurosci.,
25(31): 7210–7220.
Royer, S. and Pare, D. (2003) Conservation of total synaptic
weight through balanced synaptic depression and potentiat-
ion. Nature, 422(6931): 518–522.
Scanziani, M., Gahwiler, B.H., et al. (1991) Paroxysmal inhib-
itory potentials mediated by GABAB receptors in partially
disinhibited rat hippocampal slice cultures. J. Physiol., 444:
375–396.
Scharfman, H.E. (1992) Differentiation of rat dentate neurons
by morphology and electrophysiology in hippocampal slices:
granule cells, spiny hilar cells and aspiny ‘fast-spiking’ cells.
Epilepsy Res. Suppl., 7: 93–109.
Scharfman, H.E. (1995) Electrophysiological evidence that
dentate hilar mossy cells are excitatory and innervate both
granule cells and interneurons. J. Neurophysiol., 74(1):
179–194.
Scharfman, H.E. (2004) Functional implications of seizure-in-
duced neurogenesis. Adv. Exp. Med. Biol., 548: 192–212.
Scharfman, H.E., Kunkel, D.D., et al. (1990) Synaptic connec-
tions of dentate granule cells and hilar neurons: results of
paired intracellular recordings and intracellular horseradish
peroxidase injections. Neuroscience, 37(3): 693–707.
Sloviter, R.S., Zappone, C.A., et al. (2006) Kainic acid-induced
recurrent mossy fiber innervation of dentate gyrus inhibitory
interneurons: possible anatomical substrate of granule cell
hyperinhibition in chronically epileptic rats. J. Comp.
Neurol., 494(6): 944–960.
Staley, K.J. (2004) Role of the depolarizing GABA response in
epilepsy. Adv. Exp. Med. Biol., 548: 104–109.
Staley, K.J., Otis, T.S., et al. (1992) Membrane properties of
dentate gyrus granule cells: comparison of sharp microelec-
trode and whole-cell recordings. J. Neurophysiol., 67(5):
1346–1358.
Stell, B.M. and Mody, I. (2002) Receptors with different affin-
ities mediate phasic and tonic GABA(A) conductances in
hippocampal neurons. J. Neurosci., 22(10): RC223.
Steward, O. and Scoville, S.A. (1976) Cells of origin of ent-
orhinal cortical afferents to the hippocampus and fascia
dentata of the rat. J. Comp. Neurol., 169(3): 347–370.
Steward, O., White, W.F., et al. (1977) Potentiation of the
excitatory synaptic action of commissural, associational and
entorhinal afferents to dentate granule cells. Brain Res.,
134(3): 551–560.
Stringer, J.L. and Lothman, E.W. (1992) Reverberatory seizure
discharges in hippocampal-parahippocampal circuits. Exp.
Neurol., 116(2): 198–203.
Stringer, J.L., Williamson, J.M., et al. (1989) Induction of
paroxysmal discharges in the dentate gyrus: frequency
dependence and relationship to afterdischarge production.
J. Neurophysiol., 62(1): 126–135.
Strowbridge, B.W., Buckmaster, P.S., et al. (1992) Potentiation
of spontaneous synaptic activity in rat mossy cells. Neurosci.
Lett., 142(2): 205–210.
Strowbridge, B.W. and Schwartzkroin, P.A. (1996) Transient
potentiation of spontaneous EPSPs in rat mossy cells induced
by depolarization of a single neurone. J. Physiol., 494(Pt 2):
493–510.
Sutula, T.P., Golarai, G., et al. (1992) Assessing the functional
significance of mossy fiber sprouting. Epilepsy Res. Suppl., 7:
251–259.
Teyler, T.J. and Alger, B.E. (1976) Monosynaptic habituation
in the vertebrate forebrain: the dentate gyrus examined in
vitro. Brain Res., 115(3): 413–425.
Thompson, S.M. and Gahwiler, B.H. (1989a) Activity-depend-
ent disinhibition. I. Repetitive stimulation reduces IPSP driv-
ing force and conductance in the hippocampus in vitro.
J. Neurophysiol., 61(3): 501–511.
Thompson, S.M. and Gahwiler, B.H. (1989b) Activity-depend-
ent disinhibition. II. Effects of extracellular potassium,
furosemide, and membrane potential on ECl- in hippocam-
pal CA3 neurons. J. Neurophysiol., 61(3): 512–523.
Thompson, S.M. and Gahwiler, B.H. (1989c) Activity-depend-
ent disinhibition. III. Desensitization and GABAB receptor-
mediated presynaptic inhibition in the hippocampus in vitro.
J. Neurophysiol., 61(3): 524–533.
Thomson, A.M., Deuchars, J., et al. (1993) Single axon exci-
tatory postsynaptic potentials in neocortical interneurons
exhibit pronounced paired pulse facilitation. Neuroscience,
54(2): 347–360.

Turrigiano, G.G. and Nelson, S.B. (2000) Hebb and home-
ostasis in neuronal plasticity. Curr. Opin. Neurobiol., 10(3):
358–364.
Walther, H., Lambert, J.D., et al. (1986) Epileptiform activity
in combined slices of the hippocampus, subiculum and
entorhinal cortex during perfusion with low magnesium
medium. Neurosci. Lett., 69(2): 156–161.
Wierenga, C.J., Ibata, K., et al. (2005) Postsynaptic expression
of homeostatic plasticity at neocortical synapses. J. Neuro-
sci., 25(11): 2895–2905.
Williamson, A., Spencer, D.D., et al. (1993) Comparison be-
tween the membrane and synaptic properties of human and
rodent dentate granule cells. Brain Res., 622(1-2): 194–202.
Winson, J. and Abzug, C. (1977) Gating of neuronal
transmission in the hippocampus: efficacy of transmis-
613
sion varies with behavioral state. Science, 196(4295):
1223–1225.
Winson, J. and Abzug, C. (1978) Neuronal transmission
through hippocampal pathways dependent on behavior.
J. Neurophysiol., 41(3): 716–732.
Witter, M.P. and Amaral, D.G. (1991) Entorhinal cortex of the
monkey: V. Projections to the dentate gyrus, hippocampus,
and subicular complex. J. Comp. Neurol., 307(3): 437–459.
Wong, R.K. and Watkins, D.J. (1982) Cellular factors influ-
encing GABA response in hippocampal pyramidal cells.
J. Neurophysiol., 48(4): 938–951.
Zappone, C.A. and Sloviter, R.S. (2004) Translamellar
disinhibition in the rat hippocampal dentate gyrus after
seizure-induced degeneration of vulnerable hilar neurons.
J. Neurosci., 24(4): 853–864.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 33

Role of the dual entorhinal inputs to hippocampus:

a hypothesis based on cue/action (non-self/self)
couplets

John E. Lisman

Department of Biology and Volen Center for Complex Systems, Brandeis University, Waltham, MA 02454, USA

Abstract: The hippocampus sits at the highest level of memory processing circuits and receives two major
inputs, one coming from the lateral entorhinal cortex and one coming from the medial entorhinal cortex.
This duality must be of fundamental importance, but its functional meaning remains unclear. A compu-
tational model used for robot navigation (Verschure, P.F., et al. (2003). Nature, 425: 620–624) has a dual
information structure that may provide insight. In this model, information is stored as couplets consisting
of information about the current sensory cues and information about the current action of the robot.
Sequences of such couplets are stored in a short-term memory buffer and transferred to a long-term
memory store whenever a goal is found. The overall system enhances the ability of the robot to find reward
sites because stored sequences enable the robot to retrace the path to a goal site whenever any of the cues
along the path to a goal is subsequently encountered. A review of the literature suggests that the idea of cue/
action couplets can be usefully mapped onto the function of the entorhinal cortex. Cue information may be
supplied by the lateral entorhinal cortex whereas action (motor) information may be supplied by the medial
entorhinal cortex. However, given that self-position information is prominent in the medial pathway and
that this is not directly related to action, a modified formulation of the duality is proposed in which the
fundamental distinction is between information about non-self vs. information about self. According to this
view, the lateral entorhinal pathway carries information about external (non-self) cues and their positions
(in egocentric coordinates) whereas the medial entorhinal pathway carries information about the organism
itself, including its position (in allocentric coordinates), motor actions and goals.

Keywords: short-term memory; reward; navigation; sequence memory; dopamine

Overview and medial regions. These provide excitatory input

The duality (Fig. 1) of the connections from the
entorhinal cortex to the hippocampus is anatom-
ically striking (Burwell et al., 1995). The entorhinal
cortex contains two distinct regions, the lateral
Corresponding author. Tel.: +1 781 736 3145 or 3148;
Fax: +1 781 736 3107; E-mail: Lisman@brandeis.edu
to several parts of the hippocampus, but the du-
ality of their structure is most obvious in the dent-
ate gyrus. The outermost dendritic region of
dentate granule cells receives exclusive input from
the lateral entorhinal cortex whereas the middle
dendritic region receives exclusive input from the
medial entorhinal cortex. Each granule cell re-
ceives both types of inputs, thereby integrating the

DOI: 10.1016/S0079-6123(07)63033-7 615

616

Fig. 1. Wiring diagram of the hippocampus (interneurons excluded) showing dual inputs from the lateral and medial entorhinal cortex.
The layer 2 cortical inputs to the dentate and CA3 diverge fan out (f) widely over these networks and then provide convergent input to
individual dentate granule cells. In contrast, the layer 3 cortical inputs are specialized for individual subregions of CA1. This is an
example of a point-to-point (p-p) connection. Within the hippocampus, granule cells provide input, via mossy fibers to the mossy cells
of the dentate and to CA3 cells. CA3 cells make feedback connections to themselves and to the mossy cells of the dentate. These, in
turn, provide excitatory input to granule cells in the inner third of the granule cell dendritic tree. (See Color Plate 33.1 in color plate
section.)

two lines of cortical information. The purpose of
this review is to discuss the possible functional
basis of this mysterious duality.
A priori, what are the grand dualities that might
be considered? Here is a list of some possibilities:
what/where; specific/context; sensory/motor; past/
present; conscious/unconscious; rewarded/punished;
stimulus/response.
My interest in the last of these, stimulus/re-
sponse, was stimulated by a paper on robotic con-
trol (Verschure et al., 2003). In that paper, the
authors describe a ‘‘brain-like’’ computer program
that enables a robot to efficiently find sites at
which reward is located. The authors postulated
several levels of control. At the lowest levels, cir-
cuits support classical conditioning. In this way,
the robot learns to associate a previously neutral
stimulus with a reward that is close to the robot.
But the robot becomes much more efficient at
finding reward sites if there are additional circuits.
These make it possible for a cue that is far away
from a reward site to specify a complex path that,
according to previous experience, led to the reward
site. The circuitry that allows the robot to do this
(Fig. 2) involves a ‘‘cortical’’ multi-item ‘‘circular’’
short-term memory (STM) buffer that stores the
last five salient events, and a long-term ‘‘hippo-
campal’’ memory that stores sequences of events in
long-term memory (LTM). When the robot for-
ages and accidentally finds a reward, it incorpo-
rates the entire content of the buffer into its LTM
store. Importantly, each event is defined as a stim-
ulus/response couplet (Verschure and Voegtlin,
1998; Verschure et al., 2003). The ‘‘stimulus’’ part
of the couplet corresponds to a prototype of the
sensations from the external world at a particular
 617

Fig. 2. Diagram of circuits responsible for storing events that led to the discovery of goals according to the model of Verschure et al.
(2003). Stimulus (C) and motor action (A) for each event are stored in a circular (first in, first out) short-term memory buffer. This
multi-item buffer can store five C/A couplets. When a goal is reached, the sequence in the short-term memory buffer is transferred to
the long-term memory buffer together with a last item representing the goal (G). As illustrated, the long-term memory store has stored
six sequences.

location. The response part of the couplet corre-
sponds to the motor act that was executed to get
from that cue to the next location (cue) in the se-
quence. ‘‘Stimulus’’ and ‘‘response’’ have a partic-
ular meaning in classical conditioning that is not
appropriate in the current context; I will therefore
use the terminology cue/action (C/A) to describe a
couplet. Thus the utilization of stored couplets can
be described as follows. If the robot comes across a
previously encountered cue, Cn, it can get to the
reward site using a simple algorithm: execute the
action An, that is associated with Cn in LTM; when
Cn+1 is found, execute An+1, etc.
Several comments are in order. First, Verschure
(personal communication) was not aware of the
dual inputs to the hippocampus; the dual repre-
sentation was utilized simply because it enhanced
the robot’s efficiency at finding goals. Second the
use of a sequence of cue-driven actions to find a
goal site is an experimentally observed mode of
animal navigation (Collett et al., 2003) and is re-
lated to ideas incorporated into previous rein-
forcement driven models of behavior (Barto and
Sutton, 1981; Hasselmo, 2005). This form of nav-
igation is less flexible than map-based navigation,
which may have developed later in the evolution.
Importantly, the use of C/A sequences to specify
the route to an accidentally discovered reward site
absolutely relies on one-trial learning, a property
that remains a defining property of human hippo-
campal episodic memory. Indeed, the development
of a multi-item STM buffer may have been the
crucial evolutionary change that made one-trial
learning possible rather than a change in synaptic
plasticity mechanisms themselves. This is because
such buffers capture one-trial, brief events and
then, upon command, provide the repetitive activ-
ity patterns to LTM networks that are necessary to
produce stable synaptic modifications.
In the following sections I will address two
questions:
1. Is the idea of storing C/A couplets potentially
applicable to understanding the dual cortical
inputs to the hippocampus? In particular, is
there any way of mapping this duality onto
618
the known properties of the medial and lat-
eral entorhinal inputs to the hippocampus?
2. What is the status of the evidence for the
various building blocks postulated in the
model of Fig. 2?
Evaluation of the cue/action (C/A) couplet
hypothesis
Evidence that the hippocampus receives action
(motor) and cue (sensory) information
The idea that the hippocampus might be influ-
enced by the motor system is one that is not com-
monly discussed, but support for this idea is
actually quite long-standing. Vanderwolf (1969)
observed that the frequency and amplitude of hip-
pocampal theta oscillations is modulated by the
speed of a rat [reviewed in Bland and Oddie
(2001)]. Other work shows the influence of motor
set and running speed on place cell firing
(McNaughton et al., 1983; Foster et al., 1989;
Wiener et al., 1989). Ranck (1973) found hippo-
campal neurons that fired on ‘‘approach’’ to
objects. More recent work provides direct evi-
dence that ‘‘motor intent’’ influences hippocampal
function. Specifically, when rats run in a T-maze,
the firing of neurons that occurs when the rat is in
the vertical stem of the T can be different depend-
ing on which direction the rat will turn at the top
of the stem (Frank et al., 2000; Wood et al., 2000;
Bower et al., 2005). Because the sensory stimula-
tion in the stem is always identical, it is hard to
escape the conclusion that goals or action are what
cause the difference in firing. It might be argued
that what is represented is the goal, in some ab-
stract sense not related to the motor action needed
to get to goal, but this is not the case. In exper-
iments using a plus maze, rats were started in the
north or south arm and had to reach a goal in the
east arm; when the rat reached the east arm, many
cells fired differentially depending on whether
the approach was from the north or south
(Ferbinteanu and Shapiro, 2003). Thus, although
a common goal was involved, information about
the specific path taken was encoded. Recent results
emphasize that pathway-specific firing may occur
only when the behavioral paradigm has conse-
quences for reward (Smith and Mizumori, 2006).
The influence of goals on hippocampal responses
is not limited to rats, but is also evident in humans
(Ekstrom et al., 2003).
There is also strong experimental support for
the entry of sensory-specific information into the
hippocampus. Odor-specific cells are found in the
rat hippocampus (Wood et al., 1999), and neurons
that are scene-dependent have been observed in
monkey hippocampus (Wirth et al., 2003). Rat
hippocampal neurons become sensitive to tones
after aversive conditioning (Berger et al., 1976;
Moita et al., 2003). Particularly clear evidence for
cells responsive to particular faces has been
obtained in humans (Quiroga et al., 2005).
It thus seems clear that both motor and sensory
information can get to the hippocampus. Moreo-
ver, particularly important from the perspective of
the hypothesis developed here is the finding of cells
in both rat (Ranck, 1973) and monkey (Wirth
et al., 2003) that jointly encode sensory and motor
information. For instance, in the monkey work,
particular pictures are rewarded for movements in
a particular direction; as learning becomes appar-
ent, some hippocampal neurons fire selectively
when a particular picture is presented and the
monkey performs the correct learned movement.
Evidence that the hippocampus is necessary for
sensory/motor associations
An important series of experiments has addressed
whether the hippocampus is necessary for learning
the association between sensory cues and motor
responses. It was found that lesions to the fornix, a
major input/output tract of the hippocampus, pre-
vent a monkey from learning which direction to
move in response to a sensory cue (Rupniak and
Gaffan, 1987), and interferes with the ability of a
monkey to report the previous movement that it
made (Gaffan, 1985). Recent work (Brasted et al.,
2005) extends these observations by showing that
fornix lesions prevent learning of arbitrary sen-
sory/motor tasks in single trials, the fast learning
that is required to enhance foraging (see above).
Related work on instrumental conditioning in rats
 619

demonstrates a similar requirement for hippocam- interfere with place learning (Ferbinteanu et al.,
pal involvement (Corbit and Balleine, 2000). 1999).
Taken together, these observations indicate that
hippocampal cells code for much more than po-
Validity of the C/A duality
sition and are influenced by the goal and actions
required to get there. Moreover, the association of
In mapping the C/A duality onto the entorhinal
these two types of information in the hippocampus
cortex, we would have to suppose that the lateral
appears necessary for behavioral responses based
region is sensory (cue) and that the medial region
on this association. I now turn to the issue of
represents ‘‘action’’. However, the most abundant
which region of the entorhinal cortex supplies each
cells in the medial region are grid cells, represent-
type of information.
ing the position of the rat in the environment, and
such information does not seem directly tied
to action. At best one could argue that motor in-
Sensory properties of the lateral entorhinal cortex
formation may be used to compute position
(McNaughton et al., 2006); the computation is
The perirhinal cortex, often termed inferotemporal
not dramatically altered when important sensory
cortex, is involved in high-level aspects of object
cues, such as vision, are taken away, probably
recognition (Brown and Aggleton, 2001). This re-
because a path integration mechanism can com-
gion provides strong input to the lateral entorhinal
pute position using head direction and velocity
cortex, but little to the medial entorhinal (Burwell,
information, both of which are motor-related.
2000). Physiological evidence for sensory input to
A significant number of layer 3 cells are direction-
the lateral entorhinal cortex is its responsiveness to
dependent, providing additional suggestive evi-
stimulation of olfactory cortex in rat (Gnatkovsky
dence that goal/motor information is present in
et al., 2004) and the presence of novelty-sensitive
the medial entorhinal cortex (Sargolini et al.,
responses to visually presented objects in monkey
2006). Thus, to some extent, positional cells could
(Fahy et al., 1993). Thus, the idea that the lateral
be described as related to action, but this argument
perforant pathway is sensory driven has experi-
is not compelling.
mental support. Consistent with a sensory role,
Another major theory of the duality (what/where)
lesions of the hippocampal inputs from the lateral
also faces difficulties. If one views grid cells as pri-
entorhinal cortex produces decreased investigation
marily representing ‘‘where’’, it at first seems sen-
of novel objects (Myhrer, 1988).
sible to conclude (Hargreaves et al., 2005) that the
dual inputs to the hippocampus form a what/where
couplet, not a C/A couplet. However, does the
Motor/goal properties of the medial entorhinal
‘‘where’’ in this formulation refer to the organism
cortex
or to objects in the environment? The brain needs to
keep track of both, and it is unclear how this would
In contrast to the cells of the lateral entorhinal
be done in the context of a what/where duality.
cortex, which do not have spatial properties
(Hargreaves et al., 2005), layers 2 and 3 of the
medial entorhinal cortex contain grid cells with Cue/action as a special case of a non-self/self
robust spatial properties (Fyhn et al., 2004; Hafting duality
et al., 2005; Sargolini et al., 2006). These cells are
coded in an allocentric coordinate system (i.e. with To deal with these difficulties, I suggest a refor-
respect to an absolute reference frame in the envi- mulation of the duality as non-self information
ronment, as e.g. north/south). Consistent with the vs. self-information. According to this view, the
presence of such allocentric spatial information, lateral entorhinal pathway would carry information
lesions of the hippocampal inputs from the medial about what is in the environment (non-self) and
entorhinal (but not the lateral entorhinal cortex) the ‘‘where’’ of those objects. The coding of the
620

‘‘where’’ component would presumably be done in transfer the buffer content to the LTM store when a
the egocentric (relative to the observer) coordinates reward site is found. I will now review what is
of the parietal lobe. The medial entorhinal pathway known about the existence of such components.
would encode information about various aspects of
self that includes a general specification of where
Evidence for the circular multi-item STM buffer in
one is in the environment (in allocentric coordinates
cortex
— relative to a fixed map of the environment) and
the action being taken towards achieving goals.
A requirement of reward-dependent storage of
Importantly, the allocentric coding evident in the
long paths leading to the reward is the existence of
medial entorhinal cortex could provide an excellent
a multi-item ‘‘circular’’ buffer. When the reward
way of specifying directional action (e.g. turn
site is found, the contents of the buffer (the C/A
North) because the correct action can be triggered
sequences leading to the reward site) are trans-
by a cue approached from any direction. In con-
ferred into the LTM store. The existence of a
trast, egocentric specification (e.g. turn Right)
multi-item STM in humans has been deduced from
would work only if the cue is approached from
psychophysical studies (Atkinson and Shiffrin,
the same angle as during initial learning.
1968). This buffer is thought to have limited ca-
To give a specific example of how non-self/self
pacity (772 items) and to be ‘‘circular’’. Here
information might be encoded by a lateral/medial
‘‘circular’’ means that when the buffer is full, the
entorhinal couplet, consider the following descrip-
next arriving item knocks out the item that has
tion of a moment in my morning commute. When
been in the buffer the longest (i.e. first in, first out).
I see the following non-self information: [McDo-
In the context of the model shown in Fig. 2, the
nald’s and, to its right, Burger King (complex cue
circular property ensures that when the goal is
specified in egocentric coordinates)], this would be
reached, transfer of buffer information into LTM
coupled to the following self information: [given that
will incorporate information about the last five
I am on my way to work (my goal), when I’m near
events that occurred before the goal was found.
Brandeis (my place approximated in allocentric co-
A physiologically plausible model of a multi-
ordinates), I turn West (my action in allocentric
item STM buffer has been developed (Lisman and
coordinates)].
Idiart, 1995) and a recent variant has first in, first
Parenthetically, one can see from this example
out properties (Koene and Hasselmo, 2006). fMRI
that purely landmark-based navigation would fail
evidence points to the temporal lobe as a site of a
if there are many similar landmarks, as is the case
multi-item working memory buffer, as evidenced
for fast food establishments, and that even a crude
by the load dependence of the fMRI signal (Fie-
allocentric system (‘‘near Brandeis’’) can therefore
bach et al., 2006). There is evidence that the ent-
be helpful in disambiguating landmark cues.
orhinal cortex can maintain STM information
It would seem that the non-self/self formulation
(Otto and Eichenbaum, 1992) [reviewed in Jensen
is a plausible one and fits the data somewhat better
and Lisman (2005); Hasselmo and Stern (2006)].
than the C/A formulation. In the final sections I
Recent experiments indicate that the persistent fir-
will turn to the question of how this formulation
ing is due to a cholinergically enhanced intrinsic
might be tested.
conductance that produces an after depolarization
(Fransen et al., 2006), as postulated in the theo-
retical models (Lisman and Idiart, 1995).
Evidence for the building blocks required by the
model (Fig. 2)
Evidence for reward-mediated fixation of STM
The general model of Verschure et al. requires sev- content into hippocampal LTM
eral important building blocks: a circular, multi-
item STM buffer, a LTM capable of holding cou- It is generally thought that reward signaling is
plet sequences, and a control system that would mediated by the dopaminergic cells of the ventral

tegmental area and substantia nigra. It was orig-
inally thought that dopamine did not affect the
hippocampus, but recent work indicates that
there is dopaminergic innervation of the hippo-
campus and that it can profoundly affect trans-
mission and enhance LTP [reviewed in Lisman
and Grace (2005)]. In the context of the model of
Fig. 2, one might imagine that it is the dopamine
reward signal that stimulates reward-dependent
transfer of information from STM to LTM, but
no experiments have addressed this directly, so
not much can be said about this possibility at the
present time. One relevant observation (Foster
and Wilson, 2006) is that when rats reach a re-
ward site, a replay of the sequence of events
(places) that led the rat to the reward occurs in
the hippocampus (the replay occurs in reverse
order). Although it is suspected that this replay
occurs as a result of processes within the hippo-
campus itself, the possibility that it actually oc-
curs because of input from the cortex needs to be
directly tested.
Although the role of dopamine in transfer, is
uncertain, there is quite strong evidence that do-
pamine is necessary for the late stage of LTP (re-
viewed in Lisman and Grace, 2005). In this way,
dopamine may contribute to the reward-dependent
storage of sequences envisioned in Fig. 2.
Evidence for hippocampal LTM stores capable of
storing sequences
There have been many computational models of the
hippocampus (Burgess et al., 2001; Kunec
et al., 2005; Rolls and Kesner, 2006), but most
deal exclusively with autoassociational memories
that encode events at a given moment, and so can-
not be used to account for sequences of events.
Models developed in my laboratory, which have
been improved over time (Jensen and Lisman, 1996;
Lisman, 1999; Lisman et al., 2005), deal with how
the hippocampus can store and recall sequences (see
also Levy (1996)]. This class of models shows how
gamma and theta oscillations, standard NMDA
receptor-mediated plasticity processes and known
anatomical connections (including the feedback
connections from CA3 to dentate) can mediate
621
successful transfer of sequences from a multi-item
cortical STM buffer to hippocampal LTM.
It may be instructive to consider how such
processes might transfer C/A sequences from STM
to LTM and later use the stored information to
recall them. During learning, the elements of C
from lateral entorhinal cortex and the elements of
A from medial entorhinal cortex converge on sub-
sets of granule cells; each granule cell that fires
forms an element of the C/A representation. The
output of active granule cells excites a subset of
CA3 pyramidal cells, and these are combined into
an associative representation of the C/A couplet
by LTP in the recurrent connections of the active
CA3 cells. The CA3 output is then sent back to the
dentate, where the synapses onto granule cells
representing elements of the next C/A couplet in
the sequence become potentiated. These feedback
synapses thus form linkages between sequential
events. In this way, the entire sequence can be
transferred from a cortical buffer to the hippo-
campal memory [see Lisman et al. (2005) for de-
tails]. Now let us assume that the organism
subsequently comes across C2, and the corre-
sponding input is supplied to the dentate, which
then provides input to CA3. Through autoassoci-
ative memory of the stored couplet, the A2 com-
ponent of the couplet will be evoked. The
completed C2/A2 couplet is sent back to the dent-
ate, where it evokes the C3/A3 couplet (this is
termed a chaining step). The vagaries of synaptic
transmission will, however, lead to a slightly cor-
rupted representation of the couplet, which, if used
to chain to the next element, would lead to an even
more corrupted version of C4/A4. However, the
reciprocal connections between the dentate and
CA3 can serve to prevent such concatenation of
errors (Lisman, 1999). Specifically, C3A3 sent to
CA3 where the autoassociative properties of CA3
correct the errors in the C3/A3 representation and
it is this corrected version that is sent from CA3 to
the dentate, where it triggers C4/A4. Thus, through
sequential chaining and correction processes, the
dentate and CA3 can lead to accurate recall of the
entire sequence of couplets leading to a goal. Other
brain circuits could then use this stored informa-
tion to make a decision about whether this goal is
worthwhile in the current context and, if so, to
622
execute the specific motor plans stored in the cou-
plet sequence.
Although most computational models of the
hippocampus posit that the sole function of the
dentate is to produce orthogonalization (the gen-
eration of very different firing patterns to slightly
different input patterns), it is clear from the anat-
omy that a second function, the mixing of lateral
and medial inputs, must also be important. If the
ideas proposed here are correct, dentate and CA3
cells should display a unique code that is a mixture
of sensory cues and action.
Importantly, the mixing of cortical inputs to
CA1 is already quite different; the region close to
CA3 gets exclusive input from the medial ent-
orhinal cortex whereas the region closer to the
subiculum gets exclusive input from the lateral
entorhinal input (Fig. 1). The same segregation
holds for the CA1 projections back to cortex. This
pattern of segregation is different from dentate/
CA3, where all cells receive both lateral and medial
information, thus presumably mixing them into a
joint representation. CA1 does not appear to be
part of this coding system; it is thus likely to have a
role in converting the dentate/CA3 code back into
the separate codes of the lateral and medial ent-
orhinal cortex (Lisman, 1999).
Concluding remarks and predictions of the model
My goal in this discussion has been to examine the
ideas incorporated into the model of Fig. 2 and to
assess whether they can be usefully mapped onto
the architecture of the hippocampus (Fig. 1).
My general conclusion is that the correspondence
is promising. The basic building blocks utilized in
the model have a reasonable correspondence to the
capabilities of the cortex and hippocampus.
The more specific question, whether any of the
functional dualities inspired by the model (C/A,
non-self/self) underlie the structural duality of ent-
orhinal inputs, cannot yet be answered with any
certainty. The idea that the medial entorhinal
encodes aspects of self whereas the lateral ent-
orhinal encodes the outside world (non-self) leads
to experimentally testable predictions, as outlined
in the following paragraphs.
Representation of self-action on the T-maze in the
medial entorhinal cortex
It should be possible to test the hypothesis pro-
posed here by recording from the entorhinal cortex
during the T- and plus-maze paradigms that dem-
onstrate motor/action information in the hippo-
campus. The specific prediction is that the medial,
but not the lateral entorhinal cortex should encode
motor/goal information. Preliminary evidence in-
dicates that this prediction is correct (Lipton et al.,
2006).
Unique conjunctions of non-self/self in the dentate/
CA3; cue information in the lateral entorhinal cortex
A second set of predictions has to do with the
changes in representation. The dentate and CA3
should have cells that represent conjunction of
non-self and self (e.g. cue and action) and this
conjunction should not be present in layers 2 and 3
of the entorhinal cortex. An example of such
a conjunction would be a cell that responded to
scene, but only in conjunction with a cued action
(Wirth et al., 2003). In these experiments, the scene
and cue information should be encoded in the lat-
eral entorhinal; the action taken (which may
or may not be consistent with the cue) should be
encoded in the medial entorhinal.
Conditioned fearful stimuli vs. the emotion of fear
The output of the basolateral amygdala represents
a sensory signal, the conditioned stimulus (CS),
and may well be source of the hippocampal re-
sponse to the CS that develops after conditioning
(Berger et al., 1976). Because these signals are non-
self similar signals would be predicted in the lateral
entorhinal cortex. In contrast the further process-
ing in the amygdala that occurs in the central nu-
cleus (Pare et al., 2004; Balleine and Killcross,
2006) is thought to represent the emotion of fear
and to evoke fear-related responses in the hypo-
thalamus and brainstem. Such emotional re-
sponses are a property of self and would be
predicted to occur in the medial entorhinal cortex.

Connectivity with areas mediating personal/extra-
personal attention
If there is a fundamental segregation of information
about self and non-self in the entorhinal cortex,
might one expect to find functionally related sub-
divisions at earlier stages of cortical processing? In
this regard it is interesting that very recent work
(Committeri et al., 2007; Ortigue et al., 2006) points
to two very different forms of hemi-neglect due to
brain injuries. One form, termed ‘‘extra-personal’’,
leads to neglect of one half of the external world;
the other form, termed ‘‘personal’’ leads to neglect
of half of the body. These forms can be doubly
dissociated and involve injuries to different brain
regions. It will be of interest to determine whether
these regions have selective connections to the me-
dial and lateral entorhinal cortex.
Note added in proof
I have recently become aware of other works on
cortical and subcortical processing that posits a
separate neural system for self-referential processing
(Buckner and Carroll, 2007; Northoff et al., 2006).
Acknowledgments
I thank Andre Fenton, Menno Witter, Bruce McN-
aughton, Edvard Moser, James Knierim, Steven
Wise, Howard Eichenbaum, David Redish, and
Matthew Shapiro for useful conversations. I thank
Paul Verschure for comments on the manuscript.
This work was supported by NIH grants R01
NH027337, P50 MH060450-07A1, and R01
MH61975-01A2. Special thanks also to the Dart
Foundation and the Marine Biological Laboratory.
References
Atkinson, R. and Shiffrin, R. (1968) Human memory: a pro-
posed system and its control processes. In: Spence K. (Ed.),
The psychology of learning and motivation: advances in re-
search and theory, Vol. 2. Academic Press, New York.
Balleine, B.W. and Killcross, S. (2006) Parallel incentive
processing: an integrated view of amygdala function. Trends
Neurosci., 29: 272–279.
623
Barto, A.G. and Sutton, R.S. (1981) Landmark learning: an
illustration of associative search. Biol. Cybern., 42: 1–8.
Berger, T.W., Alger, B. and Thompson, R.F. (1976) Neuronal
substrate of classical conditioning in the hippocampus. Sci-
ence, 192: 483–485.
Bland, B.H. and Oddie, S.D. (2001) Theta band oscillation and
synchrony in the hippocampal formation and associated
structures: the case for its role in sensorimotor integration.
Behav. Brain Res., 127: 119–136.
Bower, M.R., Euston, D.R. and McNaughton, B.L. (2005) Se-
quential-context-dependent hippocampal activity is not nec-
essary to learn sequences with repeated elements. J.
Neurosci., 25: 1313–1323.
Brasted, P.J., Bussey, T.J., Murray, E.A. and Wise, S.P.
(2005) Conditional motor learning in the nonspatial do-
main: effects of errorless learning and the contribution of
the fornix to one-trial learning. Behav. Neurosci., 119:
662–676.
Brown, M.W. and Aggleton, J.P. (2001) Recognition memory:
what are the roles of the perirhinal cortex and hippocampus?
Nat. Rev. Neurosci., 2: 51–61.
Buckner, R.L. and Carroll, D.C. (2007) Self-projection and the
brain. Trends Cogn. Sci., 11: 49–57.
Burgess, N., Becker, S., King, J.A. and O’Keefe, J. (2001)
Memory for events and their spatial context: models and ex-
periments. Philos. Trans. R. Soc. Lond. B Biol. Sci., 356:
1493–1503.
Burwell, R.D. (2000) The parahippocampal region: corticocor-
tical connectivity. Ann. N.Y. Acad. Sci., 911: 25–42.
Burwell, R.D., Witter, M.P. and Amaral, D.G. (1995) Per-
irhinal and postrhinal cortices of the rat: a review of the
neuroanatomical literature and comparison with findings
from the monkey brain. Hippocampus, 5: 390–408.
Collett, T.S., Graham, P. and Durier, V. (2003) Route learning
by insects. Curr. Opin. Neurobiol., 13: 718–725.
Committeri, G., Pitzalis, S., Galati, G., Patria, F.,
Pelle, G., Sabatini, U., Castriota-Scanderbeg, A., Piccardi,
L., Guariglia, C. and Pizzamiglio, L. (2007) Neural bases of
personal and extrapersonal neglect in humans. Brain, 130:
431–441.
Corbit, L.H. and Balleine, B.W. (2000) The role of the hippo-
campus in instrumental conditioning. J. Neurosci., 20:
4233–4239.
Ekstrom, A.D., Kahana, M.J., Caplan, J.B., Fields, T.A.,
Isham, E.A., Newman, E.L. and Fried, I. (2003) Cellular
networks underlying human spatial navigation. Nature, 425:
184–188.
Fahy, F.L., Riches, I.P. and Brown, M.W. (1993) Neuronal
activity related to visual recognition memory: long-term
memory and the encoding of recency and familiarity infor-
mation in the primate anterior and medial inferior temporal
and rhinal cortex. Exp. Brain Res., 96: 457–472.
Ferbinteanu, J., Holsinger, R.M. and McDonald, R.J. (1999)
Lesions of the medial or lateral perforant path have different
effects on hippocampal contributions to place learning and
on fear conditioning to context. Behav. Brain Res., 101:
65–84.
624
Ferbinteanu, J. and Shapiro, M.L. (2003) Prospective and ret-
rospective memory coding in the hippocampus. Neuron, 40:
1227–1239.
Fiebach, C.J., Rissman, J. and D’Esposito, M. (2006)
Modulation of inferotemporal cortex activation during
verbal working memory maintenance. Neuron, 51:
251–261.
Foster, D.J. and Wilson, M.A. (2006) Reverse replay of be-
havioural sequences in hippocampal place cells during the
awake state. Nature, 440: 680–683.
Foster, T.C., Castro, C.A. and McNaughton, B.L. (1989) Spa-
tial selectivity of rat hippocampal neurons: dependence on
preparedness for movement. Science, 244: 1580–1582.
Frank, L.M., Brown, E.N. and Wilson, M. (2000) Trajectory
encoding in the hippocampus and entorhinal cortex. Neuron,
27: 169–178.
Fransen, E., Tahvildari, B., Egorov, A.V., Hasselmo, M.E. and
Alonso, A.A. (2006) Mechanism of graded persistent cellular
activity of entorhinal cortex layer v neurons. Neuron, 49:
735–746.
Fyhn, M., Molden, S., Witter, M.P., Moser, E.I. and Moser,
M.B. (2004) Spatial representation in the entorhinal cortex.
Science, 305: 1258–1264.
Gaffan, D. (1985) Hippocampus: memory, habit and voluntary
movement. Philos. Trans. R. Soc. Lond. B Biol. Sci., 308:
87–99.
Gnatkovsky, V., Uva, L. and de Curtis, M. (2004) Topographic
distribution of direct and hippocampus-mediated entorhinal
cortex activity evoked by olfactory tract stimulation. Eur. J.
Neurosci., 20: 1897–1905.
Hafting, T., Fyhn, M., Molden, S., Moser, M.B. and Moser,
E.I. (2005) Microstructure of a spatial map in the entorhinal
cortex. Nature, 436: 801–806.
Hargreaves, E.L., Rao, G., Lee, I. and Knierim, J.J.
(2005) Major dissociation between medial and lateral ent-
orhinal input to dorsal hippocampus. Science, 308:
1792–1794.
Hasselmo, M.E. (2005) A model of prefrontal cortical mech-
anisms for goal-directed behavior. J. Cogn. Neurosci., 17:
1115–1129.
Hasselmo, M.E. and Stern, C.E. (2006) Mechanisms underlying
working memory for novel information. Trends. Cogn. Sci.,
10: 487–493.
Jensen, O. and Lisman, J.E. (1996) Hippocampal CA3 region
predicts memory sequences: accounting for the phase preces-
sion of place cells. Learn. Mem., 3: 279–287.
Jensen, O. and Lisman, J.E. (2005) Hippocampal sequence-en-
coding driven by a cortical multi-item working memory
buffer. Trends Neurosci., 28: 67–72.
Koene, R.A. and Hasselmo, M.E. (2006) First-in-first-out
item replacement in a model of short-term memory
based on persistent spiking. Cereb. Cortex (epub ahead of
print).
Kunec, S., Hasselmo, M.E. and Kopell, N. (2005) En-
coding and retrieval in the CA3 region of the hippocampus:
a model of theta-phase separation. J. Neurophysiol., 94:
70–82.
Levy, W.B. (1996) A sequence predicting CA3 is a flexible
associator that learns and uses context to solve hippocampal-
like tasks. Hippocampus, 6: 579–590.
Lipton, P.A., White, J.A. and Eichenbaum, H. (2006) Di-
sambiguation of overlapping spatial sequences in medial ent-
orhinal cortex. In: Annual Society for Neuroscience Meeting.
Atlanta, GA: Society for Neuroscience, Program #66.6/
Poster #Z23.
Lisman, J.E. (1999) Relating hippocampal circuitry to function:
recall of memory sequences by reciprocal dentate-CA3 inter-
actions. Neuron, 22: 233–242.
Lisman, J.E. and Grace, A.A. (2005) The hippocampal-VTA
loop: controlling the entry of information into long-term
memory. Neuron, 46: 703–713.
Lisman, J.E. and Idiart, M.A. (1995) Storage of 7+/ 2 short-
term memories in oscillatory subcycles. Science, 267:
1512–1515.
Lisman, J.E., Talamini, L.M. and Raffone, A. (2005) Recall of
memory sequences by interaction of the dentate and CA3: a
revised model of the phase precession. Neural Netw., 18:
1191–1201.
McNaughton, B.L., Barnes, C.A. and O’Keefe, J. (1983) The
contributions of position, direction, and velocity to single
unit activity in the hippocampus of freely-moving rats. Exp.
Brain Res., 52: 41–49.
McNaughton, B.L., Battaglia, F.P., Jensen, O., Moser, E.I.
and Moser, M.B. (2006) Path integration and the neural
basis of the ‘cognitive map’. Nat. Rev. Neurosci., 7:
663–678.
Moita, MA., Rosis, S., Zhou, Y., LeDoux, J.E. and Blair, H.T.
(2003) Hippocampal place cells acquire location-specific re-
sponses to the conditioned stimulus during auditory fear
conditioning. Neuron, 37: 485–497.
Myhrer, T. (1988) Exploratory behavior and reaction to novelty
in rats with hippocampal perforant path systems disrupted.
Behav. Neurosci., 102: 356–362.
Northoff, G., Heinzel, A., de Greck, M., Bermpohl, F., Do-
browolny, H. and Panksepp, J. (2006) Self-referential
processing in our brain – a meta-analysis of imaging studies
on the self. Neuroimage, 31: 440–457.
Ortigue, S., Megevand, P., Perren, F., Landis, T. and Blanke,
O. (2006) Double dissociation between representational per-
sonal and extrapersonal neglect. Neurology, 66: 1414–1417.
Otto, T. and Eichenbaum, H. (1992) Complementary roles of
the orbital prefrontal cortex and the perirhinal-entorhinal
cortices in an odor-guided delayed-nonmatching-to-sample
task. Behav. Neurosci., 106: 762–775.
Pare, D., Quirk, G.J. and Ledoux, J.E. (2004) New vistas on
amygdala networks in conditioned fear. J. Neurophysiol., 92:
1–9.
Quiroga, R.Q., Reddy, L., Kreiman, G., Koch, C. and Fried, I.
(2005) Invariant visual representation by single neurons in
the human brain. Nature, 435: 1102–1107.
Ranck Jr., J.B. (1973) Studies on single neurons in dorsal hip-
pocampal formation and septum in unrestrained rats. I. Be-
havioral correlates and firing repertoires. Exp. Neurol., 41:
461–531.

Rolls, E.T. and Kesner, R.P. (2006) A computational theory of
hippocampal function, and empirical tests of the theory.
Prog. Neurobiol., 79: 1–48.
Rupniak, N.M. and Gaffan, D. (1987) Monkey hippocampus
and learning about spatially directed movements. J. Neuro-
sci., 7: 2331–2337.
Sargolini, F., Fyhn, M., Hafting, T., McNaughton, B.L.,
Witter, M.P., Moser, M.B. and Moser, E.I. (2006) Conjunc-
tive representation of position, direction, and velocity in
entorhinal cortex. Science, 312: 758–762.
Smith, D.M. and Mizumori, S.J. (2006) Learning-related de-
velopment of context-specific neuronal responses to places
and events: the hippocampal role in context processing. J.
Neurosci., 26: 3154–3163.
Vanderwolf, C.H. (1969) Hippocampal electrical activity and
voluntary movement in the rat. Electroencephalogr. Clin.
Neurophysiol., 26: 407–418.
Verschure, P.F. and Voegtlin, T. (1998) A bottom up approach
towards the acquisition and expression of sequential
625
representations applied to a behaving real-world device: Dis-
tributed Adaptive Control III. Neural Netw., 11: 1531–1549.
Verschure, P.F., Voegtlin, T. and Douglas, R.J. (2003) Envi-
ronmentally mediated synergy between perception and
behaviour in mobile robots. Nature, 425: 620–624.
Wiener, S.I., Paul, C.A. and Eichenbaum, H. (1989) Spatial and
behavioral correlates of hippocampal neuronal activity. J.
Neurosci., 9: 2737–2763.
Wirth, S., Yanike, M., Frank, L.M., Smith, A.C., Brown, E.N.
and Suzuki, W.A. (2003) Single neurons in the monkey hip-
pocampus and learning of new associations. Science, 300:
1578–1581.
Wood, E.R., Dudchenko, P.A. and Eichenbaum, H. (1999) The
global record of memory in hippocampal neuronal activity.
Nature, 397: 613–616.
Wood, E.R., Dudchenko, P.A., Robitsek, R.J. and Eichen-
baum, H. (2000) Hippocampal neurons encode information
about different types of memory episodes occurring in the
same location. Neuron, 27: 623–633.
Plate 33.1. Wiring diagram of the hippocampus (interneurons excluded) showing dual inputs from the lateral and medial entorhinal
cortex. The layer 2 cortical inputs to the dentate and CA3 diverge fan out (f) widely over these networks and then provide convergent
input to individual dentate granule cells. In contrast, the layer 3 cortical inputs are specialized for individual subregions of CA1. This is
an example of a point-to-point (p-p) connection. Within the hippocampus, granule cells provide input, via mossy fibers to the mossy
cells of the dentate and to CA3 cells. CA3 cells make feedback connections to themselves and to the mossy cells of the dentate. These,
in turn, provide excitatory input to granule cells in the inner third of the granule cell dendritic tree. (For B/W version, see page 616 in
the volume.)

Plate 36.1. (A) Two representative granule cells filled with 5,6 carboxyfluorescein from the dentate gyrus of a 24-month-old rat. (B)
Histograms showing the numbers of carboxyfluorescein injections that resulted in single, double, triple, or greater numbers of granule
cells filled with dye. Aged rats showed significantly increased electrotonic coupling compared to young animals. This may account for
the increased excitability of old granule cells. Adapted with modifications from Barnes et al. (1987). (For B/W version, see page 665 in
the volume.)
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 34

The CA3 ‘‘backprojection’’ to the dentate gyrus

Helen E. Scharfman

Department of Pharmacology and Neurology, Columbia University, New York, NY 10032, USA and the Center for
Neural Recovery and Rehabilitation Research, Helen Hayes Hospital, New York State Department of Health,
Rte 9W, West Haverstraw, NY 10993-1195, USA

Abstract: The hippocampus is typically described in the context of the trisynaptic circuit, a pathway that
relays information from the perforant path to the dentate gyrus, dentate to area CA3, and CA3 to area
CA1. Associated with this concept is the assumption that most hippocampal information processing occurs
along the trisynaptic circuit. However, the entorhinal cortex may not be the only major extrinsic input to
consider, and the trisynaptic circuit may not be the only way information is processed in hippocampus.
Area CA3 receives input from a variety of sources, and may be as much of an ‘‘entry point’’ to hippo-
campus as the dentate gyrus. The axon of CA3 pyramidal cells targets diverse cell types, and has com-
missural projections, which together make it able to send information to much more of the hippocampus
than granule cells. Therefore, CA3 pyramidal cells seem better designed to spread information through
hippocampus than the granule cells. From this perspective, CA3 may be a point of entry that receives
information which needs to be ‘‘broadcasted,’’ whereas the dentate gyrus may be a point of entry that
receives information with more selective needs for hippocampal processing. One aspect of the argument
that CA3 pyramidal cells have a widespread projection is based on a part of its axonal arbor that has
received relatively little attention, the collaterals that project in the opposite direction to the trisynaptic
circuit, ‘‘back’’ to the dentate gyrus. The evidence for this ‘‘backprojection’’ to the dentate gyrus is strong,
particularly in area CA3c, the region closest to the dentate gyrus, and in temporal hippocampus. The
influence on granule cells is indirect, through hilar mossy cells and GABAergic neurons of the dentate
gyrus, and appears to include direct projections in the case of CA3c pyramidal cells of ventral hippo-
campus. Physiological studies suggest that normally area CA3 does not have a robust excitatory influence
on granule cells, but serves instead to inhibit it by activating dentate gyrus GABAergic neurons. Thus,
GABAergic inhibition normally controls the backprojection to dentate granule cells, analogous to the way
GABAergic inhibition appears to control the perforant path input to granule cells. From this perspective,
the dentate gyrus has two robust glutamatergic inputs, entorhinal cortex and CA3, and two ‘‘gates,’’ or
inhibitory filters that reduce the efficacy of both inputs, keeping granule cells relatively quiescent. When
GABAergic inhibition is reduced experimentally, or under pathological conditions, CA3 pyramidal cells
activate granule cells reliably, and do so primarily by disynaptic excitation that is mediated by mossy cells.
We suggest that the backprojection has important functions normally that are dynamically regulated by
nonprincipal cells of the dentate gyrus. Slightly reduced GABAergic input would lead to increased po-
lysynaptic associative processing between CA3 and the dentate gyrus. Under pathological conditions as-
sociated with loss of GABAergic interneurons, the backprojection may support reverberatory excitatory

Corresponding author. Tel.: +1 845 398 5427;

Fax: +1 845 398 5422; E-mail: hscharfman@nki.rfmh.org

DOI: 10.1016/S0079-6123(07)63034-9 627

628

activity between CA3, mossy cells, and granule cells, possibly enhanced by mossy fiber sprouting. In this
case, the backprojection could be important to seizure activity originating in hippocampus, and help
explain the seizure susceptibility of ventral hippocampus.

Keywords: pyramidal cell; dentate gyrus; hilus; mossy cell; interneuron; granule cell; recurrent excitation

Introduction
Most explanations of hippocampal circuitry begin
with the identification of the major subfields and
the trisynaptic circuit. As shown in Fig. 1, the
hippocampus of mammals is composed of three
subfields: area CA1, containing primarily CA1 py-
ramidal cells, area CA3, also composed primarily
of pyramidal cells, and the dentate gyrus, where
the primary principal cell is the granule cell. The
trisynaptic circuit is composed of three sequential
glutamatergic synapses:
1) perforant path axons of layer II neurons in
entorhinal cortex project to the outer two-
thirds of the dentate gyrus molecular layer, the
location of the distal granule cell dendrites,
2) mossy fiber axons of granule cells project to
proximal dendrites of area CA3 pyramidal
cells, and
3) the Schaffer collateral axons of CA3 pyram-
idal cells project to stratum radiatum of CA3,
where the apical dendrites of area CA1 py-
ramidal cells are located (Amaral and Witter,
1989).
This projection is clearly fundamental, and
its importance to information processing in
hippocampus is not to be doubted here. The point
of emphasis here is, instead, that the trisynaptic
pathway is not the only circuit to consider. One
reason for hesitation is that a considerable number
of studies have now shown that nonprincipal cells
of hippocampus play critical roles in modulating
the trisynaptic pathway. They are innervated by
each of the fiber pathways mentioned above, and
have diverse projections to all hippocampal prin-
cipal cells. For example, within the dentate gyrus,
the perforant path and the mossy fibers innervate
numerous types of GABAergic neurons, as well as
glutamatergic mossy cells of the hilus. The GAB-
Aergic ‘‘interneurons’’ and mossy cells project to
granule cells at the level of the soma, axon hillock,
and at every part of the dendritic tree. Indeed, it
has been proposed that the primary targets of
mossy fibers are GABAergic neurons, and the pri-
mary effect of mossy fiber transmission under
normal conditions is inhibitory (Acsady et al.,
1998). This appears to be the case in recordings of
CA3 neurons in slices, where they dominant re-
sponse to mossy fiber stimulation is an inhibitory
postsynaptic potential (IPSP) (Scharfman, 1993a,
b), despite the large unitary excitatory postsynap-
tic potential (EPSP) produced by granule cells
(Scharfman et al., 1990) (for review, see Jaffe and

Fig. 1. The ‘‘trisynapto-centric’’ vs. ‘‘CA3-centric’’ view of hippocampal information processing. (A) The trisynaptic circuit is di-
agrammed schematically for a horizontal section through the adult rodent hippocampus. Cell layers are in gray. (B) The axonal arbor
of CA3 pyramidal cells is schematically presented to illustrate the perspective that these neurons may be a central point of information
processing in the hippocampus.

Gutierrez, this volume). It has been suggested that
other characteristics of mossy fiber transmission
may be even more important than their ability to
depolarize postsynaptic pyramidal cells (Urban
et al., 2001).
In this chapter an alternative to the trisynapto-
centric view is emphasized. It involves the axons of
CA3 pyramidal cells that project in the opposite
direction of the trisynaptic circuit (‘‘back projec-
tion’’), into the dentate gyrus (Fig. 1B).
Anatomical evidence for a ‘‘backprojection’’
What is the evidence for a ‘‘backprojection?’’ First
we will consider anatomical arguments and then
physiological data. Historically, there have been
only rare suggestions that CA3 pyramidal cells
might project into the dentate gyrus (Zimmer,
1971; Schwerdtfeger and Sarvey, 1983). These did
not seem to attract much attention, possibly be-
cause some of the experimental approaches had
technical limitations. For example, the most com-
mon approach for tract tracing at the time relied
on tracer injection, but CA3 is difficult to label
selectively with this technique.
Newer approaches provided the requisite selec-
tivity. These techniques involved dye injection, typ-
ically biocytin, into individual CA3 pyramidal cells
using intracellular microelectrodes, followed by
axon reconstruction to evaluate the axon arbor.
Using this approach, Ishizuka et al. (1990) filled
individual CA3 neurons in hippocampal slices of
the rat, and emphasized several aspects of the axon
of CA3 neurons. They found evidence for axon
collateralization within the hilus in many of their
sampled cells, particularly those with a cell body in
‘‘proximal’’ CA3 (near the dentate gyrus, i.e., CA3b
and CA3c). Li et al. (1994) injected biocytin into
CA3 pyramidal cells in vivo, also using adult rats,
and found even greater evidence for a backprojec-
tion. They found that individual CA3 neurons had
collaterals in the hilus whether the cell body of or-
igin was in CA3a, b, or c. Like Ishizuka et al.
(1990), they found the most hilar collateralization
from cells within CA3c, and they specifically iden-
tified the ventral portion of CA3c region as the area
with the greatest collateralization in the hilus.
629
Ventral CA3c pyramidal cells collateralized
approximately 3–4  as much in the hilus as dor-
sal CA3c pyramidal cells, or CA3b cells.
Li et al. (1994) also noted that CA3 pyramidal
cell axons entered the granule cell layer and col-
lateralized in the inner molecular layer, and this
was predominantly a characteristic of ventral
CA3c pyramidal cells. These axons exhibited nu-
merous varicosities, suggesting that they inner-
vated dentate gyrus granule cells, and possibly
other cell types. Therefore, the CA3c population
of ventral hippocampus appears to be a subset of
CA3 pyramidal cells with the highest potential for
dentate gyrus interactions.
These studies were the first to define hilar col-
lateralization of CA3 neurons unequivocally.
However, one caveat was that the sample sizes
were small, and it remains unclear exactly how
representative the data were of the entire CA3
population. It also is unclear whether the data
from adult rat might vary across age and species.
Due to the labor-intensive nature of the experi-
ments, these questions remain unanswered.
Comparisons of CA3a, b, and c pyramidal cells
are interesting to review in light of the findings that
their axons appear to have a different preference
for the dentate gyrus. Anatomically, CA3c pyram-
idal cells are heterogeneous neurons (Scharfman,
1993; Turner et al., 1995), sometimes reflecting
stereotypical pyramidal cell morphology, but in
other cases they may appear more similar to non-
pyramidal cells. Some of the electrophysiological
characteristics of CA3c pyramidal cells appear to
be distinct from CA3a and b, and this may be
important because it could lead to unique physi-
ological attributes to the backprojection. CA3c
neurons are not as easily activated by mossy fiber
stimulation as CA3b neurons in slices, suggesting
that the GABAergic neurons which are targets of
mossy fibers innervate CA3c more than CA3b. All
else being equal, such preferential inhibition might
quiet the backprojection normally, relative to
other CA3 projection systems.
In comparisons of CA3b and CA3c pyramidal
cells of the adult male rat, almost all CA3c neu-
rons demonstrated burst (phasic) firing in response
to current injection, whereas CA3b pyramidal cells
were much more heterogeneous, often exhibiting
630
trains of action potential (tonic firing) only. Var-
iation in CA3a and b in this respect has been re-
ported (Bilkey and Schwartzkroin, 1990). If CA3c
neurons do have a greater percentage of neurons
which intrinsically discharge in bursts, the infor-
mation processed by backprojecting axons may
differ in the way it influences granule cells from the
way information is transferred to other targets of
CA3 neurons.
CA3c neurons demonstrated the longest time
constant, which is noteworthy because this char-
acteristic might lead to a greater capacity for in-
tegration of synaptic inputs from distinct locations
along the dendrites (Scharfman, 1993b). However,
CA3c neurons appear to have a relatively short
electrotonic length, which would lead to rapid de-
cay before integration of dendritic inputs would
reach the soma. Together the data would suggest a
specialization of the dendrite as a separable
processing unit, perhaps more than other CA3
neurons. It also would suggest that somatic inputs
would have a relatively greater influence on action
potential output than dendritic inputs.
CA3c is more vulnerable to certain types of ex-
citotoxic stimulation than CA3a or b pyramidal cells
(Chang and Dyer, 1985; Freund et al., 1992; Miet-
tinen et al., 1998; Haas et al., 2001). It was proposed
that CA3 may be an area where burst generation
occurs in models of epilepsy, but others have sug-
gested CA3a/b is that location (Knowles et al., 1987;
Colom and Saggau, 1994). Selective lesions to CA3a
vs. b vs. c are difficult to make without injuring
adjacent areas, so the functional relevance of the
differences between CA3a, b, and c remain unclear.
In summary, the anatomical data suggest that
ventral hippocampus and CA3c neurons make the
most robust backprojection in normal adult rats,
and there is collateralization in the hilus, granule
cell layer and inner molecular layer. Specific phys-
iological characteristics of CA3c neurons may im-
part unique functional consequences to the
backprojection.
Physiological evidence for a ‘‘backprojection’’
Some of the first physiological evidence that CA3
axons innervated the dentate gyrus came from
studies in hippocampal slices, which showed that a
stimulating electrode placed in the hilus could ac-
tivate an antidromic population spike, recorded
extracellularly in the CA3 pyramidal cell layer.
Subsequent studies identified what cell types in
the dentate gyrus could be influenced by the CA3
projection. Several approaches established that hi-
lar neurons could be activated readily, but granule
cells could not.
First, fimbria stimulation was used to activate
CA3 pyramidal cells instead of the hilus, because
antidromic action potentials are easily elicited in
CA3 neurons by stimulation at this site, with no
evidence for direct activation of hilar neurons and
granule cells. Direct activation would be expected
because hilar cells that project commissurally have
an axon that descends into stratum oriens of
CA3b, presumably on its way to the fimbria, but in
our slices, we have found that it is severed before
reaching the fimbria. Direct activation of granule
cells would only be possible if current spread en-
tered stratum lucidum, which recordings show
does not occur after fimbria stimulation using the
methods we have employed.
Extracellular recordings showed that such fi-
mbria stimulation led to antidromic and or-
thodromic activation of CA3 b and c pyramidal
cells, followed by an extracellular negative wave in
the granule cell layer and inner molecular layer
that reversed polarity in the middle molecular
layer (Fig. 2). Due to the negative resting potential
of granule cells, the negative potential could reflect
an EPSP or depolarized IPSP, and it was shown
subsequently by intracellular recording that it was
a depolarized GABAA receptor-mediated IPSP
(Fig. 3; Scharfman, 1994a, b, c).
This fimbria stimulation site activated hilar
mossy cells and hilar GABAergic neurons disy-
naptically, and was sensitive to the AMPA recep-
tor antagonist CNQX but not the muscarinic
antagonist atropine (Scharfman, 1993a). At a
longer latency, the granule cell IPSP began. These
data suggesting that fimbria activation of CA3
pyramidal cells led to hilar neuron activation, pre-
sumably by hilar collaterals of CA3 pyramidal
cells. Comparing the latency to the response of
CA3 pyramidal cells and hilar cells supported this
hypothesis, because CA3 pyramidal cells were
 631

Fig. 2. Recordings in hippocampal slices show evidence of trisynaptic and backprojecting pathways. (A) Extracellular recordings of
evoked responses to five stimulation sites (1–5) and four recording sites (A–D). Recordings were made sequentially in the same slice in
response to a fixed stimulus for each stimulus site. (B) A schematic is used to allow better comparisons of fimbria-evoked and outer
molecular layer-evoked responses recorded at four sites in the same slices: the outer molecular layer, granule cell layer, CA3c cell layer,
and CA3b cell layer. X marks recording site locations. (C) Comparison of evoked responses to fimbria stimulation in the same slice
demonstrate the responses in the granule cell layer following pyramidal cell activation by the fimbria, but do so with a delay.
Superimposition of responses illustrates that the onset of the granule cell layer field potential begins approximately 1–2 ms after the
antidromic population spike recorded in area CA3b. This delay suggests an intermediary synapse, probably in CA3c or the hilus,
between CA3b and granule cells. Indeed, the granule cell field potential begins immediately after the CA3c orthodromic population
spike, which probably is due to CA3b recurrent excitation of CA3c pyramidal cells. At the same time, hilar cells are also activated and
innervate granule cells (see later figures), although CA3c pyramidal cells could innervate granule cells also (Li et al., 1994). The bottom
trace is an IPSP recorded intracellularly from a granule cell in the same slice in response to the same stimulus. It shows that the onset of
the IPSP is similar to the onset of the granule cell layer field potential, which likely reflects the average of many IPSPs in granule cells
situated around the extracellular electrode.

activated at short latency and hilar cells approx-
imately 2 ms later. The same stimulus evoked a
long latency IPSP in granule cells, consistent with
a CA3-hilar neuron-granule cell pathway. In the
presence of bicuculline, a GABAA receptor antag-
onist, fimbria stimulation evoked EPSPs in granule
cells, and did so with a similar latency as the
IPSPs, suggesting a CA3-mossy cell-granule cell
pathway is normally masked by the activation of
GABAergic interneurons by CA3 hilar collaterals
(Fig. 3). Additional studies revealed that the
GABAergic neurons may not only involve those
in the hilus, but also the ‘‘basket cells’’ of the
granule cell layer (Scharfman, 1994a, c; Kneisler
and Dingledine, 1995).
Use of GABAA receptor antagonists also al-
lowed another approach to examine the effects of
CA3 on the dentate gyrus, one that involved no
stimulation to activate CA3 neurons. In the pres-
ence of GABAergic receptor antagonists such as
penicillin, picrotoxin, or bicuculline, CA3 neurons
discharge rhythmically in bursts of action
632

Fig. 3. Physiological evidence for a CA3 backprojection mediated by hilar neurons. (A) A schematic illustrates the backprojection
supported by physiological evidence collected to date. It shows that CA3 pyramidal cells innervate GABAergic and glutamatergic
mossy cells of the hilus, which in turn innervate granule cells. (B) Recordings illustrate the physiological correlates of the schematic in
A. Intracellular recordings from granule cells in response to fimbria stimulation in an adult male rat slice illustrate an evoked IPSP that
reverses at 70 mV, indicated a GABAA receptor-mediated IPSP is primarily evoked under normal conditions. After the GABAA
receptor antagonist bicuculline was bath-applied, the evoked response was an EPSP followed by an IPSP that reversed at approx-
imately 80 mV, suggesting an EPSP followed by a GABAB receptor-mediated IPSP is normally masked by GABAA receptor-
mediated inhibition. From Scharfman (1994a). Reprinted with permission from the Society for Neuroscience.

potentials (Schwartzkroin and Prince, 1977;
Knowles et al., 1987; Scharfman, 1994a, c). There-
fore, one can examine simultaneously any other
cell type in the slice to determine when and if they
are concurrently influenced. Simultaneous record-
ings showed that immediately after the onset of
spontaneous bursts in CA3 pyramidal cells, hilar
mossy cells and GABAergic interneurons also
demonstrated bursts (Scharfman, 1994a, c). Gran-
ule cells depolarized and could reach threshold,
but did so at a greater delay from the CA3 burst
than hilar cells. The hilar activity and granule cell
excitation, but not CA3 bursts, were blocked by
transecting the border between the dentate gyrus
and CA3b, or by application of CNQX (Fig. 4;
Scharfman, 1994a). When the timing of the burst
discharges was examined more closely, the onset of
the depolarization of CA3 neurons immediately
preceded the onset in the mossy cells and GAB-
Aergic neurons, and the onset of hilar neuron
bursts immediately preceded granule cell depolari-
zation, suggesting single synaptic delays of a CA3-
hilus-granule cell pathway (Scharfman, 1994a, c).
The most parsimonious explanation was that CA3
pyramidal cells innervated both types of hilar
neurons, and evidence was obtained for this
by simultaneous intracellular recordings from mo-
nosynaptically connected pyramidal cells and
mossy cells (Scharfman, 1994b). The same ap-
proach showed that mossy cells innervated granule
cells and interneurons in the hilus (Scharfman,
1995), suggesting that multiple polysynaptic cir-
cuits could be set in motion by CA3 pyramidal
cells, and ultimately influence dentate granule
cells. Current source density has also provided evi-
dence of CA3–dentate gyrus interactions similar to
those discussed above (Wu et al., 1998).
In summary, there is strong evidence from stud-
ies in hippocampal slices that the CA3 backpro-
jection innervates hilar neurons, both mossy cells
and GABAergic neurons. Currently there is no
physiological evidence that CA3 directly inner-
vates the granule cells of the dentate gyrus, but
absence of evidence is not a proof of absence.
At the present time, the evidence suggests that
the granule cells, even if they receive a direct
 633

Fig. 4. The CA3 excitatory backprojection is dependent on hilar mossy cells. (A) Schematic illustrating the experimental approach for
results shown in B–C. A site in CA3c was used for extracellular recording while recording intracellularly from a granule cell. Pressure
application of CNQX was used at two distinct sites in the hilus or in the inner molecular layer. CNQX was applied in microdrops that
were barely detectable by eye, allowing preferential application to select areas of the slice. (B) In the presence of bicuculline, CA3
epileptiform discharges were evoked by fimbria stimulation, and following the onset of the burst discharge, a granule cell depolarized
and discharged two action potentials. After CNQX was pressure-applied to the hilus, the granule cell response decreased but the CA3
discharge remained unaffected. (C) In a different slice, CNQX pressure-application to the inner molecular layer, near the recorded
granule cell, reversibly decreased the EPSP of the granule cell, suggesting an inner molecular layer glutamatergic synapse was necessary
for the EPSP. From Scharfman (1994a). Reprinted with permission from the Society for Neuroscience.

projection from CA3 neurons, are primarily in-
hibited by the backprojection, not activated. How-
ever, once they are relieved of GABAA receptor-
mediated inhibition, a strong excitatory circuit is
unmasked. The excitatory circuit appears to be
primarily derived from the mossy cell projection to
the inner molecular layer.
Functional implications of the CA3 backprojection
Hippocampal information processing
Is CA3 the point of entry to hippocampus?
Given the projection of CA3 pyramidal cells inner-
vates CA1, the dentate gyrus, other CA3 neurons
by recurrent collaterals, and sends information both
ipsilaterally and contralaterally, the argument can
be made that it is a cell type that may be central to
hippocampal information processing (Fig. 1B). In
light of this, it is interesting that the general as-
sumption is ‘‘trisynapto-centric,’’ that granule cells
receive the primary input, and the mossy fiber con-
trol how the hippocampus processes the incoming
input. This view might be correct in cases that
involve layer II of entorhinal cortex specifically, but
it may be rare that layer II is ever activated in iso-
lation, given its broad input from other layers
within the entorhinal cortex, and the diverse inputs
to entorhinal cortex. In actuality, it is highly likely
that the sensory and other cortical input that comes
to entorhinal cortex would also involve subcortical
structures that at the same time might reach CA3
through the fimbria. Furthermore, the tempo-
roammonic pathway may activate CA3 neurons at
lower threshold than granule cells (see Chapter by
Derrick, this volume), and this further supports the
tenet that incoming information might first activate
the CA3 region of hippocampus, rather than do so
via dentate granule cells. Given the data that mossy
fibers primarily inhibit CA3 pyramidal cells, the
dentate may actually function to keep CA3 from
being activated excessively by the concomitant in-
puts from the cortex and subcortical zones.
The dentate gyrus has two ‘‘gates’’ rather than one
Another implication of the discussion above, par-
ticularly the physiological studies’ is that the dent-
ate gyrus actually has two ‘‘gates’’ (Fig. 5). The
634
Fig. 5. The trisynaptic circuit and CA3 backprojection sup-
ports a bi-directional gate to the dentate gyrus granule cells.
first is a gate that prevents activation by the ent-
orhinal cortex (see Chapter by Hsu, this volume).
This may not be an all-or-none function, so the
word ‘‘filter’’ may be better (see Chapter by Dudek
and Sutula, this volume). However, the concept is
appropriate regardless of the semantics: cortical
input is not very effective in activating granule
cells above their threshold.
It appears that a second ‘‘gate’’ or ‘‘filter’’ also
exists, from CA3 pyramidal cells back to the dent-
ate gyrus. Thus, the backprojection does not ap-
pear to be effective in depolarizing granule cells,
and it appears that the reason is the divergence of
CA3 hilar collaterals to both excitatory hilar
mossy cells, and inhibitory GABAergic interneu-
rons. Evidently, the inhibitory neurons exert a
stronger net influence, and it is important to note
that in vivo it may be different, because mossy cell
axons are truncated in the slice much more than
GABAergic neurons. In vivo, one would predict
that CA3 would inhibit granule cells within the
same lamella, but activate them in distal sections
of hippocampus because of the long translamellar
mossy cell axon projection.
When inhibition is reduced experimentally, CA3
discharge is extremely robust in depolarizing gran-
ule cells, and it appears to be primarily due to
inner molecular synapses of mossy cells. Thus, the
hilar neurons modulate the CA3 backprojection
under normal conditions, and any condition that
would reduce GABAergic inhibition would be ex-
pected to facilitate CA3 activation of dentate
granule cells.
If this occurs, it would be predicted that asso-
ciative processes would be enhanced. In other
words, not only could there be recurrent excitatory
circuits among CA3 pyramidal cells that perform
autoassociative functions, but polysynaptic auto-
association would also potentially develop be-
tween CA3 and the granule cells when GABAergic
inhibition is weak. One argument that makes this
even more compelling is that granule cells do not
necessarily need to reach threshold in order for
mossy fiber transmission to occur (Alle and Gei-
ger, 2006). Subthreshold depolarizations in gran-
ule cells can lead to effects in granule cell targets.
An important implication is that autoassociative
networks, particularly the complex CA3–dentate
network discussed here, is likely to be controlled
by GABAergic hilar neurons and mossy cells.
Pathological conditions
The CA3 backprojection may subserve an impor-
tant role in conditions like ischemia or epilepsy,
when hilar GABAergic neurons are injured or killed
(Johansen et al., 1987; Sloviter, 1987; see Chapter by
Tallent, this volume). In both conditions, the som-
atostatin-immunoreactive hilar neurons that project
preferentially to the outer molecular layer are lost.
In animal models of epilepsy that use status epilep-
ticus as the initial experimental manipulation to in-
duce an epileptic state, there are many types of
GABAergic neurons that may be affected, but most
laboratories suggest a relative preservation of parv-
albumin-immunoreactive basket cells. The animal
models of epilepsy are also interesting because both
the granule cell axons and the pyramidal cell axons
may sprout new collaterals in response to injury
adjacent to them. Granule cell axons sprout into the
inner molecular layer and hilus, as well as stratum
oriens in CA3. It would be of interest to determine
whether the backprojection increases in response to
injury.
Regardless of these changes, it does appear that
the backprojection is robust under epileptic condi-
tions. The data are from hippocampal slices from
animals that had status epilepticus and became ep-
ileptic, i.e., they developed recurrent spontaneous
seizures for the rest of their lifespan. In most of
 635

Fig. 6. Evidence for associative networks and reverberatory circuits in the dentate gyrus under control of GABAergic inhibition. (A) A
schematic illustrates electrode locations for the recordings shown in B. (B) Stimulation of specific sites in the slice evoked bursts of
action potentials in CA3 pyramidal cells, and simultaneous intracellular recordings from a granule cell show bi-directional activation
of CA3 and the granule cell. (C) In a different slice from B where CA3 epileptiform bursts were followed by numerous after discharges,
CA3 appeared to precede activation of the simultaneously-recorded granule cell, but during the after discharges, the converse appeared
to develop. From Scharfman (1994a). Reprinted with permission from the Society for Neuroscience.

these animals, CA3 pyramidal cell burst discharges
occur when slices are prepared, even when bicucul-
line is absent (Scharfman et al., 2001). This affords
an opportunity to examine the backprojection by
simply examining concurrent activity in hilar neu-
rons and granule cells. When this has been done, it
is clear that the backprojection to mossy cells still
exists, because mossy cells exhibit synchronized
burst discharges with CA3 pyramidal cells in slices
from rats with recurrent seizures (Scharfman et al.,
2001). Interneurons in the hilus do as well, although
the population of both mossy cells and interneurons
is reduced relative to the normal rat, and it is not
yet clear which GABAergic neurons are involved of
the many types that exist. Granule cells usually ap-
pear to be relatively quiet when CA3 burst dis-
charges occur (Scharfman and McCloskey, 2007)
but this is not always the case: some granule cells
have depolarizations during the CA3 burst dis-
charges. In this case, bicuculline may not be re-
quired because of a loss of GABAergic inhibition
when hilar inhibitory neurons are killed after status
epilepticus.
This pathway may be relevant to the seizures
that occur in the whole animal, because the CA3
discharges appear to generate reverberatory activ-
ity between the dentate and CA3 pyramidal cells
(Fig. 6). During burst discharges in CA3, we have
found that the onset of the burst discharge occurs
in pyramidal cells first, and then in granule cells,
but during after discharges that follow the primary
burst, granule cell discharge may precede the py-
ramidal cells. These data suggest CA3 may be a
generator of activity initially, but granule cells can
initiate further excitation. In the whole animal,
such reverberatory excitation could lead to sub-
stantial network activity of CA3/dentate neurons,
and possibly trigger seizures after a certain thresh-
old of network activity is reached. One would pre-
dict that this might occur in more ventral
hippocampus than dorsal, given that the backpro-
jection is more prevalent in ventral hippocampus.
This prediction is supported by findings in slices of
temporal hippocampus, where seizure activity is
most robust (Bragdon et al., 1986; Borck and
Jefferys, 1999; Scharfman et al., 2002).
Conclusions
The CA3 region is often considered to be the sec-
ond subfield that receives information from ex-
trinsic afferents to hippocampus, while the dentate
gyrus is the initial entry point. Here we suggest
636
that CA3 may be a major gateway to hippocam-
pus, particularly with respect to subcortical input.
One reason to think of CA3 as a central access
point is that it has a projection that reaches the
vast majority of hippocampal neurons, both ipsi-
laterally and contralaterally, including both CA1
and the dentate gyrus. The projection to the dent-
ate gyrus, a ‘backprojection,’ has not been studied
in as much detail as the Schaffer collateral system,
leading to more appreciation of CA1–CA3 inter-
actions when CA3–dentate networks may actually
be just as important to hippocampal function.
Physiological data suggest that the CA3 backpro-
jection engages intermediary hilar neurons to exert
its primary effect on granule cells, although direct
projections to granule cells in a subset of CA3
neurons in ventral hippocampus have been shown
by anatomical methods. GABAergic inhibition
appears to regulate the ability of CA3 pyramidal
cells to exert a primarily inhibitory or excitatory
effect on granule cells, and to control the possible
reverberatory activity between the two subfields.
CA3–dentate interactions may subserve an impor-
tant associative function in the behaving animal,
influenced by factors that depress GABAergic
neuron activity or GABAergic transmission; this
may also apply under pathological conditions
when GABAergic neurons are damaged or lost,
such as temporal lobe epilepsy.
Acknowledgments
NIH (NINDS) and the New York State Depart-
ment of Health.
References
Acsady, L., Kamondi, A., Sik, A., Freund, T. and Buzsaki, G.
(1998) GABAergic cells are the major postsynaptic targets of
mossy fibers in the rat hippocampus. J. Neurosci., 18:
3386–3403.
Alle, H. and Geiger, J.R. (2006) Combined analog and action
potential coding in hippocampal mossy fibers. Science, 311:
1290–1293.
Amaral, D.G. and Witter, M.P. (1989) The three-dimensional
organization of the hippocampal formation: a review of an-
atomical data. Neuroscience, 31: 571–591.
Bilkey, D.K. and Schwartzkroin, P.A. (1990) Variation in
electrophysiology and morphology of hippocampal CA3 py-
ramidal cells. Brain Res., 514: 77–83.
Borck, C. and Jefferys, J.G. (1999) Seizure-like events in dis-
inhibited ventral slices of adult rat hippocampus. J. Ne-
urophysiol., 82: 2130–2142.
Bragdon, A.C., Taylor, D.M. and Wilson, W.A. (1986) Potas-
sium-induced epileptiform activity in area CA3 varies mark-
edly along the septotemporal axis of the rat hippocampus.
Brain Res., 378: 169–173.
Chang, L.W. and Dyer, R.S. (1985) Septotemporal gradients of
trimethyltin-induced hippocampal lesions. Neurobehav. Tox-
icol. Teratol., 7: 43–49.
Colom, L.V. and Saggau, P. (1994) Spontaneous interictal-like
activity originates in multiple areas of the CA2-CA3 region
of hippocampal slices. J. Neurophysiol., 71: 1574–1585.
Freund, T.F., Ylinen, A., Miettinen, R., Pitkanen, A., Lah-
tinen, H., Baimbridge, K.G. and Riekkinen, P.J. (1992) Pat-
tern of neuronal death in the rat hippocampus after status
epilepticus. Relationship to calcium binding protein content
and ischemic vulnerability. Brain Res. Bull., 28: 27–38.
Haas, K.Z., Sperber, E.F., Opanashuk, L.A., Stanton, P.K. and
Moshe, S.L. (2001) Resistance of immature hippocampus to
morphologic and physiologic alterations following status ep-
ilepticus or kindling. Hippocampus, 11: 615–625.
Ishizuka, N., Weber, J. and Amaral, D.G. (1990) Organi-
zation of intrahippocampal projections originating from
CA3 pyramidal cells in the rat. J. Comp. Neurol., 295:
580–623.
Johansen, F.F., Zimmer, J. and Diemer, N.H. (1987) Early loss
of somatostatin neurons in dentate hilus after cerebral is-
chemia in the rat precedes CA-1 pyramidal cell loss. Acta
Neuropathol. (Berl.), 73: 110–114.
Kneisler, T.B. and Dingledine, R. (1995) Synaptic input from
CA3 pyramidal cells to dentate basket cells in rat hippocam-
pus. J. Physiol., 487: 125–146.
Knowles, W.D., Traub, R.D. and Strowbridge, B.W. (1987)
The initiation and spread of epileptiform bursts in the in vitro
hippocampal slice. Neuroscience, 21: 441–455.
Li, X.G., Somogyi, P., Ylinen, A. and Buzsaki, G. (1994) The
hippocampal CA3 network: an in vivo intracellular labeling
study. J. Comp. Neurol., 339: 181–208.
Miettinen, R., Kotti, T., Tuunanen, J., Toppinen, A., Ri-
ekkinen Sr., P. and Halonen, T. (1998) Hippocampal damage
after injection of kainic acid into the rat entorhinal cortex.
Brain Res., 813: 9–17.
Scharfman, H.E. (1993a) Activation of dentate hilar neurons by
stimulation of the fimbria in rat hippocampal slices. Neurosci
Lett., 156: 61–66.
Scharfman, H.E. (1993b) Spiny neurons of area CA3c in rat
hippocampal slices have similar electrophysiological charac-
teristics and synaptic responses despite morphological vari-
ation. Hippocampus, 3: 9–28.
Scharfman, H.E. (1994a) EPSPs of dentate gyrus granule cells
during epileptiform bursts of dentate hilar ‘‘mossy’’ cells and
area CA3 pyramidal cells in disinhibited rat hippocampal
slices. J. Neurosci., 14: 6041–6057.

Scharfman, H.E. (1994b) Evidence from simultaneous intracellu-
lar recordings in rat hippocampal slices that area CA3 pyram-
idal cells innervate dentate hilar mossy cells. J. Neurophysiol.,
72: 2167–2180.
Scharfman, H.E. (1994c) Synchronization of area CA3 hippo-
campal pyramidal cells and non-granule cells of the dentate
gyrus in bicuculline-treated rat hippocampal slices. Neuro-
science, 59: 245–257.
Scharfman, H.E. (1995) Electrophysiological evidence that
dentate hilar mossy cells are excitatory and innervate both
granule cells and interneurons. J. Neurophysiol., 74: 179–194.
Scharfman, H.E., Kunkel, D.D. and Schwartzkroin, P.A. (1990)
Synaptic connections of dentate granule cells and hilar neu-
rons: results of paired intracellular recordings and intracellular
horseradish peroxidase injections. Neuroscience, 37: 693–707.
Scharfman, H.E., Goodman, J.H. and McCloskey, D.P. (2007)
Ectopic granule cells of the rat dentate gyrus. Dev. Neurosci.,
29: 14–27.
Scharfman, H.E., Smith, K.L., Goodman, J.H. and Sollas, A.L.
(2001) Survival of dentate hilar mossy cells after pilocarpine-
induced seizures and their synchronized burst discharges with
area CA3 pyramidal cells. Neuroscience, 104: 741–759.
Scharfman, H.E., Sollas, A.L., Smith, K.L., Jackson, M.B. and
Goodman, J.H. (2002) Structural and functional asymmetry
637
in the normal and epileptic rat dentate gyrus. J. Comp. Ne-
urol., 454: 424–439.
Schwartzkroin, P.A. and Prince, D.A. (1977) Penicillin-induced
epileptiform activity in the hippocampal in vitro preparation.
Ann. Neurol., 1: 463–469.
Schwerdtfeger, W.K. and Sarvey, J.M. (1983) Connectivity of
the hilar region of the hippocampal formation in the rat. J.
Hirnforschung, 24: 201–207.
Sloviter, R.S. (1987) Decreased hippocampal inhibition and a
selective loss of interneurons in experimental epilepsy. Sci-
ence, 235: 73–76.
Turner, D.A., Li, X.G., Pyapali, G.K., Ylinen, A. and Buzsaki,
G. (1995) Morphometric and electrical properties of recon-
structed hippocampal CA3 neurons recorded in vivo. J.
Comp. Neurol., 356: 580–594.
Urban, N.N., Henze, D.A. and Barrionuevo, G. (2001) Revis-
iting the role of the hippocampal mossy fiber synapse. Hip-
pocampus, 11: 408–417.
Wu, K., Canning, K.J. and Leung, L.S. (1998) Functional in-
terconnections between CA3 and the dentate gyrus revealed
by current source density analysis. Hippocampus, 8: 217–230.
Zimmer, J. (1971) Ipsilateral afferents to the commissural zone
of the fascia dentata, demonstrated in decommissurated rats
by silver impregnation. J. Comp. Neurol., 142: 393–416.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 35

Modeling the dentate gyrus

Robert J. Morgan1,, Vijayalakshmi Santhakumar2 and Ivan Soltesz1

1
Department of Anatomy and Neurobiology, 193 Irvine Hall, University of California, Irvine, CA 92697, USA
2
Department of Neurology, David Geffen School of Medicine, University of California, Los Angeles, CA, USA

Abstract: Computational modeling has become an increasingly useful tool for studying complex neuronal
circuits such as the dentate gyrus. In order to effectively apply computational techniques and theories to
answer pressing biological questions, however, it is necessary to develop detailed, data-driven models.
Development of such models is a complicated process, akin to putting together a jigsaw puzzle with the
pieces being such things as cell types, cell numbers, and specific connectivity. This chapter provides a
walkthrough for the development of a very large-scale, biophysically realistic model of the dentate gyrus.
Subsequently, it demonstrates the utility of a modeling approach in asking and answering questions about
both healthy and pathological states involving the modeled brain region. Finally, this chapter discusses
some predictions that come directly from the model that can be tested in future experimental approaches.

Keywords: computational model; epilepsy; granule cell; interneuron; mossy cell; sclerosis

The dentate gyrus in the mammalian brain is a
complex neuronal circuit that is thought to serve
as a gate for activity propagation throughout the
limbic system (Heinemann et al., 1992; Lothman
et al., 1992). Scientific understanding of such com-
plex systems has traditionally been obtained
through use of animal models, culture systems,
and numerous other experimental approaches. In
the past several decades, however, computational
modeling has become an increasingly useful tool for
studying and understanding neuronal network
structure and function. In fact, the dentate gyrus
has been the subject of a number of recent compu-
tational models looking at such issues as mossy fiber
sprouting (Lytton et al., 1998; Santhakumar et al.,
2005; Dyhrfield-Johnsen et al., 2007) and hilar cell
loss (Santhakumar et al., 2005; Dyhrfjeld-Johnsen
et al., 2007), interneurons and shunting inhibition
Corresponding author. Tel.: +1 (949) 824-3967;
E-mail: rjmorgan@uci.edu
(Bartos et al., 2001; Vida et al., 2006), homeostatic
plasticity (Howard et al., 2007), and topological al-
terations during epileptogenesis (Dyhrfjeld-Johnsen
et al., 2007). These studies demonstrate the impor-
tance and utility of computational modeling in con-
tributing to our understanding of the dentate gyrus
both in healthy animals and in pathological states
such as epilepsy.
In this chapter, we will use the analogy of a
jigsaw puzzle to walk through the process of
modeling the dentate gyrus neural network. In the
first and second parts, ‘‘Finding the right pieces’’
and ‘‘Putting the pieces together,’’ we will discuss
what types of considerations are necessary when
building a large-scale, data-driven model of a nor-
mal, healthy rat dentate gyrus (all data presented
in this chapter comes from the rat dentate). Then,
in ‘‘The big picture’’ we will examine how such a
model is useful for studying the effects of struc-
tural and functional network alterations that occur
during epileptogenesis. Finally, we will conclude

DOI: 10.1016/S0079-6123(07)63035-0 639

640
with ‘‘Tackling future puzzles and problems’’ by
discussing future directions for dentate modeling
and some predictions that can be tested experi-
mentally.
Finding the right pieces
Assembling a computational model is much like
putting together a puzzle. The pieces include cell
types, cell numbers, synapses, and other cellular
interactions such as gap junctions, and distributions
of axons and dendrites. These are assembled in a
specific fashion upon a backdrop that includes the
software necessary for implementing the model and
the hardware, the physical computer and memory
used for the model calculations. Just as the number
and size of pieces in a puzzle varies, the number of
components and the details of those components
vary dramatically in computational models. The
precise amount of detail necessary in a model de-
pends directly on the question the model is designed
to answer. Examining the effects of the loss of spe-
cific subgroups of cell types, for example, requires
the model to capture the relative changes in cell
densities, necessitating the development of large-
scale models. The first two sections of this chapter
will provide an overview of the general approach
for developing such a large-scale model to study
both the healthy rat dentate gyrus and the effects of
epileptogenesis with emphasis on mossy fiber
sprouting and hilar cell loss.
With over a million cells and a billion synapses
amongst at least eight distinct cell types in the
dentate gyrus, the first step in developing a model is
to determine the anatomy of the dentate network.
This can be accomplished by creating a connection
matrix of the dentate gyrus (Patton and Mcnaugh-
ton, 1995), which can be thought of as the border or
backbone of our puzzle. The matrix itself contains a
number of distinct pieces that are outlined here and
described in detail below. The first piece is the cell
type. Eight types of dentate cells can be identified as
anatomically well-described: granule cells (GCs),
mossy cells (MCs), basket cells (BCs), axo-axonic
cells, molecular layer cells with axonal projections
to the perforant path (MOPP cells), hilar cells with
axonal projections to the perforant path (HIPP
cells), hilar cells with axonal projections to the
commissural-associational pathway (HICAP cells),
and interneuron-specific (IS) cells (Fig. 1) (San-
thakumar and Soltesz, 2004). The second piece is
the number of cells of each type, data that can be
estimated from published data (Table 1). The third
piece is the specific connectivity between cell types.
This quantifies precisely how many postsynaptic
cells among each of the eight cell types a single
presynaptic neuron of a given type innervates (e.g.,
from the third row, second column in Table 1; a
single basket cell innervates approximately 1,250
GCs; mean and ranges are indicated, with refer-
ences). The final piece required for constructing the
connection matrix of the dentate is the spatial con-
straint in cell-to-cell connectivity. For each cell
type, the extent of the axons of single cells along the
septo-temporal axis of the dentate gyrus can be de-
termined from in vivo single cell fills published in
the literature (Fig. 2).
The first piece: cell types
Determination of cell type is based on several fac-
tors including: (1) the nature of the neurotrans-
mitter released; (2) location in the laminar
structure of the dentate gyrus; (3) morphological
features such as shape of the soma, dendritic ar-
bor, and axonal distribution pattern; and (4) pres-
ence of specific markers such as calcium-binding
proteins and neuropeptides (for detailed descrip-
tion of cell type discrimination see Freund and
Buzsaki, 1996). According to these criteria, two
excitatory cell types (GCs and MCs) and at least
six distinct inhibitory interneurons can be identi-
fied in the dentate gyrus (Fig. 1).
The second piece: cell numbers
A practical approach for determining cell numbers
for various cell types involves first identifying the
cell layers and the total number of neurons in each
layer. Then, it is possible to estimate subtype num-
bers by estimating the percentage of neurons that
express specific markers (Buckmaster et al., 1996,
2002a; Nomura et al., 1997a, b). The dentate gyrus
is composed of three cell layers: the molecular layer,
 641

Fig. 1. Schematic of the basic circuitry of the dentate gyrus. Relational representation of the healthy dentate gyrus illustrating the
network connections between the eight major cell types (GC, granule cell; BC, basket cell; MC, mossy cell; AAC, axo-axonic cells;
MOPP, molecular layer interneurons with axons in perforant-path termination zone; HIPP, hilar interneurons with axons in perforant-
path termination zone; HICAP, hilar interneurons with axons in the commissural/associational pathway termination zone; IS, in-
terneuron selective cells). The schematic shows the characteristic location of the various cell types within the three layers of the dentate
gyrus. Note, however, that this diagram does not indicate the topography of axonal connectivity or the somato-dendritic location of
the synapses. Reproduced with permission from Dyhrfjeld-Johnsen et al. (2007).

granule cell layer, and the hilus (Fig. 1). The GCs
are the primary projection cells of the dentate
gyrus. These are small, densely packed cells, typi-
cally located in the granule cell layer of the dentate
gyrus. The number of GCs in the rat dentate gyrus
is approximately 1,000,000 (Gaarskjaer, 1978; Boss
et al., 1985; West, 1990; Patton and Mcnaughton,
1995; Freund and Buzsaki, 1996). The other major
excitatory cell type, the MC, is located in the hilus
and does not express the GABA synthetic enzyme
glutamic acid decarboxylase (GAD). Buckmaster
and Jongen-Relo (1999) estimated the number of
GAD-mRNA negative neurons in the dentate hilus
(presumed MCs) as approximately 30,000. The
GABAergic interneurons are located in all three
layers of the dentate gyrus. These include the BCs,
parvalbumin-positive (PV) axo-axonic cells (whose
axon terminals project exclusively to the axon initial
segment of excitatory cells), somatostatin-positive
HIPP cells (Freund and Buzsaki, 1996; Katona et
al., 1999), nitric oxide synthase-positive HICAP
cells (Freund and Buzsaki, 1996), and aspiny and
calretinin-positive hilar IS cells (Gulyas et al., 1996).
The number of GABAergic cell types along the
granule cell-hilar border and in the hilus can be
determined based on published histochemical data
642

Table 1. Connectivity matrix for the neuronal network of the control dentate gyrus

Granule cells Mossy cells Basket cells Axo-axonic ells MOPP cells HIPP cells HICAP cells IS cells

Granule cells X 9.5 15 3 X 110 40 20

(1,000,000) X 7–12 10–20 1–5 X 100–120 30–50 10–30
Ref. [1–5] Ref. [6] Ref. [7] Ref. [6–9] Ref. [6,7,9] Ref. [6] Ref. [4,10,11] Ref. [4,7,10,11] Ref. [7]
Mossy cells 32,500 350 7.5 7.5 5 600 200 X
(30,000) 30,000–35,000 200–500 5–10 5–10 5 600 200 X
Ref. [11] Ref. [4,11–13] Ref. [12,13] Ref. [13] Ref. [13] Ref. [14] Ref. [12,13] Ref. [12,13] Ref. [15]
Basket cells 1250 75 35 X X 0.5 X X
(10,000) 1000–1500 50–100 20–50 X X 0–1 X X
Ref. [16,17] Ref. [4,16–19] Ref. Ref. Ref. [18] Ref. [18] Ref. [18] Ref. [18] Ref. [10,20]
[11,16,17,19] [16,17,20,21]
Axo-axonic 3000 150 X X X X X X
cells
(2000) 2000–4000 100–200 X X X X X X
Ref. [4,22] Ref. [4,18,22] Ref. Ref. [5,18] Ref. [5,18] Ref. [5,18] Ref. [5,18] Ref. [5,18] Ref. [5,18,19]
[4,5,11,14,23]
MOPP cells 7500 X 40 1.5 7.5 X 7.5 X
(4000) 5000–10,000 X 30–50 1–2 5–10 X 5–10 X
Ref. [11,14] Ref. [14] Ref. [14,24] Ref. [14,25] Ref. [14,26] Ref. [14,25] Ref. [14,20,25] Ref. [14,25] Ref. [14,15]
HIPP cells 1550 35 450 30 15 X 15 X
(12,000) 1500–1600 20–50 400–500 20–40 10–20 X 10–20 X
Ref. [11] Ref. [4,11,20] Ref. Ref. [4,11,20] Ref. [20,25] Ref. [25] Ref. [14,20,25] Ref. [25] Ref. [15,20]
[4,11,12,27,28]
HICAP cells 700 35 175 X 15 50 50 X
(3000) 700 30–40 150–200 X 10–20 50 50 X
Ref. [5,29,30] Ref. [4,11,20] Ref. [20] Ref. [4,11,20] Ref. [20] Ref. [14,20] Ref. [20] Ref. [20]
IS cells X X 7.5 X X 7.5 7.5 450
(3000) X X 5–10 X X 5–10 5–10 100–800
Ref. [15,29,30] Ref. [15] Ref. [15] Ref. [15,19] Ref. [15] Ref. [19] Ref. [19] Ref. [15]

Notes: Cell numbers and connectivity values were estimated from published data for granule cells, mossy cells, basket cells, axo-axonic cells, molecular
layer interneurons with axonal projections to the perforant path (MOPP), hilar interneurons with axonal projections to the perforant path (HIPP), hilar
interneurons with axonal projections to the commissural/associational pathway (HICAP) and interneuron specific cells (IS). Connectivity is given as
number of connections to a postsynaptic population (row 1) from a single presynaptic neuron (column 1). The average number of connections used in the
1:1 structural model is given in bold. Note, however, that the small world structure (see Section ‘‘Structural alterations in the dentate during epile-
ptogenesis’’) was preserved even if only the extreme low or the extreme high estimates were used for the calculation of L and C. References given in table
correspond to: 1Gaarskjaer (1978), 2Boss et al. (1985), 3West (1990), 4Patton and McNaughton (1995), 5Freund and Buzsaki (1996), 6Buckmaster and
Dudek (1999), 7Acsady et al. (1998), 8Geiger et al. (1997), 9Blasco-Ibanez et al. (2000), 10Gulyas et al. (1992), 11Buckmaster and Jongen-Relo (1999),
12
Buckmaster et al. (1996), 13Wenzel et al. (1997), 14Han et al. (1993), 15Gulyas et al. (1996), 16Babb et al. (1988), 17Woodson et al. (1989), 18Halasy and
Somogyi (1993), 19Acsady et al. (2000), 20Sik et al. (1997), 21Bartos et al. (2001), 22Li et al. (1992), 23Ribak et al. (1985), 24Frotscher et al. (1991), 25Katona
et al. (1999), 26Soriano et al. (1990), 27Claiborne et al. (1990), 28Buckmaster et al. (2002a), 29Nomura et al. (1997a), 30Nomura et al. (1997b). Reproduced
with permission from Dyhrfjeld-Johnsen et al. (2007).

(Buckmaster et al., 1996, 2002a; Nomura et al.,
1997a, b). Buckmaster and Jongen-Relo (1999) es-
timated the total number of GABAergic neurons in
the molecular layer of the dentate gyrus as approx-
imately 10,000. An even distribution between inner,
medial, and outer molecular layers is assumed,
yielding 4,000 MOPP cells with somata located in
the inner molecular layer (Han et al., 1993). Note
that this estimate excludes the molecular layer
interneurons that project primarily outside of the
dentate gyrus such as the outer molecular layer
interneurons projecting to the subiculum (Ceranik
et al., 1997).
The third piece: connectivity between cell types
In the past decade, a vast amount of high-quality
data about the connectivity of the dentate gyrus
has been collected, and this data serves as an in-
valuable resource in constructing the cell-type spe-
cific connectivity matrix for the dentate gyrus. The
connectivity for each cell type is summarized in
Table 1 and described below. Each value was de-
termined by a detailed survey of the anatomical
literature and was based on an assumption of uni-
form bouton density along the axon of the presy-
naptic cell. This assumption is in agreement with
 643

Fig. 2. Gaussian fits to experimentally determined distributions of axonal branch length used in construction of the models of the
dentate gyrus. (A) Plot shows the averaged axonal distribution of 13 granule cells (Buckmaster and Dudek, 1999) and the corre-
sponding Gaussian fit. (B) Fit to the septo-temporal distribution of axonal lengths of a filled and reconstructed basket cell (Sik et al.,
1997). (C) Fit to the axonal distribution of a CA1 axo-axonic cell (Li et al., 1992). (D) Gaussian fit to the averaged axonal distributions
of three HIPP cells from gerbil (Buckmaster et al., 2002a, b). (E) Fit to averaged axonal distributions of three mossy cells illustrates the
characteristic bimodal pattern of distribution (Buckmaster et al., 1996). (F) Histogram of the axonal lengths of a HICAP cell along the
long axis of the dentate gyrus (Sik et al., 1997) and the Gaussian fit to the distribution. All distributions were based on axonal
reconstruction of cells filled in vivo. In all plots, the septal end of the dentate gyrus is on the left (indicated by negative coordinates) and
the soma is located at zero. Reproduced with permission from Dyhrfjeld-Johnsen et al. (2007).

the in vivo data of Sik et al. (1997), and it greatly
simplifies the estimation of connectivity from an-
atomical data on axonal length and synapse den-
sity per unit length of axon.
Granule cells
Mossy fibers (GC axons) in a healthy rat dentate
gyrus are primarily restricted to the hilus (97%),
with few collaterals (3%) in the GC layer (Buck-
master and Dudek, 1999). In addition to MCs
(Acsady et al., 1998), mossy fibers have also been
shown to contact BCs (Buckmaster and Schwa-
rtzkroin, 1994; Geiger et al., 1997) and PV inter-
neurons (Blasco-Ibanez et al., 2000). With a total
of 400–500 synaptic contacts made by a single
mossy fiber (Acsady et al., 1998), the 3% of each
axon located in the GC layer (Buckmaster and
644

Dudek, 1999) is estimated to contact 15 BCs and 3 postsynaptic hilar interneuron receives two synap-
axo-axonic cells. In the hilus, a single GC forms tic contacts from a single mossy cell axon (Buck-
large, complex mossy fiber boutons that innervate master et al., 1996), each mossy cell is estimated to
7–12 MCs (Acsady et al., 1998), while an estimated contact 600 HIPP and 200 HICAP cells. There is
100–150 mossy fiber terminals target hilar inter- very low mossy cell to interneuron connectivity in
neurons with approximately one synapse per the inner molecular layer (Wenzel et al., 1997);
postsynaptic interneuron (Acsady et al., 1998). MCs could contact 5–10 basket and axo-axonic
Gulyas et al. (1992) estimated that a single spiny cells and approximately 5 MOPP cells with somata
calretinin-positive cell (presumed HIPP cell) is in the inner molecular layer (Han et al., 1993).
contacted by approximately 9,000 GCs. With
12,000 HIPP cells and 1,000,000 GCs, each GC
Basket cells
can be estimated to contact approximately 110
HIPP cells and 40 HICAP cells. Additionally, in
In the CA3 region of the rat hippocampus, each
agreement with the presence of mossy fiber termi-
principal cell is contacted by approximately 200
nals on aspiny calretinin-positive interneurons
BCs (Halasy and Somogyi, 1993), but a GC in the
(Acsady et al., 1998), 15 mossy fibers are expected
dentate could be contacted by as few as 30 BCs.
to synapse onto IS cells. Since mossy fibers avoid
Assuming that each of the 1,000,000 GCs is con-
the molecular layer (Buckmaster and Dudek,
tacted by 115 BCs each making 1–20 synaptic
1999) in the healthy dentate gyrus, it is assumed
connections (Halasy and Somogyi, 1993; Acsady
that they do not contact MOPP cells. However, in
et al., 2000), it can be estimated that each BC
animal models of temporal lobe epilepsy, sprouted
contacts approximately 1,250 GCs. MCs receive
mossy fibers have been shown to contact up to 500
10–15 basket cell synapses (Acsady et al., 2000),
postsynaptic GCs (Buckmaster et al., 2002b).
resulting in an estimate of 75 BC to mossy cell
synapses per BC. Approximately 1% of the 11,000
synapses made by a single basket cell axon in the
Mossy cells
GC layer are onto other BCs (Sik et al., 1997) with
3–7 synapses per postsynaptic cell (Bartos et al.,
A single filled mossy cell axon has been reported to
2001). Consequently, each BC in the dentate gyrus
make 35,000 synapses in the inner molecular layer
contacts approximately 35 other BCs. Since hilar
(Buckmaster et al., 1996; Wenzel et al., 1997). As-
and molecular layer interneurons are not a major
suming a single synapse per postsynaptic cell, a
target of basket cells (Halasy and Somogyi, 1993),
single mossy cell is estimated to contact
a single BC may contact 0–1 HIPP cells. Similarly,
30,000–35,000 GCs. Of the 2700 synapses made
the BC synapses onto axo-axonic cells, HICAP,
by a single MC axon in the hilus, approximately
and MOPP cells are assumed to be negligible. As
40% target GABA-negative neurons (Wenzel et
PV cells preferentially contact other PV-positive
al., 1997). As each mossy cell is estimated to make
cells in the hilus (Acsady et al., 2000), we assume
1–5 synaptic contacts on a single postsynaptic
that BCs do not contact calretinin-positive IS cells
mossy cell (Buckmaster et al., 1996), it is estimated
(Gulyas et al., 1992).
that each mossy cell contacts approximately 350
other MCs (this is likely to be a generous estimate
since it is based on the assumption that all GAD Axo-axonic cells
negative somata in the hilus represent MCs). The
remaining 60% of the hilar mossy cell axons target Most synapses made by axo-axonic cell axons are
GABA-positive cells (Buckmaster et al., 1996; thought to target GC axon initial segments
Wenzel et al., 1997), with no reports demonstrat- (Halasy and Somogyi, 1993). However, a small
ing mossy cell synapses onto IS cells. Assuming fraction of axon collaterals also descend into the
that there is no preferential target selectivity be- superficial and deep hilus (Han et al., 1993; Freund
tween HIPP and HICAP cells and that each and Buzsaki, 1996). In neocortex, an axo-axonic

cell makes 4–10 synapses on the axon intial seg-
ment of a postsynaptic cell (Li et al., 1992). With
22,000,000 estimated axon initial segment synapses
in the GC layer (Halasy and Somogyi, 1993) and 4
synapses per postsynaptic cell (based on the data
from the neocortex from Li et al., 1992), each of
the 2,000 axo-axonic cells are estimated to target
approximately 3,000 GCs. MCs receive axo-axonic
cell input (Ribak et al., 1985), and with the com-
paratively small fraction of axons from axo-axonic
cells in the hilus (Han et al., 1993; Freund and
Buzsaki, 1996), it can be estimated that axo-axonic
cells target a number of MCs equal to approxi-
mately 5% of their GC targets, which results
in approximately 150 MCs. Since axo-axonic
cells primarily target the axon initial segment
of non-GABAergic cells (Halasy and Somogyi,
1993; Freund and Buzsaki, 1996), it may be
assumed that these cells do not project to inter-
neurons.
HIPP cells
HIPP cells have been estimated to contact about
1,600 GCs and 450 BCs with 1–5 synapses per
postsynaptic cell (Sik et al., 1997). MCs can have
one dendrite in the molecular layer (Amaral, 1978;
Ribak et al., 1985; Scharfman, 1991) which can be
targeted by HIPP cell axons, whereas GCs have two
primary dendrites (Claiborne et al., 1990; Lubke
et al., 1998). Since the mossy cell to GC ratio is
approximately 1:30, the mossy cell dendrites con-
stitute only approximately 1/60 of the targets for
HIPP cells. Increasing this fraction to approxi-
mately 1/45 to account for the presence of a few
HIPP cell contacts on MC in the hilus (Buckmaster
et al., 2002a) results in an estimate that each HIPP
cell contacts approximately 35 MCs. HIPP cell
axonal divergence onto HICAP and MOPP cells in
the molecular layer can be assumed to be similar to
that of somatostatin-positive cells in CA1 (Katona
et al., 1999) and estimated to be 15 connections
onto each population. The HIPP cell axonal diver-
gence to axo-axonic cells is estimated to be between
the divergence to basket and HICAP cells; therefore
the HIPP cell axon likely contacts 30 axo-axonic
cells.
645
MOPP cells
MOPP cells target an estimated 7,500 GCs in the
rat dentate gyrus. While MOPP cell axons project
in the horizontal axis to a similar extent as HIPP
cells, they show considerably less collateralization
(Han et al., 1993), resulting in an estimate of half
as many synapses onto MOPP and HICAP cells as
HIPP cells make. As MOPP cell axons are re-
stricted to the molecular layer (Han et al., 1993)
and do not target the basal dendrites of BCs, they
are assumed to contact less than 1/10 the number
of BCs targeted by HIPP cells. Likewise, MOPP
cells with axons restricted to the outer and middle
molecular layers (Han et al., 1993) would not tar-
get the hilar dendrites of axo-axonic cells (Soriano
et al., 1990) or the axo-axonic cells with somata
and proximal dendrites in the hilus (Han et al.,
1993). It is estimated that MOPP cells only contact
1–2 axo-axonic cells. As the MOPP cell axonal
arbors in the molecular layer (Han et al., 1993) do
not overlap with major parts of the dendritic ar-
borizations of MCs (Frotscher et al., 1991), HIPP
cells (Han et al., 1993; Sik et al., 1997; Katona et
al., 1999) or IS cells (Gulyas et al., 1996), the con-
nectivity to these cells is negligible.
HICAP cells
Sik et al. (1997) estimated that the septo-temporal
extent and bouton density of HICAP cell axons is
similar to the HIPP cell axons, whereas the esti-
mated axonal length of HICAP cells is approxi-
mately half of the HIPP cell axonal length. Thus, it
is estimated that HICAP cells contact about half
the number of GCs contacted by HIPP cells.
However, since HICAP cells have an additional
3% of axon collaterals in the hilus (Sik et al.,
1997), their number of postsynaptic MCs can be
assumed to be the same as that of the HIPP cells.
HICAP cells are assumed to contact less than half
the number of BCs targeted by HIPP cells (
175)
and a negligible number of axo-axonic cells. With
a total of 26,000 synapses from a single HICAP
cell axon (Sik et al., 1997), approximately 700
synapses should be present in the hilus. Assuming
2–5 synapses per postsynaptic cell, each HICAP
646
cell could contact 100–300 hilar cells. Each HI-
CAP cell is assumed to target 50 HIPP and HI-
CAP cells, which, along with 35 synapses on MCs,
is in the estimated range. Although the total ax-
onal length of HICAP cells is only about half of
that of HIPP cells, the number of MOPP cells
targeted is assumed to be the same (
10–20), as
the HICAP cell axons primarily project to the in-
ner molecular layer where both cell bodies and
proximal dendrites of MOPP cells are located
(Han et al., 1993).
IS cells
IS cells contact an estimated 100–800 other IS cells
and 5–10 (presumably CCK positive) BCs (Gulyas
et al., 1996). Acsady et al. (2000) suggested that
CCK cells would include both basket cell and HI-
CAP morphologies and that IS cells also project to
somatostatin-positive HIPP cells. These data result
in an estimate that IS cells project to 5–10 HICAP
cells and HIPP cells.
The fourth piece: spatial constraints and axonal
distributions
In addition to the number of connections made by
any given cell, the distribution of these connections
along the septo-temporal axis of the dentate gyrus
is very important. Fortunately, in vivo single cell
fills have been performed which provide a very
complete description of the axonal extent of the
eight cell types discussed above. Averages of these
in vivo axonal distributions can be fit with either a
single or double Gaussian for each cell type, which
then defines the model axonal distribution for each
cell of that type (Fig. 2). These Gaussians then
describe the probability of any given cell connect-
ing to another cell based on the distance between
the two.
Putting the pieces together
Once the border of the puzzle is complete (the
connection matrix in our analogy), it is possible to
start assembling the bulk of the picture. In the case
of modeling, this corresponds to a number of dis-
tinct obstacles. Single cell models must be created
with sufficient detail to function appropriately in
the network; network scaling must be implemented
(i.e., what percentage of the actual cell numbers,
connections, etc., will be represented in the net-
work); receptor types and synaptic data must be
included; and network topology must be deter-
mined (will the network be a strip or a ring, e.g.,
and how will connection probabilities be imple-
mented to account for the shape and size of the
network). Finally, the stimulation paradigm must
be determined so that the fully connected network
will actually start to produce data. These obsta-
cles, and ways to overcome them, are discussed in
detail in this section.
Multicompartmental single cell models
In a network model that proposes to simulate both
somatic and dendritic inputs and model synaptic
contacts on specific dendritic compartments, it is
necessary to develop multicompartmental, biophys-
ically realistic models of individual cell types. It is
not always necessary to develop these models from
scratch, however, as ModelDB (http://sense-
lab.med.yale.edu/senselab/modeldb/) is a search-
able resource that makes published models of
several cell types including dentate GCs available
for download (Aradi and Holmes, 1999a; San-
thakumar et al., 2005). These models can be readily
adapted to a wide variety of networks. Morpho-
logical properties for multicompartmental models
of other cell types not available on ModelDB, such
as some of the other dentate cell types, can be
developed based on the data reported in the liter-
ature (Buckmaster et al., 1993, 2002b; Freund and
Buzsaki, 1996; Geiger et al., 1997; Bartos et al.,
2001).
The intrinsic properties of cell types can be
modeled based on data from whole cell physio-
logical experiments obtained in the presence of
blockers of synaptic activity (Lubke et al., 1998).
Detailed descriptions of how to model passive and
active properties of individual cells can be found in
The Neuron Book (Carnevale and Hines, 2006), but
we will discuss a couple of relevant issues here

briefly. First, since GCs (Desmond and Levy,
1985) and MCs (Amaral, 1978) are rich in dendri-
tic spines, it is necessary to account for the mem-
brane area contribution of the spines (Rall et al.,
1992). This can be accomplished by decreasing
membrane resistivity (increasing leak conduct-
ance) and increasing membrane capacitance
(Aradi and Holmes, 1999b). An additional con-
sideration is the presence of spontaneous activity
in some cell types such as MCs (Ishizuka et al.,
1995; Ratzliff et al., 2004), as well as their lower
input resistance in the presence of background
synaptic activity (Ratzliff et al., 2004), compared
to the input resistance in ionotropic glutamate and
GABA receptor antagonists (Lubke et al., 1998).
Spontaneous firing rate can be simulated via a
constant direct current injection (Santhakumar
et al., 2005), and the synaptic background activity
can be simulated as a fluctuating point conduct-
ance with balanced excitation and inhibition, as
described in Destexhe et al. (2001).
Both excitatory cell types (GCs and MCs) are
well defined and have a vast amount of electro-
physiological data available for the construction of
single cell models. However, the wealth of infor-
mation available for the interneurons is substan-
tially less for many cell types. Additionally, the
role of interneurons in the network must be con-
sidered with regard to the question being asked.
For example, in a model network that is pursuing
the question of the role of mossy fiber sprouting
and hilar cell loss in epileptogenesis (Santhakumar
et al., 2005; Dyhrfjeld-Johnsen et al., 2007), several
factors must be considered. First, inhibition can
operate in two principal modes. One mode is feed-
forward inhibition, where an interneuron is acti-
vated by the same afferent input as the excitatory
neuron that it contacts. The basket cell is the pri-
mary source of feed-forward inhibition in the
dentate gurus (Freund and Buzsaki, 1996). The
second mode is feedback inhibition, where inter-
neurons are activated by excitatory cells which
they subsequently inhibit. The BCs and HIPP cells
are the two key feedback inhibitory cells in the
dentate (Freund and Buzsaki, 1996; Buckmaster et
al., 2002a). Another important aspect of inhibition
is the location of synaptic contact between inter-
neuron and principal cell. Perisomatic inhibition
647
by BCs is considered crucial in maintaining weak
GC activity in response to afferent input. The
HIPP cells, on the other hand, synapse on the
dendritic region near the afferent inputs, and they
are likely to modulate integration of dendritic in-
puts. The loss of this dendritic inhibition with an
essentially intact or increased somatic inhibition
has been proposed to be particularly relevant to
epileptogenesis (Cossart et al., 2001). Further-
more, BCs are resistant to cell death whereas HIPP
cells are extremely vulnerable in epileptic tissue
(Buckmaster et al., 2002a). These features make
the basket and HIPP cells essential constituents of
the dentate model.
The remaining four interneurons also possess
unique features of inhibition. The MOPP cells op-
erate by providing GCs with feed-forward dendri-
tic inhibition (Halasy and Somogyi, 1993). The
HICAP cells synapse on the proximal dendrites of
GCs, near where mossy cell axons terminate and
provide feedback inhibition (Freund and Buzsaki,
1996). The axo-axonic cells provide GABAergic
input at the axon initial segments of GCs and MCs
and potentially play powerful roles in modifying
spike initiation (Freund and Buzsaki, 1996; Ho-
ward et al., 2005). The IS cells connect exclusively
to other interneurons and form a well connected
network that could modulate excitability and
synchrony of the network.
Unfortunately, detailed physiological data for
the development of realistic models of MOPP,
HICAP, axo-axonic, and IS cells are not readily
available. Therefore, in an approach that avoids
errors of commission one could choose to include
only basket and HIPP cells in the network. How-
ever, this makes it essential to perform control
simulations to examine whether augmenting inhi-
bition (in order to compensate for excluded cell
types) influences the overall outcome of the net-
work simulations.
Network scaling
Once the single cell models are complete and ready
to go, it is time to start thinking about how big the
network is going to be. With the immense in-
creases in computational power over the last
648
couple of decades, very large networks are quite
feasible. However, some questions may not require
such large networks, and often computational re-
sources will be limited. Also important in deter-
mining what scale a network should be is whether
the network includes functional model cells that
require a great deal of computational power to
process. Purely structural network models (that do
not use multicompartmental functional single cell
models), also called network graphs, can provide a
great deal of useful information about neuronal
circuits. Indeed, Dyhrfjeld-Johnsen et al. (2007)
created an extraordinarily large 1:1 scale structural
model of the dentate gyrus that yielded a great
deal of new data and will be discussed briefly in
Section ‘‘The big picture’’.
Putting purely structural graphs aside for now
though, and focusing on the functional model that
we are building in this chapter, we can determine
how to scale the cell numbers. We know that the
GC:MC:BC:HIPP ratio is 1000:30:10:12. In agree-
ment with these ratios, the Lytton et al. (1998)
model that focused on the network effects of
sprouting included 50 GCs, 2 MCs, and 2 inter-
neurons. Using this relatively small network repre-
sentation of the dentate they demonstrated that the
effect of sprouting on enhancing network excitabil-
ity was strongly limited by the intrinsic firing prop-
erties of GCs (e.g., strong spike frequency
adaptation). Although this model made the predic-
tion that mossy fiber sprouting could lead to hyper-
excitability, the network was highly interconnected
as a consequence of the small size. This network
architecture, combined with a very strong perforant
path synaptic conductance, could have resulted in
artificial synchronous activity in the network. In-
creasing the size of the Lytton et al. (1998) model by
10-fold entails building a 1:2000 scale dentate with
500 GCs, 15 MCs, 6 BCs and 6 HIPP cells. This
500+ cell model can be readily simulated on a
laptop and allows for manipulating hilar cell num-
bers and connectivity patterns (Santhakumar et al.,
2005). However, even in a network of this size, ar-
tificially large conductances as well as other factors
(e.g., strong edge effects and unrealistic distance
dependent cell-type specific connectivity) could in-
fluence the results. With the current computational
power of state-of-the-art multiprocessor computers,
the size of a realistic dentate model can be scaled up
further to a 1:20 scale model with 50,000 GCs
(Dyhrfjeld-Johnsen et al., 2007). The specific con-
nection parameters that can be used for a model of
this size are shown in Table 2. In Section ‘‘The big
picture,’’ we will discuss results obtained from this
very large-scale network with direct implications for
epileptogenesis.
Receptor types and synapses
Building a model of a dentate network with hun-
dreds or thousands of biophysically realistic multi-
compartmental model cells is only a meaningful
venture if those cells have a way to talk to one an-
other. For this reason, we must consider neuro-
transmitter-receptor types that must be included in
the network. Again, the types of receptors, synapses,
and other cellular interaction devices (such as gap
junctions) to include in the network will depend on
the particular question the network is designed to
answer. For the example network we have been
building throughout this chapter, the goal is to de-
termine the effects of mossy fiber sprouting and hilar
cell loss on excitability. To examine these effects, the
minimal requirement is to incorporate ionotropic
glutamatergic AMPA synapses and GABAA
synapses. Experiments from both normal and epi-
leptic animals are constantly ongoing, and as the
data become available it will be possible to add
other receptors such as NMDA and metabotropic
receptors. It would also be interesting to incorporate
greater diversity at individual synapses through in-
corporation of mechanisms such as short-term plas-
ticity. Each addition to the network must be
carefully considered, however, as greater complex-
ity necessarily increases the load on the computa-
tional resources. Since the basic circuit effects of
sprouting and cell loss can be simulated by including
AMPA and GABAA synapses, these will be the
types included in our example model.
Postsynaptic conductances can be represented as
a sum of two exponentials (Bartos et al., 2001).
The peak conductance (gmax), rise and decay time
constants, and the synaptic delays (distinct from
the axonal conduction delay described later) for
each network connection can be estimated from
 649

Table 2. Parameters of the 50,000+ neuron functional model network

Functional model network parameters

From To –> GC MC BC HC

Granule cellsa Convergence 68.03 78.05 370.95 2266.64

(50,000) Divergence 68.03 2.34 3.71 27.19
Synapse weight (nS) 1.00 0.20 0.94 0.10
Mossy cells Convergence 243.62 87.23 5.59 375.53
(1500) Divergence 8120.82 87.23 1.86 150.21
Synapse weight (nS) 0.30 0.50 0.30 0.20
Basket cells Convergence 3.11 6.31 8.98 n/a
(500) Divergence 313.22 18.93 8.98 n/a
Synapse weight (nS) 1.60 1.50 7.60 n/a
HIPP cells Convergence 4.82 3.76 140.13 n/a
(600) Divergence 401.86 9.39 116.77 n/a
Synapse weight (nS) 0.50 1.00 0.50 n/a

Perforant pathb Synapse weight (nS) 20 17.5 10 n/a

Notes: The cell numbers (column 1) and synaptic connectivity values and strengths in the functional model network used for the activity simulations in
Fig. 4. Note that this network is smaller (50,000+ cells) than the full-scale dentate gyrus (over 1 million cells) model described in Section ‘‘Structural
alterations in the dentate during epileptogenesis’’. Therefore, the connectivity had to be adjusted from what is shown in Table 1. Convergence is given as
number of connections onto a single postsynaptic neuron (row 1) from a presynaptic neuronal population (column 1). For example, 243 mossy cells
converge on a single granule cell in this network. Divergence is given as the number of connections to a postsynaptic population (row 1) from a single
presynaptic neuron (column 1). For example, a single mossy cell makes synapses on 8120 postsynaptic granule cells in this network. The strengths of the
connections are given in nS. For example, the strength of the excitatory synapse formed by a single mossy cell on a single granule cell is 0.3 nS.
a
Granule cell to granule cell connections represent values at 100% sprouting.
b
Perforant path input to 5000 granule cells (2 synapses each), 50 basket cells (2 synapses each) and 10 mossy cells (1 synapse each) in the central 10 bins of
the network model. Reproduced with permission from Dyhrfjeld-Johnsen et al. (2007).

experimental data (Kneisler and Dingledine, 1995; Network topology

Geiger et al., 1997; Bartos et al., 2001; Santhaku-
mar et al., 2005) and details of all of the specific
cellular and synaptic properties can be found in
Santhakumar et al. (2005). When making such es-
timates, care must be taken to ensure that compa-
rable conditions were used in the experiments that
provide the data. For example, since the data for
synaptic conductances is typically obtained by
paired recording experiments from several differ-
ent groups, the temperature at which the record-
ings were performed must be accounted for. Q10
estimates can be used to convert kinetic data from
room temperature recordings to the appropriate
values at physiological temperature. Axons can be
modeled implicitly by activating a given synapse
after either a static or distance-dependent delay
once a presynaptic cell crosses a preset membrane
potential threshold (Bartos et al., 2001).
The shape of the network and the way in which
connections are made depend strongly on the size
of the network. For instance, a network of ap-
proximately 50 or 500 cells (Lytton et al., 1998;
Santhakumar et al., 2005, respectively) may re-
quire a ring structure to avoid edge effects (i.e.,
cells on the edge of a network have as few as half
as many pre- and postsynaptic targets compared
to cells in the middle). A network of 50,000+
cells (Dyhrfjeld-Johnsen et al., 2007), on the other
hand can be set up as a linear strip, more similar
to the actual topology of the biological dentate.
Similarly, small networks may require non-topo-
graphic connectivity. That is, the postsynaptic
targets of each cell are selected at random from
the pool of potential target neurons while main-
taining the cell type specific divergence and
650
convergence. Larger networks can incorporate
the spatial rules discussed above in Section 7 The
fourth piece: Spatial constraints and axonal dis-
tributions; the connectivity between cells can be
distributed according to the experimentally de-
rived axonal distributions of the various cell types
(Fig. 2).
Network stimulation
A network such as the dentate gyrus, in which
most cells are not designed to be spontaneously
active, requires some afferent input to initiate net-
work activity. This input can come in a number of
forms, again dependent on the specific question as
well as the particular network being modeled. For
the dentate gyrus as a whole, it makes sense to
simulate a perforant path input from the ent-
orhinal cortex. This input to the network is located
on both dendrites of all GCs and the apical dend-
rites of all BCs. Since 15% of MCs have also been
shown to receive direct perforant path input
(Scharfman, 1991; Buckmaster et al., 1992), the
number of MCs that should receive input in a
given model can be easily calculated. Mass stim-
ulation of the perforant path can be simulated by
activating a maximum peak AMPA conductance
in cells postsynaptic to the stimulation (Santhaku-
mar et al., 2000). The strength of the perforant
path input in the 50,000+ cell functional model
from Dyhrfjeld-Johnsen et al. (2007) is given in
Table 2 (note the very large synaptic weight for
perforant path input compared to the small
weights for cell to cell synapses). In the biological
network and in physiological studies it is unlikely
that the entire network is activated simultane-
ously. Rather, focal activation can be simulated by
activating the input to GCs and BCs located in a
model hippocampal ‘‘lamella’’. If focal network
activation is undesirable, spontaneous network ac-
tivity can be simulated instead. This can be ac-
complished by providing uncorrelated activation
(perforant path inputs with Poisson distributed
interspike intervals) to each granule cell, basket
cell, and 15% of the MCs. Other methods for in-
itiation of network activity can be devised as nec-
essary.
The big picture
Following essentially the strategy outlined in the
previous two sections, dentate gyrus modelers
have assembled a number of different puzzles,
producing dentate models of a wide range of sizes
aimed at answering a large number of important
questions. We have walked through the creation of
a very large-scale dentate model capable of han-
dling functional network simulations with greater
than 50,000 cells. However, we have not yet dis-
cussed how a model of this type can help us gain a
better understanding of disease processes and
hopefully how to treat them. Without the ability
to provide information of this nature, models are
simply incomplete representations of reality and of
little use. Fortunately, models can provide a great
deal of information because they give scientists
complete control over variables that cannot be
isolated in typical experimental situations. In this
section, we will briefly describe an example in
which using large scale dentate gyrus modeling al-
lowed Dyhrfjeld-Johnsen et al. (2007) to isolate
structural changes in the dentate during epilepto-
genesis and conclude that structural changes alone
can lead to hyperexcitability.
Structural alterations in the dentate during
epileptogenesis
As mentioned throughout this chapter, two pri-
mary structural alterations occur in the dentate
gyrus during epileptogenesis: mossy fiber sprout-
ing and hilar cell death (including hilar interneu-
rons and MC). These two processes (which we will
refer to simply as ‘‘sclerosis,’’ defined here as the
two common structural changes in the dentate as-
sociated with mesial temporal lobe sclerosis) can
be implemented in a dentate gyrus model accord-
ing to the data driven approach given above. De-
tailed descriptions of how to implement both of
these processes can be found in Santhakumar et al.
(2005) and Dyhrfjeld-Johnsen et al. (2007). Briefly,
a maximal level of sclerosis can be simulated by
adding an average of approximately 275 granule
cell to granule cell connections (Buckmaster et al.,
2002b) and by removing all hilar cells from the

network. Using this maximum value as a point of
reference, the severity of sclerosis can be approx-
imated by simulating percentages of the maximum
(e.g., at 50% sclerosis, simulating epileptogenesis
following a moderate head injury (Howard et al.,
2007), each granule cell would contact approxi-
mately 137 other GCs, and 50% of hilar cells
would be lost).
A 1:1 scale structural model of the dentate gyrus
can be constructed where each cell is simply a node
in the computer (nodes have no functional prop-
erties and are therefore much less computationally
demanding than the multicompartmental cell
models described previously) and synapses be-
tween cells are represented as non-functioning di-
rected links (Dyhrfjeld-Johnsen et al., 2007). This
structural network can then be utilized to under-
stand the basic structure of the healthy dentate
gyrus and how that structure changes during epile-
ptogenesis.
In order to understand how network structure
can be measured and later how structural changes
have functional implications, it is necessary to ex-
plain the parameters that define network topology.
Two measures that Watts and Strogatz (1998)
originally employed to assess the structure of the
nervous system of the worm C. elegans can be used
to analyze the structure of the dentate network: the
average path length, L, and the average clustering
coefficient, C. The average path length is defined
as the average number of steps required to move
from any node to any other node in the network,
and it is therefore a measure of how well connected
the network is globally. The average clustering
coefficient, on the other hand, is a measure of local
connectivity. It is defined as the fraction of all
possible connections between ‘‘postsynaptic’’
nodes of a given node that are actually formed.
The average clustering coefficient for an entire
network is simply the average of the clustering
coefficients for each node in the network.
The topological measures, C and L, can be best
thought of in terms of a social network, or a net-
work of friends. C measures the probability that
any two friends of a given person also know each
other. L measures how many steps in a chain of
friends or acquaintances would be necessary to
connect two random people. A common theory
651
that utilizes this parameter is that anyone on the
planet is separated from anyone else by up to six
degrees of separation, or six people. While this
theory has never been satisfactorily proven, it has
led to interesting experiments and even a party
game called the Six Degrees of Kevin Bacon, in
which the participants have to connect any actor
or actress to Kevin Bacon through their film roles
in as few steps as possible.
Based on these two topological measures, net-
works can be divided into three classes: (1) Reg-
ular, with a long path length and large clustering
coefficient; (2) Random, with a short path length
and small clustering coefficient; and (3) Small
World, with a short path length and large cluster-
ing coefficient. Representative graphs of these
types of networks are displayed in Fig. 3.
As shown, the regular network (Fig. 3A) is
characterized by a high local connectivity, but the
distance (via connected nodes) between any two
random nodes, especially nodes at the ends of the
graph, can be quite large. On the other hand, the
random network (Fig. 3B) contains many long
distance connections, contributing to a very short
average path length. There is little local connec-
tivity however, as nodes are just as likely to con-
nect to far away nodes as they are to their
neighbors. The small world network (Fig. 3C) acts
as a sort of compromise between the random and
regular graphs. Indeed, there is a high degree of
local connectivity combined with a number of
long-range connections, yielding a both locally
and globally well-connected network. Small world
networks can also be thought of in terms of social
networking. Humans often form a cohesive group
of friends who are well connected locally due to
various constraints such as geography, common
activities and so on. Additionally, any given per-
son in the group may know some people who are
totally unrelated to the local group, but who each
have cohesive local groups of their own, thus cre-
ating numerous long distance connections between
the locally well-connected networks.
So how do these L and C values play a role in
understanding the dentate gyrus? Analysis of the
structural model of the dentate gyrus in the healthy
state, without any mossy fiber sprouting or hilar cell
loss, indicates that the dentate gyrus is a small
652

Fig. 3. Schematics of three basic network topologies. (A) Regular network topology. The nodes in a regular network are connected to
their nearest neighbors, resulting in a high degree of local interconnectedness (high clustering coefficient C), but also requiring a large
number of steps to reach other nodes in the network from a given starting point (high average path length L). (B) Random network
topology. In a random network, there is no spatial restriction on the connectivity of the individual nodes, resulting in a network with a
low average path length L but also a low clustering coefficient C. (C) Small world network topology. Reconnection of even a few of the
local connections in a regular network to distal nodes in a random manner results in the emergence of a small world network with a
high clustering coefficient C but a low average path length L. Reproduced with permission from Dyhrfjeld-Johnsen et al. (2007).

world network. It has a very high clustering coeffi-
cient and a path length that is only 2.68, approx-
imately the same as the path length of the C.
elegans nervous system, a nervous system with only
282 cells! This finding means that any cell is con-
nected to any of the other one million cells in the
dentate network by less than three synapses, and
cells are highly connected locally, resulting in fast
local computations and the ability to efficiently re-
lay signals to distant parts of the network.
While the analysis of the healthy dentate gyrus
provides useful and novel information about its
structure, perhaps the most interesting and coun-
terintuitive structural information comes from the
analysis of networks undergoing sclerosis. These
networks demonstrate that the small world charac-
teristics of the dentate gyrus are actually enhanced
up to approximately 80% sclerosis, despite a mas-
sive loss of connections (see Section ‘‘Mossy cells’’
for MC connectivity; note the massive divergence
onto GCs that is lost as sclerosis progresses and
MCs die). In fact, the total number of links in the
network decreases by 74% while the total number
of nodes lost due to hilar cell death is only 4.5% at
maximal sclerosis, yet the network becomes seem-
ingly more locally and globally well connected. This
finding predicts that during epileptogenesis, the
dentate gyrus may become more readily able to
transmit information throughout the network and
likely increase synchronous firing as well, phenom-
ena that could very well contribute to a seizure
phenotype. It is only at nearly 100% sclerosis that
the dentate gyrus network transforms into a more
regular network structure, predicting a counterin-
tuitive decrease in hyperexcitability at maximal
sclerosis.
Functional implications of structural alterations
Now that we have discussed the structural analysis
of the healthy and sclerotic dentate gyrus, it re-
mains to be seen whether network activity is ac-
tually influenced by the structural changes in the
ways we might expect based on the predictions in
the preceding section. As in the 1:1 scale structural

model, sclerosis can be implemented in a func-
tional model of the dentate gyrus. This model
contains approximately 50,000 cells, owing to the
increased need for computational resources to
perform the calculations necessary with multicom-
partmental biophysically realistic cells (see Section
‘‘Multicompartmental single cell models’’). Simu-
lations can be performed as sclerosis progresses in
steps (e.g., perform simulations at 20 then 40%,
and so on), and activity in the network can be
quantified by a number of different measures.
Four different measures were used in Dyhrfjeld-
Johnsen et al. (2007): (1) duration of granule cell
firing (the time of the last granule cell action po-
tential in the network minus the time of the first
granule cell action potential); (2) average number
of action potentials per granule cell; (3) time until
activity of the most distant GCs from the stimu-
lation point (i.e., latency to full network activity);
and (4) synchrony of granule cell action potentials.
Results of the functional simulations reveal that
network activity very closely parallels structural al-
terations in the dentate gyrus. The healthy non-scle-
rotic dentate model shows very limited firing in
response to a single perforant path input (Fig. 4A),
in accordance with the biological data (Santhaku-
mar et al., 2001). Additionally, as sclerosis pro-
gresses, the functional network becomes increasingly
hyperexcitable up to approximately 80% sclerosis
(Figs 4B–E), again agreeing with in vitro measures
of epileptiform activity (Rafiq et al., 1995) and in
accordance with the enhanced small world features
revealed by the structural analysis. A very interest-
ing phenomenon then occurs at levels of sclerosis
exceeding 80%. Hyperexcitability of the functional
model network actually decreases (Fig. 4F)! The
level of sclerosis at which hyperexcitability dimin-
ishes corresponds perfectly to the point at which the
network transforms from a small world structure to
a more regular network. Thus, the functional net-
work follows the same pattern that is seen in the
structural network, and the functional effects are
solely due to the structural network alterations, as
no intrinsic cellular or synaptic properties are altered
in the model as sclerosis progresses. Interestingly,
network dynamics are also affected by structural
changes; a relatively uniform pattern of granule cell
activity (from 40 to 60% sclerosis; Fig. 4C, D)
653
transforms into a pattern with distinct waves of ac-
tivity (from 80 to 100% sclerosis; Fig. 4E, F) that
can collide and mutually annihilate (Netoff et al.,
2004; Roxin et al., 2004).
These findings not only support the important
role of structural alterations in determining func-
tional network activity, but they are in agreement
with experimental observations in both epileptic
animals and humans. Indeed, no studies that have
quantified hilar cell loss in animal models of epi-
lepsy ever reported 100% cell loss (Cavazos and
Sutula, 1990; Cavazos et al., 1994; Leite et al., 1996;
Buckmaster and Dudek, 1997; Mathern et al., 1997;
Buckmaster and Jongen-Relo, 1999; Gorter et al.,
2001; van Vliet et al., 2004; Zappone and Sloviter,
2004). Additionally, in surgically removed speci-
mens from pharmacologically intractable human
temporal lobe epilepsy patients (Gabriel et al.,
2004), cell counts showed that only approximately
80% of hilar cells were lost, even in patients with
severe sclerosis (Blumcke et al., 2000). This finding
coincides perfectly with the period of maximal
hyperexcitability in the functional model networks,
when sclerosis reaches 80%.
Understanding the limitations of models
It is important to consider that no model can ever
fully replicate the system that it is modeling. As a
result, it is crucial to be certain that results and
conclusions derived from models are robust under
a number of potentially confounding situations.
The modeling strategy presented in this chapter
attempts to ensure that available biological data is
represented and accounted for. However, a num-
ber of specific components of the dentate circuit
and connectivity often cannot be modeled due to
lack of precise data. Additionally, depending on
the particular question the model is designed to
answer, several factors may be purposely omitted
in order to best focus on the relevant issues. For
the model presented in this section, for example,
changes to intrinsic cellular and synaptic proper-
ties during sclerosis were omitted since the exper-
iments were designed to study structural
alterations in isolation. For these reasons, an ex-
tensive series of control simulations must be
654

Fig. 4. Effects of the sclerosis-related topological changes on granule cell activity in functional model networks. A–F. Raster plots of
the first 300 ms of action potential discharges of granule cells in the 50,000+ neuron functional model network (granule cells number
1–50,000, plotted on the y-axis; each dot corresponds to an action potential fired by a granule cell) at increasing degrees of sclerosis.
Network activity was initiated by a single stimulation of the perforant path input to granule cells number 22,500–27,499 and to 10
mossy cells and 50 basket cells (distributed in the same area as the stimulated granule cells) at t ¼ 5 ms (as in Santhakumar et al., 2005).
Note that the most pronounced hyperactivity was observed at sub-maximal (80%) sclerosis. Reproduced with permission from
Dyhrfjeld-Johnsen et al. (2007).

performed to test the effects of a wide variety of
conditions that are not well constrained by the
available experimental data. Controls will vary
from model to model, but many are needed in any
situation. Those performed for the structural
model presented here were the following:
1) Variations in cell numbers.
2) Variations in connectivity estimates.
3) Inhomogeneous distribution of neuron den-
sities along the septo-temporal axis.
4) Inhomogeneity in connectivity along the
transverse axis.
5) Altered axonal distributions at the septal and
temporal poles (the anatomical boundaries of
the dentate gyrus).
6) Offset degrees of sprouting and hilar neuron
loss.
7) Implementation of a bilateral model of the
dentate gyrus including commissural projec-
tions.
All of the controls employed for the structural
model were also used as controls for the functional
model. In addition, the functional model had its
own set of controls, including:
1) Double inhibitory synaptic strengths.
2) Axonal conduction delays.
3) Spontaneous instead of stimulation-evoked
activity.
All of the results from these control simulations
supported the basic findings of the original mod-
els, thus greatly strengthening the primary conclu-
sions.

Tackling future puzzles and problems
Modeling has come a long way in the last several
decades. Even 10 years ago, large-scale realistic
modeling was nearly impossible due to computa-
tional restrictions and the requirement of vast su-
percomputers to perform necessary calculations.
Today, however, it is possible to perform quite in-
depth modeling studies on a home desktop. Models
such as the 50,000 cell network in Dyhrfjeld-John-
sen et al. (2007) can be realistically run on a dual
processor system (albeit with 16–32 GB of RAM).
Smaller networks such as the 500 cell model in
Santhakumar et al. (2005) can be run in a few min-
utes on a typical laptop. These are wonderful ad-
vancements, as the development of large-scale
models that can simulate brain circuits with greater
realism is critical to our understanding of the ex-
perimentally determined cellular and molecular un-
derpinnings of diseases like epilepsy. Additionally,
large-scale realistic networks are necessary for our
understanding of structure–function relationships,
especially in a highly plastic system like the dentate
gyrus.
The exact structural changes that occur in many
pathological states such as epilepsy are as yet un-
known. For example, recent evidence indicates that
local connection probability in various brain areas
may be modified by intra-class correlations (Yos-
himura and Callaway, 2005; Yoshimura et al.,
2005) and over-representation of small network
motifs (Milo et al., 2002; Reigl et al., 2004; Sporns
and Kotter, 2004; Song et al., 2005). While no di-
rect evidence is present to implicate such factors in
the dentate gyrus, it is likely that unknown struc-
tural alterations have a significant impact on net-
work activity. Computational modeling studies
provide an excellent tool for determining what the
probable result of a wide variety of structural al-
terations could be. Subsequent animal model ex-
periments can then attempt to visualize the
structural alterations predicted by the computation
models. In an example of this type of approach,
Morgan and Soltesz (2006) have studied probable
connectivity patterns within the recurrent GC net-
work that results from mossy fiber sprouting. Their
results indicate that the presence of a small number
(at most 5%) of GCs with increased connectivity
655
(on average, 5–6 times the average number of con-
nections) could serve as hubs and strongly promote
hyperexcitability within the sclerotic dentate gyrus
network. Interestingly, this finding is supported by
the presence, in epileptic rats, of a small percentage
of GCs that have basal dendrites (Spigelman et al.,
1998) and which receive a vast number of addi-
tional mossy fiber contacts (Buckmaster and Thind,
2005). Computational modeling studies provide the
best way of isolating structural changes from in-
trinsic cellular and synaptic changes that occur si-
multaneously in virtually all pathological states.
Large-scale simulations will also make it possi-
ble to hold structure constant and test the circuit
effects of specific therapeutic interventions,
thereby allowing fast, cheap, efficient, and specific
analysis of treatments for a variety of pathological
processes. The fast pace at which scientists are
gleaning knowledge about the dentate from animal
models will allow for inclusion of a complete com-
plement of interneuronal subtypes, various recep-
tor and modulatory systems, realistic channel
distributions, dendritic integration processes, and
synaptic plasticity into computational models.
These additions will greatly broaden the scope of
questions that a model can address and the
model’s predictive power. In the coming years, re-
alistic modeling is only going to become more im-
portant in bridging the gaps between animals and
humans, molecules and behavior.
Acknowledgements
The authors would like to thank Dr. Jonas
Dyhrfjeld-Johnsen for his contributions to the re-
search discussed in this chapter and his helpful
comments on the chapter. This work was funded by
NIH grant NS35915 to IS and UCI MSTP to RM.
References
Acsady, L., Kamondi, A., Sik, A., Freund, T. and Buzsaki, G.
(1998) GABAergic cells are the major postsynaptic targets of
mossy fibers in the rat hippocampus. J. Neurosci., 18:
3386–3403.
Acsady, L., Katona, I., Martinez-Guijarro, F.J., Buzsaki, G.
and Freund, T.F. (2000) Unusual target selectivity of
656
perisomatic inhibitory cells in the hilar region of the rat hip-
pocampus. J. Neurosci., 20: 6907–6919.
Amaral, D.G. (1978) A Golgi study of cell types in the hilar
region of the hippocampus in the rat. J. Comp. Neurol., 182:
851–914.
Aradi, I. and Holmes, W.R. (1999a) Role of multiple calcium
and calcium-dependent conductances in regulation of hippo-
campal dentate granule cell excitability. J. Comput. Neuro-
sci., 6: 215–235.
Aradi, I. and Holmes, W.R. (1999b) Active dendrites regulate
spatio-temporal synaptic integration in hippocampal dentate
granule cells. Neurocomputing, 26–27: 45–51.
Babb, T.L., Pretorius, J.K., Kupfer, W.R. and Brown, W.J.
(1988) Distribution of glutamate-decarboxylase-immunore-
active neurons and synapses in the rat and monkey hippo-
campus: light and electron-microscopy. J. Comp. Neurol.,
278: 121–138.
Bartos, M., Vida, I., Frotscher, M., Geiger, J.R.P. and Jonas, P.
(2001) Rapid signaling at inhibitory synapses in a dentate
gyrus interneuron network. J. Neurosci., 21: 2687–2698.
Blasco-Ibanez, J.M., Martinez-Guijarro, F.J. and Freund, T.F.
(2000) Recurrent mossy fibers preferentially innervate parv-
albumin-immunoreactive interneurons in the granule cell
layer of the rat dentate gyrus. Neuroreport, 11: 3219–3225.
Blumcke, I., Suter, B., Behle, K., Kuhn, R., Schramm, J., Elger,
C.E. and Wiestler, O.D. (2000) Loss of hilar mossy cells in
Ammon’s horn sclerosis. Epilepsia 41 Suppl 6: S174–180.
Boss, B.D., Peterson, G.M. and Cowan, W.M. (1985) On the
number of neurons in the dentate gyrus of the rat. Brain Res.,
338: 144–150.
Buckmaster, P.S. and Dudek, F.E. (1997) Network properties
of the dentate gyrus in epileptic rats with hilar neuron loss
and granule cell axon reorganization. J. Neurophysiol., 77:
2685–2696.
Buckmaster, P.S. and Dudek, F.E. (1999) In vivo intracellular
analysis of granule cell axon reorganization in epileptic rats.
J. Neurophysiol., 81: 712–721.
Buckmaster, P.S. and Jongen-Relo, A.L. (1999) Highly specific
neuron loss preserves lateral inhibitory circuits in the dentate
gyrus of kainate-induced epileptic rats. J. Neurosci., 19:
9519–9529.
Buckmaster, P.S., Strowbridge, B.W., Kunkel, D.D., Schmiege,
D.L. and Schwartzkroin, P.A. (1992) Mossy cell axonal pro-
jections to the dentate gyrus molecular layer in the rat hip-
pocampal slice. Hippocampus, 2: 349–362.
Buckmaster, P.S., Strowbridge, B.W. and Schwartzkroin, P.A.
(1993) A comparison of rat hippocampal mossy cells and
CA3c pyramidal cells. J. Neurophysiol., 70: 1281–1299.
Buckmaster, P.S. and Schwartzkroin, P.A. (1994) Hippocampal
mossy cell function: a speculative view. Hippocampus, 4:
393–402.
Buckmaster, P.S. and Thind, K. (2005) American epilepsy so-
ciety meeting abstract. Epilepsia, 46: 91–131.
Buckmaster, P.S., Wenzel, H.J., Kunkel, D.D. and Schwa-
rtzkroin, P.A. (1996) Axon arbors and synaptic connections of
hippocampal mossy cells in the rat in vivo. J. Comp. Neurol.,
366: 270–292.
Buckmaster, P.S., Yamawaki, R. and Zhang, G.F. (2002a)
Axon arbors and synaptic connections of a vulnerable pop-
ulation of interneurons in the dentate gyrus in vivo. J. Comp.
Neurol., 445: 360–373.
Buckmaster, P.S., Zhang, G.F. and Yamawaki, R. (2002b)
Axon sprouting in a model of temporal lobe epilepsy creates
a predominantly excitatory feedback circuit. J. Neurosci., 22:
6650–6658.
Carnevale, N.T. and Hines, M.L. (2006) The Neuron Book.
Cambridge University Press, Cambridge, UK.
Cavazos, J.E., Das, I. and Sutula, T.P. (1994) Neuronal loss
induced in limbic pathways by kindling: evidence for induc-
tion of hippocampal sclerosis by repeated brief seizures. J.
Neurosci., 14: 3106–3121.
Cavazos, J.E. and Sutula, T.P. (1990) Progressive neuronal loss
induced by kindling: a possible mechanism for mossy fiber
synaptic reorganization and hippocampal sclerosis. Brain
Res., 527: 1–6.
Ceranik, K., Bender, R., Geiger, J.R.P., Monyer, H., Jonas, P.,
Frotscher, M. and Lubke, J. (1997) A novel type of GAB-
Aergic interneuron connecting the input and the output re-
gions of the hippocampus. J. Neurosci., 17: 5380–5394.
Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1990) Quan-
titative, three-dimensional analysis of granule cell dendrites
in the rat dentate gyrus. J. Comp. Neurol., 302: 206–219.
Cossart, R., Dinocourt, C., Hirsch, J.C., Merchan-Perez, A.,
De Felipe, J., Ben-Ari, Y., Esclapez, M. and Bernard, C.
(2001) Dendritic but not somatic GABAergic inhibition is
decreased in experimental epilepsy. Nat. Neurosci., 4: 52–62.
Desmond, N.L. and Levy, W.B. (1985) Granule cell dendritic
spine density in the rat hippocampus varies with spine shape
and location. Neurosci. Lett., 54: 219–224.
Destexhe, A., Rudolph, M., Fellous, J.M. and Sejnowski, T.J.
(2001) Fluctuating synaptic conductances recreate in vivo-
like activity in neocortical neurons. Neuroscience, 107:
13–24.
Dyhrfield-Johnsen, J., Santhakumar, V., Morgan, R.J., Huerta,
R., Tsimring, L. and Soltesz, I. (2007) Topological determi-
nants of epileptogenesis in large-scale structural and func-
tional models of the dentate gyrus derived from experimental
data. J. Neurophysiol., 97(2): 1566–1587.
Freund, T.F. and Buzsaki, G. (1996) Interneurons of the hip-
pocampus. Hippocampus, 6: 347–470.
Frotscher, M., Seress, L., Schwerdtfeger, W.K. and Buhl, E.
(1991) The mossy cells of the fascia dentata: a comparative
study of their fine structure and synaptic connections in ro-
dents and primates. J. Comp. Neurol., 312: 145–163.
Gaarskjaer, F.B. (1978) Organization of mossy fiber system of
rat studied in extended hippocampi.1. Terminal area related
to number of granule and pyramidal cells. J. Comp. Neurol.,
178: 49–71.
Gabriel, S., Njunting, M., Pomper, J.K., Merschhemke, M.,
Sanabria, E.R.G., Eilers, A., Kivi, A., Zeller, M., Meencke,
H.-J., Cavalheiro, E.A., Heinemann, U. and Lehmann, T.-N.
(2004) Stimulus and potassium-induced epileptiform activity
in the human dentate gyrus from patients with and without
hippocampal sclerosis. J. Neurosci., 24: 10416–10430.

Gorter, J.A., van Vliet, E.A., Aronica, E. and Lopes da Silva,
F.H. (2001) Progression of spontaneous seizures after
status epilepticus is associated with mossy fibre sprouting
and extensive bilateral loss of hilar parvalbumin and som-
atostatin-immunoreactive neurons. Eur. J. Neurosci., 13:
657–669.
Geiger, J.R.P., Lubke, J., Roth, A., Frotscher, M. and Jonas, P.
(1997) Submillisecond AMPA receptor-mediated signaling at
a principal neuron-interneuron synapse. Neuron, 18:
1009–1023.
Gulyas, A.I., Hajos, N. and Freund, T.F. (1996) Interneurons
containing calretinin are specialized to control other inter-
neurons in the rat hippocampus. J. Neurosci., 16: 3397–3411.
Gulyas, A.I., Miettinen, R., Jacobowitz, D.M. and Freund,
T.F. (1992) Calretinin is present in nonpyramidal cells of the
rat hippocampus .1. A new type of neuron specifically asso-
ciated with the mossy fiber system. Neuroscience, 48: 1–27.
Halasy, K. and Somogyi, P. (1993) Subdivisions in the multiple
gabaergic innervation of granule cells in the dentate gyrus of
the rat hippocampus. Eur. J. Neurosci., 5: 411–429.
Han, Z.S., Buhl, E.H., Lorinczi, Z. and Somogyi, P. (1993) A
high-degree of spatial selectivity in the axonal and dendritic
domains of physiologically identified local-circuit neurons in
the dentate gyrus of the rat hippocampus. Eur. J. Neurosci.,
5: 395–410.
Heinemann, U., Beck, H., Dreier, J.P., Ficker, E., Stabel, J. and
Zhang, C.L. (1992) The dentate gyrus as a regulated gate for
the propagation of epileptiform activity. Epilepsy Res. Sup-
pl., 7: 273–280.
Howard, A., Tamas, G. and Soltesz, I. (2005) Lighting the
chandelier: new vistas for axo-axonic cells. Trends Neurosci.,
28: 310–316.
Howard, A.L., Neu, A., Morgan, R.J., Echegoyen, J.C. and
Soltesz, I. (2007) Opposing Modifications in Intrinsic Cur-
rents and Synaptic Inputs in Post-Traumatic Mossy Cells:
Evidence for Single-Cell Homeostasis in a Hyperexcitable
Network. J. Neurophysiol., 97(3): 2394–2409.
Ishizuka, N., Cowan, W.M. and Amaral, D.G. (1995) A quan-
titative analysis of the dendritic organization of pyramidal
cells in the rat hippocampus. J. Comp. Neurol., 362: 17–45.
Katona, I., Acsady, L. and Freund, T.F. (1999) Postsynaptic
targets of somatostatin-immunoreactive interneurons in the
rat hippocampus. Neuroscience, 88: 37–55.
Kneisler, T.B. and Dingledine, R. (1995) Synaptic input from
CA3 pyramidal cells to dentate basket cells in rat hippocam-
pus. J. Physiol., 487(Pt 1): 125–146.
Leite, J.P., Babb, T.L., Pretorius, J.K., Kuhlman, P.A., Yeo-
man, K.M. and Mathern, G.W. (1996) Neuron loss, mossy
fiber sprouting, and interictal spikes after intrahippocampal
kainate in developing rats. Epilepsy Res., 26: 219–231.
Li, X.G., Somogyi, P., Tepper, J.M. and Buzsaki, G. (1992)
Axonal and dendritic arborization of an intracellularly labe-
led chandelier cell in the ca1 region of rat hippocampus. Exp.
Brain Res., 90: 519–525.
Lothman, E.W., Stringer, J.L. and Bertram, E.H. (1992) The
dentate gyrus as a control point for seizures in the hippo-
campus and beyond. Epilepsy Res. Suppl., 7: 301–313.
657
Lubke, J., Frotscher, M. and Spruston, N. (1998) Specialized
electrophysiological properties of anatomically identified
neurons in the hilar region of the rat fascia dentata. J. Ne-
urophysiol., 79: 1518–1534.
Lytton, W.W., Hellman, K.M. and Sutula, T.P. (1998) Com-
puter models of hippocampal circuit changes of the kindling
model of epilepsy. Artif. Intell. Med., 13: 81–97.
Mathern, G.W., Bertram III, E.H., Babb, T.L., Pretorius, J.K.,
Kuhlman, P.A., Spradlin, S. and Mendoza, D. (1997) In
contrast to kindled seizures, the frequency of spontaneous
epilepsy in the limbic status model correlates with greater
aberrant fascia dentata excitatory and inhibitory axon
sprouting, and increased staining for N-methyl-D-aspartate,
AMPA and GABA(A) receptors. Neuroscience, 77:
1003–1019.
Milo, R., Shen-Orr, S., Itzkovitz, S., Kashtan, N., Chklovskii,
D. and Alon, U. (2002) Network motifs: simple building
blocks of complex networks. Science, 298: 824–827.
Morgan, R.J. and Soltesz, I. (2006) American epilepsy society
meeting abstract. Epilepsia, 47: 18–28.
Netoff, T.I., Clewley, R., Arno, S., Keck, T. and White, J.A.
(2004) Epilepsy in small-world networks. J. Neurosci., 24:
8075–8083.
Nomura, T., Fukuda, T., Aika, Y., Heizmann, C.W., Emson,
P.C., Kobayashi, T. and Kosaka, T. (1997a) Laminar distri-
bution of non-principal neurons in the rat hippocampus, with
special reference to their compositional difference among
layers. Brain Res., 764: 197–204.
Nomura, T., Fukuda, T., Aika, Y., Heizmann, C.W., Emson,
P.C., Kobayashi, T. and Kosaka, T. (1997b) Distribution of
nonprincipal neurons in the rat hippocampus, with special
reference to their dorsoventral difference. Brain Res., 751:
64–80.
Patton, P.E. and Mcnaughton, B. (1995) Connection matrix of
the hippocampal-formation .1. The dentate gyrus. Hippo-
campus, 5: 245–286.
Rafiq, A., Zhang, Y.F., DeLorenzo, R.J. and Coulter, D.A.
(1995) Long-duration self-sustained epileptiform activity in
the hippocampal-parahippocampal slice: a model of status
epilepticus. J. Neurophysiol., 74: 2028–2042.
Rall, W., Burke, R.E., Holmes, W.R., Jack, J.J., Redman, S.J.
and Segev, I. (1992) Matching dendritic neuron models to
experimental data. Physiol. Rev., 72: S159–S186.
Ratzliff, A.H., Howard, A.L., Santhakumar, V., Osapay, I. and
Soltesz, I. (2004) Rapid deletion of mossy cells does not result
in a hyperexcitable dentate gyrus: implications for epilepto-
genesis. J. Neurosci., 24: 2259–2269.
Reigl, M., Alon, U. and Chklovskii, D.B. (2004) Search for com-
putational modules in the C. elegans brain. BMC Biol., 2: 25.
Ribak, C.E., Seress, L. and Amaral, D.G. (1985) The develop-
ment, ultrastructure and synaptic connections of the mossy
cells of the dentate gyrus. J. Neurocytol., 14: 835–857.
Roxin, A., Riecke, H. and Solla, S.A. (2004) Self-sustained ac-
tivity in a small-world network of excitable neurons. Phys.
Rev. Lett., 92: 198101.
Santhakumar, V., Aradi, I. and Soltesz, I. (2005) Role of mossy
fiber sprouting and mossy cell loss in hyperexcitability: a
658
network model of the dentate gyrus incorporating cell types
and axonal topography. J. Neurophysiol., 93: 437–453.
Santhakumar, V., Bender, R., Frotscher, M., Ross, S.T., Ho-
llrigel, G.S., Toth, Z. and Soltesz, I. (2000) Granule cell
hyperexcitability in the early post-traumatic rat dentate
gyrus: the ‘irritable mossy cell’ hypothesis. J. Physiol.,
524(Pt 1): 117–134.
Santhakumar, V., Ratzliff, A.D., Jeng, J., Toth, Z. and
Soltesz, I. (2001) Long-term hyperexcitability in the hippo-
campus after experimental head trauma. Ann. Neurol., 50:
708–717.
Santhakumar, V. and Soltesz, I. (2004) Plasticity of intern-
euronal species diversity and parameter variance in neuro-
logical diseases. Trends Neurosci., 27: 504–510.
Scharfman, H.E. (1991) Dentate hilar cells with dendrites in the
molecular layer have lower thresholds for synaptic activation
by perforant path than granule cells. J. Neurosci., 11:
1660–1673.
Sik, A., Penttonen, M. and Buzsaki, G. (1997) Interneurons in
the hippocampal dentate gyrus: an in vivo intracellular study.
Eur. J. Neurosci., 9: 573–588.
Song, S., Sjostrom, P.J., Reigl, M., Nelson, S. and Chklovskii,
D.B. (2005) Highly nonrandom features of synaptic connec-
tivity in local cortical circuits. PLoS Biol., 3: e68.
Soriano, E., Nitsch, R. and Frotscher, M. (1990) Axo-axonic
chandelier cells in the rat fascia dentata: Golgi-electron mi-
croscopy and immunocytochemical studies. J. Comp. Ne-
urol., 293: 1–25.
Spigelman, I., Yan, X.X., Obenaus, A., Lee, E.Y., Wasterlain,
C.G. and Ribak, C.E. (1998) Dentate granule cells form
novel basal dendrites in a rat model of temporal lobe epi-
lepsy. Neuroscience, 86: 109–120.
Sporns, O. and Kotter, R. (2004) Motifs in brain networks.
PLoS Biol., 2: e369.
Vida, I., Bartos, M. and Jonas, P. (2006) Shunting inhibition
improves robustness of gamma oscillations in hippocampal
interneuron networks by homogenizing firing rates. Neuron,
49: 107–117.
van Vliet, E.A., Aronica, E., Tolner, E.A., Lopes da Silva, F.H.
and Gorter, J.A. (2004) Progression of temporal lobe epilepsy
in the rat is associated with immunocytochemical changes in
inhibitory interneurons in specific regions of the hippocampal
formation. Exp. Neurol., 187: 367–379.
Watts, D.J. and Strogatz, S.H. (1998) Collective dynamics of
‘small-world’ networks. Nature, 393: 440–442.
Wenzel, H.J., Buckmaster, P.S., Anderson, N.L., Wenzel, M.E.
and Schwartzkroin, P.A. (1997) Ultrastructural localization
of neurotransmitter immunoreactivity in mossy cell axons
and their synaptic targets in the rat dentate gyrus. Hippo-
campus, 7: 559–570.
West, M.J. (1990) Stereological studies of the hippocampus: a
comparison of the hippocampal subdivisions of diverse spe-
cies including hedgehogs, laboratory rodents, wild mice and
men. Prog. Brain Res., 83: 13–36.
Woodson, W., Nitecka, L. and Benari, Y. (1989) Organization
of the gabaergic system in the rat hippocampal-formation: a
quantitative immunocytochemical study. J. Comp. Neurol.,
280: 254–271.
Yoshimura, Y. and Callaway, E.M. (2005) Fine-scale specificity
of cortical networks depends on inhibitory cell type and
connectivity. Nat. Neurosci., 8: 1552–1559.
Yoshimura, Y., Dantzker, J.L. and Callaway, E.M. (2005) Ex-
citatory cortical neurons form fine-scale functional networks.
Nature, 433: 868–873.
Zappone, C.A. and Sloviter, R.S. (2004) Translamellar disin-
hibition in the rat hippocampal dentate gyrus after seizure-
induced degeneration of vulnerable hilar neurons. J. Neuro-
sci., 24: 853–864.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 36

Hippocampal granule cells in normal aging: insights

from electrophysiological and functional imaging
experiments

Monica K. Chawla1,2 and Carol A. Barnes1,2,3,

1
Arizona Research Laboratories Division of Neural Systems, Memory and Aging, University of Arizona, Life Sciences
North, Room 384, P.O. Box 245115, Tucson, AZ 85724, USA
2
Evelyn F. McKnight Brain Institute, University of Arizona, Tucson, AZ 85724, USA
3
Departments of Psychology and Neurology, University of Arizona, Tucson, AZ 85724, USA

Abstract: Normal aging, in the absence of neurodegenerative disease, can provide important insights into
the mechanisms by which the brain can maintain cognitive abilities across the lifespan. Hippocampal-
dependent memory processes can become vulnerable as age advances. The focus of this chapter is the
contribution of hippocampal granule cells to cognitive impairments that are observed during aging. A
number of alterations in structure, function, and gene expression have been observed in aged granule cells,
any of which may lead to adaptive, compensatory or detrimental consequences to hippocampal function.
As the average life span of humans continues to increase, those who reach 100 years or beyond is more
common. Individuals that have aged successfully, and exhibit high levels of cognitive ability can provide
useful clues into the enormous potential possessed by the mammalian brain.

Keywords: fascia dentata; dentate gyrus; LTP; fluorescence in situ hybridization; FMRI

Introduction appear to be involved in cell maintenance include

Aging is accompanied by numerous molecular, cel-
lular, and structural changes that have functional
consequences. Successful aging depends upon mul-
tiple factors including genetic background, and
positive or negative environmental or social condi-
tions. The observation that the brain is able to
adapt to a variety of insults, suggests a large degree
of heterogeneity in the expression of processes that
can be employed by neurons or glia to optimize
function. Among the physiological factors that
Corresponding author. Tel.: +1 520-626-2616;
Fax: +1 520-626-2618; E-mail: carol@nsma.arizona.edu
cytokine molecules, expression of antiapoptotic
proteins, and generation of various survival-related
proteins like chaperones and antioxidants. As
human populations continue to shift towards older
average ages, it is becoming increasingly important
to understand the factors that can ‘‘tip the balance’’
so that more individuals age successfully and fewer
develop pathological conditions associated with ag-
ing, including Alzheimer’s and Parkinson’s diseases.
Normal aging in the absence of disease can provide
insight into the stable, plastic, and adaptive strat-
egies adopted by the brain in these high functioning
aged individuals. Indeed, an important goal for re-
search in the neurobiology of aging is to identify

DOI: 10.1016/S0079-6123(07)63036-2 661

662
mechanisms by which some aged individuals show
minimal decrement in cognitive, motor and sensory
skills across their lifespan.
There is now a general consensus that some as-
pects of learning and memory functions decline
with advancing age (for reviews see, Landfield,
1988; Rosenzweig and Barnes, 2003; Driscoll and
Sutherland, 2005; Burke and Barnes, 2006). Mem-
ory processes that rely on hippocampal integrity
are clearly vulnerable to advancing age in humans
(e.g., Cabeza et al., 2005), monkeys (e.g., Rapp
and Amaral, 1991; Keuker et al., 2000) and in ro-
dent models of aging (e.g., Barnes, 1979; Geinis-
man et al., 1986; Barnes, 1988, 1994; Kadar et al.,
1994; Gallagher and Rapp, 1997; Bach et al., 1999;
Rapp et al., 1999; Driscoll et al., 2006).
This chapter focuses on the contribution of hip-
pocampal granule cells to cognitive impairments of
aged animals. First, the basic anatomical, bio-
chemical and biophysical properties of young
granule cells will be compared with older granule
cells. Whether these modifications in aging are
adaptive, compensatory or detrimental to the
hippocampus will be discussed. Next, functional
connectivity, inferred by extracellular electrophys-
iological recording methods, is examined in an at-
tempt to relate functional synaptic changes to
cognitive decline in aged rats. Additionally, since
aging is a multidimensional process that can affect
many complex cellular mechanisms, transcrip-
tional regulation of immediate early genomic re-
sponses are also examined here. The triggers and
consequences of altered gene responses are poorly
understood at present. Recent advances in imaging
technologies have enabled both the identification
of regions of brain that are activated by specific
behavioral experiences in monkeys and humans
(Small et al., 2002, 2004) as well as specific cells
that are actively engaged in a particular behavioral
experience (Guzowski et al., 1999). The examples
that follow will illustrate the complex pattern of
changes that occur during aging with a focus
on granule cells. These patterns are not only
subtle but highly sub-region specific and point
towards compensatory mechanisms engaged by
the brain to adapt to the complexities of the mul-
tidimensional, enigmatic process called normal
aging.
Granule cell numbers in aging
Early ideas suggested that cognitive decline ob-
served during aging is due to global loss of neurons
in rats, monkeys, and humans is now known to be
incorrect. Brody (1955) first reported that the re-
duction in brain weight observed in older individ-
uals was due to massive neuronal loss across the
cortex. While these observations were later repli-
cated in rats and humans in other laboratories
(Ball, 1977; Coleman and Flood, 1987; Coleman et
al., 1987), these early studies used cell counting
methods without stereological techniques, and were
found to be incorrect in later experiments that used
unbiased cell counting methods (Sterio, 1984; West,
1993a; West et al., 1994; Morrison and Hof, 1997).
It is now generally accepted that age-related cog-
nitive decline is not due to massive neuronal loss.
Using very detailed electron microscopic meth-
ods and serial reconstructions Geinisman and
Bondareff (1976) first reported that granule cell
numbers are not decreased in aged rat, but rather,
they lose about a third of their medial entorhinal
synaptic input. A number of more recent investi-
gations have replicated these results using modern
counting methods and have found no age-related
neuronal loss of CA pyramidal cells or dentate
gyrus granule cells (Rapp and Gallagher, 1996; Ra-
smussen et al., 1996) in rats. Similarly, in aged
monkeys (Peters et al., 1996), and in humans (West,
1993b), there appears to be no hippocampal gran-
ule or CA pyramidal cell loss in aging. Moreover,
entorhinal cortical layers II and III cell numbers do
not change during aging in rats or monkeys (Merrill
et al., 2000, 2001; Rapp et al., 2002).
Granule cell neurogenesis and aging
Granule cells are unusual, in that they are among
the few neuronal cells that undergo neurogenesis
in adulthood (Altman and Das, 1965; Gould and
Cameron, 1996; Eriksson, 2003). Neurogenesis has
also been observed in olfactory bulb neurons
across the lifespan of mammals. The newborn
granule cells originate from the subgranular zone
of the dentate gyrus and migrate into the granular
layer (Kuhn et al., 1996), develop dendritic

processes (Ribak et al., 2004) and mature into
functional neurons displaying some of the same
synaptic responsiveness and other electrophysio-
logical properties as do existing granule cells (van
Praag et al., 2002; Song et al., 2005). Continous
neurogenesis in the dentate gyrus throughout the
life span of mammals may be of great functional
significance. For example, Shors et al. (2001) have
shown that numbers of newborn granule cells cor-
relate with effective hippocampal-dependent mem-
ory function. Whether, newly born granule cells
are integrated into functional networks mediating
spatial information processing was recently inves-
tigated by Ramirez-Amaya et al. (2006). Using
triple immunohistochemistry and confocal micros-
copy, they were able to show that granule cells
born in the mature dentate gyrus (5-month old),
newborn granule cells demonstrated by the mitotic
marker-BrdU express the plasticity-related imme-
diate-early gene Arc protein in response to spatial
exploration. Additionally, the proportion of Arc
protein-containing cells was significantly higher in
the newly born granule cell population compared
to the preexisting population of granule cells
(2.8% vs. 1.6%, respectively).
Recently it was shown that exercise-induced in-
creases in neurogenesis are associated with better
performance on the Morris water maze (van Praag
et al., 2005), and a correlation between higher lev-
els of neurogenesis and preserved spatial memory
in old rats has been observed by Drapeau et al.
(2003). Other studies, however, have not found
correlations between neurogenesis and behavior
(Bizon and Gallagher, 2003), or in some instances
a negative correlation (Bizon et al., 2004) in the
aged rodent. An interesting hypothesis recently
proposed for the role of adult neurogenesis, sug-
gests that the newborn granule cells participate in
forming time-tagged associations for laying down
new memories (Aimone et al., 2006).
Aged rats do show granule cell neurogenesis and
old granule cell survival rates can be modified by
enriched environments (Kempermann and Gage,
1999; Kempermann et al., 2002). Generation of
newly born granule cells from progenitor cells,
however, show a marked decline beginning in mid-
dle age (Kuhn et al., 1996; Rao et al., 2005), but the
numbers of newly born granule cells that become
663
neurons remain stable from middle age into older
ages (Rao et al., 2005). Moreover, the increase in
neurogenesis that is induced by specific types of
learning (contextual fear conditioning) is reduced in
aged rats, as well as the survival of new granule cells
(Wati et al., 2006). The reduction in integration of
newborn granule cells into hippocampal circuitry
with age may contribute to the cognitive impair-
ment in aged animals.
Age-related changes in granule cell synaptic
contacts
The constancy in granule cell numbers (Geinisman
and Bondareff, 1976) and the volume of the gran-
ule cell layer (Coleman and Flood, 1987) with in-
creasing age in rodents stands in contrast to
reports of a decrease in synapse numbers. As was
the case with cell counts in early literature, synapse
counts conducted across age groups produced var-
iable results, possibly due to methods that were
less accurate than current techniques (Geinisman
and Bondareff, 1976; Geinisman et al., 1977;
Geinisman, 1979; Hoff et al., 1982; McWilliams
and Lynch, 1984; Bertoni-Freddari et al., 1986;
Badiali de Giorgi et al., 1987). Using unbiased
stereological techniques, Geinisman et al. (1992)
demonstrated a loss of axospinous synapse num-
bers in the middle third of the molecular layer of
the dentate gyrus in memory-impaired old rats.
When the entire hippocampus was homogenized
in preparation for protein analysis to examine pos-
sible age-related differences in levels of synapse-
related proteins (synaptogamin-, synaptophysin-, or
synaptosomal-associated protein 25), no age effects
were observed (Nicolle et al., 1999). However, with
the use of methods that can detect circuit-specific
differences in synaptic protein levels, age-related
differences are revealed. Smith et al. (2000) exam-
ined the layer II entorhinal cortical projection that
inervates both CA3 stratum lacunosum moleculare
and dentate gyrus molecular layer. There was a
significant loss of synaptophysin in both areas in
memory-impaired old rats. Studies measuring
synaptic markers, however, cannot elucidate the
possible functional changes that result from alter-
ations in these proteins, such as synaptic strength.
664

The answer to these kinds of functional questions, Electrophysiological comparisons of granule cells
require electrophysiological investigation (see of young and old rats have been made using both in
Synaptic Response section below). vivo and in vitro preparations. These studies reveal
that most of the basic granule cell electrical prop-
erties remain constant over the lifespan of rodents,
Granule cell dendrites in aging
mice, and rabbits (Barnes et al., 1994). For exam-
ple, studies investigating resting membrane poten-
As discussed above, since numbers of granule cells
tial, membrane time constant, the relation between
do not change with age, it follows that the num-
applied current and input resistance, the amplitude
bers of primary dendritic shafts should also remain
and duration of the Na+-dependent action poten-
unchanged. One study, however, did report a re-
tial, threshold of action potential following depo-
duction in numbers of granule cell shafts in old
larizing current injection, and rise time and half
rats (Geinisman et al., 1978). Because nonstereo-
width of the intracellular EPSP (Barnes and McN-
logical methods were used in that study, the results
aughton, 1979, 1980a; Barnes et al., 1987; Baskys et
may most easily be explained by sampling bias. An
al., 1987; Niesen et al., 1988; Foster et al., 1991)
investigation in young and old rhesus monkeys,
have not revealed age-associated changes (for re-
using stereological techniques, revealed no signifi-
view, see Barnes, 1994; Rosenzweig and Barnes,
cant difference in total dendritic length, number of
2003).
dendritic branch points and total segment number
Interestingly, the action potential threshold to
between young and old granule cells (Luebke and
orthodromic stimulation is decreased in aged gran-
Rosene, 2003). They did, however, find that the
ule cells (Barnes and McNaughton, 1980a). The
molecular layer was decreased in width (between
observation that threshold to action potential dis-
the granule cell layer of the upper blade and the
charge to direct current does not change during
hippocampal fissure) in old (>24 years) vs. young
aging, while the action potential threshold to or-
(o11 years) monkeys.
thodromic stimulation does change, might be ex-
Can spatial memory deficits with advancing age
plained in several ways. Among these was the
be explained in part by a decrease in spine numbers?
hypothesis that there is increased electrotonic cou-
von Bohlen und Halbach et al., 2006 recently re-
pling between granule cells with age (Barnes et al.,
ported no significant reduction in spine density in
1987). If there were more gap junctional connec-
the apical dendrites of the upper blade DG and area
tions between granule cells, current would flow
CA1 of old mice, although the mean spine length
from the granule cell with the lowest threshold into
did show a reduction in both areas in young (6–7
adjacent cells that had not reached threshold.
months) compared to old (21–22 months) mice. In-
Overall, this would result in a lowering of the
terestingly, Samsonovich and Ascoli (2006) have
threshold for the entire population resulting from
recently concluded, using statistical analysis of dig-
stimulation of the perforant path. Previously, it
itally reconstructed neurons (from several different
had been shown by MacVicar and Dudek (1982)
brain regions and different experimental manipula-
that 60% of granule cells were electrotonically
tions) that there is morphological homeostasis in
coupled, using Lucifer yellow injections and freeze
apical and basal dendritic processes of neurons.
fracture methods. In the Barnes et al. (1987), study
Therefore, dendritic changes in one area may be
a low molecular weight fluorescence dye, 5,6-car-
compensated by changes in other areas in order to
boxyfluorescein, was used to explore intercellular
preserve the dendritic tree, spine numbers, and
connections via gap junctions in in vitro slice
other characteristics of granule cells.
preparations. Increased dye coupling was observed
in all three hippocampal subregions with age. In
Biophysical properties of aged granule cells fact, granule cells showed the most extensive cou-
pling (i.e., 64% of old granule cells were coupled
Does the absence of gross granule cell loss in to at least one other cell compared to 49% of
aging mean a general preservation of function? young granule cells). Accompanying this increased
 665

Fig. 1. (A) Two representative granule cells filled with 5,6 carboxyfluorescein from the dentate gyrus of a 24-month-old rat. (B)
Histograms showing the numbers of carboxyfluorescein injections that resulted in single, double, triple, or greater numbers of granule
cells filled with dye. Aged rats showed significantly increased electrotonic coupling compared to young animals. This may account for
the increased excitability of old granule cells. Adapted with modifications from Barnes et al. (1987). (See Color Plate 36.1 in color plate
section.)

electrotonic coupling, old granule cells also exhib-
ited an apparent increase in postsynaptic excita-
bility, as assessed by the ratio of the population
spike area to EPSP amplitude (26%). Figure 1A
shows an example of two granule cells filled with
5,6-carboxyfluorescein taken from the dentate
gyrus of a 24-month-old rat. The bar graph shown
in Fig. 1B shows that older animals have signifi-
cantly higher electrotonic coupling probabilities
compared to young granule cells. The increase in
electrotonic coupling and excitability may contrib-
ute to compensatory processes acting to keep the
probability of discharge of individual granule cells
constant in spite of a considerable loss (20–30%)
of entorhinal cortical afferents to the hippocampal
dentate gyrus.
Single unit recordings in granule cells
Very few studies have investigated the spontaneous
firing characteristics of granule cells in freely be-
having animals. This is partially due to the fact that
few granule cells fire during any given behavioral
condition in the dentate gyrus (Jung and McN-
aughton, 1993; Shen et al., 1998; Gothard et al.,
2001). Additionally, granule cells were originally
misclassified as having ‘‘theta-cell-like’’ firing char-
acteristics (Deadwyler et al., 1976; Rose et al.,
1983). It is now known, however, that granule cells
have low firing rates and selective place fields in
environments, similar to hippocampal pyramidal
cell place fields. For example, Jung and McNaugh-
ton (1993) recorded single unit activity from gran-
ule cells while rats performed a spatial task on an
eight-arm radial maze. It was found that granule
cells showed both spatial and direction specificity in
their firing properties, although their place fields
were smaller and more ‘‘discontigous’’ than pyram-
idal cell place fields. Additionally, single granule
cells have multiple place fields that were stable over
at least several days. It was also observed that fewer
granule cells, compared with hippocampal pyram-
idal cells, showed place-specific firing within any
given environment (Barnes et al., 1990).
These electrophysiological data are consistent
with theoretical ideas suggesting that it is advan-
tageous for hippocampal autoassociative memory
666
processing for information to be coded sparsely
(Marr, 1971; McNaughton and Morris, 1987;
Rolls, 1990). If the dentate gyrus performs the
function of a pattern separator, the formation of
distinct representations would be facilitated by
amplification of small differences in the input to
the granule cell network. Although few granule
cells may fire in any given situation, consistent
with a pattern separation mechanism, granule cell
firing patterns undergo substantial modification
after only minimal changes in an environment
(Leutgeb et al., 2007).
As discussed above, earlier studies mistakenly
identified granule cells to possess firing character-
istics similar to inhibitory interneurons. This may
be explained in part, because of the ease with
which higher firing rate cells are encountered along
a recording tract trajectory. Mizumori et al. (1989)
were able to identify characteristics that could
differentiate between interneurons (basket cells)
and granule cells. For example, interneurons have
high spontaneous discharge rates and a relative
absence of spike accommodation, while granule
cells have lower firing rates. There have been no
systematic age comparisons of the firing charac-
teristics of granule cells; however, a radial maze
spatial working memory task was used to assess
the possible contribution of age-related changes in
the firing characteristics of theta cells (putative in-
terneurons) (Mizumori et al., 1992). The mean
discharge rates of theta cells were significantly
higher in the stratum granulosum in old rats com-
pared to young animals. This region-specific
change in the firing characteristics of theta cells
may have important consequences for information
processing with increasing age.
Perforant path granule cell synaptic responses in
vivo and in vitro
Depending on the location of termination of acti-
vated perforant path synaptic contacts onto the
granule cell dendritic tree, the resulting field poten-
tial responses have distinctive rise time and shape
characteristics. Because of the curved geometry of
the dentate gyrus, field potentials recorded within
the hilar region are nearly proportional to the
current transients at the granule cell bodies. Thus,
the waveform of the population synaptic responses
recorded below the granule cell layer, should reflect
the passive electrical properties of the dendrites,
and location of the active synapses as in an equiv-
alent cylinder model. Because a square root rela-
tionship between synaptic distance and the rise time
is empirically robust (Jefferys, 1975), the square
root of the rise time of the extracellular synaptic
response recorded from the hilus was calculated
and used as an electrotonic location parameter for
comparison across the lifespan of the rat (Barnes
and McNaughton, 1979). By selectively stimulating
small subsets of perforant path fibers along its
medio-lateral axis, synapses at increasing distances
from the soma were activated and the response
waveform assessed (as illustrated in Fig. 2). When
the mean electrotonic location parameter was com-
pared across age groups, the rise time of the extra-
cellular synaptic response was similar across the
lifespan. When the medians of the frequency distri-
butions (as shown in Fig. 3) of the electrotonic lo-
cation parameter were compared in three age
groups of rats, there was a small shift towards
lower values with advancing age. This small shift in
the distribution of extracellularly recorded synaptic
response rise times may suggest atrophy of finer
dendritic branches, which was, in fact an observa-
tion made by Geinisman et al. (1978) in the supra-
granular zone of old rats.
Barnes (1979) first noted, using extracellular
field potential recordings in young and old rats,
that, for a given stimulus intensity delivered to the
perforant path afferent fibers, old Long Evans rats
showed reduced field EPSP amplitudes compared
with young Long Evans rats. Using both in vivo
and in vitro preparations to record extracellularly
in young and old rats, Barnes and McNaughton
(1980a) found that while there was no difference in
the EPSP size at threshold stimulation intensities,
at higher stimulus strengths, the EPSP was con-
sistently smaller in old rats, over a wide range of
intensities (see Fig. 4).
It is also possible to record the compound action
potentials of incoming perforant path fibers (i.e.,
the presynaptic fiber potential) to estimate the
numbers of perforant path axon collaterals that
terminate in the granule cell molecular layer.
 667

Fig. 2. (A) Schematic diagram of a granule cell in the dentate gyrus showing the extracellular recording electrode placed slightly below
the cell body layer (in the hilus). (B) Representative extracellularly recorded synaptic responses evoked from stimulation of different
subpopulations of perforant path fibers are shown adjacent to their approximately corresponding input location on the granule cell
dendrites. Note that the rise time characteristics of the synaptic responses vary as a function of the dendritic location of the activated
synapses. Adapted with modifications from Barnes and McNaughton (1979).

When the presynaptic fiber potential was meas-
ured in young and old rats, again there was no
change in the response amplitude at threshold
stimulus intensity levels, however, above thresh-
old, younger animals showed larger fiber potential
response amplitudes. This finding was consistent
with the reported decline in anatomical synaptic
contacts, as shown by Geinisman (1979).
A smaller presynaptic fiber potential implies
that fewer axon collaterals are activated by the
stimulus, and is therefore a possible explanation
for the smaller extracellular field EPSP observed in
old rats. When field EPSP amplitudes were plotted
against fiber potential amplitudes, however, the
synaptic response was larger for a given input size.
Thus, smaller fiber potentials gave rise to larger
synaptic responses in old rats, inspite of the fact
that at stimulus intensities above threshold fiber
potential amplitudes were smaller.
One explanation for this result would be that
individual synapses that survive in the old rat
would be more effective in depolarizing the
granule cell, i.e., have stronger weights. The later
observation is presumed to be due to axon collat-
eral pruning in the perforant path projection from
the entorhinal cortex. This hypothesis was tested
in vitro, using minimal stimulation methods and
quantal analysis techniques while recording in-
tracellularly from granule cells of aged rats (Foster
et al., 1991). Foster et al. tested rats of three
different ages (5, 9, and 24 months) and found that
the increased EPSP amplitude of granule cells in
old rats was due to an increase in quantal size,
which implicates an increase in postsynaptic sen-
sitivity. Since the magnitude of paired-pulse facil-
itation was unchanged between age groups, the
increase in synaptic strength could not be ex-
plained by an increase in the probability or
number of transmitter quanta released. These post
synaptic changes in synaptic strength have impor-
tant functional implications for maintaining gran-
ule cell activity within a functional range and may
suggest a compensatory mechanism employed by
the aging dentate gyrus.
668

Fig. 3. (A) The frequency distributions of the electrotonic location parameter from adolescent (2 months), middle-aged (12 months), or
old (24 months) rats as percentage of number of observations for each age group. (B) Bar graph shows the medians of the frequency
distributions of the electrotonic parameter in adolescent, middle-aged, or old rats. Note the small shift towards lower values with
advancing age. This small shift in the distribution of extracellularly recorded synaptic response rise times towards lower values with
advancing age may suggest atrophy of the finer dendritic branches. Adapted and modified from Barnes and McNaughton (1979).

Cholinergic responses in vitro in aging granule cells
When cholinergic afferent fibers are stimulated at
40 Hz, a slow cholinergic (atropine-sensitive)
EPSP can be elicited in all three hippocampal sub-
regions. When the response to this type of stim-
ulus, delivered to stratum granulosum, was
evaluated in animals from three different age
groups (3, 9, or 24 months) there was a signifi-
cant age-related decline (50%) in the amplitude
of the slow cholinergic EPSP (Shen and Barnes,
1996). There was also a similar decine in choliner-
gic transmission in the CA3 and CA1 regions of
old rats. Therefore, the neuromodulatory role of
the cholinergic system in aging may alter
excitability of granule cells, which could signifi-
cantly influence hippocampal network behavior
(Huerta and Lisman, 1993; Hasselmo et al., 1996;
Hasselmo, 1999; Barnes et al., 2000b; Ikonen et al.,
2002).
Modification of synaptic weights at the perforant
path-granule cell synapse
The classic studies published in 1973 by Bliss and
Gardner-Medwin (Bliss and Gardner-Medwin,
1973; Bliss and Lomo, 1973) demonstrate that
synaptic projections from the entorhinal cortex
could be modified, both in vivo (Bliss and
 669

Fig. 4. The relation between field EPSP amplitude and perforant path stimulus intensity is shown in young (K) and old (J) rats for in
vivo (A) and in vitro (D) preparations. The response waveforms (averaged across age groups) are shown for young (B) and old (C)
animals from an in vivo experiment. Superimposed traces at various stimulus levels recorded simultaneously from the granule cell layer
(E) and the molecular layer (F) are shown for an in vitro slice experiment. The dashed lines in B and C and E indicate the time of
measurement of the field EPSP (2 ms after stimulus onset). The sharp negative deflections in B, C, and E are population spikes
(asterisks) whereas the early negative deflection in F represents the presynaptic fiber response (indicated by arrow). Adapted from
Barnes and McNaughton (1980a).

Gardner-Medwin, 1973) and in vitro (Bliss and
Lomo, 1973) by patterned activation of perforant
path afferents to hippocampal granule cells. This
phenomenon, now referred to as long-term po-
tentiation or LTP (Douglas and Goddard, 1975),
required convergence of inputs onto granule cells
or ‘‘cooperativity’’ (McNaughton et al., 1978),
consistent with Hebb’s neural postulate for asso-
ciation (Hebb, 1949). It was subsequently found
that this electrophysiological demonstration of
coincidence detection worked through a novel
glutamatergic receptor, the NMDA receptor (Col-
lingridge, 1985). The NMDA receptor was found
to be voltage-dependent and to allow influx of
calcium, presumed at that time to be necessary for
initiating durable forms of plasticity. In fact, Mor-
ris et al. (1986) showed that NMDA-receptor an-
tagonism (via APV administration) was able to
disrupt performance of the spatial, hippocampal-
dependent version of the Morris swim task, while
leaving the nonspatial version intact.
The relationship between the durability of LTP
and spatial memory was first reported in 1979
(Barnes, 1979), with the demonstration that spatial
performance accuracy on the circular platform
task was correlated with the durablility of LTP, as
assessed over a number of days. This experiment
was conducted in young and old Long Evans rats,
trained on the circular platform spatial memory
task, where the goal was to escape from a brightly
illuminated open surface by finding the one hole
(identified by the distal visual cues in the room)
that would allow the rat to escape into a dark box.
Following behavioral testing, 32 young and 32 old
animals were implanted bilaterally with indwelling
electrodes that could monitor field potentials in the
dentate gyrus elicited by stimulation of the medial
perforant path to granule cells. The results of this
experiment were able to show, for the first time,
that within each age group there was a significant
correlation between the decay time constant of late
phase LTP and spatial memory. The same
670
relationship between spatial behavior and LTP
maintenance has been found in young and old
mice at the Schaffer collateral — CA1 synapse in
vitro (Bach et al., 1999). In these experiments a
nonspatial version of the circular platform task
was also administered, and the performance on the
cued task did not correlate with LTP durability.
To examine the LTP induction process over
days, as well as its maintenance after induction
levels were saturated, young and old rats were
prepared for long-term field potential recordings
from the dentate gyrus. After baseline evoked re-
sponse stability to low frequency test pulses had
been established, bursts of high frequency stimu-
lation were delivered once per day for 7 consec-
utive days. Although the old rats reached
asymptotic levels of LTP following this treatment
more slowly than did the younger rats, they did
achieve the same proportional levels of synaptic
strengthening. Decay of LTP was then monitored
for several weeks. Old rats showed faster decay
over days than did young rats (time constants of
17 and 37 days, respectively); Barnes and McN-
aughton (1980a).
As discussed above, if LTP is induced using ro-
bust stimulus parameters at the perforant path-
granule cell synapse, there the degree of LTP is
equivalent between age groups (Barnes, 1979;
Barnes and McNaughton, 1980a). One question
is whether there is an altered threshold for induc-
tion of LTP at this synapse. Certainly there are
fewer synaptic contacts in the termination zone of
the medial perforant path synapses (Geinisman et
al., 1992b) and thus, less synaptic convergence
from perforant path stimulation would be
achieved in older rats.
To address the possibility that the threshold for
LTP was altered in old rats, Barnes et al. (2000a)
directly depolarized the granule cells with intra-
cellular current injection paired with orthodromic
stimulation, thereby eliminating the necessity for
synaptic convergence, which is typically required
to induce LTP (McNaughton et al., 1978). Using
this paradigm more current was required to induce
LTP in aged animals. One possible mechanism for
this apparent deficit in LTP induction in the aged
dentate gyrus could be alterations in NMDA re-
ceptors. The biochemical data reported for the
dentate gyrus are complex during aging (Magnus-
son et al., 2002, 2003), and it is not clear whether
changes in receptor-related functions would result
in the changes in plasticity (e.g., Wenk and Barnes,
2000).
While, the examination of spatial memory and
LTP in the hippocampus of old rats has provided
important insights into age-related functional de-
cline, it has also raised a number of questions.
Among these is, what causes the faster decay in
LTP in old animals? If LTP decay could be ma-
nipulated, memory function might be facilitated in
older animals. The mechanisms involved in the
maintenance of LTP are, unfortunately, less well
understood than those involved in induction of
LTP. As discussed in the following section, some
insight may be gained through an examination of
the role played by immediate early genes in dura-
ble forms of synaptic changes that may underlie
memory changes in old mammals.
Immediate early genes
The induction and translation of immediate early
genes, lead to changes in multiple intracellular
signaling cascades, and may play an important
role in changes observed in age-related plasticity
mechanisms. A number of IEGs are powerfully
activated by stimulation parameters that lead to
LTP induction. As discussed in the previous sec-
tion, old animals, in addition to changes in LTP
induction thresholds, also exhibit decreases in LTP
maintenance that are not detectable until several
days after robust induction protocols have been
applied. Studies examining the basis of the main-
tenance phase of long-term plasticity have dem-
onstrated the requirement for RNA and protein
synthesis following the conditioning stimulus
(Montarolo et al., 1986; Nguyen et al., 1994; La-
nahan et al., 1997). Therefore, it is plausible that
age-dependent memory decline could result from
alterations in pathways subsequent to transcrip-
tional responses mediated by IEGs.
Cole et al. (1989) established a relationship be-
tween LTP induction and activation of several
IEGs (including, zif268, c-fos, c-jun, jun-B, and
Arc). In that study, when LTP at the perforant

path-granule cell synapse was induced by high fre-
quency stimulation, levels of zif 268 mRNA in-
creased significantly. This increase in zif 268
mRNA could be blocked both by NMDA-recep-
tor antagonists and by convergent synaptic inhib-
itory inputs known to block LTP, thus establishing
a correlation between IEG expression and LTP.
Similar findings were reported in un-anesthetized
rats by Dragunow et al. (1989). The early studies
that linked IEG expression in granule cells with
LTP, were however, called into question by sug-
gestions that the thresholds for LTP induction and
gene activation are different. To address this issue,
Worley et al. (1993) used a chronic in vivo record-
ing approach to monitor two different patterns of
LTP-inducing stimuli to perforant path afferents.
The RNA expression pattern and the protein
products of several IEGs (zif 268, jun-B, c-jun, and
c-fos) were monitored in young and old animals,
following the two different electrical stimulation
protocols (10 pulse train or 50 pulse train, deliv-
ered at 400 Hz). Although, different patterns of
IEG expression emerged, only zif 268 was detect-
able at thresholds that induced LTP. Additionally,
when the in situ autoradiograms were quantified
using densitometry, there were no significant age-
related differences in zif 268 mRNA levels between
groups.
Using the more sensitive reverse Northern
method, Lanahan et al. (1997), examined a panel
of several genes (c-fos, jun B, c-jun, fos B, fra1, and
zif 268) known to be induced by LTP. In that
study, an LTP induction protocol (high frequency
stimulation of the perforant fibers) known to result
in LTP maintenance deficits in old rats (Barnes
and McNaughton, 1980b) was used. The dentate
gyrus was extracted from young (9 month) and old
(26 month) animals and reverse Northern blots
were quantified using a phosphoimager. Of all the
genes examined (including zif 268) only c-fos RNA
level differed between young and old rats. Inter-
estingly, while both young and old rats showed
elevated c-fos expression following LTP induction,
the older rats showed greater levels of expression
than did the young animals. This indicated that
changes in c-fos, may contribute to altered signa-
ling mechanisms that are involved in age-related
hippocampal plasticity deficits observed in aging.
671
As the studies of specific IEGs in activity-depend-
ent synaptic plasticity continued, an increase in a
unique gene, activity-regulated cytoskeleton-associ-
ated gene (Arc, also known as Arg 3.1) was iden-
tified, following seizures, or LTP induced at the
perforant path granule cell synapse (Link et al.,
1995; Lyford et al., 1995). It seemed logical to in-
vestigate whether Arc had a role in the maintenance
phase of LTP. Guzowski et al. (2000) used intra-
hippocampal infusions of antisense oligodeoxynuc-
leotides to block the synthesis of Arc protein and
monitored the effect on LTP and spatial memory.
Blocking the synthesis of Arc protein in the dorsal
hippocampus impaired the maintenance phase of
LTP induced at the perforant path-granule cell
synapse, and the long-term consolidation of memory
for a spatial location (Morris water maze) without
interfering with either the induction phase of LTP or
water maze acquisition. Thus, a role for Arc in hip-
pocampal-dependent plasticity was established.
The IEG Arc proved to be a useful tool to clarify
behavior-related synaptic plasticity as well, as it is
rapidly induced when animals explore their envi-
ronment. Guzowski et al. (1999) showed that Arc
can be induced in hippocampal CA1 neurons fol-
lowing two episodes of spatial experience and that
the proportions of cells expressing Arc matched
predictions derived from electrophysiological re-
cordings. In that investigation it was also shown
that Arc RNA is transcribed within 1–2 min fol-
lowing behavior, and that it appears as discrete
transcriptional foci in the nucleus (for 2–15 min).
Within 15–30 min the RNA translocates to the cy-
toplasm, and can be subsequently found in the
dendrites (45 min). This unique property of Arc
formed the basis for the development of a powerful
new technique called compartment analysis of tem-
poral activity by fluorescence in situ hybridization
or catFISH. With this technique, the behavior-in-
duced activity history of neuronal ensembles could
be inferred at two different time points, throughout
the brain (e.g., Guzowski et al., 1999; Chawla et al.,
2004). Although, many IEGs can be used in the
catFISH technique, Arc provides distinct advan-
tages in that it is dynamically regulated in many
brain regions.
When the pattern of spatial exploration-induced
Arc expression was examined in hippocampal
672

granule cells of young rats, it was found to be

predominantly located in the upper blade of the
dentate gyrus (Chawla et al., 2005). No major an-
atomical differences in afferents to the upper vs.
lower blade easily account for this interesting ob-
servation, since only minor differences in anatom-
ical connectivity exist between these regions. The
proportion of active cells following spatial experi-
ence is low for granule cells compared to CA1 py-
ramidal cells (4% vs. 40%), and is consistent
with the sparse firing pattern observed in the dent-
ate gyrus in electrophysiological studies (Barnes et
al., 1990; Jung and McNaughton, 1993; Shen et
al., 1997; Gothard et al., 2001). Figure 5 illustrates
the expression of behaviorally induced Arc in
young rats (Fig. 5A), which is significantly ele-
vated compared to baseline levels of expression in
young animals that have not performed the spatial
foraging task (Fig. 5B). The difference in upper
and lower blade Arc expression cannot be ex-
plained by the inability of the lower blade to ex-
press Arc, because as early as 5 min following
maximal electroconvulsive shock (MECS) treat-
ment, both the upper and lower blades show ro-
bust Arc expression. It is possible that granule cells
in the lower blade are largely inactive for a variety
of experiences.

Vulnerability of dentate gyrus in aging

The information gathered over the past several
decades from electrophysiological recording and
anatomical studies, has been critical in furthering
our understanding of the kinds of learning and
memory deficits that should be expected in normal
aging. Anatomical and physiological studies can
be complimented by functional imaging ap-
proaches that can directly identify behavioral cor-
relates of neuronal function. Several imaging
techniques are now available that provide cellular
and temporal resolution in this regard. Among
them is the functional imaging method based on
measurement of cerebral blood flow volumes using
magnetic resonance imaging (MRI) that have been
shown to correlate with regional energy metabo-
lism (e.g., Gonzalez et al., 1995; Hyder et al.,
2001). Recently, Small et al. (2004) measured CBV
Fig. 5. Image of granule cells (blue) following in situ hybrid-
ization using a c-RNA probe for the IEG Arc. (A) Arc positive
cells (shown in red) in the granule cell layer from the upper
blade of a young rat following two 5 min (separated by a 20 min
rest period in the home cage) episodes of spatial exploratory
behavior. (B) Granule cells from a caged control animal of the
same age, that did not perform the exploratory behavior, are
shown, illustrating baseline or negligible expression of Arc
mRNA. (See Color Plate 36.5 in color plate section.)
values in 14 young and old rhesus monkeys (7–31
years) that also performed a memory assessment
task (delayed nonmatching-to-sample, DNMS).
Representative CBV meaures from a young and
old monkey is shown in Fig. 6A. The CBV meas-
ures indicated a significant inverse correlation with
age only in the dentate gyrus (Fig. 6B). In addi-
tion, when CBV measurements were plotted
 673

Fig. 6. (A) Examples of gadolinium-induced change in MRI signal, (a measure of cerebral blood flow volume, CBV, that is a correlate
of brain metabolism) from a young and old rhesus monkey. Warm colors indicate higher CBV values. The white circle, identified in
precontrast images, overlies the dentate gyrus. (B) CBV measurements plotted vs. age in the dentate gyrus indicates a significant decline
with advancing age. (C) CBV measurements in the dentate gyrus of old monkeys plotted against memory performance on the delayed
nonmatching-to-sample test. A significant relationship between CBV and memory performance is observed, selectively in the old
animals. Adapted with modifications from Small et al. (2004). (See Color Plate 36.6 in color plate section.)

against memory performance on a delayed non-
matching-to-sample task; again, only the dentate
gyrus in old monkeys showed a significant corre-
lation with CBV activity (Fig. 6C). As discussed
previously, catFISH can provide both temporal
and cellular resolution in determining the activity
history of individual neurons. When cellular Arc
RNA was determined using catFISH combined
with high-resolution confocal microscopy in
young, middle-aged, and old rats, only the dent-
ate gyrus showed age-related changes. As shown in
Fig. 7B, a significant age-related decline in Arc-
RNA-containing granule cells was observed in
aged animals. Importantly, as was seen in mon-
keys with MRI imaging, the pyramidal subregion
in rats was unaffected by advancing age, i.e., there
was no change in Arc-expressing pyramidal cell
numbers.
Furthermore, using MRI studies in humans,
Small et al. (2002) have also reported that
subiculum and dentate gyrus show selective
changes in normal aging that could be correlated
with memory decline in these individuals. Thus, a
cross species consensus is now emerging in that the
dentate gyrus in rats, monkeys, and humans is the
hippocampal subregion that is particularly vulner-
able to advancing age.
Conclusions and summary
Aging is accompanied by selective memory
changes that depend on proper hippocampal func-
tion. Investigation of hippocampal anatomy, phys-
iology, and gene expression in normal aging
provides a foundation for understanding age-re-
lated cognitive decline. A common theme that
emerges from examination of the neurobiological
changes in aging granule cells, is that cell loss can-
not account for the cognitive decline, as granule
674

Fig. 7. Behavior-induced Arc mRNA expression pattern in young and old rat granule cells of the dentate gyrus. (A) Individual
examples of behaviorally induced Arc RNA. Fluorescence-tagged CY3 Arc RNA containing cells are green, and granule cell nuclei are
shown in red (counterstained with propydium iodide, a nucleic acid stain). (B) The average proportion of Arc-positive neurons (% of
total cells) measured from the three different age groups (9, 15, and 24 months) is shown for the dentate gyrus. Note that the dentate
gyrus of old rats contains a significantly lower proportion of behaviorally induced Arc-positive granule cells compared to young or
middle-aged animals. (See Color Plate 36.7 in color plate section.)

cell numbers are preserved during aging. Although
there is an absence of neuronal degeneration and a
general preservation of basic biophysical proper-
ties, there is a significant loss of axodentritic
synapses onto granule cells from entorhinal affer-
ents. In addition to the synapse loss, there are a
number of other changes that occur in aged ani-
mals. Among them, is an altered threshold for
LTP induction, an increase in strength of the sur-
viving synapses, and increased electrotonic cou-
pling between the old granule cells. Together, these
factors may result in an attenuation of the dele-
terious impact of aging on cognition.
In spite of the compensatory changes men-
tioned above, granule cells appear to be partic-
ularly vulnerable to the effects of aging compared
to hippocampal pyramidal cells. For example,
there is a reduction in the numbers of granule
cells that show behavior-induced gene expression
in old rats. Furthermore, older monkeys that
have difficulty performing a hippocampal-de-
pendent learning task also have lower functional
activity in the dentate gyrus as assessed by cer-
ebral blood volume measures using MRI tech-
niques. Additionally, MRI studies in humans also
indicate that dentate gyrus is also a hippocampal
subregion with selective changes in normal aging
that could be correlated with memory decline in
these individuals.
With the human population living well into their
eighties or nineties, it is becoming increasingly im-
portant to find ways to preserve cognitive func-
tion, and maintain an optimal quality of life
during the last decades. New imaging techniques,
pharmacological interventions or genetic manipu-
lations, can advance our knowledge regarding
normal aging. A multidisciplinary effort will be
essential for identifying early interventions to
ameliorate declining cognition with advancing age.
Acknowledgements
We thank Michelle Carroll for her excellent techni-
cal assistance in preparing the figures. The authors
and some of this work were supported by grants
from the National Institute of Health AG009219
and AG003376, the state of Arizona and ADHS
and the McKnight Brain Research Foundation.
References
Aimone, J.B., Wiles, J. and Gage, F.H. (2006) Potential role for
adult neurogenesis in the encoding of time in new memories.
Nat. Neurosci., 9: 723–727.
Altman, J. and Das, G.D. (1965) Autoradiographic and histo-
logical evidence of postnatal hippocampal neurogenesis in
rats. J. Comp. Neurol., 124: 319–335.

Bach, M.E., Barad, M., Son, H., Zhuo, M., Lu, Y.F., Shih, R.,
Mansuy, I., Hawkins, R.D. and Kandel, E.R. (1999) Age-
related defects in spatial memory are correlated with defects in
the late phase of hippocampal long-term potentiation in vitro
and are attenuated by drugs that enhance the cAMP signaling
pathway. Proc. Natl. Acad. Sci. U.S.A., 96: 5280–5285.
Badiali de Giorgi, L., Bonvicini, F., Bianchi, D., Bossoni, G.
and Laschi, R. (1987) Ultrastructural aspects of ageing rat
hippocampus and effects of L-acetyl-carnitine treatment.
Drugs Exp. Clin. Res., 13: 185–189.
Ball, M.J. (1977) Neuronal loss, neurofibrillary tangles and
granulovacuolar degeneration in the hippocampus with age-
ing and dementia. A quantitative study. Acta Neuropathol.
(Berl.), 37: 111–118.
Barnes, C.A. (1979) Memory deficits associated with senes-
cence: a neurophysiological and behavioral study in the rat. J.
Comp. Physiol. Psychol., 93: 74–104.
Barnes, C.A. (1988) Spatial learning and memory processes: the
search for their neurobiological mechanisms in the rat.
Trends Neurosci., 11: 163–169.
Barnes, C.A. (1994) Normal aging: regionally specific changes
in hippocampal synaptic transmission. Trends Neurosci., 17:
13–18.
Barnes, C.A. and McNaughton, B.L. (1979) Neurophysiolog-
ical comparison of dendritic cable properties in adolescent,
middle-aged, and senescent rats. Exp. Aging Res., 5: 195–206.
Barnes, C.A. and McNaughton, B.L. (1980a) Physiological
compensation for loss of afferent synapses in rat hippocam-
pal granule cells during senescence. J. Physiol., 309: 473–485.
Barnes, C.A. and McNaughton, B.L. (1980b) Spatial memory
and hippocampal synaptic plasticity in senescent and middle-
aged rats. In: Stein D. (Ed.), The Psychobiology of Aging:
Problems and Perspectives. Elsevier Press, New York,
pp. 253–272.
Barnes, C.A., McNaughton, B.L., Mizumori, S.J., Leonard,
B.W. and Lin, L.H. (1990) Comparison of spatial and tem-
poral characteristics of neuronal activity in sequential stages
of hippocampal processing. Prog. Brain Res., 83: 287–300.
Barnes, C.A., Meltzer, J., Houston, F., Orr, G., McGann, K.
and Wenk, G.L. (2000b) Chronic treatment of old rats with
donepezil or galantamine: effects on memory, hippocampal
plasticity and nicotinic receptors. Neuroscience, 99: 17–23.
Barnes, C.A., Rao, G. and Houston, F.P. (2000a) LTP induc-
tion threshold change in old rats at the perforant path —
granule cell synapse. Neurobiol. Aging, 21: 613–620.
Barnes, C.A., Rao, G. and McNaughton, B.L. (1987) Increased
electrotonic coupling in aged rat hippocampus: a possible
mechanism for cellular excitability changes. J. Comp. Ne-
urol., 259: 549–558.
Barnes, C.A., Treves, A., Rao, G. and Shen, J. (1994) Electro-
physiological markers of cognitive aging: region specificity and
computational consequences. Semin. Neurosci., 6: 359–367.
Baskys, A., Niesen, C.E. and Carlen, P.L. (1987) Altered mod-
ulatory actions of serotonin on dentate granule cells of aged
rats. Brain Res., 419: 112–118.
Bertoni-Freddari, C., Giuli, C., Pieri, C. and Paci, D. (1986)
Quantitative investigation of the morphological plasticity of
675
synaptic junctions in rat dentate gyrus during aging. Brain
Res., 366: 187–192.
Bizon, J.L. and Gallagher, M. (2003) Production of new cells in
the rat dentate gyrus over the lifespan: relation to cognitive
decline. Eur. J. Neurosci., 18: 215–219.
Bizon, J.L., Lee, H.J. and Gallagher, M. (2004) Neurogenesis in
a rat model of age-related cognitive decline. Aging Cell, 3:
227–234.
Bliss, T.V.P. and Gardner-Medwin, A.R. (1973) Long-lasting
potentiation of synaptic transmission in the dentate area of
the unanaesthetised rabbit following stimulation of the per-
forant path. J. Physiol., 232: 357–374.
Bliss, T.V.P. and Lomo, T. (1973) Long-lasting potentiation of
synaptic transmission in the dentate area of the anesthetized
rabbit following stimulus of perforant path. J. Physiol., 232:
331–356.
von Bohlen und Halbach, O., Zacher, C., Gass, P. and Un-
sicker, K. (2006) Age-related alterations in hippocampal
spines and deficiencies in spatial memory in mice. J. Neuro-
sci. Res., 83: 525–531.
Brody, H. (1955) Organization of the cerebral cortex. III. A
study of aging in the human cerebral cortex. J. Comp. Ne-
urol., 102: 511–556.
Burke, S.N. and Barnes, C.A. (2006) Neural plasticity in the
ageing brain. Nat. Rev. Neurosci., 7: 30–40.
Cabeza, R., Nyberg, L. and Park, D. (2005) Cognitive Neuro-
science of Aging. Oxford University Press, New York.
Chawla, M.K., Guzowski, J.F., Ramirez-Amaya, V., Lipa, P.,
Hoffman, K.L., Marriott, L.K., Worley, P.F., McNaughton,
B.L. and Barnes, C.A. (2005) Sparse, environmentally selec-
tive expression of Arc RNA in the upper blade of the rodent
fascia dentata by brief spatial experience. Hippocampus, 15:
579–586.
Chawla, M.K., Lin, G., Olson, K., Vazdarjanova, A., Burke,
S.N., McNaughton, B.L., Worley, P.F., Guzowski, J.F.,
Roysam, B. and Barnes, C.A. (2004) 3D-catFISH: a system
for automated quantitative three-dimensional compartmental
analysis of temporal gene transcription activity imaged by flu-
orescence in situ hybridization. J. Neurosci. Methods, 139:
13–24.
Cole, A.J., Saffen, D.W., Baraban, J.M. and Worley, P.F.
(1989) Rapid increase of an immediate early gene messenger
RNA in hippocampal neurons by synaptic NMDA receptor
activation. Nature, 340: 474–476.
Coleman, P.D. and Flood, D.G. (1987) Neuron numbers and
dendritic extent in normal aging and Alzheimer’s disease.
Neurobiol. Aging, 8: 521–545.
Coleman, P.D., Flood, D.G. and West, M.J. (1987) Volumes of
the components of the hippocampus in the aging F344 rat. J.
Comp. Neurol., 266: 300–306.
Collingridge, G.L. (1985) Long term potentiation in the hip-
pocampus: Mechanisms of initiation and modulation by ne-
urotransmitters. Trends Pharmacol. Sci., 6: 407–411.
Deadwyler, S.A., Gribkoff, V., Cotman, C. and Lynch, G.
(1976) Long lasting changes in the spontaneous activity of
hippocampal neurons following stimulation of the entorhinal
cortex. Brain. Res. Bull., 1: 1–7.
676
Douglas, R.M. and Goddard, G.V. (1975) Long-term potent-
iation of the perforant path-granule cell synapse in the rat
hippocampus. Brain Res., 86: 205–215.
Dragunow, M., Abraham, W.C., Goulding, M., Mason, S.E.,
Robertson, H.A. and Faull, R.L. (1989) Long-term potent-
iation and the induction of c-fos mRNA and proteins in the
dentate gyrus of unanesthetized rats. Neurosci. Lett., 101:
274–280.
Drapeau, E., Mayo, W., Aurousseau, C., Le Moal, M., Piazza,
P.V. and Abrous, D.N. (2003) Spatial memory performances
of aged rats in the water maze predict levels of hippocampal
neurogenesis. Proc. Natl. Acad. Sci. U.S.A., 100: 14385–14390.
Driscoll, I., Howard, S.R., Stone, J.C., Monfils, M.H., To-
manek, B., Brooks, W.M. and Sutherland, R.J. (2006) The
aging hippocampus: a multi-level analysis in the rat. Neuro-
science, 139: 1173–1185.
Driscoll, I. and Sutherland, R.J. (2005) The aging hippocam-
pus: navigating between rat and human experiments. Rev.
Neurosci., 16: 87–121.
Eriksson, P.S. (2003) Neurogenesis and its implications for re-
generation in the adult brain. J. Rehabil. Med., 41(Suppl.):
17–19.
Foster, T.C., Barnes, C.A., Rao, G. and McNaughton, B.L.
(1991) Increase in perforant path quantal size in aged F-344
rats. Neurobiol. Aging, 12: 441–448.
Gallagher, M. and Rapp, P.R. (1997) The use of animal models
to study the effects of aging on cognition. Annu. Rev. Psy-
chol., 48: 339–370.
Geinisman, Y. (1979) Loss of axosomatic synapses in the dent-
ate gyrus of aged rats. Brain Res., 168: 485–492.
Geinisman, Y. and Bondareff, W. (1976) Decrease in the
number of synapses in the senescent brain: A quantitative
electron microscopic analysis of the dentate gyrus molecular
layer in the rat. Mech. Aging Dev., 5: 11–23.
Geinisman, Y., Bondareff, W. and Dodge, J.T. (1977) Partial
deafferntation of neurons in the dentate gyrus of the senes-
cent rat. Brain Res., 134: 541–545.
Geinisman, Y., Bondareff, W. and Dodge, J.T. (1978) Dendritic
atrophy in the dentate gyrus of the senescent rat. Am. J.
Anat., 152: 321–329.
Geinisman, Y., de Toledo-Morrell, L. and Morrell, F. (1986)
Loss of perforated synapses in the dentate gyrus: morpho-
logical substrate of memory deficit in aged rats. Proc. Natl.
Acad. Sci. U.S.A., 83: 3027–3031.
Geinisman, Y., de Toledo-Morrell, L., Morrell, F., Persina, I.S.
and Rossi, M. (1992) Age-related loss of axospinous synapses
formed by two afferent systems in the rat dentate gyrus as
revealed by the unbiased stereological dissector technique.
Hippocampus, 2: 437–444.
Gonzalez, R.G., Fischman, A.J., Guimaraes, A.R., Carr, C.A.,
Stern, C.E., Halpern, E.F., Growdon, J.H. and Rosen, B.R.
(1995) Functional MR in the evaluation of dementia: corre-
lation of abnormal dynamic cerebral blood volume measure-
ments with changes in cerebral metabolism on positron
emission tomography with fludeoxyglucose F 18. AJNR Am.
J. Neuroradiol., 16: 1763–1770.
Gothard, K.M., Hoffman, K.L., Battaglia, F.P. and McN-
aughton, B.L. (2001) Dentate gyrus and ca1 ensemble activity
during spatial reference frame shifts in the presence and ab-
sence of visual input. J. Neurosci., 21: 7284–7292.
Gould, E. and Cameron, H.A. (1996) Regulation of neuronal
birth, migration and death in the rat dentate gyrus. Dev.
Neurosci., 18: 22–35.
Guzowski, J.F., Lyford, G.L., Stevenson, G.D., Houston,
F.P., McGaugh, J.L., Worley, P.F. and Barnes, C.A. (2000)
Inhibition of activity-dependent arc protein expression in the
rat hippocampus impairs the maintenance of long-term po-
tentiation and the consolidation of long-term memory.
J. Neurosci., 20: 3993–4001.
Guzowski, J.F., McNaughton, B.L., Barnes, C.A. and Worley,
P.F. (1999) Environment-specific expression of the immedi-
ate-early gene Arc in hippocampal neuronal ensembles. Nat.
Neurosci., 2: 1120–1124.
Hasselmo, M.E. (1999) Neuromodulation: acetylcholine and
memory consolidation. Trends Cogn. Sci., 3: 351–359.
Hasselmo, M.E., Wyble, B.P. and Wallenstein, G.V. (1996)
Encoding and retrieval of episodic memories: role of
cholinergic and GABAergic modulation in the hippocam-
pus. Hippocampus, 6: 693–708.
Hebb, D.O. (1949) The Organization of Behavior. Wiley, New
York.
Hoff, S.F., Scheff, S.W., Bernardo, L.S. and Cotman, C.W.
(1982) Lesion-induced synaptogenesis in the dentate gyrus of
aged rats: I. Loss and reaquisition of normal synaptic den-
sity. J. Comp. Neurol., 205: 246–252.
Huerta, P.T. and Lisman, J.E. (1993) Heightened synaptic
plasticity of hippocampal CA1 neurons during a cholinergic-
ally induced rhythmic state. Nature, 364: 723–725.
Hyder, F., Kida, I., Behar, K.L., Kennan, R.P., Maciejewski,
P.K. and Rothman, D.L. (2001) Quantitative functional im-
aging of the brain: towards mapping neuronal activity by
BOLD fMRI. NMR Biomed., 14: 413–431.
Ikonen, S., McMahan, R., Gallagher, M., Eichenbaum, H. and
Tanila, H. (2002) Cholinergic system regulation of spatial
representation by the hippocampus. Hippocampus, 12:
386–397.
Jefferys, J.G.R. (1975) Propagation of action potentials into the
dendrite of hippocampal granule cells in vitro. J. Physiol.,
249: 16–18.
Jung, M.W. and McNaughton, B.L. (1993) Spatial selectivity of
unit activity in the hippocampal granular layer. Hippocam-
pus, 3:165–182.
Kadar, T., Arbel, I., Silbermann, M. and Levy, A. (1994) Mor-
phological hippocampal changes during normal aging and
their relation to cognitive deterioration. J. Neural. Transm.
Suppl., 44: 133–143.
Kempermann, G. and Gage, F.H. (1999) Experience-dependent
regulation of adult hippocampal neurogenesis: effects of
long-term stimulation and stimulus withdrawal. Hippocam-
pus, 9: 321–332.
Kempermann, G., Gast, D. and Gage, F.H. (2002) Neuroplas-
ticity in old age: sustained fivefold induction of hippocampal

neurogenesis by long-term environmental enrichment. Ann.
Neurol., 52: 135–143.
Keuker, J.I., Luiten, P.G. and Fuchs, E. (2000) Capillary
changes in hippocampal CA1 and CA3 areas of the aging
rhesus monkey. Acta Neuropathol. (Berl.), 100: 665–672.
Kuhn, H.G., Dickinson-Anson, H. and Gage, F.H. (1996) Ne-
urogenesis in the dentate gyrus of the adult rat: age-related
decrease of neuronal progenitor proliferation. J. Neurosci.,
16: 2027–2033.
Lanahan, A., Lyford, G., Stevenson, G.S., Worley, P.F. and
Barnes, C.A. (1997) Selective alteration of long-term potent-
iation-induced transcriptional response in hippocampus of
aged, memory-impaired rats. J. Neurosci., 17: 2876–2885.
Landfield, P.W. (1988) Hippocampal neurobiological mecha-
nisms of age-related memory dysfunction. Neurobiol. Aging,
9: 571–579.
Leutgeb, J.K., Leutgeb, S., Moser, M.B. and Moser, E.I. (2007)
Pattern separation in the dentate gyrus and CA3 of the
hippocampus. Science, 315: 961–966.
Link, W., Konietzko, U., Kauselmann, G., Krug, M., Schwa-
nke, B., Frey, U. and Kuhl, D. (1995) Somatodendritic ex-
pression of an immediate early gene is regulated by synaptic
activity. Proc. Natl. Acad. Sci. U.S.A., 92: 5734–5738.
Luebke, J.I. and Rosene, D.L. (2003) Aging alters dendritic
morphology, input resistance, and inhibitory signaling in
dentate granule cells of the rhesus monkey. J. Comp. Neurol.,
460: 573–584.
Lyford, G.L., Yamagata, K., Kaufmann, W.E., Barnes, C.A.,
Sanders, L.K., Copeland, N.G., Gilbert, D.J., Jenkins, N.A.,
Lanahan, A.A. and Worley, P.F. (1995) Arc, a growth factor
and activity-regulated gene, encodes a novel cytoskeleton-
associated protein that is enriched in neuronal dendrites.
Neuron, 14: 433–445.
MacVicar, B.A. and Dudek, F.E. (1982) Electrotonic coupling
between granule cells of rat dentate gyrus: Physiological and
anatomical evidence. J. Neurophysiol., 47: 579–592.
Magnusson, K.R., Nelson, S.E. and Young, A.B. (2002) Age-
related changes in the protein expression of subunits of the
NMDA receptor. Brain Res. Mol. Brain Res., 99: 40–45.
Magnusson, K.R., Scruggs, B., Aniya, J., Wright, K.C., Ontl,
T., Xing, Y. and Bai, L. (2003) Age-related deficits in mice
performing working memory tasks in a water maze. Behav.
Neurosci., 117: 485–495.
Marr, D. (1971) Simple memory: a theory for archicortex. Phi-
los. Trans. R. Soc. B, 262: 23–81.
McNaughton, B.L., Douglas, R.M. and Goddard, G.V. (1978)
Synaptic enhancement in fascia dentata: cooperativity among
coactive afferents. Brain Res., 157: 277–293.
McNaughton, B.L. and Morris, R.G.M. (1987) Hippocampal
synaptic enhancement and information storage within a dis-
tributed memory system. Trends Neurosci., 10: 408–415.
McWilliams, J.R. and Lynch, G. (1984) Synaptic density and
axonal aprouting in rat hippocampus: stability in adulthood
abd decline in late adulthood. Brain Res., 294: 152–156.
Merrill, D.A., Chiba, A.A. and Tuszynski, M.H. (2001) Con-
servation of neuronal number and size in entorhinal cortex of
677
behaviorally characterized aged rats. J. Comp. Neurol., 438:
445–456.
Merrill, D.A., Roberts, J.A. and Tuszynski, M.H. (2000) Con-
servation of neuron number and size in entorhinal cortex
layers, I.I., III, and V/VI of aged primates. J. Comp. Neurol.,
422: 396–401.
Mizumori, S.J., Barnes, C.A. and McNaughton, B.L. (1989)
Reversible inactivation of the medial septum: selective effects
on the spontaneous unit activity of different hippocampal cell
types. Brain Res., 500: 99–106.
Mizumori, S.J., Barnes, C.A. and McNaughton, B.L. (1992)
Differential effects of age on subpopulations of hippocampal
theta cells. Neurobiol. Aging, 13: 673–679.
Montarolo, P.G., Goelet, P., Castellucci, V.F., Morgan, J.,
Kandel, E.R. and Schacher, S. (1986) A critical period for
macromolecular synthesis in long-term heterosynaptic facil-
itation in Aplysia. Science, 234: 1249–1254.
Morris, R.G.M., Anderson, E., Lynch, G.S. and Baudry, M.
(1986) Selective impairment of learning and blockade of long-
term potentiation by an N-methyl-D-aspartate receptor an-
tagonist, AP5. Nature, 319: 774–776.
Morrison, J.H. and Hof, P.R. (1997) Life and death of neurons
in the aging brain. Science, 278: 412–419.
Nguyen, P.V., Abel, T. and Kandel, E.R. (1994) Requirement
of a critical period of transcription for induction of a late
phase of LTP. Science, 265: 1104–1107.
Nicolle, M.M., Gallagher, M. and McKinney, M. (1999) No
loss of synaptic proteins in the hippocampus of aged, be-
haviorally impaired rats. Neurobiol. Aging, 20: 343–348.
Niesen, C.E., Baskys, A. and Carlen, P.L. (1988) Reversed et-
hanol effects on potassium conductances in aged hippocam-
pal dentate granule neurons. Brain Res., 445: 137–141.
Peters, A., Rosene, D.L., Moss, M.B., Kemper, T.L., Abraham,
C.R., Tigges, J. and Albert, M.S. (1996) Neurobiological
bases of age-related cognitive decline in the rhesus monkey. J.
Neuropathol. Exp. Neurol., 55: 861–874.
van Praag, H., Schinder, A.F., Christie, B.R., Toni, N., Palmer,
T.D. and Gage, F.H. (2002) Functional neurogenesis in the
adult hippocampus. Nature, 415: 1030–1034.
van Praag, H., Shubert, T., Zhao, C. and Gage, F.H. (2005)
Exercise enhances learning and hippocampal neurogenesis in
aged mice. J. Neurosci., 25: 8680–8685.
Ramirez-Amaya, V., Gage, F.H., Worley, P.F. and Barnes,
C.A. (2006) Integration of new neurons into functional neu-
ral networks. J. Neurosci., 26: 12237–12241.
Rao, M.S., Hattiangady, B., Abdel-Rahman, A., Stanley, D.P.
and Shetty, A.K. (2005) Newly born cells in the ageing dent-
ate gyrus display normal migration, survival and neuronal
fate choice but endure retarded early maturation. Eur. J.
Neurosci., 21: 464–476.
Rapp, P.R. and Amaral, D.G. (1991) Recognition memory
deficits in a subpopulation of aged monkeys resemble the
effects of medial temporal lobe damage. Neurobiol. Aging,
12: 481–486.
Rapp, P.R., Deroche, P.S., Mao, Y. and Burwell, R.D. (2002)
Neuron number in the parahippocampal region is preserved
678
in aged rats with spatial learning deficits. Cereb. Cortex, 12:
1171–1179.
Rapp, P.R. and Gallagher, M. (1996) Preserved neuron number
in the hippocampus of aged rats with spatial learning deficits.
Proc. Natl. Acad. Sci. U.S.A., 93: 9926–9930.
Rapp, P.R., Stack, E.C. and Gallagher, M. (1999) Morpho-
metric studies of the aged hippocampus: I. Volumetric anal-
ysis in behaviorally characterized rats. J. Comp. Neurol., 403:
459–470.
Rasmussen, T., Schliemann, T., Sorensen, J.C., Zimmer, J. and
West, M.J. (1996) Memory impaired aged rats: no loss of
principal hippocampal and subicular neurons. Neurobiol.
Aging, 17: 143–147.
Ribak, C.E., Korn, M.J., Shan, Z. and Obenaus, A.
(2004) Dendritic growth cones and recurrent basal dend-
rites are typical features of newly generated dentate gran-
ule cells in the adult hippocampus. Brain Res., 1000:
195–199.
Rolls, E.T. (1990) Functions of the primate hippocampus in
spatial processing and memory. In: Kesner R.P. and Olton
D.S. (Eds.), Neurobiology of Comparative Cognition. Com-
parative Cognition and Neuroscience. Lawrence Erlbaum
Associates, Inc., Hillsdale, MI, pp. 339–362.
Rose, G., Diamond, D. and Lynch, G.S. (1983) Dentate gran-
ule cells have the behavioral characteristics of theta neurons.
Brain Res., 266: 29–37.
Rosenzweig, E.S. and Barnes, C.A. (2003) Impact of aging on
hippocampal function: plasticity, network dynamics, and
cognition. Prog. Neurobiol., 69: 143–179.
Samsonovich, A.V. and Ascoli, G.A. (2006) Morphological
homeostasis in cortical dendrites. Proc. Natl. Acad. Sci.
U.S.A., 103: 1569–1574.
Shen, J. and Barnes, C.A. (1996) Age-related decrease in
cholinergic synaptic transmission in three hippocampal sub-
fields. Neurobiol. Aging, 17: 439–451.
Shen, J., Barnes, C.A., McNaughton, B.L., Skaggs, W.E. and
Weaver, K.L. (1997) The effect of aging on experience-de-
pendent plasticity of hippocampal place cells. J. Neurosci.,
17: 6769–6782.
Shen, J., Kudrimoti, H.S., McNaughton, B.L. and Barnes, C.A.
(1998) Reactivation of neuronal ensembles in hippocampal
dentate gyrus during sleep after spatial experience. J. Sleep
Res., 7(Suppl 1): 6–16.
Shors, T.J., Miesegaes, G., Beylin, A., Zhao, M., Rydel, T. and
Gould, E. (2001) Neurogenesis in the adult is involved in the
formation of trace memories. Nature, 410: 372–376.
Small, S.A., Chawla, M.K., Buonocore, M., Rapp, P.R. and
Barnes, C.A. (2004) Imaging correlates of brain function in
monkeys and rats isolates a hippocampal subregion differ-
entially vulnerable to aging. Proc. Natl. Acad. Sci. U.S.A.,
101: 7181–7186.
Small, S.A., Tsai, W.Y., DeLaPaz, R., Mayeux, R. and Stern,
Y. (2002) Imaging hippocampal function across the human
life span: is memory decline normal or not? Ann. Neurol., 51:
290–295.
Smith, T.D., Adams, M.M., Gallagher, M., Morrison, J.H. and
Rapp, P.R. (2000) Circuit-specific alterations in hippocampal
synaptophysin immunoreactivity predict spatial learning im-
pairment in aged rats. J. Neurosci., 20: 6587–6593.
Song, H., Kempermann, G., Overstreet Wadiche, L., Zhao, C.,
Schinder, A.F. and Bischofberger, J. (2005) New neurons in
the adult mammalian brain: synaptogenesis and functional
integration. J. Neurosci., 25: 10366–10368.
Sterio, D.C. (1984) The unbiased estimation of number and
sizes of arbitrary particles using the disector. J. Microsc., 134:
127–136.
Wati, H., Kudo, K., Qiao, C., Kuroki, T. and Kanba, S. (2006)
A decreased survival of proliferated cells in the hippocampus
is associated with a decline in spatial memory in aged rats.
Neurosci Lett., 399: 171–174.
Wenk, G.L. and Barnes, C.A. (2000) Regional changes in the
hippocampal density of AMPA and NMDA receptors across
the lifespan of the rat. Brain Res., 885: 1–5.
West, M.J. (1993a) New stereological methods for counting
neurons. Neurobiol. Aging, 14: 275–285.
West, M.J. (1993b) Regionally specific loss of neurons in the
aging human hippocampus. Neurobiol. Aging, 14: 287–293.
West, M.J., Coleman, P.D., Flood, D.G. and Troncoso, J.C.
(1994) Differences in the pattern of hippocampal neuronal
loss in normal ageing and Alzheimer’s disease. Lancet, 344:
769–772.
Worley, P.F., Bhat, R.V., Baraban, J.M., Erickson, C.A.,
McNaughton, B.L. and Barnes, C.A. (1993) Thresholds for
synaptic activation of transcription factors in hippocampus:
correlation with long-term enhancement. J. Neurosci., 13:
4776–4786.
Plate 33.1. Wiring diagram of the hippocampus (interneurons excluded) showing dual inputs from the lateral and medial entorhinal
cortex. The layer 2 cortical inputs to the dentate and CA3 diverge fan out (f) widely over these networks and then provide convergent
input to individual dentate granule cells. In contrast, the layer 3 cortical inputs are specialized for individual subregions of CA1. This is
an example of a point-to-point (p-p) connection. Within the hippocampus, granule cells provide input, via mossy fibers to the mossy
cells of the dentate and to CA3 cells. CA3 cells make feedback connections to themselves and to the mossy cells of the dentate. These,
in turn, provide excitatory input to granule cells in the inner third of the granule cell dendritic tree. (For B/W version, see page 616 in
the volume.)

Plate 36.1. (A) Two representative granule cells filled with 5,6 carboxyfluorescein from the dentate gyrus of a 24-month-old rat. (B)
Histograms showing the numbers of carboxyfluorescein injections that resulted in single, double, triple, or greater numbers of granule
cells filled with dye. Aged rats showed significantly increased electrotonic coupling compared to young animals. This may account for
the increased excitability of old granule cells. Adapted with modifications from Barnes et al. (1987). (For B/W version, see page 665 in
the volume.)
Plate 36.5. Image of granule cells (blue) following in situ hybridization using a c-RNA probe for the IEG Arc. (A) Arc positive cells
(shown in red) in the granule cell layer from the upper blade of a young rat following two 5 min (separated by a 20 min rest period in
the home cage) episodes of spatial exploratory behavior. (B) Granule cells from a caged control animal of the same age, that did not
perform the exploratory behavior, are shown, illustrating baseline or negligible expression of Arc mRNA. (For B/W version, see page
672 in the volume.)
Plate 36.6. (A) Examples of gadolinium-induced change in MRI signal, (a measure of cerebral blood flow volume, CBV, that is a
correlate of brain metabolism) from a young and old rhesus monkey. Warm colors indicate higher CBV values. The white circle,
identified in precontrast images, overlies the dentate gyrus. (B) CBV measurements plotted vs. age in the dentate gyrus indicates a
significant decline with advancing age. (C) CBV measurements in the dentate gyrus of old monkeys plotted against memory per-
formance on the delayed nonmatching-to-sample test. A significant relationship between CBV and memory performance is observed,
selectively in the old animals. Adapted with modifications from Small et al. (2004). (For B/W version, see page 673 in the volume.)

Plate 36.7. Behavior-induced Arc mRNA expression pattern in young and old rat granule cells of the dentate gyrus. (A) Individual
examples of behaviorally induced Arc RNA. Fluorescence-tagged CY3 Arc RNA containing cells are green, and granule cell nuclei are
shown in red (counterstained with propydium iodide, a nucleic acid stain). (B) The average proportion of Arc-positive neurons (% of
total cells) measured from the three different age groups (9, 15, and 24 months) is shown for the dentate gyrus. Note that the dentate
gyrus of old rats contains a significantly lower proportion of behaviorally induced Arc-positive granule cells compared to young or
middle-aged animals. (For B/W version, see page 674 in the volume.)
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 37

The effects of aging on dentate circuitry and function

Peter R. Patrylo1, and Anne Williamson2

1
Department of Physiology, Southern Illinois University School of Medicine Carbondale, IL 62901, USA
2
Department of Neurosurgery, Yale University School of Medicine New Haven, CT 06520, USA

Abstract: The central nervous system (CNS) undergoes a variety of anatomic, physiologic, and behavioral
changes during aging. One region that has received a great deal of attention is the hippocampal formation
due to the increased incidence of impaired spatial learning and memory with age. The hippocampal
formation is also highly susceptible to Alzheimer’s disease, ischemia/hypoxia, and seizure generation, the
three most common aging-related neurological disorders. While data reveal that the dentate gyrus plays a
key role in hippocampal function and dysfunction, the majority of electrophysiological studies that have
examined the effects of age on the hippocampal formation have focused on CA3 and CA1. We perceive this
to be an oversight and consequently will highlight data in this review which demonstrate an age-related
disruption in dentate circuitry and function, and propose that these changes contribute to the decline in
hippocampal-dependent behavior seen with ‘‘normal’’ aging.

Keywords: dentate gating; mossy fibers; recurrent excitation; IPSP; epileptogenicity; spatial learning and
memory; aging

Dentate filter function function comes from several experimental findings.

‘‘Normal’’ and pathologic hippocampal function
(e.g., spatial learning and memory, and epileptiform
activity, respectively) are both associated with an
increase in excitatory activity between principal cells.
Determining whether ‘‘normal’’ or pathological
function occurs has been proposed to depend, in
part, on the dentate gyrus due to it’s strategic po-
sition within the tri-synaptic circuit of the hippo-
campal formation that allows it to ‘‘filter/gate’’ the
majority of excitatory input entering hippocampal
region CA3, a region highly susceptible to epilepti-
form activity (Johnson and Brown, 1984; Miles and
Wong, 1987; Christian and Dudek, 1988). Evidence
that the dentate gyrus subserves a gating/filter
Corresponding author. Tel.: +1618 453 6743;
Fax: +1618 453 1517; E-mail: ppatrylo@siumed.edu
First, 2-deoxyglucose studies and electrophysiolog-
ical recordings have shown that seizures are unlikely
to invade the hippocampus unless epileptiform ac-
tivity is generated within the dentate gyrus itself
(Heinemann et al., 1992; Lothman et al., 1992).
Second, low frequency orthodromic activation of
dentate granule cells normally suppresses CA3 py-
ramidal cell activity (Bragin et al., 1995; Penttonen
et al., 1997) since granule cell axons preferentially
form synapses with inhibitory interneurons in CA3
(Acsady et al., 1998). If granule cell activity is in-
creased, however, glutamatergic excitation can be
evoked in CA3 due to greater frequency facilitation
at synapses between granule cell axons (mossy
fibers) and pyramidal neurons (Henze et al., 2002;
Lawrence and McBain, 2003). Third, several groups
have shown that, if the intrinsic properties or
connectivity of the dentate gyrus are altered

DOI: 10.1016/S0079-6123(07)63037-4 679

680
experimentally or naturally (e.g., insult or genetic
variability), hippocampal-dependent function is
affected adversely including, learning and memory
(Lipp et al., 1984; Crusio et al., 1987; McNaughton
et al., 1989; Nanry et al., 1989; Bernasconi-Guastalla
et al., 1994; Schuster et al., 1997), synaptic plasticity
(Beck et al., 2000), and seizure susceptibility (Grimes
et al., 1990; Cronin et al., 1992; Grecksch et al.,
1995; Wuarin and Dudek, 1996; Patrylo and Dudek,
1998).
The cellular and circuit properties that bestow
the dentate gyrus with the capacity to ‘‘filter/gate’’
excitatory activity are unknown. However this re-
gion does exhibit several unique physiological
properties that are likely to contribute to this
function. Among these properties are: (1) the prin-
cipal cells (granule cells) having a relatively hyper-
polarized resting membrane potential and strong
spike frequency adaptation (Fricke and Prince,
1984; Staley et al., 1992); (2) the capacity of dent-
ate granule cells to fire normally in single spike
mode following disinhibition versus a graded burst
discharge (Schwartzkroin and Prince, 1978; Fricke
and Prince, 1984; Lynch et al., 2000); (3) the pres-
ence of strong feedfoward and feedback inhibitory
circuits (Andersen, 1975; Freund and Buzsáki,
1996); (4) a paucity of recurrent excitatory inter-
connections (Schwartzkroin and Prince, 1978;
Fricke and Prince, 1984; Wuarin and Dudek,
1996; Nadler, 2003); and (5) the dynamic proper-
ties of mossy fiber — CA3 synaptic transmission
(e.g., Lawrence and McBain, 2003).
Aging and dentate filter function
Spatial learning and memory, synaptic plasticity,
and seizure susceptibility are all affected adversely
with age. Moreover, available data suggest that
several of the cellular and circuit properties of the
dentate gyrus are altered during aging (see below)
and consequently it raises the question whether
dentate filter function is also comprised. To begin
to determine the effects of age on dentate filter
function, Patrylo et al. (2007) examined field po-
tential activity elicited in hippocampal region CA3
with dentate molecular layer stimulation in brain
slices from adult (2–11 months) and aged Fisher
344 (21–32 months) rats. The Fischer 344 rat is one
of the most commonly used strain of rodents in
neurobiology of aging research, and the ages used
reflects the convention of classifying rodents as
aged when approximately 50% of their cohorts
have perished. To ensure that the activity elicited
in CA3 was disynaptic in nature, and thus a re-
flection of dentate filtering, only responses with a
delay Z2 ms were examined (monosynaptic inter-
connections normally exhibit a delay r2 ms).
These experiments revealed that during a train of
5 Hz molecular layer stimulation (theta frequency)
hyperexcitable activity (multiple population
spikes) occurs in area CA3 in approximately
30% of aged slices (Fig. 1C). These slices also fre-
quently exhibited spontaneous interictal-like ac-
tivity in CA3 following termination of the train. In
contrast, 100% of adult slices and the remaining
70% of aged slices exhibited either suppression or
no change in the amplitude of the activity evoked
during the train and no subsequent epileptiform
activity. Further, this compromise in dentate filter
function appeared to be associated with an im-
paired capacity of the dentate gyrus to respond to
afferent input, since 5 Hz molecular layer stimula-
tion evoked hyperexcitable activity in the dentate
gyrus in approximately the same percentage of
aged slices as were noted to generate hyperexcit-
able activity in CA3 (Patrylo et al., 2007). These
data were also proposed to reflect an aging-related
compromise in inhibitory activity in CA3 since
modeling studies indicate that epileptiform activity
is generated in hippocampal region CA3 if granule
cell activity is increased and inhibitory tone is
compromised (Lawrence and McBain, 2003).
This aging-related breakdown of dentate filter
function poses two questions. (1) What are the un-
derlying mechanisms for this decline in function? (2)
Does this disruption in dentate function contribute
to aging-related hippocampal dysfunction (e.g., im-
paired spatial learning and memory)? While, the
answers to these questions are currently unclear, in
the subsequent sections we will highlight aging-re-
lated changes that are likely to contribute to this
decline in dentate filter function, and propose that
this change in function contributes to the decline in
spatial learning and memory and enhanced suscep-
tibility for complex partial seizures during aging.
 681

Fig. 1. Dentate filter function is impaired in a subpopulation of aged rodents. (A) Schematic diagram illustrating the position of the
recording and stimulating electrodes used to assess dentate filter function. Note that the recordings were performed in CA3. (B) The
experimental paradigm illustrated schematically. (C) In adult slices, 5 Hz molecular layer stimulation elicited minimal change in the
evoked response, recorded from CA3. In contrast, similar stimulation could elicit hyperexcitability in approximately 30% of aged
slices. (D) The majority of aged slices exhibiting multiple population spikes during 5 Hz molecular layer stimulation developed
spontaneous epileptiform activity subsequently.

Potential mechanisms contributing to the aging-
related decline in dentate filter function
Numerous changes have been reported to occur in
the dentate gyrus during aging and a partial listing
is illustrated in Table 1. To summarize briefly, ag-
ing is associated with: (1) a change in synaptic
connectivity (e.g., Geinisman et al., 1992; Patrylo
et al., 2007); (2) an increase in dye coupling be-
tween granule cells (Barnes et al., 1987); (3) a
682
change in receptor and channel properties (e.g.,
Gutierrez et al., 1996; Magnusson et al., 2006); (4)
a compromise in synaptic plasticity (e.g., Barnes
and McNaughton, 1980b; de Toledo-Morrell
et al., 1988; Barnes, 2000; Barnes et al., 2000);
(5) a decrease in extracellular volume (Sykova
et al., 1998); and (6) a change in the levels of
growth factors that can affect neuronal survival,
metabolism, and activity (e.g., Katoh-Semba et al.,
1998; Sonntag et al., 1999). While all of these
changes would be expected to alter dentate net-
work activity and thus filter function, it is beyond
the scope of this article to provide a detailed re-
view of each change. Consequently, we will focus
on data demonstrating an aging-related change
in local inhibitory and excitatory circuitry (GAB-
Aergic and glutamatergic respectively) since sim-
ilar alterations have been reported in models of
epilepsy and have been associated with a break-
down of dentate filter function (e.g., Behr et al.,
1998; Finnerty et al., 2001; Nadler, 2003).
GABAergic circuitry
Interneurons help regulate principal cell excitabil-
ity, receptive field size, and plasticity, and conse-
quently play a critical role in information
processing as well as pathophysiology (Freund
and Buzsáki, 1996; Paulsen and Moser, 1998). In
the dentate gyrus a tremendous diversity of inter-
neurons are found (Freund and Buzsáki, 1996)
that can be divided into two main categories based
on the distribution of their axonal arbor and sub-
sequently physiological function. The first group
exhibits an axonal arbor that forms synapses with
the somata, initial dendritic tree, and axon hillock
of granule cells and are involved with regulating
the capacity of granule cells to discharge. The sec-
ond group exhibits axonal projections that form
synapses with the distal segments of granule cell
dendrites and subsequently help regulate afferent
input (i.e., perforant path input). These two types
of interneurons play distinct roles in ‘‘normal’’
hippocampal function (Austin et al., 1989; Moser,
1996) and are affected differentially in specific ne-
uropathological conditions (e.g., Kobayashi and
Buckmaster, 2003).
During aging, the inhibitory system in the dent-
ate gyrus exhibits a variety of anatomic and bio-
chemical changes that suggest a compromise in
function. Among these changes are: (1) a decrease
in glutamic acid decarboxylase (GAD; Shetty and
Turner, 1998; Vela et al., 2003; Ling et al., 2005)
and extracellular GABA levels (Almaguer-Melian
et al., 2005); (2) an increase in GABA transami-
nase (Lumeng and Li, 1974; Kugler and Baier,
1992; Hwang et al., 2004); and (3) a decrease in the
number of neurons immunopositive for GAD,
and/or other peptides/neuromodulators that co-
localize in inhibitory neurons (Shetty and Turner,
1998; Cadacio et al., 2003; Vela et al., 2003;
Stanley and Shetty, 2004). It is important to note
however that several other changes occur during
aging that have been suggested to be compensa-
tory. For example, an increase in the levels of the
GABA a1 subunit has been noted in the hippo-
campus and dentate gyrus of aged rodents and has
been proposed to be an adaptive response to the
decrease in GAD levels (e.g., Gutierrez et al., 1996)
since incorporation of the a1 subunit into a GABA
receptor/channel complex results in a current that
has a prolonged decay constant (e.g., Fisher,
2004). Examination of the anatomical data on
the distribution and number of GAD immunore-
active neurons suggest that not all inhibitory cir-
cuits are affected equally with age. Specifically,
Shetty and Turner (1998) reported that the
number of GAD-immunoreactive neurons within
and subjacent to the granule cell layer is preserved
with age, while GAD immunopositive neuron
number in the hilus is decreased. Based on the
characteristic laminar location of specific types of
interneurons (e.g., Freund and Buzsaki, 1996),
these findings suggest that the basket and chande-
lier cells, those primarily responsible for somatic
and axonal inhibition, are preserved with age and
subsequently suggest that feedback inhibition
should be preserved during aging. In agreement
with this suggestion, the number of parvalbumin-
immunoreactive neurons within and adjacent to
the granule cell layer is preserved with age (Shetty
and Turner, 1998) and basket and chandelier cells
co-localize this calcium binding protein (e.g.,
Soriano et al., 1990; Ribak, 1992; Ribak et al.,
1993; Freund and Buzsáki, 1996). Further,
Table 1. Aging-related changes in the dentate gyrus

Findings Species Exemplary reference

Glutamatergic NT
Entorhinal Decreased axospinous synapse number in the MML Rat – F344 Geinisman et al. (1992)
Decreased ratio of Timm staining volume in the MML/ Rat – Long Evans Rapp et al. (1999)
ML
Decreased pre-synaptic volley Rat – F344 Barnes and McNaughton (1980a)
Assoc/Com Decreased axospinous synapse number in the IML Rat – F344 Geinisman et al. (1992)
Expanded zone of Timm staining in the IML Rat – Long Evans Rapp et al. (1999)
Mossy fibers Elaboration of Timm stained mossy fibers into the GCL Guinea pig Wolfer and Lipp (1995)
and IML — can decrease slightly in very old subjects
Elaboration of Timm stained mossy fibers into the IML Human Cassell and Brown (1984)
AMPA receptors Decreased GluR1 and GluR2 mRNA levels Rat – Wistar Pagliusi et al. (1994)
No significant change in [3H] AMPA binding density Rat – Long Evans Nicolle et al. (1996)
Decreased [3H] AMPA binding density Rat – SD Wenk and Barnes (2000)
KA receptors Decreased [3H]KA binding density in the hilus Rat – Long Evans Nagahara et al. (1993)
Expanded zone, but not density, of [3H]KA binding in the Rat – Long Evans Nagahara et al. (1993); Nicolle et
IML al. (1996)
NMDA receptors Non-significant trend for decreased [3H]NMDA receptor Rat – Long Evans Nicolle et al. (1996)
binding in ML
No significant change in [3H] glutamate binding (AP-5 Rat – SD Wenk and Barnes (2000)
subtraction)
Non-significant trend for decreased NR2B mRNA Rhesus monkey Bai et al. (2004)
Decreased NR2B mRNA; decreased NR1 mRNA in the Mice – C57Bl/6 Magnusson (2000)
ventral DG
Decreased [3H] glutamate and [3H] CPP binding density Mice – C57Bl/6 Magnusson (2000)
Decreased NR2B, but not NR2A and NR1, binding Mice – C57BL/6J Magnusson et al. (2006)
density; NR2B decreased preferentially in the intermediate
DG
Decreased NR1-IR Rhesus monkey Gazzaley et al. (1996)
Miscellaneous Increased probability for prolonged EPSPs in BIC Rat – F344 Patrylo et al. (2007)

683
 684
Table 1 (continued )

Findings Species Exemplary reference

GABAergic NT
Cell number Decreased number of GAD-immunopositive neurons Rat – F344 Shetty and Turner, (1998)
Synapse number Decreased axodendritic synapse number in OML (on Rhesus monkey Tigges et al. (1995)
shaft)
Receptors Preserved a1 receptor subunit IR Rhesus monkey Rissman et al. (2006)
Binding Increased a1 (but not b2, b3, and g2) mRNA levels and IR Rat – F344 & SD Gutierrez et al. (1996)
Increased [3H] zolpidem binding Rat – Wistar Ruano et al. (1995)
Electrophysiology Preserved paired-pulse inhibition and polysynaptic evoked Rat – F344 Barnes (1979); Patrylo et al. (2007)
IPSPs (conductances and reversal potentials)
Decreased sIPSP frequency Rat – F344 Patrylo et al. (2007)
Increased mIPSC decay time constant Rhesus monkey Luebke and Rosene (2003)
Increased benzodiazepine potentiation of mIPSC decay Rhesus monkey Luebke and Rosene (2003)
constant
Plasticity
Decreased LTP maintenance Rat – F344 Barnes (1979); Barnes and
McNaughton (1980b)
Decreased capacity for LTP induction at low, but not Rat – F344 Barnes (1979); Barnes et al. (2000);
high, stimulus intensities Orr et al. (2001)
Decreased capacity for perforant path — kindling Rat – F344 de Toledo-Morell et al. (1984)
Decreased BDNF-induced LTP Rat – Wistar Gooney et al. (2004)
Ion channels and intrinsic
properties
Ca2+ Decreased calcium currents (e.g., L-type) Rat – F344 Reynolds and Carlen (1989)
(however see Herman et al., 1998;
Thibault et al., 2001, regarding
CA1)
Intrinsic membrane No change in RMP, Rin, spike width, AP amplitude, and Rat – F344 Barnes and McNaughton (1980a);
properties time constant Niesen et al. (1988); Patrylo et al.
(2007)
No change in AHP and spike frequency adaptation Rat – F344 Niesen et al. (1988)
Other NTs and neuropeptides
NPY Decreased number of NPY immunopositive neurons Rat – F344 Hattiangady et al. (2005); Cadacio
et al. (2003)
Somatostatin Decreased somatostatin mRNA levels and IR Rat – Wistar Vela et al. (2003)
ACh Decreased vesicular ACh transporter IR Rhesus monkey Calhoun et al. (2004)
Decreased ChAT IR Rat – Wistar Lukoyanov et al. (1999)
Decreased amplitude of the slow cholinergic EPSP Rat – F344 Shen and Barnes (1996)
Decreased M2 and M1 receptor binding; trend for Rat – F344 Tayebati et al. (2002)
increased M3 receptor binding
Increased [3H]AF-DX 384 binding (increased M2 levels) Rat – Long Evans Aubert et al. (1995)
Decreased nAChRa4 IR in polymorphic region Mouse – CBA/J Rogers et al. (1998)
Growth factors
BDNF Decreased BDNF activity and IR Rat – F344 Hattiangady et al. (2005)
Decreased BDNF IR Rhesus monkey Hayashi et al. (2001)
Increased BDNF IR in the hilus Rat – SD Katoh-Semba et al. (1998)
Insulin and IGF No change in binding density for the IGF1, IGF2, and Rat – Long Evans Dore et al. (1997)
insulin receptors
NGF and NT3 Unknown whether changes are specifically seen within the
dentate gyrus
Neurogenesis
Decreased rate of basal neurogenesis Rat – F344 Kuhn et al. (1996); Merrill et al.
(2003)
Maturation of newly generated neurons retarded Rat – F344 Rao et al. (2005)
Miscellaneous
Decreased extracellular volume Rat – Wistar Sykova et al. (1998)
Increased dye coupling Rat – F344 Barnes et al. (1987)

ACh, acetylcholine; AF-DX, 5,11-dihydro-11-[[(2-(2-[(dipropylamino) methyl]-1-piperidinyl) ethyl) amino] carbonyl]-6H-pyrido [2, 3-b] [1,4]-benzodiazepin-6-one; AHP, after
hyperpolarization; AMPA, alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; AP, action potential; AP-5, amino phosphonovalerate; Assoc/Com, associational/commissural; BDNF,
brain-derived neurotrophic factor; BIC, bicuculline methiodide; CPP, [(7)-2-carboxypiperazin-4-yl] propyl-1-phosphonic acid; DG, dentate gyrus; EPSP, excitatory post-synaptic potential;
F344, Fischer 344; GAD, glutamic acid decarboxylase; GCL, granule cell layer; IGF1, insulin-like growth factor-1; IGF2, insulin-like growth factor-2; IML, inner molecular layer; IPSP,
inhibitory post-synaptic potential; IR, immunoreactivity; KA, kainate; LTP, long term potentiation; mIPSC, miniature inhibitory post-synaptic current; ML, molecular layer; MML, middle
molecular layer; NGF, nerve growth factor; NMDA, n-methyl-d-aspartate; NPY, neuropeptide Y; NT, neurotransmission; NT3, neurotrophin-3; OML, outer molecular layer; Rin, input
resistance; RMP, resting membrane potential; SD, Sprague-Dawleys; IPSP, Spontaneous IPSP.

685
686
electrophysiological data from hippocampal slices
from aged rodents have shown that paired-pulse
inhibition is preserved in the aged dentate gyrus
(Barnes 1979; Patrylo et al., 2007), and whole cell
current clamp data reveal no significant change in
the reversal potentials and conductances of fast
and slow inhibitory post-synaptic potentials
(IPSPs) evoked in granule cells from aged versus
adult rodents (Patrylo et al., 2007).
Regarding the aging-related decrease in number
of GAD immunopositive neurons within the hilus,
anatomic and electrophysiological data suggest that
these interneurons are likely to be those involved
with inhibiting the distal dendrites of granule
cells and subsequently modulating afferent input.
Specifically, anatomic studies have shown that the
number of somatostatin (SOM) and neuropeptide-Y
(NPY) immunoreactive interneurons are decreased
within the hilus with age (Cadacio et al., 2003; Vela
et al., 2003) and these types of interneurons are
known to have an axonal arborization that targets
the distal dendrites of granule cells (e.g., Leranth
et al., 1990; Freund and Buzsáki, 1996; Katona
et al., 1999; Maccaferri et al., 2000). Further, whole
cell patch recordings from hippocampal slices from
adult and aged Fisher 344 rats have shown that the
overall frequency of spontaneous IPSPs (sIPSPs) is
decreased in aged granule cells, although the fre-
quency of sIPSPs with a large amplitude and pro-
longed duration is preserved (Patrylo et al., 2007).
Since large amplitude composite inhibitory events
are believed to arise from input impinging on the
soma and proximal dendrites of granule cells, while
smaller unitary events are believed to result from
more distal input (Soltèsz et al., 1995; Williams et
al., 1998), this aging-related decrease in sIPSP fre-
quency is likely to reflect a reduction of inhibitory
input onto the distal dendrites of granule cells (Pat-
rylo et al., 2007). In vitro studies on hippocampal
slices from aged primates have also noted a trend for
reduced IPSC frequency, although this change did
not reach significance (Luebke and Rosene, 2003).
It is interesting to note that afferent input is
suppressed during theta wave activity (Moser,
1996), while feedback inhibition onto the soma/
axon of granule cells is decreased (Austin et al.,
1989; Moser, 1996). Consequently, it has been
proposed that inhibitory input onto distal
dendrites and presynaptic afferent fibers play a
critical role in dentate filter function and encoding
of new spatial information due to increasing the
signal to noise ratio of incoming cortical input
(Paulsen and Moser, 1998). These findings also
suggest that if such dendritic/afferent inhibition is
compromised, both functions would be affected
adversely. Based on the data presented above, we
propose that inhibitory input onto the distal dend-
rites of granule cells, and potentially afferent ter-
minals, is decreased during aging and results in
abnormal processing of afferent input during spa-
tial exploration, and subsequently decreases the
signal to noise ratio below the level required for
transmitting new spatial information. Experiments
examining dendritic inhibition in behaviorally
characterized aged animals are required to address
this possibility, however.
Whether the decrease in the number of GAD-IR
neurons within the hilus of the dentate gyrus of aged
rodents reflects a loss of SOM and NPY containing
interneurons, a change in their properties (decreased
protein levels), or both, is unclear with currently
available data supporting either possibility. For
example, studies examining total hilar neuron
number reveal that there is no aging-related change
(Rasmussen et al., 1996), while studies examining
the number of NPY immunoreactive neurons sug-
gest that a relatively small subpopulation of neurons
may be lost (Cadacio et al., 2003; Vela et al., 2003).
It should be noted, however, that attention to ac-
curate immunohistochemical cell quantification is
critical, since the level of an immunohistochemical
marker may simply decrease below the level re-
quired for visual detection without a loss of the cell
expressing the protein. In this regard, it is interesting
to note that Stanley and Shetty (2004) have reported
that the loss of GAD-IR neurons in hippocampal
regions CA3 and CA1 does not correspond with a
decrease in total neuron number, and that several
groups have shown that GAD, SOM, and NPY
protein expression can be regulated by afferent
input and neurotrophic levels (Akhtar and Land,
1991; Hyman et al., 1994; Carnahan and Nawa,
1995; Patel et al., 1995; Aamodt et al., 2000;
Marty, 2000; Reibel et al., 2000; de Almeida et al.,
2002) which may be altered with age (e.g., Geinis-
man et al., 1992; Katoh-Semba et al., 1998; Patrylo

et al., 2007). Consequently, additional studies are
required to determine whether aging is associated
with a loss of interneurons in the dentate gyrus, or a
change in protein expression and/or function.
Regardless, these aging-related changes in in-
hibitory efficacy are likely to affect granule cell
activity and synchronization, and consequently
their capacity to respond to afferent input. If this
decline in inhibitory efficacy is further coupled
with an enhancement of NMDA-receptor medi-
ated activity, or an increase in recurrent excitatory
interconnections, one would predict a breakdown
of the dentate gate (e.g., Lynch et al., 2000; Finn-
erty et al., 2001; Nadler, 2003). During aging it is
unlikely that an enhancement of NMDA receptor-
mediated activity in the dentate gyrus contributes
to the decline in filter function, since decreases in
the protein levels of certain NMDA receptor sub-
units (Gazzaley et al., 1996; Nicolle et al., 1996;
Magnusson, 2000; Bai et al., 2004; Magnusson et
al., 2006), the strength of NMDA-receptor medi-
ated activity (Rao et al., 1994), and NMDA-re-
ceptor mediated plasticity have been reported with
age (Barnes and McNaughton, 1980b; de Toledo-
Morrell et al., 1988; Barnes et al., 2000). As for a
change (increase) in local excitatory interconnec-
tivity contributing to the compromise in dentate
filter function, we will review data in the next sec-
tion that address this possibility.
Glutamatergic circuitry
In the early 1990s, Geinisman et al. (1992) re-
ported that axospinous synapses number is de-
creased in the molecular layer (middle and inner
molecular layers) of the dentate gyrus in aged ro-
dents and is associated with the cognitive status of
the animal (loss occurred in cognitively impaired
subjects). This loss of axospinous synapses is be-
lieved to reflect a decrease in entorhinal and asso-
ciational/commissural input onto granule cell
dendrites, because electrophysiological studies in-
dicate that the amplitude of the pre-synaptic volley
and resulting EPSP are decreased in the dentate
gyrus of aged rats (e.g., Barnes and McNaughton,
1980a, b). Further, a loss of presynaptic neurons
(layer II neurons in the entorhinal cortex) is
687
unlikely to contribute to this change since neuron
number in the entorhinal cortex is preserved in
aged animals (Merrill et al., 2001; Rapp et al.,
2002). Subsequent studies using synaptophysin
immunohistochemistry (Smith et al., 2000) have
provided corroboratory findings. Specifically,
Rapp and colleagues noted that there is a decrease
in the level of synaptophysin immunostaining in
the molecular layer of the dentate gyrus, as well as
stratum lacunosum moleculare of CA3 in aged
rodents with cognitive impairment. Although this
study did not determine whether this change in
synaptophysin immunohistochemistry reflects a
loss or modification of synapses, the results do re-
veal an aging-related change in a pre-synaptic
marker that is found preferentially in cognitively
impaired rodents. Based on these results, Rapp
and colleagues proposed that the decline in spatial
learning and memory during aging results from the
disruption in information flow from the entorhinal
cortex to the hippocampus. It should be noted
however, that the decrease in axospinous synapse
number in the molecular layer of aged rodents
could also reflect a decrease in excitatory input
onto inhibitory interneurons with dendritic spines
(Soriano and Frotscher, 1993; Freund and Buz-
sáki, 1996; Mott et al., 1997), although one could
argue that these are few in number compared to
granule cell dendrites. Further, a decrease of in-
hibitory input onto granule cell dendritic spines is
unlikely to be a significant contributor because
GABAergic synapses are not typically axospinous
(Halasy and Somogyi, 1993; Soriano and Frotsc-
her, 1993; Freund and Buzsáki, 1996; Katona et
al., 1999).
In the adult CNS, a loss of entorhinal input into
the dentate gyrus (e.g., unilateral lesion) results in
a transient compromise in cognitive function and
an increased propensity for epileptiform activity
(e.g., Steward, 1982; Kelley and Steward, 1996).
These effects eventually reverse, however, due to
reorganization of the remaining excitatory circuits,
in particular the temporodentate pathway arising
from the contralateral entorhinal cortex (e.g.,
Steward, 1982; Reeves and Smith, 1987). Asso-
ciational/commissural fibers and the mossy fibers
have also been shown to reorganize following a
loss of entorhinal input (e.g., Lynch et al., 1976;
688
Steward et al., 1976; Steward, 1982, 1992; West
and Dewey, 1984; Reeves and Smith, 1987). Dur-
ing aging, the loss of entorhinal input is bilateral in
nature, and therefore suggests that the capacity for
restorative reorganization is likely to be reduced.
This raises the question whether there is evidence
for reorganization of the associational/commis-
sural fibers and/or mossy fibers in the aged dentate
gyrus. Reorganization of these pathways would
increase recurrent excitatory interconnections
within the hippocampus and thus potentially affect
dentate filter function as has been suggested in
models of temporal lobe epilepsy (e.g., Behr et al.,
1998; Finnerty et al., 2001; Nadler, 2003).
Electrophysiological studies suggest that local
excitatory circuits do reorganize within the dentate
gyrus with age. Current source density (CSD) stud-
ies have shown that antidromic stimulation elicits
an additional peak of net inward current in the
granule cell and inner molecular layer of the dentate
gyrus in slices from aged, but not adult, rodents
(Barnes and McNaughton, 1983). While the reason
for this additional current sink is unknown, it is
consistent with the presence of additional synaptic
input onto the soma and proximal dendrites of
granule cells. Furthermore, whole cell patch clamp
recordings from transverse hippocampal slices have
demonstrated that granule cells from aged rodents
have an increased propensity for generating spon-
taneous, prolonged, large amplitude EPSPs follow-
ing exposure to the GABAA receptor antagonist
bicuculline methiodide (Patrylo et al., 2007). These
events resemble the giant synaptic potentials that
underlie paroxysmal depolarizing shifts, and sug-
gest that aberrant excitatory circuits are present in
the aged dentate gyrus, since granule cells do not
typically exhibit reverberatory activity following
disinhibition (Schwartzkroin and Prince, 1978;
Fricke and Prince, 1984; Wuarin and Dudek, 1996).
Although the cellular player(s) that contribute
to this aberrant excitatory circuit are unclear, an-
atomic data suggest several possibilities. First, the
volume of [3H] kainate binding (Nicolle et al.,
1996) and Timm staining (Rapp et al., 1999)
within the inner molecular layer relative to the
middle/outer molecular layers is increased in the
aged dentate gyrus, and thus it could reflect an
expansion of the associational/commissural fibers.
With regard to the CSD studies (Barnes and
McNaughton, 1983) however, this expansion is
unlikely to produce the novel net inward current in
the granule cell and inner molecular layer of the
aged dentate gyrus since even though granule
cells form a feedback circuit with hilar mossy
cells and CA3 pyramidal neurons in the in vivo
dentate gyrus (Ishizuka et al., 1990; Li et al., 1994;
Buckmaster et al., 1996; Wenzel et al., 1997) the
majority of these interconnections are severed dur-
ing the preparation of transverse hippocampal
slices (Ishizuka et al., 1990; however see Scharf-
man, 1994; Buckmaster et al., 1996; Wenzel et al.,
1997). Thus, unless the distribution of the associ-
ation/commissural fibers is altered along the lon-
gitudinal axis of the hippocampus during aging, it
is unlikely that this reorganization accounts for the
increase in excitatory activity in hippocampal
slices from aged rodents. Moreover, Geinisman
et al. (1992) suggest that there is a decrease in
associational/commissural — granule cell axospi-
nous synapses within the inner molecular layer. An
alternate explanation for the net inward current in
the granule cell and inner molecular layer is that
the distribution of mossy fibers changes with age.
In this regard, studies that have labeled the mossy
fibers selectively in tissue from humans and ro-
dents using the Timm stain have shown that mossy
fibers can extend up into the granule cell and inner
molecular layers in the transverse plane of the
hippocampus in aged subjects (Cassell and Brown,
1984; Wolfer and Lipp, 1995). While this distri-
bution pattern of mossy fibers corresponds with
the location of the novel net current flux in the
aged dentate gyrus (Barnes and McNaughton,
1983), no data are currently available to verify that
they lead to increased granule cell to granule cell
synapses within the aged dentate gyrus.
Recent data suggest an additional mechanism
that could contribute to the aberrant excitatory
interconnections in the aged dentate gyrus (Rao et
al., 2005). Specifically, it was demonstrated that
newly generated granule cells in middle-aged and
aged rodents exhibit a slowing of developmental
maturation and an increased incidence of basal
dendrites relative to granule cells born in younger
animals. While it is unclear whether these basal
dendrites persist in the newly generated granule

cells, it is interesting to note that GABAA recep-
tor-mediated inhibition plays a critical role in in-
tegrating newly generated granule cells into
dentate circuitry (Ge et al., 2006) and inhibitory
tone is disrupted during aging (see above). Several
groups have reported basal dendrites in animal
models of temporal lobe epilepsy (Spigelman et al.,
1998; Buckmaster and Dudek, 1999; Ribak et al.,
2000) and have provided evidence that these ab-
errant processes can create anomalous excitatory
circuits capable of positive feedback (e.g., Ribak
et al., 2000).
In summary, the efficacy of the inhibitory sys-
tem is compromised within the dentate gyrus dur-
ing aging, and excitatory interconnections appear
to be increased. This disruption in the balance of
inhibition relative to excitation could impair dent-
ate filter function, as has been shown in animal
models of epilepsy that exhibit similar, although
more severe, changes in dentate circuitry.
The dentate gyrus, spatial learning and memory,
and aging
One of the most common aging-related neurological
dysfunctions is a decline in spatial learning and
memory, which has been reported to occur in
humans, non-human primates, and rodents (Barnes
1979; Gage et al., 1984; de Toledo-Morrell et al.,
1984; Barnes and McNaughton, 1985; Gallagher
and Pelleymonter, 1988; Uttl and Graf, 1993; Rapp
and Gallagher, 1996; Wilkniss et al., 1997). While
the dentate gyrus exhibits several physiological
characteristics that are believed to be critical for
spatial learning and memory, including burst dis-
charges during theta (Rose et al., 1983; Buzsaki and
Czeh, 1992), phase precession (Skaggs et al., 1996),
and reactivation during sleep (Shen et al., 1998), and
data indicate that alternations of dentate intrinsic
properties or connectivity can have detrimental
effects on learning and memory, it is not entirely
clear whether aging-related changes in dentate cir-
cuitry and function contribute to the decline in cog-
nitive function seen with age.
There is good reason to define the potential role
that the dentate gyrus plays in aging-related cogni-
tive decline, because recent data suggest that aging
689
is not only associated with a decline in memory re-
tention, but also a reduced capacity to encode new
information rapidly (Tanila et al., 1997; Rosenzweig
and Barnes, 2003; Wilson et al., 2005), and encod-
ing depends on the entorhinal–dentate gyrus–CA3
pathway (e.g., Lee and Kesner, 2004; Jerman et al.,
2006; Rolls and Kesner, 2006). Recent data reveal
that aged rodents with a decreased ability to encode
new spatial information exhibit an increase in spon-
taneous activity in CA3 (Wilson et al., 2005). It has
been suggested that this increase in basal activity in
CA3 decreases the signal-to-noise ratio by increas-
ing the ‘‘noise’’, and thus obscures the influx of new
pertinent spatial information. The hyperexcitability
observed in CA3 in aged rodents during and after
5 Hz perforant path stimulation (Patrylo et al.,
2007) could be an in vitro equivalent of the increase
in basal CA3 activity reported in vivo by Wilson et
al. (2005). Consequently, we postulate that the ag-
ing-related disruption in dentate gating, an effect
that occurs in a fraction of aged subjects, contrib-
utes to a decreased ability to encode new informa-
tion and thus the compromise in spatial learning
and memory seen in a subpopulation of aged sub-
jects. This increase in basal activity could also affect
cognitive function negatively by changing the ca-
pacity for synaptic plasticity, due to a metaplastic
shift in network dynamics (Abraham and Bear,
1996).
The dentate gyrus, seizure susceptibility, and aging
Epidemiologic studies indicate that aging is also
associated with an increase in the incidence and
prevalence for seizure disorders (Loiseau et al.,
1990; Hauser et al., 1991, 1996; Hauser, 1992). In-
deed, seizure disorders are the third most common
neurological disorder seen in the elderly, with up to
5% of individuals Z65 years of age affected (Tallis
et al., 1991). The most common type of seizure
observed in this age group is complex-partial in
nature (Hauser, 1997; Rowan et al., 2005), and it is
well established that limbic structures (i.e., hippo-
campus) are frequently involved. Furthermore,
50% of newly identified seizure disorders in the
elderly have no identifiable antecedent (i.e., are
cryptogenic), which has led to the suggestion that
690
aging itself may be epileptogenic (Ng et al., 1985;
Hauser et al., 1996; Hauser, 1997).
Data from laboratory animals support and ex-
tend these clinical observations with an enhanced
susceptibility for hippocampal-dependent seizures
seen in aged rodents. For example, aged rodents
exhibit longer duration afterdischarges during
kindling, and faster seizure propagation to the cont-
ralateral hemisphere (Chiba et al., 1992). The rate
of kindling is slower in aged rodents however
(de Toledo-Morrell et al., 1984; Chiba et al.,
1992), although this is likely to reflect a decreased
capacity for plasticity during aging, rather than a
decreased susceptibility to seizures, since the pro-
gression of kindling is dependent on NMDA recep-
tor-mediated plasticity (Mody et al., 1988; Sayin
et al., 1999) and decreases in NMDA receptor-
mediated properties occur with age (e.g., Rao et al.,
1994; Barnes et al., 2000; Magnusson et al., 2006).
Aged rodents also demonstrate increased seizure
susceptibility when tested with the convulsant
kainic acid. Systemic treatment with kainic acid
results in exacerbated pre-seizure behavioral man-
ifestations in aged rodents, compared to adults
(e.g., wet-dog shakes), and a decreased latency to
onset for limbic seizures and sustained seizures
(Darbin et al., 2004). Since the dentate gyrus plays
a role in generating wet-dog shakes (Damiano and
Connor, 1984; Frush and McNamara, 1986; Barn-
es and Mitchell, 1990; Grimes et al., 1990; Mitchell
et al., 1990), and regulating limbic seizures (Dash-
eiff and McNamara, 1982; Lothman et al., 1992;
Grecksch et al., 1995), it is plausible that a break-
down of dentate filter function results in an en-
hanced propensity for dentate–CA3 activation and
thus an increased susceptibility to seizures. Sup-
port for this hypothesis comes from two sources.
First, the mossy fiber–CA3 synapse is preferen-
tially susceptible to kainic acid (Ben-Ari and Coss-
art, 2000), and preliminary data suggest that aged
hippocampal slices exhibit a lower threshold for
kainic acid-induced epileptiform activity (Willing-
ham et al., 2006). Second, a similar, although more
severe, disruption of dentate circuitry and gating
has been observed in the dentate gyrus in models
of epilepsy that are associated with hippocampal
activation (e.g., Behr et al., 1998; Finnerty et al.,
2001; Nadler, 2003). Taken together, the data from
kindling and kainic acid models of epilepsy suggest
that aging-related changes in dentate circuitry and
function are likely to contribute to the increased
incidence for complex-partial seizures with age.
Conclusions
In conclusion, extensive changes occur in the dent-
ate gyrus at the structural and functional levels
during aging, and we suggest that one critical con-
sequence is the breakdown in filter function.
Changes in dentate gyrus gating could explain the
increase in hippocampal-dependent seizure suscep-
tibility that occurs during the aging process. These,
and additional effects in associated structures such
as area CA3, may also contribute to the decline in
cognitive function with age. However, the precise
role of the dentate gyrus in age-related functional
changes (i.e., whether it is the cause or the effect of
the increase in seizures susceptibility or cognitive
decline) remains to be addressed completely.
Acknowledgments
Supported by a grant from the NIH (AG00795)
and the Epilepsy Foundation through the gener-
ous support of the CDC (PRP).
References
Aamodt, S.M., Shi, J., Colonnese, M.T., Veras, W. and Cons-
tantine-Paton, M. (2000) Chronic NMDA exposure acceler-
ates development of GABAergic inhibition in the superior
colliculus. J. Neurophysiol., 83: 1580–1591.
Abraham, W.C. and Bear, M.F. (1996) Metaplasticity: the plas-
ticity of synaptic plasticity. Trends Neurosci., 19: 126–130.
Acsady, L., Kamondi, A., Sik, A., Freund, T. and Buzsaki, G.
(1998) GABAergic cells are the major postsynaptic targets of
mossy fibers in the rat hippocampus. J. Neurosci., 18:
3386–3403.
Akhtar, N.D. and Land, P.W. (1991) Activity-dependent reg-
ulation of glutamic acid decarboxylase in the rat barrel cor-
tex: effects of neonatal versus adult sensory deprivation. J.
Comp. Neurol., 307: 200–213.
Almaguer-Melian, W., Cruz-Aguado, R., Riva Cde, L., Kend-
rick, K.M., Frey, J.U. and Bergado, J. (2005) Effect of LTP-
reinforcing paradigms on neurotransmitter release in the
dentate gyrus of young and aged rats. Biochem. Biophys.
Res. Commun., 327: 877–883.

de Almeida, O.M., Gardino, P.F., Loureiro dos Santos, N.E.,
Yamasaki, E.N., de Mello, M.C., Hokoc, J.N. and de Mello,
F.G. (2002) Opposite roles of GABA and excitatory amino
acids on the control of GAD expression in cultured retina
cells. Brain Res., 925: 89–99.
Andersen, P. (1975) Organization of hippocampal neurons and
their interconnections. In: Isaacson R.L. and Pribram K.H.
(Eds.), The Hippocampus, Vol. 1. Plenum Press, New York,
pp. 155–175.
Aubert, I., Rowe, W., Meaney, M.J., Gauthier, S. and Quirion,
R. (1995) Cholinergic markers in aged cognitively impaired
Long-Evans rats. Neuroscience, 67: 277–292.
Austin, K.B., Bronzino, J.D. and Morgane, P.J. (1989) Paired-
pulse facilitation and inhibition in the dentate gyrus is de-
pendent on behavioral state. Exp. Brain Res., 77(3): 594–604.
Bai, L., Hof, P.R., Standaert, D.G., Xing, Y., Nelson, S.E.,
Young, A.B. and Magnusson, K.R. (2004) Changes in the
expression of the NR2B subunit during aging in macaque
monkeys. Neurobiol. Aging, 25: 201–208.
Barnes, C.A. (1979) Memory deficits associated with senes-
cence: a neurophysiological and behavioral study in the rat. J.
Comp. Psychol., 93: 74–104.
Barnes, C.A. (2000) Plasticity in the aging central nervous sys-
tem. Int. Rev. Neurobiol., 45: 339–354.
Barnes, C.A. and McNaughton, B.L. (1980a) Physiological
compensation for loss of afferent synapses in rat hippocam-
pal granule cells during senescence. J. Physiol. Lond., 309:
473–485.
Barnes, C.A. and McNaughton, B.L. (1980b) Spatial memory
and hippocampal synaptic plasticity in middle-aged and se-
nescent rats. In: Stein D. (Ed.), Psychobiology of Aging:
Problems and Perspectives. Elsevier/North-Holland, New
York, pp. 253–272.
Barnes, C.A. and McNaughton, B.L. (1983) Excitability
changes in hippocampal granule cells of senescent rat. In:
Changeux J.P., Glowinski J., Imbert M. and Bloom F.E.
(Eds.), Molecular and Cellular Interactions Underlying
Higher Brain Function, Progress in Brain Research. Elsevi-
er, New York, pp. 445–451.
Barnes, C.A. and McNaughton, B.L. (1985) An age-compar-
ison of the rates of acquisition and forgetting of spatial in-
formation in relation to long-term enhancement of
hippocampal synapses. Behav. Neurosci., 6: 563–571.
Barnes, C.A., Rao, G. and Houston, F.P. (2000) LTP induction
threshold change in old rats at the perforant path — granule
cell synapse. Neurobiol. Aging, 21: 613–620.
Barnes, C.A., Rao, G. and McNaughton, B.L. (1987) Increased
electrotonic coupling in aged rat hippocampus: a possible
mechanism for cellular excitability changes. J. Comp. Ne-
urol., 259: 549–558.
Barnes, M.I. and Mitchell, C.L. (1990) Differential effects of
colchicine lesions of dentate granule cells on wet dog shakes
and seizures elicited by direct hippocampal stimulation.
Physiol. Behav., 48: 131–138.
Beck, H., Goussakov, I.V., Lie, A., Helmstaedter, C. and Elger,
C.E. (2000) Synaptic plasticity in the human dentate gyrus. J.
Neurosci., 20: 7080–7086.
691
Behr, J., Lyson, K.J. and Mody, I. (1998) Enhanced propaga-
tion of epileptiform activity through the kindled dentate
gyrus. J. Neurophysiol., 79: 1726–1732.
Ben-Ari, Y. and Cossart, R. (2000) Kainate, a double agent that
generates seizures: two decades of progress. Trends Neuro-
sci., 23: 580–587.
Bernasconi-Guastalla, S., Wolfer, D.P. and Lipp, H.P. (1994)
Hippocampal mossy fibers and swimming navigation in mice:
correlations with size and left-right asymmetries. Hippocam-
pus, 4: 53–63.
Bragin, A., Jando, G., Nadasdy, Z., van Landeghem, M. and
Buzsaki, G. (1995) Dentate EEG spikes and associated in-
terneuronal population bursts in the hippocampal hilar re-
gion of the rat. J. Neurophysiol., 73: 1691–1705.
Buckmaster, P.S. and Dudek, F.E. (1999) In vivo intracellular
analysis of granule cell axon reorganization in epileptic rats.
J. Neurophysiol., 81: 712–721.
Buckmaster, P.S., Wenzel, H.J., Kunkel, D.D. and Schwa-
rtzkroin, PA. (1996) Axon arbors and synaptic connections
of hippocampal mossy cells in the rat in vivo. J. Comp.
Neurol., 366: 271–292.
Buzsaki, G. and Czeh, G. (1992) Physiological function
of granule cells: a hypothesis. Epilepsy Res. Suppl., 7:
281–290.
Cadacio, C.L., Milner, T.A., Gallagher, M. and Pierce, J.P.
(2003) Hilar neuropeptide Y interneuron loss in the aged rat
hippocampal formation. Exp. Neurol., 183: 147–158.
Calhoun, M.E., Mao, Y., Roberts, J.A. and Rapp, P.R. (2004)
Reduction in hippocampal cholinergic innervation is unre-
lated to recognition memory impairment in aged rhesus
monkeys. J. Comp. Neurol., 475: 238–246.
Carnahan, J. and Nawa, H. (1995) Regulation of neuropeptide
expression in the brain by neurotrophins. Potential role in
vivo. Mol. Neurobiol., 10: 135–149.
Cassell, M.D. and Brown, M.W. (1984) The distribution of
Timm’s stain in the nonsulphide-perfused human hippocam-
pal formation. J. Comp. Neurol., 222: 461–471.
Chiba, S., Muneoka, Y., Sato, Y. and Miyagishi, T. (1992)
Seizure susceptibility in aged rats — pentylenetetrazol-in-
duced seizures and amygdaloid kindling seizures. No To
Shinkei [Brain Nerve], 44: 559–564.
Christian, E.P. and Dudek, F.E. (1988) Characteristics of local
excitatory circuits studied with glutamate microapplication in
the CA3 area of rat hippocampal slices. J. Neurophysiol., 59:
90–109.
Cronin, J., Obenaus, A., Houser, C.R. and Dudek, F.E. (1992)
Electrophysiology of dentate granule cells after kainate-in-
duced synaptic reorganization of the mossy fibers. Brain
Res., 573: 305–310.
Crusio, W.E., Schwegler, H. and Lipp, H.P. (1987) Radial-
maze performance and structural variation of the hippocam-
pus in mice: a correlation with mossy fibre distribution. Brain
Res., 425: 182–185.
Damiano, B.P. and Connor, J.D. (1984) Hippocampal
mediation of shaking behavior induced by electrical stimu-
lation of the perforant path in the rat. Brain Res., 308:
383–386.
692
Darbin, O., Naritoku, D. and Patrylo, P.R. (2004) Aging alters
electroencephalographic and clinical manifestations of kain-
ate-induced status epilepticus. Epilepsia, 45: 1219–1227.
Dasheiff, R.M. and McNamara, J.O. (1982) Intradentate col-
chicine retards the developments of amygdala kindling. Ann.
Neurol., 11: 347–352.
Dore, S., Kar, S., Rowe, W. and Quirion, R. (1997) Distribu-
tion and levels of [125I]IGF-I, [125I]IGF-II and [125I]insulin
receptor binding sites in the hippocampus of aged memory-
unimpaired and -impaired rats. Neuroscience, 80: 1033–1040.
Finnerty, G.T., Whittington, M.A. and Jefferys, J.G. (2001)
Altered dentate filtering during the transition to seizure in the
rat tetanus toxin model of epilepsy. J. Neurophysiol., 86:
2748–2753.
Fisher, J.L. (2004) The alpha 1 and alpha 6 subunit subtypes of
the mammalian GABA(A) receptor confer distinct channel
gating kinetics. J. Physiol., 561: 433–448.
Freund, T.F. and Buzsáki, G. (1996) Interneurons of the hip-
pocampus. Hippocampus, 6: 347–470.
Fricke, R.A. and Prince, D.A. (1984) Electrophysiology of
dentate gyrus granule cells. J. Neurophysiol., 51: 195–209.
Frush, D.P. and McNamara, J.O. (1986) Evidence implicating
dentate granule cells in wet dog shakes produced by kindling
stimulations of entorhinal cortex. Exp. Neurol., 92: 102–113.
Gage, F.H., Bjorklund, A., Stenevi, U., Dunnett, S.B. and
Kelly, P.A.T. (1984) Intrahippocampal septal grafts amelio-
rate learning impairments in aged rats. Science, 225: 533–536.
Gallagher, M. and Pelleymonter, M.A. (1988) Spatial learning
deficits in old rats: a model for memory decline in the aged.
Neurobiol. Aging, 9: 549–556.
Gazzaley, A.H., Siegel, S.J., Kordower, J.H., Mufson, E.J. and
Morrison, J.H. (1996) Circuit-specific alterations of N-me-
thyl-D-aspartate receptor subunit 1 in the dentate gyrus of
aged monkeys. Proc. Natl. Acad. Sci. U.S.A., 93: 3121–3125.
Ge, S., Goh, E.L., Sailor, K.A., Kitabatake, Y., Ming, G.L.
and Song, H. (2006) GABA regulates synaptic integration of
newly generated neurons in the adult brain. Nature, 439:
589–593.
Geinisman, Y., de Toledo-Morrell, L., Morrell, F., Persina, I.S.
and Rossi, M. (1992) Age-related loss of axospinous synapses
formed by two afferent systems in the rat dentate gyrus as
revealed by the unbiased stereological dissector technique.
Hippocampus, 2: 437–444.
Gooney, M., Messaoudi, E., Maher, F.O., Bramham, C.R. and
Lynch, M.A. (2004) BDNF-induced LTP in dentate gyrus is
impaired with age: analysis of changes in cell signaling events.
Neurobiol. Aging, 25: 1323–1331.
Grecksch, G., Ruethrich, H., Bernstein, H.G. and Becker, A.
(1995) PTZ-kindling after colchicine lesion in the dentate
gyrus of the rat hippocampus. Physiol. Behav., 58: 695–698.
Grimes, L.M., Earnhardt, T.S., Mitchell, C.L., Tilson, H.A.
and Hong, J.S. (1990) Granule cells in the ventral, but not
dorsal, dentate gyrus are essential for kainic acid-induced wet
dog shakes. Brain Res., 514: 167–170.
Gutierrez, A., Khan, Z.U., Ruano, D., Miralles, C.P., Vitorica,
J. and De Blas, A.L. (1996) Aging-related subunit expression
changes of the GABAA receptor in the rat hippocampus.
Neuroscience, 74: 341–348.
Halasy, K. and Somogyi, P. (1993) Distribution of GABAergic
synapses and their targets in the dentate gyrus of rat: a
quantitative immunoelectron microscopic analysis. J. Hi-
rnforsch., 34: 299–308.
Hattiangady, B., Rao, M.S., Shetty, G.A. and Shetty, A.K.
(2005) Brain-derived neurotrophic factor, phosphorylated
cyclic AMP response element binding protein and neuropep-
tide Y decline as early as middle age in the dentate gyrus and
CA1 and CA3 subfields of the hippocampus. Exp. Neurol.,
195: 353–371.
Hauser, W.A. (1992) Seizure disorders: the changes with age.
Epilepsia, 33: S6–S14.
Hauser, W.A. (1997) Epidemiology of seizures and epilepsy in
the elderly. In: Rowan A.J. and Ramsay R.E. (Eds.), Seizures
and Epilepsy in the Elderly. Butterworth-Heinemann, Bos-
ton, MA, pp. 7–18.
Hauser, W.A., Annegers, J.F. and Kurland, L.T. (1991) Prev-
alence of epilepsy in Rochester, Minnesota. Epilepsia, 32:
429–445.
Hauser, W.A., Annegers, J.F. and Rocca, W.A. (1996) De-
scriptive epidemiology of epilepsy: contributions of popula-
tion-based studies from Rochester, Minnesota. Mayo Clin.
Proc., 71: 576–586.
Hayashi, M., Mistunaga, F., Ohira, K. and Shimizu, K. (2001)
Changes in BDNF-immunoreactive structures in the hippo-
campal formation of the aged macaque monkey. Brain Res.,
918: 191–196.
Heinemann, U., Beck, H., Dreier, J.P., Ficker, E., Stabel, J. and
Zhang, C.L. (1992) The dentate gyrus as a regulated gate for
the propagation of epileptiform activity. Epilepsy Res. Sup-
pl., 7: 273–280.
Henze, D.A., Wittner, L. and Buzsaki, G. (2002) Single granule
cells reliably discharge targets in the hippocampal CA3 net-
work in vivo. Nat. Neurosci., 5: 790–795.
Herman, J.P., Chen, K.C., Booze, R. and Landfield, P.W.
(1998) Up-regulation of alpha1D Ca2+ channel subunit
mRNA expression in the hippocampus of aged F344 rats.
Neurobiol. Aging, 19: 581–587.
Hwang, I.K., Kim, D.W., Yoo, K.Y., Kim, D.S., Kim, K.S.,
Kang, J.H., Choi, S.Y., Kim, Y.S., Kang, T.C. and Won,
M.H. (2004) Age-related changes of the gamma-amino-
butyric acid transaminase immunoreactivity in the hippo-
campus and dentate gyrus of the Mongolian gerbil. Brain
Res., 1017: 77–84.
Hyman, C., Juhasz, M., Jackson, C. and Wright, P. (1994) Ip
NY, Lindsay RM. Overlapping and distinct actions of the
neurotrophins BDNF, NT-3, and NT-4/5 on cultured do-
paminergic and GABAergic neurons of the ventral me-
sencephalon. J. Neurosci., 14: 335–347.
Ishizuka, N., Weber, J. and Amaral, D.G. (1990) Organization
of intrahippocampal projections originating from CA3 py-
ramidal cells in the rat. J. Comp. Neurol., 295: 580–623.
Jerman, T., Kesner, R.P. and Hunsaker, M.R. (2006) Discon-
nection analysis of CA3 and DG in mediating encoding but

not retrieval in a spatial maze learning task. Learn. Mem., 13:
458–464.
Johnson, D. and Brown, T.H. (1984) Mechanisms of neuronal
burst generation. In: Schwartzkroin P.A. and Wheal H.V.
(Eds.), Electrophysiology of Epilepsy. Academic Press, New
York, pp. 277–301.
Katoh-Semba, R., Semba, R., Takeuchi, I.K. and Kato, K.
(1998) Age-related changes in levels of brain-derived ne-
urotrophic factor in selected brain regions of rats, normal
mice and senescence-accelerated mice: a comparison to those
of nerve growth factor and neurotrophin-3. Neurosci. Res.,
31: 227–234.
Katona, I., Acsady, L. and Freund, T.F. (1999) Postsynaptic
targets of somatostatin-immunoreactive interneurons in the
rat hippocampus. Neuroscience, 88: 37–55.
Kelley, M.S. and Steward, O. (1996) The process of reinner-
vation in the dentate gyrus of adult rats: physiological events
at the time of the lesion and during the early postlesion
period. Exp. Neurol., 139: 73–82.
Kobayashi, M. and Buckmaster, P.S. (2003) Reduced inhibition
of dentate granule cells in a model of temporal lob epilepsy. J.
Neurosci., 23: 2440–2452.
Kugler, P. and Baier, G. (1992) Mitochondrial enzymes related
to glutamate and GABA metabolism in the hippocampus
of young and aged rats: a quantitative histochemical study.
Neurochem. Res., 17: 179–185.
Kuhn, H.G., Dickinson-Anson, H. and Gage, F.H. (1996)
Neurogenesis in the dentate gyrus of the adult rat: age-related
decrease of neuronal progenitor proliferation. J. Neurosci.,
16: 2027–2033.
Lawrence, J.J. and McBain, C.J. (2003) Interneuron diver-
sity series: containing the detonation — feedforward in-
hibition in the CA3 hippocampus. Trends Neurosci., 26:
631–640.
Lee, I. and Kesner, R.P. (2004) Encoding versus retrieval of
spatial memory: double dissociation between the dentate
gyrus and the perforant path inputs into CA3 in the dorsal
hippocampus. Hippocampus, 14: 66–76.
Leranth, C., Malcolm, A.J. and Frotscher, M. (1990) Afferent
and efferent synaptic connections of somatostatin-immuno-
reactive neurons in the rat fascia dentate. J. Comp. Neurol.,
295: 111–122.
Li, X.G., Somogyi, P., Ylinen, A. and Buzsaki, G. (1994) The
hippocampal CA3 network: an in vivo intracellular labeling
study. J. Comp. Neurol., 339: 181–208.
Ling, L.L., Hughes, L.F. and Caspary, D.M. (2005) Age-related
loss of the GABA synthetic enzyme glutamic acid dec-
arboxylase in rat primary auditory cortex. Neuroscience, 132:
1103–1113.
Lipp, H.-P., Schwegler, H. and Driscoll, P. (1984) Postnatal
modification of hippocampal circuitry alters avoidance learn-
ing in adult rats. Science, 225: 80–82.
Loiseau, J., Louiseau, P., Duche, B., Guyot, M., Dartigues, J.F.
and Aublet, B. (1990) A survey of epileptic disorders in
southwest France: seizures in elderly patients. Ann. Neurol.,
27: 232–237.
693
Lothman, E.W., Stringer, J.L. and Bertram, E.H. (1992) The
dentate gyrus as a control point for seizures in the hippo-
campus and beyond. Epilepsy Res. Suppl., 7: 301–313.
Luebke, J.I. and Rosene, D.L. (2003) Aging alters dendritic
morphology, input resistance, and inhibitory signaling in
dentate granule cells of the rhesus monkey. J. Comp. Neurol.,
460: 573–584.
Lukoyanov, N.V., Andrade, J.P., Dulce Madeira, M. and
Paula-Barbosa, M.M. (1999) Effects of age and sex on the
water maze performance and hippocampal cholinergic fibers
in rats. Neurosci. Lett., 269: 141–144.
Lumeng, L. and Li, T.K. (1974) Vitamin B6 metabolism in
chronic alcohol abuse. Pyridoxal phosphate levels in plasma
and the effects of acetaldehyde on pyridoxal phosphate syn-
thesis and degradation in human erythrocytes. J. Clin. In-
vest., 53: 693–704.
Lynch, G., Gall, C., Rose, G. and Cotman, C. (1976) Changes
in the distribution of the dentate gyrus associational system
following unilateral or bilateral entorhinal lesions in the adult
rat. Brain Res., 110: 57–71.
Lynch, M., Sayin, U., Golarai, G. and Sutula, T. (2000)
NMDA receptor-dependent plasticity of granule cell spiking
in the dentate gyrus of normal and epileptic rats. J. Ne-
urophysiol., 84: 2868–2879.
Maccaferri, G., Roberts, J.D., Szucs, P., Cottingham, C.A. and
Somogyi, P. (2000) Cell surface domain specific postsynaptic
currents evoked by identified GABAergic neurons in rat hip-
pocampus in vitro. J. Physiol., 524: 91–116.
Magnusson, K.R. (2000) Declines in mRNA expression of
different subunits may account for differential effects of ag-
ing on agonist and antagonist binding to the NMDA recep-
tor. J. Neurosci., 20: 1666–1674.
Magnusson, K.R., Kresge, D. and Supon, J. (2006) Differential
effects of aging on NMDA receptors in the intermediate versus
the dorsal hippocampus. Neurobiol. Aging, 27: 324–333.
Marty, S. (2000) Differences in the regulation of neuropeptide
Y, somatostatin and parvalbumin levels in hippocampal in-
terneurons by neuronal activity and BDNF. Prog. Brain
Res., 128: 1.
McNaughton, B.L., Barnes, C.A., Meltzer, J. and Sutherland,
R.J. (1989) Hippocampal granule cells are necessary for nor-
mal spatial learning but not for spatially-selective pyramidal
cell discharge. Exp. Brain Res., 76: 485–496.
Merrill, D.A., Chiba, A.A. and Tuszynski, M.H. (2001) Con-
servation of neuronal number and size in the entorhinal cor-
tex of behaviorally characterized aged rats. J. Comp. Neurol.,
438: 445–456.
Merrill, D.A., Karim, R., Darraq, M., Chiba, A.A. and
Tuszynski, M.H. (2003) Hippocampal cell genesis does not
correlate with spatial learning ability in aged rats. J. Comp.
Neurol., 459: 201–207.
Miles, R. and Wong, R.K.S. (1987) Inhibitory control of local
excitatory circuits in the guinea pig hippocampus. J. Physiol.,
388: 611–629.
Mitchell, C.L., Grimes, L.M. and Hong, J.S. (1990) Granule
cells in the ventral dentate gyrus are essential for kainic
694
acid-induced wet dog shakes but not those induced by pre-
cipitated abstinence in morphine-dependent rats. Brain Res.,
511(2): 338–340.
Mody, I., Stanton, P.K. and Heinemann, U. (1988) Activation
of N-methyl-D-aspartate receptors parallels changes in cellu-
lar and synaptic properties of dentate gyrus granule cells after
kindling. J. Neurophysiol., 59: 1033–1054.
Moser, E.I. (1996) Altered inhibition of dentate granule cells
during spatial learning in an exploration task. J. Neurosci.,
16: 1247–1259.
Mott, D.D., Turner, D.A., Okazaki, M.M. and Lewis, D.V.
(1997) Interneurons of the dentate-hilus border of the rat
dentate gyrus: morphological and electrophysiological heter-
ogeneity. J. Neurosci., 17: 3990–4005.
Nadler, J.V. (2003) The recurrent mossy fiber pathway of the
epileptic brain. Neurochem. Res., 28: 1649–1658.
Nagahara, A.H., Nicolle, M.M. and Gallagher, M. (1993)
Alterations in [3H]-kainate receptor binding in the hippocam-
pal formation of aged Long-Evans rats. Hippocampus, 3:
269–277.
Nanry, K.P., Mundy, W.R. and Tilson, H.A. (1989) Colchicine-
induced alterations of reference memory in rats: role of spa-
tial versus non-spatial task components. Behav. Brain Res.,
35: 45–53.
Ng, S.K., Hauser, W.A., Brust, J.C.M., Healton, E.B. and
Susser, M.W. (1985) Risk factors for adult-onset first sei-
zures. Ann. Neurol., 18: 153.
Nicolle, M.M., Bizon, J.L. and Gallagher, M. (1996) In vitro
autoradiography of ionotropic glutamate receptors in hip-
pocampus and striatum of aged Long-Evans rats: relation-
ship to spatial learning. Neuroscience, 74: 741–756.
Niesen, C.E., Baskys, A. and Carlen, P.L. (1988) Reversed et-
hanol effects on potassium conductances in aged hippocam-
pal dentate granule neurons. Brain Res., 445: 137–141.
Orr, G., Rao, G., Houston, F.P., McNaughton, B.L. and
Barnes, C.A. (2001) Hippocampal synaptic plasticity is mod-
ulated by theta rhythm in the fascia dentata of adult and aged
freely behaving rats. Hippocampus, 11: 647–654.
Pagliusi, S.R., Gerrard, P., Abdallah, M., Talabot, D. and
Catsicas, S. (1994) Age-related changes in expression of
AMPA-selective glutamate receptor subunits: is calcium-per-
meability altered in hippocampal neurons? Neuroscience, 61:
429–433.
Patel, Y.C., Liu, J.L., Warszynska, A., Kent, G., Papachristou,
D.N. and Patel, S.C. (1995) Differential stimulation of som-
atostatin but not neuropeptide Y gene expression by
quinolinic acid in cultured cortical neurons. J. Neurochem.,
65: 998–1006.
Patrylo, P.R. and Dudek, F.E. (1998) Physiological unmasking
of new glutamatergic pathways in the dentate gyrus of hip-
pocampal slices from kainate-induced epileptic rats. J. Ne-
urophysiol., 79: 418–429.
Patrylo, P.R., Tyagi, I., Willingham, A.L., Lee, S. and Will-
iamson, A. (2007) Dentate filter function is altered in a pro-
epileptic fashion during aging. Epilepsia (in press) doi:
10.1111/j.1528-1167.2007.01139.X
Paulsen, O. and Moser, E.I. (1998) A model of hippocampal
memory encoding and retrieval: GABAergic control of
synaptic plasticity. Trends Neurosci., 21: 273–278.
Penttonen, M., Kamondi, A., Sik, A., Acsady, L. and Buzsaki,
G. (1997) Feed-forward and feed-back activation of the
dentate gyrus in vivo during dentate spikes and sharp wave
bursts. Hippocampus, 7: 437–450.
Rao, G., Barnes, C.A., McNaughton, B.L. and Shen, J. (1994)
Age-related decrease in the NMDA-mediated EPSP in FD.
Soc. Neurosci. Abstr., 20: 1207.
Rao, M.S., Hattiangady, B., Abdel-Rahman, A., Stanley, D.P.
and Shetty, A.K. (2005) Newly born cells in the ageing dent-
ate gyrus display normal migration, survival and neuronal
fate choice but endure retarded early maturation. Eur. J.
Neurosci., 21: 464–476.
Rapp, P.R., Deroche, P.S., Mao, Y. and Burwell, R.D. (2002)
Neuron number in the parahippocampal region is preserved
in aged rats with spatial learning deficits. Cereb. Cortex, 12:
1171–1179.
Rapp, P.R. and Gallagher, M. (1996) Preserved neuron number
in the hippocampus of aged rats with spatial learning deficits.
Proc. Natl. Acad. Sci., 93: 9926–9930.
Rapp, P.R., Stack, E.C. and Gallagher, M. (1999) Morpho-
metric studies of the aged hippocampus: I. Volumetric anal-
ysis in behaviorally characterized rats. J. Comp. Neurol., 403:
459–470.
Rasmussen, T., Schliemann, T., Sorensen, J.C., Zimmer, J. and
West, M.J. (1996) Memory impaired aged rats: no loss of
principal hippocampal and subicular neurons. Neurobiol.
Aging, 17: 143–147.
Reeves, T.M. and Smith, D.C. (1987) Reinnervation of the
dentate gyrus and recovery of alternation behavior following
entorhinal cortex lesions. Behav. Neurosci., 101: 179–186.
Reibel, S., Vivien-Roels, B., Le, B.T., Larmet, Y., Carnahan, J.,
Marescaux, C. and Depaulis, A. (2000) Overexpression of
neuropeptide Y induced by brain-derived neurotrophic factor
in the rat hippocampus is long lasting. Eur. J. Neurosci., 12:
595–605.
Reynolds, J.N. and Carlen, P.L. (1989) Diminished calcium
currents in aged hippocampal dentate gyrus granule neu-
rones. Brain Res., 479: 384–390.
Ribak, C.E. (1992) Local circuitry of GABAergic basket cells in
the dentate gyrus. Epilepsy Res. Suppl., 7: 29–47.
Ribak, C.E., Seress, L. and Leranth, C. (1993) Electron micro-
scopic immunocytochemical study of the distribution of
parvalbumin-containing neurons and axon terminals in the
primate dentate gyrus and Ammon’s horn. J. Comp. Neurol.,
327: 298–321.
Ribak, C.E., Tran, P.H., Spigelman, I., Okazaki, M.M. and
Nadler, J.V. (2000) Status epilepticus-induced hilar basal
dendrites on rodent granule cells contribute to recurrent ex-
citatory circuitry. J. Comp. Neurol., 428: 240–253.
Rissman, R.A., Nocera, R., Fuller, L.M., Kordower, J.H. and
Armstrong, D.M. (2006) Age-related alterations in
GABA(A) receptor subunits in the nonhuman primate hip-
pocampus. Brain Res., 1073-1074: 120–130.

Rogers, S.W., Gahring, L.C., Collins, A.C. and Marks, M.
(1998) Age-related changes in neuronal nicotinic acetylcho-
line receptor subunit alpha4 expression are modified by long-
term nicotine administration. J. Neurosci., 18: 4825–4832.
Rolls, E.T. and Kesner, R.P. (2006) A computational theory of
hippocampal function, and empirical tests of the theory.
Prog. Neurobiol., 79: 1–48.
Rose, G., Diamond, D. and Lynch, G.S. (1983) Dentate gran-
ule cells in the rat hippocampal formation have the behavi-
oral characteristics of theta neurons. Brain Res., 266: 29–37.
Rosenzweig, E.S. and Barnes, C.A. (2003) Impact of aging on
hippocampal function: plasticity, network dynamics, and
cognition. Prog. Neurobiol., 69: 143–179.
Rowan, A.J., Ramsay, R.E., Collins, J.F., Pryor, F., Board-
man, K.D., Uthman, B.M., Spitz, M., Frederick, T., Towne,
A., Carter, G.S., Marks, W., Felicetta, J., Tomyanovich,
M.L. and VA Cooperative Study 428 Group. (2005)
New onset geriatric epilepsy: a randomized study of gaba-
pentin, lamotrigine, and carbamazepine. Neurology, 64:
1868–1873.
Ruano, D., Benavides, J., Machado, A. and Vitorica, J. (1995)
Aging-associated changes in the pharmacological properties
of the benzodiazepine (omega) receptor isotypes in the rat
hippocampus. J. Neurochem., 64: 867–873.
Sayin, Ü., Rutecki, P. and Sutula, T. (1999) NMDA-dependent
currents in granule cells of the dentate gyrus contribute to
induction but not permanence of kindling. J. Neurophysiol.,
81: 564–574.
Scharfman, H.E. (1994) Synchronization of area CA3 hippo-
campal pyramidal cells and non-granule cells of the dentate
gyrus in bicuculline-treated rat hippocampal slices. Neuro-
science, 59: 245–257.
Schuster, G., Cassel, J.C. and Will, B. (1997) Comparison of the
behavioral and morphological effects of colchicine- or neu-
tral fluid-induced destruction of granule cells in the dentate
gyrus of the rat. Neurobiol. Learn. Mem., 68: 86–91.
Schwartzkroin, P.A. and Prince, D.A. (1978) Cellular and field
potential properties of epileptogenic hippocampal slices.
Brain Res., 147: 117–130.
Shen, J. and Barnes, C.A. (1996) Age-related decrease in
cholinergic synaptic transmission in three hippocampal sub-
fields. Neurobiol. Aging, 17: 439–451.
Shen, J., Kudrimoti, H.S., McNaughton, B.L. and Barnes, C.A.
(1998) Reactivation of neuronal ensembles in hippocampal
dentate gyrus during sleep after spatial experience. J. Sleep
Res., 7: 6–16.
Shetty, A.K. and Turner, D.A. (1998) Hippocampal interneu-
rons expressing glutamic acid decarboxylase and calcium-
binding proteins decrease with aging in Fischer 344 rats. J.
Comp. Neurol., 394: 252–269.
Skaggs, W.E., McNaughton, B.L., Wilson, M.A. and Barnes,
C.A. (1996) Theta phase precession in hippocampal neuronal
populations and the compression of temporal sequences.
Hippocampus, 6: 149–172.
Smith, T.D., Adams, M.M., Gallagher, M., Morrison, J.H. and
Rapp, P.R. (2000) Circuit-specific alterations in hippocampal
695
synaptophysin immunoreactivity predict spatial learning im-
pairment in aged rats. J. Neurosci., 20: 6587–6593.
Soltèsz, I., Smetters, D.K. and Mody, I. (1995) Tonic inhibition
originates from synapses close to the soma. Neuron, 14:
1273–1283.
Sonntag, W.E., Lynch, C.D., Bennett, S.A., Khan, A.S.,
Thornton, P.L., Cooney, P.T., Ingram, R.L., McShane, T.
and Brunso-Bechtold, J.K. (1999) Alterations in insulin-like
growth factor-1 gene and protein expression and type 1 in-
sulin-like growth factor receptors in the brains of ageing rats.
Neuroscience, 88: 269–279.
Soriano, E. and Frotscher, M. (1993) GABAergic innervation
of the rat fascia dentata: a novel type of interneuron in the
granule cell layer with extensive axonal arborization in the
molecular layer. J. Comp. Neurol., 334: 385–396.
Soriano, E., Nitsch, R. and Frotscher, M. (1990) Axo-axonic
chandelier cells in the rat fascia dentata: golgi-electron mi-
croscopy and immunocytochemical studies. J. Comp. Ne-
urol., 293: 1–25.
Spigelman, I., Yan, X.X., Obenaus, A., Lee, E.Y., Wasterlain,
C.G. and Ribak, C.E. (1998) Dentate granule cells form
novel basal dendrites in a rat model of temporal lobe epi-
lepsy. Neuroscience, 86: 109–120.
Staley, K.J., Otis, T.S. and Mody, I. (1992) Membrane prop-
erties of dentate gyrus granule cells: comparison of sharp
microelectrode and whole-cell recordings. J. Neurophysiol.,
67: 1346–1358.
Stanley, D.P. and Shetty, A.K. (2004) Aging in the rat hippo-
campus is associated with widespread reductions in the
number of glutamate decarboxylase-67 positive interneurons
but not interneuron degeneration. J. Neurochem., 89:
204–216.
Steward, O. (1982) Assessing the functional significance of le-
sion-induced neuronal plasticity. Int. Rev. Neurobiol., 23:
197–254.
Steward, O. (1992) Lesion-induced synapse reorganization in
the hippocampus of cats: sprouting of entorhinal, commis-
sural/associational, and mossy fiber projections after unilat-
eral entorhinal cortex lesions, with comments on the normal
organization of these pathways. Hippocampus, 2: 247–268.
Steward, O., Loesche, J. and Horton, W.C. (1976) Behavioral
correlates of denervation and reinnervation of the hippo-
campal formation of the rat: open field activity and cue uti-
lization following bilateral entorhinal cortex lesions. Brain
Res. Bull., 2: 41–48.
Sykova, E., Mazel, T. and Simonova, Z. (1998) Diffusion con-
straints and neuron-glia interaction during aging. Exp. Ger-
ontol., 33: 837–851.
Tallis, R., Hall, G., Craig, I. and Dean, A. (1991) How common
are epileptic seizures in old age? Age Ageing, 20: 442–448.
Tanila, H., Shapiro, M., Gallagher, M. and Eichenbaum, H.
(1997) Brain aging: changes in the nature of information
coding by the hippocampus. J. Neurosci., 17: 5155–5166.
Tayebati, S.K., Amenta, F., El-Assouad, D. and Zaccheo, D.
(2002) Muscarinic cholinergic receptor subtypes in the hip-
pocampus of aged rats. Mech. Ageing Dev., 123: 521–528.
696
Thibault, O., Hadley, R. and Landfield, P.W. (2001) Elevated
postsynaptic [Ca2+]i and L-type calcium channel activity in
aged hippocampal neurons: relationship to impaired synaptic
plasticity. J. Neurosci., 21: 9744–9756.
Tigges, J., Herndon, J.G. and Rosene, D.L. (1995) Mild age-
related changes in the dentate gyrus of adult rhesus monkeys.
Acta Anat. (Basel), 153: 39–48.
de Toledo-Morrell, L., Geinisman, Y. and Morrell, F. (1988)
Age-dependent alterations in hippocampal synaptic plastic-
ity: relation to memory disorders. Neurobiol. Aging, 9:
581–590.
de Toledo-Morrell, L., Morrell, F. and Fleming, S. (1984) Age-
dependent deficits in spatial memory are related to impaired
hippocampal kindling. Behav. Neurosci., 98: 902–907.
Uttl, B. and Graf, P. (1993) Episodic spatial memory in adult-
hood. Psychol. Aging, 8: 257–273.
Vela, J., Gutierrez, A., Vitorica, J. and Ruano, D. (2003) Rat
hippocampal GABAergic molecular markers are differen-
tially affected by ageing. J. Neurochem., 85: 368–377.
Wenk, G.L. and Barnes, C.A. (2000) Regional changes in the
hippocampal density of AMPA and NMDA receptors across
the lifespan of the rat. Brain Res., 885: 1–5.
Wenzel, H.J., Buckmaster, P.S., Anderson, N.L., Wenzel, M.E.
and Schwartzkroin, P.A. (1997) Ultrastructural localization
of neurotransmitter immunoreactivity in mossy cell axons
and their synaptic targets in the rat dentate gyrus. Hippo-
campus, 7: 559–570.
West, J.R. and Dewey, S.L. (1984) Mossy fiber sprouting in the
fascia dentata after unilateral entorhinal lesions: quantitative
analysis using computer-assisted image processing. Neuro-
science, 13: 377–384.
Wilkniss, S.M., Jones, M.G., Korol, D.L., Gold, P.E. and
Manning, C.A. (1997) Age-related differences in an ecolog-
ically based study of route learning. Psychol. Aging, 12:
372–375.
Williams, S.R., Buhl, E.H. and Mody, I. (1998) The dynamics
of synchronized neurotransmitter release determined from
compound spontaneous IPSCs in rat dentate granule neu-
rones in vitro. J. Physiol., 510: 477–497.
Willingham, A., Tokhi, A., Jones, R. and Patrylo, P.R. (2006)
Hippocampal slices from aged rats have an enhanced suscep-
tibility to kainate-induced epileptiform activity. Abstract #
3.102, 60th annual meeting of the American Epilspy Society.
Wilson, I.A., Ikonen, S., Gallagher, M., Eichenbaum, H. and
Tanila, H. (2005) Age-associated alterations of hippocampal
place cells are subregion specific. J. Neurosci., 25: 6877–6886.
Wolfer, D.P. and Lipp, H.P. (1995) Evidence for physiological
growth of hippocampal mossy fiber collaterals in the guinea
pig during puberty and adulthood. Hippocampus, 5:
329–340.
Wuarin, J.P. and Dudek, F.E. (1996) Electrographic seizures
and new recurrent excitatory circuits in the dentate gyrus of
hippocampal slices from kaintae-treated epileptic rats. J. Ne-
urosci., 16: 4438–4448.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 38

Dentate gyrus neurogenesis and depression

Amar Sahay1,2,, Michael R. Drew1,2 and Rene Hen1,2,3,

1
Departments of Neuroscience and Psychiatry, Columbia University, New York, NY 10032, USA
2
Division of Integrative Neuroscience, Columbia University, New York, NY 10032, USA
3
Department of Pharmacology, Columbia University, New York, NY 10032, USA

Abstract: Major depressive disorder (MDD) is a debilitating and complex psychiatric disorder that in-
volves multiple neural circuits and genetic and non-genetic risk factors. In the quest for elucidating the
neurobiological basis of MDD, hippocampal neurogenesis has emerged as a candidate substrate, both for
the etiology as well as treatment of MDD. This chapter critiques the advances made in the study of
hippocampal neurogenesis as they relate to the neurogenic hypothesis of MDD. While an involvement of
neurogenesis in the etiology of depression remains highly speculative, preclinical studies have revealed a
novel and previously unrecognized role for hippocampal neurogenesis in mediating some of the behavioral
effects of antidepressants. The implications of these findings are discussed to reevaluate the role of
hippocampal neurogenesis in MDD.

Keywords: dentate gyrus; depression; neurogenesis; serotonin; antidepressants; hippocampus

Introduction ability to concentrate, diminished interest in pleas-

Understanding the neurobiological basis of major
depressive disorder (MDD) is one of the most
pressing challenges for today’s society. Severe
forms of depression affect 2–5% of the U.S. pop-
ulation, and mood disorders impact 7% of the
world’s population and rank among the top ten
causes of disability (Murray and Lopez, 1996). The
diagnosis of MDD based on the criteria established
by the Diagnostics and Statistical Manual of Men-
tal Disorders (American Psychological Association,
2000) includes the persistence of depressed mood,
low self esteem, feelings of hopelessness, decreased
Corresponding author: Tel.: +1 212-543-5477;
Fax: +1 212-543-5074;
E-mail: as2619@columbia.edu (A. Sahay)
Corresponding author: Tel: +1 212-543-5328;
Fax: +1 212-543-5074; E-mail: rh95@columbia.edu (R. Hen)
urable activities, daily insomnia or hypersomnia,
weight loss or gain, and recurrent suicidal ideation.
The diagnostic criteria for MDD convey the com-
plexity of the disease and suggest that multiple
neural circuits subserving distinct cognitive and
affective processes are likely to be involved.
Our comprehension of the mechanisms under-
lying the pathogenesis of MDD has evolved con-
siderably since the formulation of the monoamine
hypothesis (Bunney and Davis, 1965; Schildkraut,
1965; Nestler et al., 2002). The recent emphasis on
neural circuits as opposed to a chemical imbalance
catalyzed a fundamental shift in our conceptual-
ization of MDD and psychiatric disorders. It pro-
vided a framework to understand how genes,
through their effects on neural circuits, influence
our ability to encode experience and adapt to en-
vironmental stimuli and stressors. Implicit in this
idea is that genes moderate vulnerability to the

DOI: 10.1016/S0079-6123(07)63038-6 697

698
effects of environmental stress during particularly
sensitive or critical periods in brain development
by determining the optimal range of neuronal
circuit function for the organism. Indeed, the ne-
urotrophic, neuroplasticity and network hypothe-
ses of MDD all reflect the biology of gene products
in the context of synaptic and structural plasticity
of neural circuitry (Duman et al., 1997; Duman,
2002; Nestler et al., 2002; Castren, 2005).
Dentate gyrus neurogenesis has gained consid-
erable attention as both a form of structural plas-
ticity and as a neural substrate for the patho-
physiology of MDD. The neurogenic hypothesis
posits that a decrease in the production of new-
born dentate granule cells in the hippocampus
causally relates to the pathogenesis and patho-
physiology of MDD and that enhanced neurogen-
esis is necessary for treatment of depression
(Duman et al., 2000; Jacobs et al., 2000). The hy-
pothesis, when first proposed, was predicated on
the following observations, which are reviewed in
greater detail in subsequent sections. First, stress,
which is widely recognized as a major causal factor
in MDD, is known to suppress neurogenesis. Sec-
ond, most antidepressant (AD) treatments increase
hippocampal neurogenesis. Third, imbalance in
the serotonin system influences hippocampal ne-
urogenesis. Fourth, the induction of neurogenesis
is contingent upon chronic but not subchronic
(acute) selective serotonin reuptake inhibitor
(SSRI) treatment, paralleling the time course for
therapeutic actions of ADs. Finally, the therapeu-
tic lag in the response to SSRIs in patients with
MDD mirrors the timeline of maturation and
integration of newborn dentate granule cells.
Consequently, the dentate gyrus and neurogenesis
therein are potential substrates for the AD
response.
Central to the neurogenic hypothesis is the as-
sumption that the dentate gyrus plays an impor-
tant role in mediating cognitive and affective
processes. Moreover, since levels of neurogenesis
change during the lifetime of the organism,
changes in dentate neurogenesis may contribute
to dentate gyrus function in different ways. An-
other assumption is that neurogenesis represents
a potentially adaptive mechanism or form of plas-
ticity. Deficits in neurogenesis during critical
periods in brain development could, therefore, be
pathogenic in that they profoundly impact the
trajectory of emotional development. Deficits in
adult hippocampal neurogenesis could compro-
mise hippocampal-dependent functions and con-
tribute to the pathophysiology of MDD.
In this chapter, we focus on hippocampal ne-
urogenesis as it relates to MDD. Our aim is to
distill the observations made in the rapidly grow-
ing field of hippocampal neurogenesis and to crit-
ically assess the putative role of neurogenesis in the
etiology and treatment of MDD. We begin by de-
fining a framework for the reader to understand
how neurogenesis can contribute to dentate gyrus
function. Within this framework, we will first eval-
uate evidence for deficits in hippocampal neuro-
genesis in patients with MDD. We will then
examine the role of the serotonergic system in
hippocampal neurogenesis because the best char-
acterized genetic risk alleles for MDD encode
components of this system. Because susceptibility
to MDD conferred by genes is likely to be revealed
by environmental risk factors such as stress, we
will discuss the relationship between stress and
neurogenesis. We then review the considerable ev-
idence linking the effects of ADs with increased
hippocampal neurogenesis. Finally, we will turn to
evidence provided by studies using preclinical
models that attempt to establish a causal link be-
tween hippocampal neurogenesis and the etiology,
and pathophysiology of MDD and the require-
ment for neurogenesis in mediating the behavioral
effects of ADs.
Neurogenesis and MDD
A general framework for neurogenesis and dentate
gyrus function
Since the seminal findings of Altman and Das in
1965, it is now well accepted that the adult hip-
pocampus is host to the birth and integration of
newborn dentate granule cells in the dentate gyrus
(Altman and Das, 1965). In the rat, the species for
which the best data are available, it is estimated
that 9000 new cells are born each day in the DG,
and, of these, approximately 50% go on to express

neuron-specific markers. At this rate, the number
of new granule neurons born each month is equal
to 6% of the mature granule cell population
(Cameron and McKay, 2001). In non-human pri-
mates, the rate of neurogenesis may be lower than
the rate documented in rodents (Kornack and
Rakic, 1999; Gould et al., 1999b). One should bear
in mind that these data reflect neurogenesis under
laboratory housing conditions, and given the in-
crease in neurogenesis with environmental enrich-
ment (Kempermann et al., 1997; Gould et al.,
1999a), could underestimate the rates of neuro-
genesis in the normal habitat. Likewise, rates of
neurogenesis in man maybe underestimated by
available data, because human data were based on
a single study in which tissue samples were taken
from cancer patients injected with a mitotic
marker, bromo-deoxyuridine (BrdU) before death
(Eriksson et al., 1998). The number of BrdU-labe-
led neurons entering the neuronal lineage was
lower than that reported for marmosets and ro-
dents, but the age of subjects could explain the
difference because they were old, and neurogenesis
declines with age (Seki and Arai, 1995; Kuhn et al.,
1996; Rao and Shetty, 2004). Therefore, it is un-
clear to what extent the relatively low level of ne-
urogenesis observed in the human subjects was due
to real species difference.
The study of adult hippocampal neurogenesis
has revealed it to be a robust phenomenon that is
capable of conferring previously unrecognized
forms of plasticity to the dentate gyrus. For ex-
ample, it is clear that both net addition of newly
generated neurons and replacement of mature cells
occur in the adult dentate gyrus and that the extent
to which these processes occur may vary with the
animals age, and environmental and physiological
parameters (Bayer et al., 1982; West, 1993; Kem-
permann et al., 1998; Nottebohm, 2002; Amrein
et al., 2004; Wiskott et al., 2006). Modeling and
computational approaches have revealed merits of
both net addition and replacement in optimizing
hippocampal network function (Chambers et al.,
2004; Becker, 2005; Meltzer et al., 2005; Wiskott
et al., 2006). Figure 1 illustrates the distinct, but
potentially interrelated, ways by which neurogen-
esis can modify the cellular composition of the
dentate gyrus.
699
1. Increase the number of mature dentate gran-
ule cells
The integration of newborn neurons can re-
sult in an increase in the granule cell layer of
the dentate gyrus. It is conceivable that a net
increase in size is possible only within a cer-
tain period in an animal’s life. An increase in
cell number can result from an enhancement
in the rate of proliferation or the percentage
of newborn neurons that survive.
2. Provide a reservoir of highly plastic immature
neurons in the adult dentate gyrus
Newly generated dentate granule cells also
exhibit forms of synaptic plasticity distinct
from those of mature cells in the adult hip-
pocampus. Newborn dentate granule cells
show unique physiological properties such as
lower thresholds for induction of long-term
potentiation and long-term depression than
do mature neurons (Schmidt-Hieber et al.,
2004; Song et al., 2005). Moreover, newborn
dentate granule cells, unlike mature granule
cells, are able to undergo LTP under condi-
tions of increased GABAergic inhibition
(Wang et al., 2000; Snyder et al., 2001; Saxe
et al., 2006). Thus, in addition to conferring
structural plasticity to the dentate gyrus, ne-
urogenesis also creates a transient reservoir of
excitable, highly plastic cells that may serve a
unique biological function, distinct from that
of mature granule neurons. The size of such a
reservoir can by influenced by numerous
physiological and environmental factors and
contingencies.
3. Generate multiple cell types in the dentate
gyrus
While it is widely agreed that neurogenesis in
the subgranular zone (SGZ) results in gener-
ation of dentate granule cells, there is one
report showing that GABAergic basket cells
in the dentate gyrus incorporate BrdU and
form functional inhibitory synapses with
dentate granule cells (Liu et al., 2003). Thus,
it is plausible that the generation of interneu-
rons may occur under certain conditions to
influence network activity. Clearly more
evidence is needed to support this possibility.
In addition to the generation of neurons in
700

Fig. 1. A schematic of the dentate gyrus granule cell layer (GCL) illustrating the different ways by which neurogenesis can influence its
structure and function. Boxed panel reveals a cross section of the dentate GCL with the different populations that reside within it:
mature granule cells born during development (light blue), adult-generated mature granule cells (dark blue), adult-born immature
neurons (red) and interneurons (green). Over the lifespan, the GCL may increase in size due to a net addition of new neurons (A) or
may remain unchanged due to a net replacement of developmentally generated granule cells (B). Changes in neurogenesis can result in
increased representation of interneurons (C), a larger pool of adult generated immature neurons (D) or the generation of mature
neurons with distinct physiological and biochemical properties (E). Conceivably, neurogenesis may be altered in any one of these ways
in MDD. Conversely, AD drugs may influence DG function in more than one way to exert their behavioral effects. (See Color Plate
38.1 in color plate section.)

the dentate gyrus, proliferation in the SGZ
also generates glial cells. The emerging role
for glial cells in modulating synaptic function
in health and disease underscores the need to
understand how newly generated glial cells
contribute to hippocampal physiology and
function (Ma et al., 2005; Haydon and
Carmignoto, 2006).
4. Drive turnover and replacement of mature
dentate granule cells
The integration of newborn neurons can oc-
cur to replace the death of mature neurons.
Such a mechanism, when predominant,
would not increase the size of the dentate
gyrus but ensure replacement of cells, whose
functions are impaired, and rejuvenate the
network with new cells (Nottebohm, 2002).
There is some evidence to suggest that adult-
generated mature dentate granule cells, while
sharing electrophysiological properties with
their early-development-born counterparts,
exhibit greater plasticity in response to be-
haviorally relevant stimuli (Laplagne et al.,
2006; Ramirez-Amaya et al., 2006).

Independent of the balance between the inte-
gration of new neurons and the death of mature
neurons, it is conceivable that a specific form of
experience can result in a larger representation
of specific granule cells selected for by that kind
of experience. Such a representation may manifest
in a distinct pattern of biochemical and electro-
physiological properties found in one cohort of
newborn cells versus another. Functional hetero-
geneity within the hippocampus and dentate gyrus
supports the possibility that subsets of neurons
within different regions of the dentate gyrus could
reflect distinct experiences (Moser and Moser,
1998; Scharfman et al., 2002; Silva et al., 2006).
The aforementioned ways by which neurogene-
sis contributes to the structure and function of the
dentate gyrus convey the complexity of the phe-
nomenon of hippocampal neurogenesis. They also
remind us of the many ways by which neurogenesis
may be altered in pathological conditions.
Hippocampal dysfunction and atrophy in MDD
Hippocampal dysfunction in MDD is well sup-
ported by clinical studies which have shown that
MDD is often accompanied by deficits in declar-
ative learning and memory and diminished cogni-
tive flexibility that are dissociable from changes in
motivation (Austin et al., 2001; Fossati et al.,
2002). The anatomical and functional segregation
of the hippocampus along its septotemporal axis
suggests roles for the hippocampus in both cogni-
tive and emotional processes (Moser and Moser,
1998; Strange and Dolan, 1999; Strange et al.,
1999; Bannerman et al., 2003). As a first step to
thinking about the contribution of the dentate
gyrus and dentate neurogenesis to the pathophys-
iology of MDD, one must consider the direct ev-
idence for hippocampal dysfunction and atrophy
in MDD.
While longitudinal studies linking changes in
hippocampal structure and function with the etio-
logy of depression are lacking, we have made
progress in identifying changes in hippocampal
function, cellular structure and volume that are
associated with pathophysiology of depression.
Direct measurements of hippocampal function in
701
the depressed brain are made using neuroimaging
techniques such as positron emission tomography
(PET), a powerful way of identifying neural struc-
tures with altered metabolic activity. Studies on
patients with MDD have revealed alterations, but
only in a very small number of studies and with
conflicting results. This is partly due to the limited
resolution of PET. Using cerebral blood flow PET,
one group reported increased blood flow in the
hippocampus of acutely depressed patients with a
short duration of illness (Videbech et al., 2001,
2002; Videbech and Ravnkilde, 2004). By contrast,
two other studies have shown either a decrease or
no change in metabolism in the hippocampus of
patients with MDD using fluorodeoxyglucose
(FDG)–PET imaging (Saxena et al., 2001; Dre-
vets et al., 2002; Kimbrell et al., 2002). Differences
in patient profile with regards to severity and du-
ration of illness and treatment could explain these
differences. Alternatively, it has been suggested
that increased activity, if untreated, may result in
hippocampal atrophy and decreased metabolism.
Atrophy would occlude detection of changes in
activity.
A consistent finding that has emerged from
magnetic resonance imaging (MRI) studies on pa-
tients with MDD is a reduction in hippocampal
volume. Despite a few studies that failed to report
any differences between patients with MDD and
control groups using MRI, there is consensus for
reduced hippocampal volume in MDD (Sheline,
1996; Sheline et al., 1999; Bremner et al., 2000; von
Gunten et al., 2000; Vakili et al., 2000; Neumeister
et al., 2005). Two recent meta-analyses of studies
measuring temporal lobe structures in MDD com-
pellingly demonstrate a reduction in hippocampus
in people with recurrent depression relative to age-
and sex-matched controls (Campbell et al., 2004;
Videbech and Ravnkilde, 2004). Interestingly,
frequency of depressive episodes and the dura-
tion for which depression is untreated corre-
late with magnitude of reduction in hippocampal
volume (MacQueen et al., 2003; Sheline et al.,
2003). Taken together, the evidence argues for
reduced hippocampal volume in MDD and that
such changes are likely to be a result of depres-
sion rather than a cause. However, it is worth
mentioning here that a smaller hippocampus is
702
thought to be a predisposing factor for, rather
than a consequence of, post-traumatic stress dis-
order (Gilbertson et al., 2002).
The significance of hippocampal volume change
in the context of cognitive deficits is conveyed by a
recent study that shows that healthy individuals
who complained of memory impairments had
smaller hippocampal volumes than non-impaired
controls (van der Flier et al., 2004). Another study
showed a correlation between deficits in recollec-
tion memory performance and reduced hippocam-
pal volume in elderly depressed patients
(von Gunten and Ron, 2004). However, these
studies are limited in number and comprised
elderly individuals who are likely to have other
brain changes that may contribute to the memory
impairments.
Changes in hippocampal volume can be ex-
plained by several different mechanisms that may
operate in concert at the cellular and circuit levels
including: (i) Increased apoptosis of mature neu-
rons or glial cells. (ii) A loss of neuropil which may
involve changes in dendritic complexity, spine
density, and number and size of afferent and effer-
ent axonal projections. (iii) Reduced neurogenesis
or gliogenesis in the SGZ of dentate gyrus. The
link between neurogenesis and hippocampal vol-
ume has been addressed in histological analyses
of postmortem tissue obtained from brains of
patients with MDD, and will be discussed next.
Neurogenesis and cell death in MDD
Pathohistological studies of postmortem tissue of
patients, while small in number, have provided
some clues about the nature of cellular changes in
the hippocampus of a depressed individual. One
study examined synaptic density and glial cell
number using synaptophysin and GFAP-immuno-
reactivity, respectively, and found no differences in
the hippocampus of medicated patients with MDD
relative to controls (Muller et al., 2001). Another
study revealed low levels of apoptosis in the dent-
ate gyrus, CA1, CA4, subiculum and entorhinal
cortex of patients with MDD (Lucassen et al.,
2001). A third study showed a significant increase
in cell density of granule and glial cells in the
dentate gyrus and pyramidal neurons and glial
cells in the CA fields (Stockmeier et al., 2004). In
addition, the authors reported a reduction in soma
size of pyramidal neurons and a trend towards the
same in dentate granule cells. Finally, one group
directly examined the proliferation of cells in the
adult dentate gyrus of MDD patients using an
M-phase marker, Ki-67 (Reif et al., 2006). Their
results showed no changes in Ki-67 immunopos-
itive cells in hippocampus of depressed patients.
While this study is informative and is the first to
estimate levels of proliferation in the MDD brain,
it must be interpreted with several caveats in mind.
First, a reduction in neurogenesis in patients with
MDD could be masked by AD-mediated increase
in cell proliferation. Second, and for obvious rea-
sons, the study could not measure changes in sur-
vival of newborn cells or examine the kinetics
of turnover and maturation of newborn dentate
granule cells.
The pathohistological analyses suggest that
changes in neuropil, rather than neurogenesis,
may account for reductions in hippocampal vol-
ume. Indeed, the effects of stress on hippocampal
white matter are well documented. Preclinical
studies have shown that volumetric changes result
from reduced dendritic complexity and not abla-
tion of hippocampal neurogenesis (Santarelli et al.,
2003; McEwen, 2005). It should be noted that
while these data argue against the possibility that
reduced cell number owing to extensive cell death
or decreased neurogenesis is a primary mediator of
hippocampal volume change, they do not directly
address the possibility that neurogenesis is altered
in MDD. Since most of the patients were on med-
ication at the time of death, and since AD drugs
potently upregulate hippocampal neurogenesis, it
is possible that depression-related alterations in
neurogenesis could have been masked in these
studies. Moreover, since hippocampal neurogene-
sis in humans is likely to change with age, small
differences in proliferation or survival of newborn
neurons in postmortem analyses of older patients
could be difficult to detect. Further studies on
postmortem tissue of medicated and non-medi-
cated individuals are needed to identify specific
changes in hippocampal neurogenesis associated
with MDD.

Genes, environment and MDD
Depression is a complex and multifactorial illness
with genetic and non-genetic underpinnings. The
heritability of MDD is likely to be in the range of
40–50% and there is substantial evidence to sug-
gest that the phenotypic expression of MDD is
contingent upon interactions between the genetic
make-up of the individual and environmental fac-
tors, an interaction that has a dramatic effect on
the formation and functioning of neural circuitry
(Sullivan et al., 2000; Kendler et al., 2001; Caspi
and Moffitt, 2006; Leonardo and Hen, 2006;
Levinson, 2006). Here, it must be emphasized that
human susceptibility to MDD as revealed by en-
vironmental factors is tremendously magnified in
early life. For example, adults who had experi-
enced four out of seven traumatic events in early
life had a 4.6-fold increased risk of developing de-
pressive symptoms later in life and were 12.2-fold
more likely to commit suicide (Felitti et al., 1998;
Chapman et al., 2004). These studies underscore
the idea of a critical period in ‘‘emotional devel-
opment’’ when mechanisms mediating neural cir-
cuit synaptic- and structural plasticity are
particularly susceptible to environmental input,
and if compromised by a vulnerability conferred
by genes, can result in maladaptive alterations in
neural circuit function and pathological behavior.
While the search for candidate genes for MDD has
yielded some convergence from linkage studies
that certain genetic loci are involved, more data
are clearly needed to conclusively implicate a spe-
cific gene in the etiology of MDD. A notable
exception is the serotonin transporter (5-HTT)
polymorphism. In a landmark study, Caspi and
colleagues found that a functional polymorphism
in the 5-HTT gene moderated the sensitivity of
individuals to the depressogenic effects of early life
stress, a finding recently replicated by Kendler and
colleagues (Caspi et al., 2003; Kendler et al., 2005).
Caspi and colleagues found that people who car-
ried one or two copies of the ‘‘short’’ allele of
5-HTT, associated with lower levels of 5-HTT
and impaired reuptake of serotonin (5-HT) at
synapses, had more depressive symptoms and su-
icidal behavior in relation to stressful life events
than did people who had the ‘‘long allele’’.
703
Importantly, the association between the 5-HTT
polymorphism and depression is only observed
in individuals who had experienced significant
stressful life events. These findings argue that the
serotonergic system has a critical influence on ne-
urodevelopmental processes that lead to MDD. If
hippocampal neurogenesis is to be a considered as
a candidate neural substrate for depression, then
the effects of serotonergic dysregulation on it war-
rants comment. The following section addresses
the role of the serotonergic system in modulating
dentate gyrus structure and function.
The 5-HT system and hippocampal neurogenesis
Serotonergic terminals originating from the dorsal
raphe nucleus (DRN) and median raphe nuclei
(MRN) diffusely innervate multiple structures in
the vertebrate forebrain and reach the ventricles
via the supra-ependymal plexus (Azmitia and
Segal, 1978; Freund et al., 1990). The serotonergic
innervation of the hilus, molecular layers of the
dentate gyrus and the SGZ supports the possibility
that 5-HT signaling may influence adult neuro-
genesis (Oleskevich et al., 1991). The idea that
serotonin can influence neurogenesis was first pro-
posed three decades ago (Lauder and Krebs,
1978). However, the specific ways by which the
5-HT system influences adult dentate neurogenesis
was established only relatively recently. The first
studies to address the role of 5-HT in adult hip-
pocampal neurogenesis used serotonin depletion
analyses. Injection of the serotonin neurotoxin
5,7-dihydroneurotoxin (5,7-DHT) or 5-HT syn-
thesis inhibitor parachlorophenylalanine (PCPA)
into the raphe of young female rats resulted in a
reduction of dentate granule cell proliferation and
the number of immature neurons as assessed by
BrdU uptake and PSA-NCAM immunostaining,
respectively (Brezun and Daszuta, 1999, 2000).
Moreover, the same group showed that they could
rescue the deficit in hippocampal proliferation in
these rats following intrahippocampal grafts of
embryonic 5-HT neurons (Brezun and Daszuta,
2000). A potential role for 5-HT in influencing
the maturation of dentate granule cells comes
from studies in which rat pups were treated
704
with phenylchloromethamphetamine (PCA) or
5,7-DHT to reduce serotonergic innervation of
the forebrain. Analysis of dentate granule cells in
rodents with reduced serotonergic innervation re-
vealed fewer dendritic spines and synapses, but
otherwise normal dendritic complexity (Yan et al.,
1997; Faber and Haring, 1999), suggesting that
serotonergic signaling is important for selected,
and not all, aspects of the neuronal maturation
process. While the limitations inherent to these
pharmacological lesion studies must be acknowl-
edged, these studies illustrate how deficits in the
5-HT system can have consequences for the mat-
uration of dentate granule cells. Given the striking
recapitulation in adult neurogenesis of the early-
developmental neuronal maturation process, it is
likely that changes in 5-HT signaling have similar
consequences on neurons born in adulthood
(Esposito et al., 2005; Laplagne et al., 2006; Over-
street-Wadiche and Westbrook, 2006). Indeed, the
well-characterized effects of 5-HT receptor agonist
and antagonists and SSRIs on adult neurogenesis
(Malberg et al., 2000) solidify the link between
5-HT and adult hippocampal neurogenesis. Im-
portantly, studies on 5-HT receptors, which are
discussed next, offer a glimpse into the ways by
which altered serotonin levels as a consequence of
a genetic polymorphism, such as the 5-HTT poly-
morphism, can influence the birth and maturation
of newborn dentate granule cells.
The effects of 5-HT levels on neurogenesis re-
flect the sum of interactions between the synthesis
of 5-HT, its release and its actions at different
5-HT postsynaptic receptors acting in both a cell
autonomous and non-cell autonomous manner.
The effects of 5-HT on a newborn neuron depend
on the repertoire of 5-HT receptors that it ex-
presses. Since the maturation of newborn neurons
is intimately connected with the activity of the
network, it is also influenced by the actions of
different 5-HT receptors expressed within the hip-
pocampal formation in interneurons, mature dent-
ate granule cells and afferent projections arising
in the entorhinal cortex. The 5-HT1A receptor
(5-HT1AR) is the best-studied 5-HT receptor in
the context of adult hippocampal neurogenesis.
Acute administration of 5-HT1A antagonists re-
sults in decreased cell proliferation in the adult
dentate gyrus (Radley and Jacobs, 2002). Consist-
ent with these findings, acute or chronic treatment
with the 5-HT1A agonist 8-OH DPAT increases
proliferation in the SGZ and the number of adult
born neurons (Santarelli et al., 2003; Banasr et al.,
2004). The effects of activating 5-HT1AR appear
to be restricted to proliferation and do not affect
the differentiation of newborn progenitors into
neurons or glial cells. The increase in proliferation
could reflect a change in rate of progression
through the cell cycle or an increase in the size of
the proliferative pool in the SGZ. It is unclear,
given the experimental design employed in these
studies, whether activation of 5-HT1AR also
influences the survival of newborn neurons. Inter-
estingly, mice lacking the 5-HT1AR fail to re-
spond to the neurogenic effects of chronic
fluoxetine (Santarelli et al., 2003).
Two other 5-HT receptors, the 5-HT2A and
5-HT1B receptors, have also been implicated in
cell proliferation in the adult SGZ. While neither
5-HT1B agonists nor antagonists affect baseline
cell proliferation, the former can, however, restore
normal levels of proliferation in PCPA pretreated
rats. These data suggest that effects of the 5-HT1B
receptor on cell proliferation are small under
physiological conditions, but can become impor-
tant when 5-HT levels are decreased. The phar-
macology of the 5-HT2A receptor is also complex.
5-HT2A antagonists decrease cell proliferation
but agonists have no effect (Banasr et al., 2004),
suggesting that under physiological conditions,
the 5-HT2A-dependent signaling pathways
that modulate neurogenesis may be saturated.
Taken together, these observations reveal the
differential effects of recruiting different post-
synaptic 5-HT receptors on hippocampal neuro-
genesis.
Central to understanding how changes in 5-HT
levels influence neurogenesis is knowledge of the
expression of different 5-HT receptors in neural
progenitors and at different stages of their matu-
ration. Conspicuously absent from the pharmaco-
logical studies is precise information for 5-HT
receptor distribution in the adult SGZ. The
5-HT1AR is an exception to the rule. We know
that the 5-HT1AR is expressed at very low levels,
if any, in the SGZ of the rodent hippocampus.

In both rat and mouse, 5-HT1AR expression is
restricted to the mature dentate granule cells
rather than the immature population of cells dur-
ing development of the dentate gyrus (Patel and
Zhou, 2005; Sahay and Hen, unpublished data).
The effect of 5-HT1AR agonists on cell prolifer-
ation is, therefore, likely to be non-cell autono-
mous. It is possible that the 5-HT1AR is required
in hilar interneurons or mature dentate granule
cells to mediate the effects of 5-HT on cell prolif-
eration. Cell-type specific ablation and overexpres-
sion of the 5-HT1AR will reveal its precise
contribution in different cell types to adult hippo-
campal neurogenesis.
The specific role of different 5-HT receptors in
the maturation and integration of newborn neu-
rons is still to be elucidated. It is also unclear how
5-HT may impact the turnover of mature dentate
granule cells or influence the survival of newborn
dentate granule cells. The role for distinct 5-HT
receptors in regulation of developmental processes
such as dendritic development, synaptogenesis and
glutamate receptor trafficking (Kondoh et al.,
2004; Kvachnina et al., 2005; Yuen et al., 2005a,
b) suggests that 5-HT receptor-dependent mecha-
nisms during development may be conserved in
neurogenesis in the adult brain depending on
which 5-HT receptors are expressed and when
during neuronal maturation. In addition, 5-HT
signaling can induce the production of ne-
urotrophins and growth factors known to regu-
late hippocampal neurogenesis.
The role of stress in MDD and its effects on
neurogenesis
Stressful life experiences play a pivotal role in de-
velopment of MDD in individuals with a genetic
vulnerability (Holsboer and Barden, 1996; Gold
and Chrousos, 2002). Major stressors precede the
appearance of the first symptoms, and dysregula-
tion of the hypothalamic-pituitary-adenocortical
(HPA) system is often observed in patients with
MDD (Carroll et al., 1968a). An optimally func-
tional HPA system enables the organism to re-
spond appropriately to stressful stimuli by
controlling the production of adrenal steroids,
705
the glucocorticoids, which mobilize energy, in-
crease cardiovascular tone and influence immune
and nervous system functions. In response to a
stressful event, for example, a neuroendocrine cas-
cade is initiated in which corticotropin releasing
hormone (CRH) is released from neurons in the
paraventricular nucleus (PVN) in the hypo-
thalamus, which then triggers release of corticotro-
pin by the anterior pituitary to stimulate
glucorticoid secretion by the adrenal cortex. A hy-
peractive HPA, on the other hand, results in the
oversecretion of glucocorticoids with deleterious
consequences for the physiology of the viscera and
brain (Sapolsky, 2000). In addition, increased
glucocorticoid levels can impair serotonergic
signaling (Joels and van Riel, 2004). It is there-
fore not surprising that the stress response is
tightly regulated by efferents from multiple brain
regions that converge onto the PVN, where
incoming information is integrated to elicit an
adaptive response. Afferent inputs of the PVN are
inhibitory or facilitative in nature and arise in
brain stem nuclei, amygdala, cortex, septum and
the hippocampus, and are in turn regulated by a
negative feedback system. The hippocampus, for
example, exercises a powerful inhibitory influence
on HPA function to terminate the stress response,
and is in turn regulated by glucocorticoids acting
on cognate receptors expressed within the hippo-
campal formation. Dysregulation of such control
can, therefore, result in hypersecretion of gluco-
corticoids and an exaggerated stress response
that can severely impact neural circuit function.
Consistent with preclinical findings, about half of
all depressed patients show a blunted response to
the dexamethasone suppression test (Carroll et al.,
1968b; Holsboer et al., 1982), which measures the
ability of an exogenous glucocorticoid receptor
agonist to suppress endogenous stress hormone
release. Moreover, patients with MDD show ele-
vated levels of CRH in the cerebrospinal fluid,
increased numbers of CRH containing cells in the
PVN, and decreased CRH binding in the prefron-
tal cortex. Thus, the optimal function of the
hippocampal formation is a critical factor in mod-
ulation of the stress response, and an exaggerated
stress response can, in turn, negatively impact hip-
pocampal function.
706
One effect of stress on neural circuitry within the
hippocampal formation is the suppression of ne-
urogenesis by glucocorticoids. Stress-induced sup-
pression of cell proliferation in the DG has been
reported in several different mammalian species
and there is considerable evidence arguing for a
role for glucocorticoids as the mediators of the
stress response. Elimination of circulating adrenal
steroids by adrenalectomy, for example, increases
cell proliferation and neurogenesis in the adult
dentate gyrus (Cameron and McKay, 1999).
Exogenous administration of corticosterone, on
the other hand, suppresses proliferation (Cameron
and Gould, 1994). Glucocorticoids have been
shown to inhibit the proliferation and differenti-
ation of neural progenitors, and also the survival
of young neurons (Wong and Herbert, 2004,
2006). These effects are likely to be mediated di-
rectly through high affinity mineralocorticoid re-
ceptors (MR) and low affinity glucocorticoid
receptors (GR) that are expressed at various stages
of maturation and also indirectly through changes
in network activity of the hippocampal formation
(see chapter by Joels in this volume). Analysis of
receptor distribution in the dentate gyrus of ro-
dents reveals that GRs are expressed in both neu-
ral progenitors and mature dentate granule cells,
while MRs are expressed only in the latter.
Throughout most of adulthood, neither GRs
nor MRs are expressed in immature neurons
(Cameron et al., 1993; Garcia et al., 2004a). How-
ever, in aged rodents GR and MR expression is
seen in immature neurons, suggesting that imma-
ture neurons at this stage, but not earlier in the
animal’s life, may show increased sensitivity to
corticosterone action.
The dynamic expression of GR and MR during
neurogenesis and the ability of corticosteroids to
regulate the expression of growth factors such as
IGF-1, BDNF and EGF (Islam et al., 1998; Schaaf
et al., 2000), which have distinct effects on prolif-
eration and survival (Kuhn et al., 1997; Aberg
et al., 2000; Sairanen et al., 2005), illustrate the cell
autonomous and non-cell autonomous ways by
which corticosteroids can affect neurogenesis. It
should be noted that the sensitivity of neurons or
neural progenitors to corticosteroids must be stud-
ied with the state of the neuron or network in
mind. For example, glucocorticoids have deleteri-
ous effects on neurogenesis during stressful epi-
sodes but not during physical activity. This may be
due to the fact that physical activity, unlike stress,
elicits the production of growth factors that may
buffer the effects of corticosteroids on neurons.
Likewise, given the differences in the developing
and adult brain, an increase in glucocorticoids
during early postnatal life may have profoundly
different effects from those in adulthood. A recent
study modeling early life stress using a maternal
separation paradigm in rodents supports this idea
(Mirescu et al., 2004). Maternal separation during
the early postnatal period in rodents leads to per-
sistent changes in the HPA axis, protracted release
of corticosterone in response to mild stressors,
increased anxiety behavior (Huot et al., 2001),
impaired maternal care (Lovic et al., 2001) and
impaired spatial navigation (Huot et al., 2002).
Rats subjected to prolonged maternal deprivation
during the early postnatal period also showed re-
duced proliferation in the dentate gyrus of neural
progenitors in adulthood. Interestingly, the early
life stressor did not affect the number of mature
neurons, suggesting a compensatory increase in
survival of newly generated neurons. The authors
in this study showed that the change in prolifer-
ation could be rescued in adulthood by adrenal-
ectomy and by reducing corticosteroid production.
Without adrenalectomy, the corticosterone levels
were normal in stressed animals, not only under
baseline conditions but also in response to a
stressor, suggesting that the suppressed prolifera-
tion was likely to be result of increased sensitivity
to corticosteroids rather than increased levels
(Mirescu et al., 2004). It is tempting to speculate
that a shift in the temporal pattern of MR and GR
distribution during neurogenesis may contribute to
this increased sensitivity.
It is possible that changes in neurogenesis as a
result of HPA axis dysregulation contribute to the
pathophysiology of MDD by affecting the role of
the hippocampus in learning and other cognitive-
emotional processes. Preclinical studies (discussed
later) have proven tremendously informative in
defining the physiological relevance of adult hip-
pocampal neurogenesis to these processes, and
have greatly facilitated our understanding of how

stress-mediated suppression of neurogenesis may
be relevant to the pathophysiology of MDD.
A second possible consequence of the stress-medi-
ated suppression of neurogenesis is that the ability
of the hippocampus to regulate HPA activity be-
comes compromised. However, evidence directly
linking changes in dentate gyrus function with
HPA activity is scarce. Lesions of the hippocam-
pus and fimbria-fornix transections reduce the
ability of dexamethasone to inhibit stress induced
adrenocortical responses and results in hyperse-
cretion of glucocorticoids and ACTH following
stressful stimuli (Knigge, 1961; Sapolsky et al.,
1989; Herman et al., 1992; Feldman and
Weidenfeld, 1993; Bratt et al., 2001; Goursaud
et al., 2006). Antagonism of GR in the rat hippo-
campus results in hypersecretion of ACTH
and corticosterone following stressful stimuli
(Sapolsky, 1994; Feldman and Weidenfeld, 1999).
Analysis of subfields within the hippocampal for-
mation confirms the differential contributions
of CA fields and the dentate gyrus in the ventral
hippocampus in regulating HPA reactivity. In-
triguingly, lesion studies indicate that damage of
ventral subiculum but not ventral CA1 or dentate
gyrus results in prolonged glucocorticoid re-
sponses (Herman et al., 1992, 1995). By contrast,
electrical stimulation of DG in anesthesized ani-
mals inhibits corticosteroid secretion (Dunn and
Orr, 1984). These studies suggest that lesions
of DG in adulthood may be compensated for by
activity in structures downstream such as the sub-
iculum. Genetic manipulations that specifically
impair glucocorticoid signaling in the DG both in
adulthood and in early postnatal life will prove
critical in establishing the link between hippocam-
pal neurogenesis and HPA reactivity.
In sum, the well-documented effects of stress on
neurogenesis suggest that patients with MDD are
likely to show reductions in neuronal proliferation.
While it is plausible that a reduction in neurogen-
esis in turn contributes to HPA dysregulation,
there is as yet no evidence for this.
Neurogenesis and antidepressants
It is well recognized that all of the major classes of
ADs are associated with a several week delay in
707
onset. This delay is likely to reflect changes in
structural and synaptic plasticity in the brain
mediated by multiple mechanisms involving
monoaminergic signaling and neurotrophins.
PET imaging studies on MDD patients treated
with SSRIs such as paroxetine and fluoxetine have
helped define a neuroanatomical basis comprising
corticolimbic circuits (Seminowicz et al., 2004).
Structures that showed changes in metabolic ac-
tivity included the subgenual cingulate, hippocam-
pus and prefrontal cortex (Mayberg et al., 2000;
Kennedy et al., 2001). One form of structural
plasticity within the hippocampus that is consist-
ent with the delayed onset of ADs is the birth and
subsequent integration of newborn dentate gran-
ule cells in the adult dentate gyrus. Moreover,
almost all ADs known to date increase adult
neurogenesis. Therefore, the idea that ADs may
work through enhancing neurogenesis has received
abundant attention and is now considered central
to the neurogenic hypothesis of MDD (Malberg
and Schechter, 2005).
Neurogenic effects of antidepressant treatments
Numerous groups have shown that different
classes of ADs including 5-HT and norepinephrine
selective reuptake inhibitors, tricyclics, monoa-
mine oxidase inhibitors, phosphodiesterase inhib-
itors and electroconvulsive shock therapy increase
neurogenesis (Madsen et al., 2000; Malberg et al.,
2000; Manev et al., 2001; Nakagawa et al., 2002b).
AD treatment does not appear to affect the ratio
of newly generated neurons to glial cells, with the
majority of newborn cells adopting the neuronal
fate. The neurogenic effects of ADs are specific to
the SGZ, and are not observed in other compo-
nents of the ventricular system such as the lateral
ventricles or the subventricular zone. Moreover,
administration of non-AD psychotropic drugs
such as haloperidol does not increase hippocam-
pal neurogenesis (Eisch, 2002). Other treatments
reported to have AD effects, including exercise
(Babyak et al., 2000; Singh et al., 2001; Motl et al.,
2004), environmental enrichment and estrogen,
have also been shown to increase neurogenesis
(van Praag et al., 1999; Tanapat et al., 1999;
708
Rhodes et al., 2003; Meshi et al., 2006). In addi-
tion, also lithium, which is used in the treatment of
bipolar disorder increases neurogenesis (Chen
et al., 2000). The one AD treatment that has not
been shown to enhance neurogenesis is repetitive
transcranial magnetic stimulation or rTMS, which
is still awaiting FDA approval (Loo and Mitchell,
2005). While rTMS has been shown to reverse the
effects of chronic psychosocial stress on stress
hormone levels, it does not upregulate neurogen-
esis (Czeh et al., 2002).
That ADs block the behavioral effects of stress
and restore normal levels of neurogenesis in the
adult hippocampus lends further credence to the
possibility that ADs may work by increasing ne-
urogenesis to exert their behavioral effects. In tree
shrews, chronic exposure to psychosocial conflict
results in a decrease in cell proliferation, which is
blocked by treatment with the atypical AD tianep-
tine (Czeh et al., 2001). In another model of de-
pression, the learned helplessness (LH) paradigm,
exposure to inescapable shock engenders prode-
pressive behavior and a reduction in hippocampal
cell proliferation, both of which, are reversed by
AD treatment (Cryan et al., 2002; Malberg and
Duman, 2003). In addition, ECS enhances cell
proliferation after chronic corticosterone treat-
ment (Hellsten et al., 2002).
It is well known that ADs have pleiotropic
effects on neuronal circuits. That a diverse range
of ADs appears to enhance neurogenesis indicates
that the dentate gyrus may be a neuroanatomical
substrate to target for the development of novel
AD treatments. In the next section, we describe
recent preclinical studies elucidating how SSRIs
modulate adult hippocampal neurogenesis, and
then in the following sections we describe work
from animal models aimed at identifying whether
neurogenesis is necessary for the behavioral effects
of AD treatments.
Serotonin-dependent ADs and hippocampal
neurogenesis
SSRIs represent the most successful class of ADs
identified to date. Based on our understanding of
neurogenesis in the SGZ, it is conceivable that
SSRIs act directly on progenitors or immature
neurons to influence processes such as prolifera-
tion, differentiation, maturation and survival. In
addition, SSRIs are also likely to modulate net-
work activity within the dentate gyrus and as a
result, regulate neurogenesis indirectly. By virtue
of their effects on neurogenesis, SSRIs may be ca-
pable of driving replacement or turnover within
the dentate gyrus and catalyzing the insertion of
newly generated neurons with distinct electrophys-
iological and biochemical properties. The best-
characterized effect of SSRIs to date is the increase
in proliferation of neural progenitors in the SGZ.
Studies in rodents indicate that a 14-day, but not
shorter-term, administration of fluoxetine (1–5
days) is sufficient to upregulate cell proliferation
(Malberg et al., 2000). By contrast, a longer treat-
ment regimen is required to enhance survival
of newly generated neuroblasts. Fluoxetine treat-
ment for 28 days, but not 14 days, following
BrdU injection resulted in an increase in cell sur-
vival (Malberg et al., 2000; Nakagawa et al.,
2002a).
The delay with which the neurogenic effects
emerge after the initiation of SSRI treatment
provides a potential mechanism to explain the
therapeutic lag in the effects of these drugs.
Namely, it could be that the therapeutic effects
depend on the increase in proliferation, which it-
self requires several weeks of treatment. However,
closer inspection of this hypothesis reveals several
shortcomings. It is unlikely that a boost in prolif-
eration would produce an immediate psychologi-
cal effect. Rodent studies indicate that newborn
neurons do not become functionally integrated
until approximately 2–4 weeks after exiting the cell
cycle, so it would seem that any functional effects
elicited by the increase in proliferation would not
appear until that time (Esposito et al., 2005). This
means that behavioral effects of SSRIs mediated
by increased proliferation should not manifest
until about 4 weeks after treatment initiation. In
contrast, some behavioral effects of AD drugs in
animal models begin immediately following acute
treatment, and virtually all the behavioral effects
manifest within 4 weeks of SSRI treatment. In
primates the time required for maturation of new
neurons is greater than in rodents (Kohler et al.,

2006), so the predicted therapeutic lag would be
even longer. A recent meta-analysis of clinical data
indicates that some of the therapeutic effects of
SSRIs may commence very rapidly after treatment
initiation (within 1–2 weeks) (Taylor et al., 2006).
Thus, increases in proliferation are unlikely to un-
derlie the early onset effects of these drugs, but it
remains plausible that neurogenesis contributes to
the more long-term effects.
Consistent with this interpretation, remarkable
preliminary data from Meltzer, Deisseroth and
colleagues suggest that mature adult-born neurons
contribute to the behavioral effects of SSRIs
(Meltzer et al., 2006). In this study, rats were
given 7 days of fluoxetine treatment and then
tested behaviorally 1 month later. At that time
point, the previously treated rats showed enhanced
performance in the forced swim test, and histology
revealed an increase in the number of newly gen-
erated neurons. The results suggest that the pro-
liferative effects of SSRIs may appear sooner than
once thought (after 1 week rather than 2 weeks
of treatment), and that some behavioral effects of
these drugs depend not on the acute presence of
the drug but rather on slow, time-dependent proc-
esses initiated by drug treatment, such as neuro-
genesis and circuit reorganization.
The studies that examined the effects of fluoxe-
tine on proliferation do not identify the specific
types of neural progenitors that respond to
changes in 5-HT levels. A recent study addressed
this question using transgenic mice in which ex-
pression of a fluorescent reporter gene was regu-
lated by a nestin promoter fragment. Because
nestin is expressed in multiple progenitor cell
types, this approach allowed the visualization
and quantification of the distinct sub types of
neural progenitors that reside within the SGZ
(Encinas et al., 2006). The results of this study
showed that only a specific proliferative cell type,
the transiently amplifying neural progenitor
(ANP) that exists as an intermediate between the
type I radial glial-like neural progenitor and the
type II cell (Filippov et al., 2003; Tozuka et al.,
2005; Encinas et al., 2006), directly responds to
fluoxetine. Moreover, the study confirmed that
once exposure to fluoxetine ends, the rate of pro-
genitor cell division is restored to baseline and that
709
the increase in ANPs translates into a net increase
in the number of new neurons.
A net increase in the number of mature neurons
implies that SSRIs also increases the population of
immature neurons. It is presently not clear how
SSRIs influence the maturation of immature neu-
rons with regards to their physiological properties,
synaptic connectivity and dendritic complexity.
What is well appreciated, however, is that SSRI
treatment results in the induction of growth fac-
tors and neurotrophins whose effects on matura-
tion and survival are well understood (Carlezon
et al., 2005; Duman and Monteggia, 2006) and,
importantly, whose receptors are expressed in im-
mature neurons. It is also possible that SSRIs may
induce the secretion of growth factors from neural
progenitors, which then influence the function of
neighboring mature granule cells in a paracrine
manner. There is no evidence for this as yet.
The consequences of increased proliferation and
survival of newly generated cells following fluoxe-
tine treatment for the net size of the granule cell
population of the dentate gyrus have not been as-
certained. One study suggested that there may not
be a net change in the size of the dentate gyrus
because the increase in newly generated neurons is
offset by death of previously born mature granule
cells (Sairanen et al., 2005). Based on their data
showing that chronic fluoxetine treatment in-
creases not only proliferation and survival of
newly generated neurons but also the rate of
apoptosis, the authors argue that the net size of the
dentate gyrus does not change with chronic AD
treatment.
Finally, it is plausible that SSRI treatment alters
the physiological properties of newly generated
neurons, generating a cohort of cells within the
dentate that are unique. These unique cells may
confer greater adaptive potential to the dentate
gyrus than naı̈ve newly generated neurons. Much
work remains to be done in this area to address the
different ways by which SSRIs influence neuro-
genesis.
Delineating the pattern of expression of distinct
5-HT receptor subtypes within different cell types
in the SGZ and the DG will shed light on how
changes in 5-HT can elicit the diverse range of
effects discussed here. It follows that the
710
identification of genes downstream of 5-HT recep-
tors that mediate the behavioral effects of ADs will
pave the way for developing neurogenic non-
monoamine based therapeutics. Preclinical studies
have proven invaluable in defining the contribu-
tion of neurogenesis to the etiology and treatment
of MDD as they allow for discernment between
correlation and causality.
Preclinical studies: in the search for causality
Ultimately, whether neurogenesis is causally re-
lated to the etiology or treatment of depression
requires the use of animal models in which neuro-
genesis and emotional state can be experimentally
manipulated. If reduced neurogenesis contributes
to depression, it should be possible to produce a
depressive phenotype by experimentally reducing
neurogenesis. Conversely, behavioral manipula-
tions that produce a depressive phenotype should
reduce neurogenesis and do so before the behavi-
oral manifestations of depression develop. Of
course, the predictive value of such experiments
will depend on the specificity of the methods for
manipulating neurogenesis and the validity of the
animal models of depression. This section will re-
view recent work that addresses the role of neuro-
genesis in the etiology and treatment of MDD
assessed in preclinical models.
Lessons from stress based depression paradigms and
neurogenesis
One prediction drawn from the neurogenic hy-
pothesis of MDD is that behavioral manipulations
that produce depressed behavior in animals should
produce a concomitant decrease in neurogenesis.
Several methods have been used to produce de-
pression-like behavior in rodents, all involving the
exposure to inescapable stress. One such procedure
is LH, originally developed by Seligman and col-
leagues (Overmier and Seligman, 1967; Seligman
and Beagley, 1975). LH is a relatively short-term
procedure in which animals are exposed to ines-
capable shock inside a conditioning chamber.
Subjects are then given an escape/avoidance task
in which shock is controllable (e.g., shuttling from
one side of the chamber to the other cancels or
terminates the shock). Exposure to inescapable
shock impairs subsequent acquisition of the es-
cape/avoidance task (relative to naı̈ve subjects),
arguably because the animal has learned it is help-
less (Willner and Mitchell, 2002). LH training also
reduces cell proliferation in the DG. Treatment
with AD drugs alleviates both the behavioral help-
lessness (Malberg and Duman, 2003; Chourbaji et
al., 2005) and the reduction in cell proliferation
(Malberg and Duman, 2003). However, there are
several reasons why reductions in proliferation are
not a likely mechanism of the behavioral helpless-
ness. First, behavioral helplessness manifests im-
mediately with exposure to inescapable shock, and
it is unlikely that a reduction in proliferation could
so rapidly give rise to a behavioral effect. If the
reduction in proliferation were to impact behavior,
the effects would more likely manifest 1–3 weeks
after the onset of the reduction in proliferation, at
the time when the newborn cells would be becom-
ing functionally integrated into DG circuits.
Similarly, acute dosing with AD drugs is sufficient
to alleviate behavioral helplessness (Malberg and
Duman, 2003; Chourbaji et al., 2005), but the
neurogenic effect of these drugs requires chronic
treatment (Malberg et al., 2000). Finally, a recent
study has demonstrated that LH training in rats
reduces cell proliferation in all subjects, but only a
subset of subjects display behavioral helplessness
(Vollmayr et al., 2003). One unexplored possibility
invoked by these studies is that the extant popu-
lation of immature neurons, rather than the pro-
liferative pool, is a substrate for the depressogenic
effects of stress. Studies are underway to test this
hypothesis.
Neurogenesis can more plausibly be linked to
the effects of chronic exposure to stress, which
have been studied extensively in animals. This
work typically involves exposing animals to vari-
ety of mild stressors over a period of several weeks.
Stressors include food and water deprivation, tem-
perature changes, restraint and tail suspension
(Strekalova et al., 2004; Willner, 2005; Mineur
et al., 2006). There are variations in the methods
used for chronic stress in rats (Willner et al., 1987,
1992; Willner, 2005) and mice (Mineur et al., 2003,
2006; Strekalova et al., 2004). The effects of

chronic stress include a reduction in sucrose pref-
erence, which has been interpreted as anhedonia
(Willner et al., 1987; Strekalova et al., 2004), al-
terations in sleep–wake cycle, reduced sexual and
self-care behavior, and increases in anxiety-like
behavior in traditional tests such as the elevated-
plus maze (Willner, 2005). Unlike the effects of
LH, the behavioral effects of chronic stress are
ameliorated by chronic but not acute treatment
with ADs (Willner et al., 1987; Yalcin et al., 2005),
suggesting that chronic stress may be a better
model of depression because it captures the ther-
apeutic lag seen in human patients. An abundance
of research demonstrates that hippocampal neuro-
genesis is reduced by a variety of chronic stress
procedures (for review, see Duman, 2004), and this
reduction in neurogenesis is blocked by chronic
treatment with AD drugs (Alonso et al., 2004).
Moreover, a recent study has shown that among
rats exposed to chronic stress, only a subset re-
spond behaviorally to SSRI treatment (Jayatissa
et al., 2006). Interestingly, neurogenesis was re-
stored to normal levels only in the behaviorally
identified responders.
Does blockade of neurogenesis produce symptoms of
depression?
The animal research on chronic stress evidences an
intriguing correlation between the rate of neuro-
genesis and emotional state. Chronic stress pro-
duces a behavioral phenotype analogous to aspects
of depression while reducing neurogenesis. AD
drugs restore neurogenesis and ameliorate the be-
havioral phenotype. Does this correlation reflect
that neurogenesis is a causal mechanism for these
behavioral changes? Or is neurogenesis simply a
marker for changes in other biological pathways
or network activity?
We have begun to address this question exper-
imentally by blocking hippocampal neurogenesis
and then examining the behavioral consequences.
One method of arresting neurogenesis is to subject
the brain to low doses of X-irradiation. Irradiation
kills mitotic cells such that neurogenesis is arrested
virtually completely (Monje et al., 2002, 2003;
Wojtowicz, 2006). In our laboratory, a shield is
711
used to target X-rays specifically over the hippo-
campus, while protecting regions anterior and
posterior, including the subventricular zone.
Blocking neurogenesis with this procedure has no
effect in a number of relevant behavioral tasks,
including the novelty-suppressed feeding test
(Santarelli et al., 2003) and the novelty-induced
hypophagia test (our unpublished data). Both of
these tests are AD screens that measure the latency
of a mouse to venture into the center of an open
field or novel context to obtain food. Latency-to-
feed is decreased by chronic AD drug treatment
and acute anxiolytic treatment, but not by acute
AD drug treatment (Bodnoff et al., 1989; Dulawa
et al., 2004). Irradiation does not affect behavior in
two traditional anxiety tests, the elevated-plus
maze and light-dark choice test (our unpublished
data), nor does it increase the susceptibility of mice
to the effects of chronic stress (Santarelli et al.,
2003).
We have also examined some of these behaviors
in a transgenic mouse line in which neuronal pro-
liferation can be blocked conditionally. The mouse
line expresses herpes-simplex virus thymidine kin-
ase (HSV-TK) under control of the GFAP pro-
moter (GFAP-TK) (Garcia et al., 2004b). Mitotic
cells expressing HSV-TK are killed by the antiviral
drug ganciclovir. Thus, in this mouse, the dividing
neuronal progenitors, which express GFAP, are
killed after ganciclovir is administered, and as
a result, neurogenesis is reduced to low levels
(Garcia et al., 2004b; Saxe et al., 2006). As with
irradiation, blocking neurogenesis in this mouse
line had no effect on anxiety-like behavior in sev-
eral tests, including the open field, NSF test, or
light-dark choice test (Saxe et al., 2006). Thus, we
have not found any evidence that blocking neuro-
genesis in adult mice produces a depression-like
phenotype.
A role for hippocampal neurogenesis in learning
The only domain in which direct behavioral effects
of arresting neurogenesis have been reported is in
learning and memory. Indeed, the very first evi-
dence for a behavioral function of neurogenesis in
mammals came from studies of learning and
712
memory. These studies showed that participating
in a hippocampus-dependent classical condition-
ing procedure (trace eyeblink conditioning) en-
hances the survival of newborn neurons in the
SGZ (Gould et al., 1999a) and that reducing ne-
urogenesis with the anti-mitotic compound me-
thylazoxymethanol acetate (MAM) impairs
acquisition of the same task (Shors et al., 2001).
The use of MAM is encumbered by deleterious
side effects and moreover, it does not completely
block neurogenesis (Dupret et al., 2005). Subse-
quent experiments using more targeted methods
for arresting neurogenesis have confirmed that ne-
urogenesis is required for some hippocampus-
dependent learning tasks (Snyder et al., 2005;
Saxe et al., 2006; Winocur et al., 2006). A recent
study from our laboratory has also revealed a
paradoxical role for hippocampal neurogenesis
in a hippocampus-dependent working memory
paradigm, where ablation of newly generated
neurons results in improved performance (Saxe
et al., 2007).
Perhaps the most compelling of these findings is
the requirement of neurogenesis for contextual
fear conditioning, which has been demonstrated in
two species (rats and mice) using two different
methods for arresting neurogenesis (Saxe et al.,
2006; Winocur et al., 2006). Contextual fear con-
ditioning is a form of Pavlovian conditioning pro-
duced by pairing a distinctive context (spatial
location) with footshock. As a result of the pair-
ing, animals exhibit characteristic fear responses
(e.g., freezing, defecation, potentiation of the star-
tle response) when re-exposed to the context.
Lesions to the hippocampus often impair this
form of learning (Phillips and LeDoux, 1992;
Matus-Amat et al., 2004), presumably because the
hippocampus participates in mnemonic encoding
of the spatial context. Blocking neurogenesis prior
to training in this procedure reduces the amount of
the contextual fear expressed when rodents are re-
exposed to the training context. Importantly,
blocking neurogenesis does not impair fear condi-
tioning to a discrete tone stimulus, indicating that
shock sensitivity and motor control of the fear
response are not impaired. The impairment of
contextual fear conditioning may thus reflect a
requirement of neurogenesis for the encoding of
novel contexts and/or for assigning emotional va-
lence to contexts.
How (and whether) these learning impairments
relate to depression is unclear. Cognitive impair-
ments have been reported in depression, but cog-
nitive impairment is not a cardinal feature of the
disease, in contrast to some other psychiatric ill-
nesses, namely schizophrenia. Still, it is certainly
the case that depressed patients have an impaired
ability to ‘‘contextualize’’ negative emotions, in
that these emotions are overgeneralized across ex-
periences and situations. Much more research into
the putative role of neurogenesis in contextual
learning will be needed to determine whether re-
duced neurogenesis could give rise to these features
of depression.
Neurogenesis is required for some behavioral
effects of AD drugs
Although animal models have not provided evi-
dence that reduced neurogenesis causes depres-
sion-like symptoms, these models have produced
evidence that neurogenesis is involved in the ther-
apeutic effects of AD drugs. Three recent studies
have used the targeted irradiation procedure de-
scribed above to test whether DG neurogenesis is
required for the behavioral effects of AD drugs in
rodent models. A study conducted in our labora-
tory demonstrated that neurogenesis is required
for the effects of both imipramine, a classic tricy-
clic AD, and fluoxetine in two mouse behavioral
screens for AD activity (Santarelli et al., 2003), the
NSF test and a chronic stress procedure. Impor-
tantly, this study has been replicated in our lab-
oratory using the GFAP-TK mice (unpublished
results). In a separate series of experiments con-
ducted by another laboratory, the synthetic can-
nabinoid HU210 was shown to have AD-like
effects in the NSF paradigm following 10 days of
treatment, and, interestingly, this effect was
blocked by X-ray irradiation (Jiang et al., 2005).
In addition, there is a recent preliminary report
using rats (Meltzer et al., 2006) that irradiation
blocks the behavioral effects of fluoxetine in the

forced swim test. Thus, a neurogenic dependence
for the behavioral effects of ADs has been revealed
for three different drugs in three different AD
screens and using two different ways to ablate ne-
urogenesis.
The above work suggests that neurogenesis may
be a critical substrate for AD efficacy. However,
an important limitation of this work is the reliance
on a very limited number of animal models and
AD treatments. It is unlikely that the three be-
havioral assays used in these experiments capture
all the clinically relevant features of AD treat-
ments, and consequently it remains possible that
some clinically important features of these treat-
ments are neurogenesis-independent. Indeed, two
more recent studies from our lab confirm that this
caveat is valid.
One study from our laboratory (Holick et al.,
2007) examined the effects of AD drugs on be-
havior and DG neurogenesis in the Balb-c mouse
strain, a strain that exhibits high anxiety in be-
havioral tests. In this strain, chronic fluoxetine
treatment reduced anxiety-like and depressive be-
havior in the novelty-induced hypophagia and
forced-swim paradigms but failed to increase neu-
ronal proliferation. Not surprisingly, the behavi-
oral effects of these drugs are not blocked by
irradiation in this strain. A second study found
that the anxiolytic effects of environmental en-
richment do not require neurogenesis (Meshi et al.,
2006).
In sum, these studies suggest that AD-like
effects can be achieved through at least two differ-
ent pathways, one that is neurogenesis-dependent
and one that is neurogenesis-independent. AD
drugs and cannabinoids appear to use a neuro-
genesis-dependent pathway in some circumstances
but not in others; chronic stress may be one factor
that governs this dichotomy. Enrichment appears
to use a neurogenesis-independent pathway either
alone or in combination with the neurogenesis-
dependent pathway. An important outstanding
question not addressed by these experiments is
whether the upregulation of neurogenesis is suffi-
cient to produce AD effects. Testing this hypoth-
esis will require new methods for very specifically
upregulating aspects of neurogenesis.
713
Summary
General insights: is neurogenesis a missing link or
is the link still missing?
Research on MDD in the last decade has led to
considerable maturation of our understanding of
how different neural circuits function in a normal
brain and in the context of pathology. Several no-
table findings have emerged from studies in hu-
mans with MDD and preclinical models of MDD.
First, the AD response and the pathogenesis
of MDD may have different neural substrates.
Second, the pathogenic mechanisms may differ
from those that underlie the pathophysiology of
MDD. Third, any model explaining the etiology of
MDD must incorporate genetic vulnerability,
stressors, critical periods and multiple neural cir-
cuits. In other words, MDD is more likely than
not, a result of multiple ‘‘hits’’ to the brain (deficits
in multiple neural substrates). This last finding
underscores the need for more refined preclinical
models for depression. In this section, we revisit
the neurogenic hypothesis of MDD with attention
to the evidence discussed thus far in the context of
etiology, pathophysiology and treatment.
The neurogenic hypothesis and the etiology of MDD
Only recently the neurogenic hypothesis of MDD
joined the neurotrophic and network hypotheses
as a candidate to explain the neurobiological basis
of MDD.
Unlike the neurotrophic and network hypothe-
sis, however, the neurogenic hypothesis implicates
only one brain region, the dentate gyrus, as the
primary neural substrate for the etiology of MDD
and the AD response.
As this review makes clear, the evidence for ne-
urogenesis as an etiological factor in MDD is scarce
at best. Pathological analyses of postmortem tissue
obtained from patients with MDD have not re-
vealed alterations in the size of the dentate gyrus,
decreased number of proliferative cells, or changes
in the degree of ongoing apoptosis. The data on
apoptosis do not indicate the age of neurons that
714
Fig. 2. A model to evaluate the role of neurogenesis as a substrate in the etiology, pathophysiology and treatment of MDD. MDD is
likely to arise from the synergistic effects of stress, biological vulnerability conferred by risk alleles and deficits in multiple neural
circuits. Whether hippocampal neurogenesis is involved in the etiology of MDD is at present unclear. Altered hippocampal neuro-
genesis may result as a consequence of pathogenic mechanisms and contribute to the pathophysiology of MDD. The treatment of
some, but not all, symptoms of MDD may rely on neurogenesis.
are dying and therefore, preclude an assessment of
changes in rates of turnover or survival. As noted
earlier in this chapter, more studies on postmortem
tissue obtained from non-medicated patients are
required to conclusively characterize dentate gyrus
neurogenesis in depressed individuals. Neverthe-
less, these data establish that changes in hippo-
campal volume are unlikely to result from changes
in hippocampal neurogenesis. Preclinical studies
have yielded data more directly controverting the
role of neurogenesis in the etiology of MDD: ab-
lation of neurogenesis in adult otherwise normal
animals does not engender a depression-like phe-
notype. However, there are some important cave-
ats to these preclinical studies. It is almost
certainly the case that MDD involves the simul-
taneous presence of multiple risk factors or ‘‘hits’’
(Fig. 2). If this were the case, ablation of neuro-
genesis in wild-type adult animals would not be
sufficient to elicit a depression-like phenotype. The
ablation of neurogenesis might create a diathesis
for depression that would only be revealed in the
presence of other genetic or environmental insults.
Moreover, the timing with respect to when the
brain experiences an insult, whether it be genetic
or environmental, is also critical. Clearly, more
studies are needed that incorporate the effects of
stressors and assess the consequences of altering
neurogenesis during the early postnatal period in
rodents with genetic backgrounds that harbor
different risk alleles. These studies will reveal
whether changes in neurogenesis lend a diathesis
for MDD, and inform us about the complex in-
terplay between multiple neural circuits in
directing the emotional trajectory. Finally, longi-
tudinal neuroimaging studies in humans are
needed to reveal how the hippocampal landscape
changes with the appearance of the first symptoms
of MDD and over time.
The neurogenic hypothesis and the pathophysiology
of MDD
The increasingly appreciated role for neurogenesis
in hippocampus-dependent learning as defined by
work from several different laboratories using ro-
dents provides evidence that neurogenesis makes a
functionally significant contribution to hippocam-
pal circuits. It is thus plausible that the putative
reductions in neurogenesis associated with MDD
have important psychological implications. These
could include cognitive deficits, which are a pos-
sible mechanism of the emotional symptoms of the
disease (Beck, 2005). In addition, it is now appre-
ciated that the hippocampus, particularly its
ventral (anterior, in humans) extent, has a central
role in emotional regulation. The development of
higher resolution neuroimaging techniques will
enable us to visualize changes in dentate gyrus ac-
tivity in patients with MDD and during learning.
The use of novel genetic approaches to selectively
manipulate the maturation of newborn neurons
influence their survival, and drive turnover of ma-
ture dentate granule cells will undoubtedly en-
hance our understanding of how neurogenesis
contributes to dentate function and to the deficits
seen in MDD.

The neurogenic hypothesis and the treatment of
MDD
The finding that neurogenesis is one mechanism
used by ADs to exert their behavioral effects has
now been repeated by several different laborato-
ries using rodents. This is in striking contrast to
the absence of data implicating neurogenesis in the
pathogenesis of MDD. One interpretation of this
apparent paradox (besides the limitations of cur-
rent preclinical models highlighted in section ‘‘The
neurogenic hypothesis and the etiology of MDD’’)
is that the etiology and treatment of MDD may
have different neural substrates. Such a dissocia-
tion is suggested by a recent association study that
looked at 21 candidate polymorphisms and
showed that the genetic basis for the capacity to
respond to monoamine-based ADs differed from
that of susceptibility to MDD (Garriock et al.,
2006). An interesting parallel was demonstrated by
preclinical studies on BDNF and depression,
which showed that increased BDNF signaling in
the hippocampus is sufficient to induce AD-like
effects, but genetic ablation of BDNF on its own
does not elicit a depression-like phenotype
(Duman and Monteggia, 2006). While the data
from preclinical studies and ADs are encouraging,
more work is needed to solidify the requirement
for neurogenesis in mediating the behavioral
effects of ADs. Conspicuously absent from the
roster of experiments are those that show that in-
creased neurogenesis is sufficient for the behavioral
effects of ADs. Experiments along these lines using
genetic approaches to specifically increase the
number of newly generated neurons are currently
underway in our laboratory. Moreover, given the
pleiotropic effects of ADs on neurogenesis, it at
present unclear whether immature neurons or the
adult-generated mature dentate granule cells are
required to induce behavioral change. Inducible
genetic approaches using promoters specific to
these different cell types will allow for cell type
specific manipulations to unequivocally identify
the cellular substrates and mechanisms underlying
the neurogenic-dependent AD response.
In interpreting the data on the neurogenic de-
pendence of ADs, we must remind ourselves of the
715
potentially different ways by which ADs may rely
on neurogenesis for their behavioral effects (Fig. 1,
Drew and Hen, in press). It is plausible that
the short-term effects of ADs, for example, are
mediated by ADs acting on the extant reservoir
of adult-generated immature neurons or by accel-
erating the maturation process of an extant pop-
ulation of adult-generated immature neurons with
concomitant replacement/addition to the dentate
gyrus. Likewise, the rapid effects of ADs in re-
versing behavioral changes induced by stress in
paradigms such as LH discussed earlier could also
be mediated through the extant population of im-
mature neurons or through adult-generated ma-
ture dentate granule cells. In this regard, it is
worthy to note that inducing BDNF and CREB
expression in the DG but not other hippocampal
subfields (or other brain regions) is sufficient
to elicit an AD response (Chen et al., 2001;
Shirayama et al., 2002; Duman and Monteggia,
2006). Thus, the DG is an attractive neural
substrate for the AD response and these studies
highlight a previously unrecognized function for
the dentate gyrus. Given our present understand-
ing of DG function (see chapter by Kesner in
this volume), it is unclear how enhancement in
processes such as pattern separation and conjunc-
tive encoding contribute to AD-like behavior
assessed in these depression paradigms. Whether
cognitive behavior therapy, an endorsed line of
treatment often used in parallel with AD drugs,
increases activity in the DG is yet to be deter-
mined.
In conclusion, there is much work to be done on
hippocampal neurogenesis to ascertain its role in
the brain and still much more to establish a role
for it in MDD.
The convergence of insights from preclinical
studies and neuroimaging studies and identifi
cation of novel genetic risk alleles will undoubt-
edly help establish whether the neurogenic
hypothesis can explain facets of this complex
and debilitating psychiatric disorder. At the
very least, the neurogenic hypothesis, like any
elegant hypothesis, has succeeded in generating
the momentum needed to rigorously test its
tenets.
716
Acknowledgments
The authors would like to thank members of the
Hen laboratory for helpful discussions. Funding
support was provided by NARSAD (A.S, M.R.D
and R.H) and by Charles H. Revson Foundation
Senior Fellowship in Biomedical Science (M.R.D).
References
Aberg, M.A., Aberg, N.D., Hedbacker, H., Oscarsson, J. and
Eriksson, P.S. (2000) Peripheral infusion of IGF-I selectively
induces neurogenesis in the adult rat hippocampus. J. Ne-
urosci., 20: 2896–2903.
Alonso, R., Griebel, G., Pavone, G., Stemmelin, J., Le Fur, G.
and Soubrie, P. (2004) Blockade of CRF(1) or V(1b) recep-
tors reverses stress-induced suppression of neurogenesis in a
mouse model of depression. Mol. Psychiatry, 9: 278–286 224.
Altman, J. and Das, G.D. (1965) Autoradiographic and histo-
logical evidence of postnatal hippocampal neurogenesis in
rats. J. Comp. Neurol., 124: 319–335.
Amrein, I., Slomianka, L. and Lipp, H.P. (2004) Granule cell
number, cell death and cell proliferation in the dentate gyrus
of wild-living rodents. Eur. J. Neurosci., 20: 3342–3350.
American Psychiatric Association (2000) Diagnostic and Sta-
tistical Manual of Mental Disorders. Fourth Edition—Text
Revision (DSMIV-TR).
Austin, M.P., Mitchell, P. and Goodwin, G.M. (2001) Cogni-
tive deficits in depression: possible implications for functional
neuropathology. Br. J. Psychiatry, 178: 200–206.
Azmitia, E.C. and Segal, M. (1978) An autoradiographic anal-
ysis of the differential ascending projections of the dorsal and
median raphe nuclei in the rat. J. Comp. Neurol., 179:
641–667.
Babyak, M., Blumenthal, J.A., Herman, S., Khatri, P., Dorai-
swamy, M., Moore, K., Craighead, W.E., Baldewicz, T.T.
and Krishnan, K.R. (2000) Exercise treatment for major de-
pression: maintenance of therapeutic benefit at 10 months.
Psychosom. Med., 62: 633–638.
Banasr, M., Hery, M., Printemps, R. and Daszuta, A. (2004)
Serotonin-induced increases in adult cell proliferation and
neurogenesis are mediated through different and common 5-
HT receptor subtypes in the dentate gyrus and the subven-
tricular zone. Neuropsychopharmacology, 29: 450–460.
Bannerman, D.M., Grubb, M., Deacon, R.M., Yee, B.K., Fel-
don, J. and Rawlins, J.N. (2003) Ventral hippocampal lesions
affect anxiety but not spatial learning. Behav. Brain Res.,
139: 197–213.
Bayer, S.A., Yackel, J.W. and Puri, P.S. (1982) Neurons in the
rat dentate gyrus granular layer substantially increase during
juvenile and adult life. Science, 216: 890–892.
Beck, A.T. (2005) The current state of cognitive therapy: a
40-year retrospective. Arch. Gen. Psychiatry, 62: 953–959.
Becker, S. (2005) A computational principle for hippocampal
learning and neurogenesis. Hippocampus, 15: 722–738.
Bodnoff, S.R., Suranyi-Cadotte, B., Quirion, R. and Meaney,
M.J. (1989) A comparison of the effects of diazepam versus
several typical and atypical anti-depressant drugs in an an-
imal model of anxiety. Psychopharmacology (Berl.), 97:
277–279.
Bratt, A.M., Kelley, S.P., Knowles, J.P., Barrett, J., Davis, K.,
Davis, M. and Mittleman, G. (2001) Long term modulation
of the HPA axis by the hippocampus. Behavioral, biochem-
ical and immunological endpoints in rats exposed to chronic
mild stress. Psychoneuroendocrinology, 26: 121–145.
Bremner, J.D., Narayan, M., Anderson, E.R., Staib, L.H.,
Miller, H.L. and Charney, D.S. (2000) Hippocampal volume
reduction in major depression. Am. J. Psychiatry, 157:
115–118.
Brezun, J.M. and Daszuta, A. (1999) Depletion in serotonin
decreases neurogenesis in the dentate gyrus and the subven-
tricular zone of adult rats. Neuroscience, 89: 999–1002.
Brezun, J.M. and Daszuta, A. (2000) Serotonin may stimulate
granule cell proliferation in the adult hippocampus, as ob-
served in rats grafted with foetal raphe neurons. Eur. J. Ne-
urosci., 12: 391–396.
Bunney Jr., W.E. and Davis, J.M. (1965) Norepinephrine in
depressive reactions. A review. Arch. Gen. Psychiatry, 13:
483–494.
Cameron, H.A. and Gould, E. (1994) Adult neurogenesis is
regulated by adrenal steroids in the dentate gyrus. Neurosci-
ence, 61: 203–209.
Cameron, H.A. and McKay, R.D. (1999) Restoring production
of hippocampal neurons in old age. Nat. Neurosci., 2:
894–897.
Cameron, H.A. and McKay, R.D. (2001) Adult neurogenesis
produces a large pool of new granule cells in the dentate
gyrus. J. Comp. Neurol., 435: 406–417.
Cameron, H.A., Woolley, C.S. and Gould, E. (1993) Adrenal
steroid receptor immunoreactivity in cells born in the adult
rat dentate gyrus. Brain Res., 611: 342–346.
Campbell, S., Marriott, M., Nahmias, C. and Macqueen, G.M.
(2004) Lower hippocampal volume in patients suffering
from depression: a meta-analysis. Am. J. Psychiatry, 161:
598–607.
Carlezon Jr., W.A., Duman, R.S. and Nestler, E.J. (2005) The
many faces of CREB. Trends Neurosci., 28: 436–445.
Carroll, B.J., Martin, F.I. and Davies, B. (1968a) Pituitary-
adrenal function in depression. Lancet, 1: 1373–1374.
Carroll, B.J., Martin, F.I. and Davies, B. (1968b) Resistance to
suppression by dexamethasone of plasma 11-O. H.C.S. levels
in severe depressive illness. Br. Med. J., 3: 285–287.
Caspi, A. and Moffitt, T.E. (2006) Gene-environment interac-
tions in psychiatry: joining forces with neuroscience. Nat.
Rev. Neurosci., 7: 583–590.
Caspi, A., Sugden, K., Moffitt, T.E., Taylor, A., Craig, I.W.,
Harrington, H., Mcclay, J., Mill, J., Martin, J., Braithwaite,
A. and Poulton, R. (2003) Influence of life stress on depres-
sion: moderation by a polymorphism in the 5-HTT gene.
Science, 301: 386–389.

Castren, E. (2005) Is mood chemistry? Nat. Rev. Neurosci., 6:
241–246.
Chambers, R.A., Potenza, M.N., Hoffman, R.E. and Miranker,
W. (2004) Simulated apoptosis/neurogenesis regulates learn-
ing and memory capabilities of adaptive neural networks.
Neuropsychopharmacology, 29: 747–758.
Chapman, D.P., Whitfield, C.L., Felitti, V.J., Dube, S.R., Ed-
wards, V.J. and Anda, R.F. (2004) Adverse childhood expe-
riences and the risk of depressive disorders in adulthood. J.
Affect. Disord., 82: 217–225.
Chen, A.C., Shirayama, Y., Shin, K.H., Neve, R.L. and
Duman, R.S. (2001) Expression of the cAMP response ele-
ment binding protein (CREB) in hippocampus produces an
antidepressant effect. Biol. Psychiatry, 49: 753–762.
Chen, G., Rajkowska, G., Du, F., Seraji-Bozorgzad, N. and
Manji, H.K. (2000) Enhancement of hippocampal neurogen-
esis by lithium. J. Neurochem., 75: 1729–1734.
Chourbaji, S., Zacher, C., Sanchis-Segura, C., Dormann, C.,
Vollmayr, B. and Gass, P. (2005) Learned helplessness: va-
lidity and reliability of depressive-like states in mice. Brain
Res. Brain Res. Protoc., 16: 70–78.
Cryan, J.F., Markou, A. and Lucki, I. (2002) Assessing anti-
depressant activity in rodents: recent developments and fu-
ture needs. Trends Pharmacol. Sci., 23: 238–245.
Czeh, B., Michaelis, T., Watanabe, T., Frahm, J., de Biurrun,
G., van Kampen, M., Bartolomucci, A. and Fuchs, E. (2001)
Stress-induced changes in cerebral metabolites, hippocampal
volume, and cell proliferation are prevented by antidepres-
sant treatment with tianeptine. Proc. Natl. Acad. Sci. U.S.A.,
98: 12796–12801.
Czeh, B., Welt, T., Fischer, A.K., Erhardt, A., Schmitt, W.,
Muller, M.B., Toschi, N., Fuchs, E. and Keck, M.E. (2002)
Chronic psychosocial stress and concomitant repetitive trans-
cranial magnetic stimulation: effects on stress hormone levels
and adult hippocampal neurogenesis. Biol. Psychiatry, 52:
1057–1065.
Drevets, W.C., Bogers, W. and Raichle, M.E. (2002) Func-
tional anatomical correlates of antidepressant drug treatment
assessed using PET measures of regional glucose metabolism.
Eur. Neuropsychopharmacol., 12: 527–544.
Drew, M.R. and Hen, R. (in press) Adult hippocampal neuro-
genesis as a target for the treatment of depression. CNS Ne-
urol. Disord. Drug Targets.
Dulawa, S.C., Holick, K.A., Gundersen, B. and Hen, R. (2004)
Effects of chronic fluoxetine in animal models of anxiety and
depression. Neuropsychopharmacology, 29: 1321–1330.
Duman, R.S. (2002) Structural alterations in depression: cellu-
lar mechanisms underlying pathology and treatment of mood
disorders. CNS Spectrosc., 7: 140–142 144–147.
Duman, R.S. (2004) Depression: a case of neuronal life and
death? Biol. Psychiatry, 56: 140–145.
Duman, R.S., Heninger, G.R. and Nestler, E.J. (1997) A mo-
lecular and cellular theory of depression. Arch. Gen. Psychi-
atry, 54: 597–606.
Duman, R.S., Malberg, J., Nakagawa, S. and D’sa, C. (2000)
Neuronal plasticity and survival in mood disorders. Biol.
Psychiatry, 48: 732–739.
717
Duman, R.S. and Monteggia, L.M. (2006) A neurotrophic
model for stress-related mood disorders. Biol. Psychiatry, 59:
1116–1127.
Dunn, J.D. and Orr, S.E. (1984) Differential plasma corticos-
terone responses to hippocampal stimulation. Exp. Brain
Res., 54: 1–6.
Dupret, D., Montaron, M.F., Drapeau, E., Aurousseau, C., Le
Moal, M., Piazza, P.V. and Abrous, D.N. (2005) Methyl-
azoxymethanol acetate does not fully block cell genesis in the
young and aged dentate gyrus. Eur. J. Neurosci., 22:
778–783.
Eisch, A.J. (2002) Adult neurogenesis: implications for psychi-
atry. Prog. Brain Res., 138: 315–342.
Encinas, J.M., Vaahtokari, A. and Enikolopov, G. (2006) Flu-
oxetine targets early progenitor cells in the adult brain. Proc.
Natl. Acad. Sci. U.S.A., 103: 8233–8238.
Eriksson, P.S., Perfilieva, E., Bjork-Eriksson, T., Alborn, A.M.,
Nordborg, C., Peterson, D.A. and Gage, F.H. (1998) Ne-
urogenesis in the adult human hippocampus. Nat. Med., 4:
1313–1317.
Esposito, M.S., Piatti, V.C., Laplagne, D.A., Morgenstern,
N.A., Ferrari, C.C., Pitossi, F.J. and Schinder, A.F. (2005)
Neuronal differentiation in the adult hippocampus recapit-
ulates embryonic development. J. Neurosci., 25:
10074–10086.
Faber, K.M. and Haring, J.H. (1999) Synaptogenesis in the
postnatal rat fascia dentata is influenced by 5-HT1a receptor
activation. Brain Res. Dev. Brain Res., 114: 245–252.
Feldman, S. and Weidenfeld, J. (1993) The dorsal hippocampus
modifies the negative feedback effect of glucocorticoids on
the adrenocortical and median eminence CRF-41 responses
to photic stimulation. Brain Res., 614: 227–232.
Feldman, S. and Weidenfeld, J. (1999) Glucocorticoid receptor
antagonists in the hippocampus modify the negative feedback
following neural stimuli. Brain Res., 821: 33–37.
Felitti, V.J., Anda, R.F., Nordenberg, D., Williamson, D.F.,
Spitz, A.M., Edwards, V., Koss, M.P. and Marks, J.S. (1998)
Relationship of childhood abuse and household dysfunction
to many of the leading causes of death in adults. The Adverse
Childhood Experiences (ACE) Study. Am. J. Prev. Med., 14:
245–258.
Filippov, V., Kronenberg, G., Pivneva, T., Reuter, K., Steiner,
B., Wang, L.P., Yamaguchi, M., Kettenmann, H. and Kem-
permann, G. (2003) Subpopulation of nestin-expressing pro-
genitor cells in the adult murine hippocampus shows
electrophysiological and morphological characteristics of as-
trocytes. Mol. Cell Neurosci., 23: 373–382.
van der Flier, W.M., van Buchem, M.A., Weverling-Rijnsburg-
er, A.W., Mutsaers, E.R., Bollen, E.L., Admiraal-Behloul,
F., Westendorp, R.G. and Middelkoop, H.A. (2004) Mem-
ory complaints in patients with normal cognition are asso-
ciated with smaller hippocampal volumes. J. Neurol., 251:
671–675.
Fossati, P., Coyette, F., Ergis, A.M. and Allilaire, J.F.
(2002) Influence of age and executive functioning on verbal
memory of inpatients with depression. J. Affect. Disord., 68:
261–271.
718
Freund, T.F., Gulyas, A.I., Acsady, L., Gorcs, T. and Toth, K.
(1990) Serotonergic control of the hippocampus via local in-
hibitory interneurons. Proc. Natl. Acad. Sci. U.S.A., 87:
8501–8505.
Garcia, A., Steiner, B., Kronenberg, G., Bick-Sander, A. and
Kempermann, G. (2004a) Age-dependent expression of
glucocorticoid- and mineralocorticoid receptors on neural
precursor cell populations in the adult murine hippocampus.
Aging Cell, 3: 363–371.
Garcia, A.D., Doan, N.B., Imura, T., Bush, T.G. and So-
froniew, M.V. (2004b) GFAP-expressing progenitors are the
principal source of constitutive neurogenesis in adult mouse
forebrain. Nat. Neurosci., 7: 1233–1241.
Garriock, H.A., Delgado, P., Kling, M.A., Carpenter, L.L.,
Burke, M., Burke, W.J., Schwartz, T., Marangell, L.B.,
Husain, M., Erickson, R.P. and Moreno, F.A. (2006)
Number of risk genotypes is a risk factor for major depres-
sive disorder: a case control study. Behav. Brain Funct., 2:
24.
Gilbertson, M.W., Shenton, M.E., Ciszewski, A., Kasai, K.,
Lasko, N.B., Orr, S.P. and Pitman, R.K. (2002) Smaller
hippocampal volume predicts pathologic vulnerability to
psychological trauma. Nat. Neurosci., 5: 1242–1247.
Gold, P.W. and Chrousos, G.P. (2002) Organization of the
stress system and its dysregulation in melancholic and atyp-
ical depression: high vs. low CRH/NE states. Mol. Psychi-
atry, 7: 254–275.
Gould, E., Beylin, A., Tanapat, P., Reeves, A. and Shors, T.J.
(1999a) Learning enhances adult neurogenesis in the hippo-
campal formation. Nat. Neurosci., 2: 260–265.
Gould, E., Reeves, A.J., Fallah, M., Tanapat, P., Gross, C.G.
and Fuchs, E. (1999b) Hippocampal neurogenesis in adult
Old World primates. Proc. Natl. Acad. Sci. U.S.A., 96:
5263–5267.
Goursaud, A.P., Mendoza, S.P. and Capitanio, J.P. (2006) Do
neonatal bilateral ibotenic acid lesions of the hippocampal
formation or of the amygdala impair HPA axis responsive-
ness and regulation in infant rhesus macaques (Macaca mu-
latta)? Brain Res., 1071: 97–104.
von Gunten, A., Fox, N.C., Cipolotti, L. and Ron, M.A. (2000)
A volumetric study of hippocampus and amygdala in de-
pressed patients with subjective memory problems. J. Ne-
uropsychiatry Clin. Neurosci., 12: 493–498.
von Gunten, A. and Ron, M.A. (2004) Hippocampal volume
and subjective memory impairment in depressed patients.
Eur. Psychiatry, 19: 438–440.
Haydon, P.G. and Carmignoto, G. (2006) Astrocyte control of
synaptic transmission and neurovascular coupling. Physiol.
Rev., 86: 1009–1031.
Hellsten, J., Wennstrom, M., Mohapel, P., Ekdahl, C.T., Ben-
gzon, J. and Tingstrom, A. (2002) Electroconvulsive seizures
increase hippocampal neurogenesis after chronic corticoster-
one treatment. Eur. J. Neurosci., 16: 283–290.
Herman, J.P., Cullinan, W.E., Morano, M.I., Akil, H. and
Watson, S.J. (1995) Contribution of the ventral subiculum to
inhibitory regulation of the hypothalamo-pituitary-adreno-
cortical axis. J. Neuroendocrinol., 7: 475–482.
Herman, J.P., Cullinan, W.E., Young, E.A., Akil, H. and Wat-
son, S.J. (1992) Selective forebrain fiber tract lesions impli-
cate ventral hippocampal structures in tonic regulation of
paraventricular nucleus corticotropin-releasing hormone
(CRH) and arginine vasopressin (AVP) mRNA expression.
Brain Res., 592: 228–238.
Holick, K.A., Lee, D.C., Hen, R. and Dulawa, S.C. (2007)
Behavioral effects of chronic fluoxetine in BALB/cJ mice do
not require adult hippocampal neurogenesis or the serotonin
1A receptor. Neuropsychopharmacology.
Holsboer, F. and Barden, N. (1996) Antidepressants and hypo-
thalamic-pituitary-adrenocortical regulation. Endocr. Rev.,
17: 187–205.
Holsboer, F., Liebl, R. and Hofschuster, E. (1982) Repeated
dexamethasone suppression test during depressive illness.
Normalisation of test result compared with clinical improve-
ment. J. Affect. Disord., 4: 93–101.
Huot, R.L., Plotsky, P.M., Lenox, R.H. and Mcnamara, R.K.
(2002) Neonatal maternal separation reduces hippocampal
mossy fiber density in adult Long Evans rats. Brain Res., 950:
52–63.
Huot, R.L., Thrivikraman, K.V., Meaney, M.J. and Plotsky,
P.M. (2001) Development of adult ethanol preference and
anxiety as a consequence of neonatal maternal separation in
Long Evans rats and reversal with antidepressant treatment.
Psychopharmacology (Berl.), 158: 366–373.
Islam, A., Ayer-Lelievre, C., Heigenskold, C., Bogdanovic, N.,
Winblad, B. and Adem, A. (1998) Changes in IGF-1 recep-
tors in the hippocampus of adult rats after long-term adre-
nalectomy: receptor autoradiography and in situ
hybridization histochemistry. Brain Res., 797: 342–346.
Jacobs, B.L., Praag, H. and Gage, F.H. (2000) Adult brain
neurogenesis and psychiatry: a novel theory of depression.
Mol. Psychiatry, 5: 262–269.
Jayatissa, M.N., Bisgaard, C., Tingstrom, A., Papp, M. and
Wiborg, O. (2006) Hippocampal cytogenesis correlates to
escitalopram-mediated recovery in a chronic mild stress rat
model of depression. Neuropsychopharmacology, 31:
2395–2404.
Jiang, W., Zhang, Y., Xiao, L., van Cleemput, J., Ji, S.P., Bai,
G. and Zhang, X. (2005) Cannabinoids promote embryonic
and adult hippocampus neurogenesis and produce anxiolytic-
and antidepressant-like effects. J. Clin. Invest., 115:
3104–3116.
Joels, M. and van Riel, E. (2004) Mineralocorticoid and
glucocorticoid receptor-mediated effects on serotonergic
transmission in health and disease. Ann. N.Y. Acad. Sci.,
1032: 301–303.
Kempermann, G., Kuhn, H.G. and Gage, F.H. (1997) More
hippocampal neurons in adult mice living in an enriched en-
vironment. Nature, 386: 493–495.
Kempermann, G., Kuhn, H.G. and Gage, F.H. (1998) Expe-
rience-induced neurogenesis in the senescent dentate gyrus. J.
Neurosci., 18: 3206–3212.
Kendler, K.S., Kuhn, J.W., Vittum, J., Prescott, C.A. and
Riley, B. (2005) The interaction of stressful life events and a
serotonin transporter polymorphism in the prediction of

episodes of major depression: a replication. Arch. Gen. Psy-
chiatry, 62: 529–535.
Kendler, K.S., Thornton, L.M. and Gardner, C.O. (2001) Ge-
netic risk, number of previous depressive episodes, and
stressful life events in predicting onset of major depression.
Am. J. Psychiatry, 158: 582–586.
Kennedy, S.H., Evans, K.R., Kruger, S., Mayberg, H.S.,
Meyer, J.H., Mccann, S., Arifuzzman, A.I., Houle, S. and
Vaccarino, F.J. (2001) Changes in regional brain glucose
metabolism measured with positron emission tomography
after paroxetine treatment of major depression. Am. J. Psy-
chiatry, 158: 899–905.
Kimbrell, T.A., Ketter, T.A., George, M.S., Little, J.T., Ben-
son, B.E., Willis, M.W., Herscovitch, P. and Post, R.M.
(2002) Regional cerebral glucose utilization in patients with a
range of severities of unipolar depression. Biol. Psychiatry,
51: 237–252.
Knigge, K.M. (1961) Adrenocortical response to stress in rats
with lesions in hippocampus and amygdala. Proc. Soc. Exp.
Biol. Med., 108: 18–21.
Kohler, S.J., Jennings, V., Boklewski, J.L., Degrush, B.J.,
Williams, N.I., Rockcastle, N.J., Cameron, J.L. and Green-
ough, W.T. (2006) Maturation of new granule cells in the
dentate gyrus of adult macaque monkeys. Soc. Neurosci.
Abstr.
Kondoh, M., Shiga, T. and Okado, N. (2004) Regulation of
dendrite formation of Purkinje cells by serotonin through
serotonin1A and serotonin2A receptors in culture. Neurosci.
Res., 48: 101–109.
Kornack, D.R. and Rakic, P. (1999) Continuation of neuro-
genesis in the hippocampus of the adult macaque monkey.
Proc. Natl. Acad. Sci. U.S.A., 96: 5768–5773.
Kuhn, H.G., Dickinson-Anson, H. and Gage, F.H. (1996) Ne-
urogenesis in the dentate gyrus of the adult rat: age-related
decrease of neuronal progenitor proliferation. J. Neurosci.,
16: 2027–2033.
Kuhn, H.G., Winkler, J., Kempermann, G., Thal, L.J. and
Gage, F.H. (1997) Epidermal growth factor and fibroblast
growth factor-2 have different effects on neural progenitors
in the adult rat brain. J. Neurosci., 17: 5820–5829.
Kvachnina, E., Liu, G., Dityatev, A., Renner, U., Dumuis, A.,
Richter, D.W., Dityateva, G., Schachner, M., Voyno-
Yasenetskaya, T.A. and Ponimaskin, E.G. (2005) 5-HT7 re-
ceptor is coupled to G alpha subunits of heterotrimeric
G12-protein to regulate gene transcription and neuronal
morphology. J. Neurosci., 25: 7821–7830.
Laplagne, D.A., Esposito, M.S., Piatti, V.C., Morgenstern,
N.A., Zhao, C., van Praag, H., Gage, F.H. and Schinder,
A.F. (2006) Functional convergence of neurons generated
in the developing and adult hippocampus. PLoS Biol., 4:
e409.
Lauder, J.M. and Krebs, H. (1978) Serotonin as a differenti-
ation signal in early neurogenesis. Dev. Neurosci., 1: 15–30.
Leonardo, E.D. and Hen, R. (2006) Genetics of affective and
anxiety disorders. Annu. Rev. Psychol., 57: 117–137.
Levinson, D.F. (2006) The genetics of depression: a review.
Biol. Psychiatry, 60: 84–92.
719
Liu, S., Wang, J., Zhu, D., Fu, Y., Lukowiak, K. and Lu, Y.M.
(2003) Generation of functional inhibitory neurons in the
adult rat hippocampus. J. Neurosci., 23: 732–736.
Loo, C.K. and Mitchell, P.B. (2005) A review of the efficacy of
transcranial magnetic stimulation (TMS) treatment for de-
pression, and current and future strategies to optimize effi-
cacy. J. Affect. Disord., 88: 255–267.
Lovic, V., Gonzalez, A. and Fleming, A.S. (2001) Maternally
separated rats show deficits in maternal care in adulthood.
Dev. Psychobiol., 39: 19–33.
Lucassen, P.J., Muller, M.B., Holsboer, F., Bauer, J., Holtrop,
A., Wouda, J., Hoogendijk, W.J., De Kloet, E.R. and Swaab,
D.F. (2001) Hippocampal apoptosis in major depression is a
minor event and absent from subareas at risk for glucocorti-
coid overexposure. Am. J. Pathol., 158: 453–468.
Ma, D.K., Ming, G.L. and Song, H. (2005) Glial influences on
neural stem cell development: cellular niches for adult ne-
urogenesis. Curr. Opin. Neurobiol., 15: 514–520.
Macqueen, G.M., Campbell, S., Mcewen, B.S., Macdonald, K.,
Amano, S., Joffe, R.T., Nahmias, C. and Young, L.T. (2003)
Course of illness, hippocampal function, and hippocampal
volume in major depression. Proc. Natl. Acad. Sci. U.S.A.,
100: 1387–1392.
Madsen, T.M., Treschow, A., Bengzon, J., Bolwig, T.G.,
Lindvall, O. and Tingstrom, A. (2000) Increased neurogen-
esis in a model of electroconvulsive therapy. Biol. Psychiatry,
47: 1043–1049.
Malberg, J.E. and Duman, R.S. (2003) Cell proliferation in
adult hippocampus is decreased by inescapable stress: re-
versal by fluoxetine treatment. Neuropsychopharmacology,
28: 1562–1571.
Malberg, J.E., Eisch, A.J., Nestler, E.J. and Duman, R.S.
(2000) Chronic antidepressant treatment increases neuro-
genesis in adult rat hippocampus. J. Neurosci., 20:
9104–9110.
Malberg, J.E. and Schechter, L.E. (2005) Increasing hippocam-
pal neurogenesis: a novel mechanism for antidepressant
drugs. Curr. Pharm. Des., 11: 145–155.
Manev, H., Uz, T., Smalheiser, N.R. and Manev, R. (2001)
Antidepressants alter cell proliferation in the adult brain in
vivo and in neural cultures in vitro. Eur. J. Pharmacol., 411:
67–70.
Matus-Amat, P., Higgins, E.A., Barrientos, R.M. and Rudy,
J.W. (2004) The role of the dorsal hippocampus in the ac-
quisition and retrieval of context memory representations. J.
Neurosci., 24: 2431–2439.
Mayberg, H.S., Brannan, S.K., Tekell, J.L., Silva, J.A., Mah-
urin, R.K., Mcginnis, S. and Jerabek, P.A. (2000) Regional
metabolic effects of fluoxetine in major depression: serial
changes and relationship to clinical response. Biol. Psychia-
try, 48: 830–843.
Mcewen, B.S. (2005) Glucocorticoids, depression, and mood
disorders: structural remodeling in the brain. Metabolism, 54:
20–23.
Meltzer, L.A., Roy, M., Parente, V. and Deisseroth, K. (2006)
Hippocampal neurogenesis is required for lasting antidepres-
sant effects of fluoxetine. Soc. Neurosci. Abstr.
720
Meltzer, L.A., Yabaluri, R. and Deisseroth, K. (2005) A role
for circuit homeostasis in adult neurogenesis. Trends Neuro-
sci., 28: 653–660.
Meshi, D., Drew, M.R., Saxe, M., Ansorge, M.S., David, D.,
Santarelli, L., Malapani, C., Moore, H. and Hen, R. (2006)
Hippocampal neurogenesis is not required for behavioral
effects of environmental enrichment. Nat. Neurosci., 9:
729–731.
Mineur, Y.S., Belzung, C. and Crusio, W.E. (2006) Effects of
unpredictable chronic mild stress on anxiety and depression-
like behavior in mice. Behav. Brain Res., 175: 43–50.
Mineur, Y.S., Prasol, D.J., Belzung, C. and Crusio, W.E. (2003)
Agonistic behavior and unpredictable chronic mild stress in
mice. Behav. Genet., 33: 513–519.
Mirescu, C., Peters, J.D. and Gould, E. (2004) Early life ex-
perience alters response of adult neurogenesis to stress. Nat.
Neurosci., 7: 841–846.
Monje, M.L., Mizumatsu, S., Fike, J.R. and Palmer, T.D.
(2002) Irradiation induces neural precursor-cell dysfunction.
Nat. Med., 8: 955–962.
Monje, M.L., Toda, H. and Palmer, T.D. (2003) Inflammatory
blockade restores adult hippocampal neurogenesis. Science,
302: 1760–1765.
Moser, M.B. and Moser, E.I. (1998) Functional differentiation
in the hippocampus. Hippocampus, 8: 608–619.
Motl, R.W., Birnbaum, A.S., Kubik, M.Y. and Dishman, R.K.
(2004) Naturally occurring changes in physical activity are
inversely related to depressive symptoms during early ado-
lescence. Psychosom. Med., 66: 336–342.
Muller, M.B., Lucassen, P.J., Yassouridis, A., Hoogendijk,
W.J., Holsboer, F. and Swaab, D.F. (2001) Neither major
depression nor glucocorticoid treatment affects the cellular
integrity of the human hippocampus. Eur. J. Neurosci., 14:
1603–1612.
Murray, C.J. and Lopez, A.D. (1996) Evidence-based health
policy — lessons from the Global Burden of Disease Study.
Science, 274: 740–743.
Nakagawa, S., Kim, J.E., Lee, R., Chen, J., Fujioka, T., Ma-
lberg, J., Tsuji, S. and Duman, R.S. (2002a) Localization of
phosphorylated cAMP response element-binding protein in
immature neurons of adult hippocampus. J. Neurosci., 22:
9868–9876.
Nakagawa, S., Kim, J.E., Lee, R., Malberg, J.E., Chen, J.,
Steffen, C., Zhang, Y.J., Nestler, E.J. and Duman, R.S.
(2002b) Regulation of neurogenesis in adult mouse hippo-
campus by cAMP and the cAMP response element-binding
protein. J. Neurosci., 22: 3673–3682.
Nestler, E.J., Barrot, M., Dileone, R.J., Eisch, A.J., Gold, S.J.
and Monteggia, L.M. (2002) Neurobiology of depression.
Neuron, 34: 13–25.
Neumeister, A., Wood, S., Bonne, O., Nugent, A.C., Luckenb-
augh, D.A., Young, T., Bain, E.E., Charney, D.S. and Dre-
vets, W.C. (2005) Reduced hippocampal volume in
unmedicated, remitted patients with major depression versus
control subjects. Biol. Psychiatry, 57: 935–937.
Nottebohm, F. (2002) Neuronal replacement in adult brain.
Brain Res. Bull., 57: 737–749.
Oleskevich, S., Descarries, L., Watkins, K.C., Seguela, P. and
Daszuta, A. (1991) Ultrastructural features of the serotonin
innervation in adult rat hippocampus: an immunocytochem-
ical description in single and serial thin sections. Neurosci-
ence, 42: 777–791.
Overmier, J.B. and Seligman, M.E. (1967) Effects of inescap-
able shock upon subsequent escape and avoidance respond-
ing. J. Comp. Physiol. Psychol., 63: 28–33.
Overstreet-Wadiche, L.S. and Westbrook, G.L. (2006) Func-
tional maturation of adult-generated granule cells. Hippo-
campus, 16: 208–215.
Patel, T.D. and Zhou, F.C. (2005) Ontogeny of 5-HT1A re-
ceptor expression in the developing hippocampus. Brain Res.
Dev Brain Res., 157: 42–57.
Phillips, R.G. and LeDoux, J.E. (1992) Differential contribu-
tion of amygdala and hippocampus to cued and contextual
fear conditioning. Behav. Neurosci., 106: 274–285.
van Praag, H., Kempermann, G. and Gage, F.H. (1999) Run-
ning increases cell proliferation and neurogenesis in the adult
mouse dentate gyrus. Nat. Neurosci., 2: 266–270.
Radley, J.J. and Jacobs, B.L. (2002) 5-HT1A receptor antag-
onist administration decreases cell proliferation in the dent-
ate gyrus. Brain Res., 955: 264–267.
Ramirez-Amaya, V., Marrone, D.F., Gage, F.H., Worley, P.F.
and Barnes, C.A. (2006) Integration of new neurons into
functional neural networks. J. Neurosci., 26: 12237–12241.
Rao, M.S. and Shetty, A.K. (2004) Efficacy of doublecortin as a
marker to analyse the absolute number and dendritic growth
of newly generated neurons in the adult dentate gyrus. Eur. J.
Neurosci., 19: 234–246.
Reif, A., Fritzen, S., Finger, M., Strobel, A., Lauer, M., Sch-
mitt, A. and Lesch, K.P. (2006) Neural stem cell proliferation
is decreased in schizophrenia, but not in depression. Mol.
Psychiatry, 11: 514–522.
Rhodes, J.S., van Praag, H., Jeffrey, S., Girard, I., Mitchell,
G.S., Garland Jr., T. and Gage, F.H. (2003) Exercise in-
creases hippocampal neurogenesis to high levels but does not
improve spatial learning in mice bred for increased voluntary
wheel running. Behav. Neurosci., 117: 1006–1016.
Sairanen, M., Lucas, G., Ernfors, P., Castren, M. and Castren,
E. (2005) Brain-derived neurotrophic factor and antidepres-
sant drugs have different but coordinated effects on neuronal
turnover, proliferation, and survival in the adult dentate
gyrus. J. Neurosci., 25: 1089–1094.
Santarelli, L., Saxe, M., Gross, C., Surget, A., Battaglia, F.,
Dulawa, S., Weisstaub, N., Lee, J., Duman, R., Arancio, O.,
Belzung, C. and Hen, R. (2003) Requirement of hippocampal
neurogenesis for the behavioral effects of antidepressants.
Science, 301: 805–809.
Sapolsky, R.M. (1994) The physiological relevance of gluco-
corticoid endangerment of the hippocampus. Ann. N.Y.
Acad. Sci., 746: 294–304 discussion 304–307.
Sapolsky, R.M. (2000) Glucocorticoids and hippocampal atro-
phy in neuropsychiatric disorders. Arch. Gen. Psychiatry, 57:
925–935.
Sapolsky, R.M., Armanini, M.P., Sutton, S.W. and Plotsky,
P.M. (1989) Elevation of hypophysial portal concentrations

of adrenocorticotropin secretagogues after fornix transec-
tion. Endocrinology, 125: 2881–2887.
Saxe, M.D., Malleret, G., Vronskaya, S., Mendez, I., Garcia,
A.D., Sofroniew, M.V., Kandel, E.R. and Hen, R. (2007)
Paradoxical influence of hippocampal neurogenesis on work-
ing memory. Proc. Natl. Acad. Sci. U.S.A., 104: 4642–4646.
Saxe, M.D., Battaglia, F., Wang, J.W., Malleret, G., David,
D.J., Monckton, J.E., Garcia, A.D., Sofroniew, M.V., Kan-
del, E.R., Santarelli, L., Hen, R. and Drew, M.R. (2006)
Ablation of hippocampal neurogenesis impairs contextual
fear conditioning and synaptic plasticity in the dentate gyrus.
Proc. Natl. Acad. Sci. U.S.A., 103: 17501–17506.
Saxena, S., Brody, A.L., Ho, M.L., Alborzian, S., Ho, M.K.,
Maidment, K.M., Huang, S.C., Wu, H.M., Au, S.C. and
Baxter Jr., L.R. (2001) Cerebral metabolism in major de-
pression and obsessive-compulsive disorder occurring sepa-
rately and concurrently. Biol. Psychiatry, 50: 159–170.
Schaaf, M.J., de Kloet, E.R. and Vreugdenhil, E. (2000)
Corticosterone effects on BDNF expression in the hippo-
campus. Implications for memory formation. Stress, 3:
201–208.
Scharfman, H.E., Sollas, A.L., Smith, K.L., Jackson, M.B. and
Goodman, J.H. (2002) Structural and functional asymmetry
in the normal and epileptic rat dentate gyrus. J. Comp.
Neurol., 454: 424–439.
Schildkraut, J.J. (1965) The catecholamine hypothesis of affec-
tive disorders: a review of supporting evidence. Am. J. Psy-
chiatry, 122: 509–522.
Schmidt-Hieber, C., Jonas, P. and Bischofberger, J. (2004)
Enhanced synaptic plasticity in newly generated granule cells
of the adult hippocampus. Nature, 429: 184–187.
Seki, T. and Arai, Y. (1995) Age-related production of new
granule cells in the adult dentate gyrus. Neuroreport, 6:
2479–2482.
Seligman, M.E. and Beagley, G. (1975) Learned helplessness in
the rat. J. Comp. Physiol. Psychol., 88: 534–541.
Seminowicz, D.A., Mayberg, H.S., Mcintosh, A.R., Goldapple,
K., Kennedy, S., Segal, Z. and Rafi-Tari, S. (2004) Limbic-
frontal circuitry in major depression: a path modeling met-
analysis. Neuroimage, 22: 409–418.
Sheline, Y.I. (1996) Hippocampal atrophy in major depression:
a result of depression-induced neurotoxicity? Mol. Psychia-
try, 1: 298–299.
Sheline, Y.I., Gado, M.H. and Kraemer, H.C. (2003) Untreated
depression and hippocampal volume loss. Am. J. Psychiatry,
160: 1516–1518.
Sheline, Y.I., Sanghavi, M., Mintun, M.A. and Gado, M.H.
(1999) Depression duration but not age predicts hippocampal
volume loss in medically healthy women with recurrent major
depression. J. Neurosci., 19: 5034–5043.
Shirayama, Y., Chen, A.C., Nakagawa, S., Russell, D.S. and
Duman, R.S. (2002) Brain-derived neurotrophic factor pro-
duces antidepressant effects in behavioral models of depres-
sion. J. Neurosci., 22: 3251–3261.
Shors, T.J., Miesegaes, G., Beylin, A., Zhao, M., Rydel, T. and
Gould, E. (2001) Neurogenesis in the adult is involved in the
formation of trace memories. Nature, 410: 372–376.
721
Silva, R., Lu, J., Wu, Y., Martins, L., Almeida, O.F. and Sousa,
N. (2006) Mapping cellular gains and losses in the postnatal
dentate gyrus: implications for psychiatric disorders. Exp.
Neurol., 200: 321–331.
Singh, N.A., Clements, K.M. and Singh, M.A. (2001) The effi-
cacy of exercise as a long-term antidepressant in elderly sub-
jects: a randomized, controlled trial. J. Gerontol. A Biol. Sci.
Med. Sci., 56: M497–M504.
Snyder, J.S., Hong, N.S., Mcdonald, R.J. and Wojtowicz, J.M.
(2005) A role for adult neurogenesis in spatial long-term
memory. Neuroscience, 130: 843–852.
Snyder, J.S., Kee, N. and Wojtowicz, J.M. (2001) Effects of
adult neurogenesis on synaptic plasticity in the rat dentate
gyrus. J. Neurophysiol., 85: 2423–2431.
Song, H., Kempermann, G., Overstreet Wadiche, L., Zhao, C.,
Schinder, A.F. and Bischofberger, J. (2005) New neurons in
the adult mammalian brain: synaptogenesis and functional
integration. J. Neurosci., 25: 10366–10368.
Stockmeier, C.A., Mahajan, G.J., Konick, L.C., Overholser,
J.C., Jurjus, G.J., Meltzer, H.Y., Uylings, H.B., Friedman, L.
and Rajkowska, G. (2004) Cellular changes in the postmor-
tem hippocampus in major depression. Biol. Psychiatry, 56:
640–650.
Strange, B. and Dolan, R. (1999) Functional segregation within
the human hippocampus. Mol. Psychiatry, 4: 508–511.
Strange, B.A., Fletcher, P.C., Henson, R.N., Friston, K.J. and
Dolan, R.J. (1999) Segregating the functions of human hip-
pocampus. Proc. Natl. Acad. Sci. U.S.A., 96: 4034–4039.
Strekalova, T., Spanagel, R., Bartsch, D., Henn, F.A. and
Gass, P. (2004) Stress-induced anhedonia in mice is associ-
ated with deficits in forced swimming and exploration. Ne-
uropsychopharmacology, 29: 2007–2017.
Sullivan, P.F., Neale, M.C. and Kendler, K.S. (2000) Genetic
epidemiology of major depression: review and meta-analysis.
Am. J. Psychiatry, 157: 1552–1562.
Tanapat, P., Hastings, N.B., Reeves, A.J. and Gould, E. (1999)
Estrogen stimulates a transient increase in the number of new
neurons in the dentate gyrus of the adult female rat. J. Ne-
urosci., 19: 5792–5801.
Taylor, M.J., Freemantle, N., Geddes, J.R. and Bhagwagar, Z.
(2006) Early onset of selective serotonin reuptake inhibitor
antidepressant action: systematic review and meta-analysis.
Arch. Gen. Psychiatry, 63: 1217–1223.
Tozuka, Y., Fukuda, S., Namba, T., Seki, T. and Hisatsune, T.
(2005) GABAergic excitation promotes neuronal differenti-
ation in adult hippocampal progenitor cells. Neuron, 47:
803–815.
Vakili, K., Pillay, S.S., Lafer, B., Fava, M., Renshaw, P.F.,
Bonello-Cintron, C.M. and Yurgelun-Todd, D.A. (2000)
Hippocampal volume in primary unipolar major depression:
a magnetic resonance imaging study. Biol. Psychiatry, 47:
1087–1090.
Videbech, P. and Ravnkilde, B. (2004) Hippocampal volume
and depression: a meta-analysis of MRI studies. Am. J. Psy-
chiatry, 161: 1957–1966.
Videbech, P., Ravnkilde, B., Pedersen, A.R., Egander, A.,
Landbo, B., Rasmussen, N.A., Andersen, F., Stodkilde-
722
Jorgensen, H., Gjedde, A. and Rosenberg, R. (2001) The
Danish PET/depression project: PET findings in patients
with major depression. Psychol. Med., 31: 1147–1158.
Videbech, P., Ravnkilde, B., Pedersen, T.H., Hartvig, H.,
Egander, A., Clemmensen, K., Rasmussen, N.A., Andersen,
F., Gjedde, A. and Rosenberg, R. (2002) The Danish PET/
depression project: clinical symptoms and cerebral blood
flow. A regions-of-interest analysis. Acta Psychiatr. Scand.,
106: 35–44.
Vollmayr, B., Simonis, C., Weber, S., Gass, P. and Henn, F.
(2003) Reduced cell proliferation in the dentate gyrus is not
correlated with the development of learned helplessness. Biol.
Psychiatry, 54: 1035–1040.
Wang, S., Scott, B.W. and Wojtowicz, J.M. (2000) Heteroge-
nous properties of dentate granule neurons in the adult rat. J.
Neurobiol., 42: 248–257.
West, M.J. (1993) Regionally specific loss of neurons in the
aging human hippocampus. Neurobiol. Aging, 14: 287–293.
Willner, P. (2005) Chronic mild stress (CMS) revisited: consist-
ency and behavioural-neurobiological concordance in the
effects of CMS. Neuropsychobiology, 52: 90–110.
Willner, P. and Mitchell, P.J. (2002) The validity of animal
models of predisposition to depression. Behav. Pharmacol.,
13: 169–188.
Willner, P., Muscat, R. and Papp, M. (1992) Chronic mild
stress-induced anhedonia: a realistic animal model of depres-
sion. Neurosci. Biobehav. Rev., 16: 525–534.
Willner, P., Towell, A., Sampson, D., Sophokleous, S. and
Muscat, R. (1987) Reduction of sucrose preference by chronic
unpredictable mild stress, and its restoration by a tricyclic
antidepressant. Psychopharmacology (Berl.), 93: 358–364.
Winocur, G., Wojtowicz, J.M., Sekeres, M., Snyder, J.S. and
Wang, S. (2006) Inhibition of neurogenesis interferes with
hippocampus-dependent memory function. Hippocampus,
16: 296–304.
Wiskott, L., Rasch, M.J. and Kempermann, G. (2006) A func-
tional hypothesis for adult hippocampal neurogenesis: avoid-
ance of catastrophic interference in the dentate gyrus.
Hippocampus, 16: 329–343.
Wojtowicz, J.M. (2006) Irradiation as an experimental tool in
studies of adult neurogenesis. Hippocampus, 16: 261–266.
Wong, E.Y. and Herbert, J. (2004) The corticoid environment:
a determining factor for neural progenitors’ survival in the
adult hippocampus. Eur. J. Neurosci., 20: 2491–2498.
Wong, E.Y. and Herbert, J. (2006) Raised circulating corticos-
terone inhibits neuronal differentiation of progenitor cells in
the adult hippocampus. Neuroscience, 137: 83–92.
Yalcin, I., Aksu, F. and Belzung, C. (2005) Effects of desipra-
mine and tramadol in a chronic mild stress model in mice are
altered by yohimbine but not by pindolol. Eur. J. Pharma-
col., 514: 165–174.
Yan, W., Wilson, C.C. and Haring, J.H. (1997) Effects of
neonatal serotonin depletion on the development of rat
dentate granule cells. Brain Res. Dev. Brain Res., 98:
177–184.
Yuen, E.Y., Jiang, Q., Chen, P., Gu, Z., Feng, J. and Yan, Z.
(2005a) Serotonin 5-HT1A receptors regulate NMDA recep-
tor channels through a microtubule-dependent mechanism.
J. Neurosci., 25: 5488–5501.
Yuen, E.Y., Jiang, Q., Feng, J. and Yan, Z. (2005b) Microtu-
bule regulation of N-methyl-D-aspartate receptor channels in
neurons. J. Biol. Chem., 280: 29420–29427.
Plate 38.1. A schematic of the dentate gyrus granule cell layer (GCL) illustrating the different ways by which neurogenesis can
influence its structure and function. Boxed panel reveals a cross section of the dentate GCL with the different populations that reside
within it: mature granule cells born during development (light blue), adult-generated mature granule cells (dark blue), adult-born
immature neurons (red) and interneurons (green). Over the lifespan, the GCL may increase in size due to a net addition of new neurons
(A) or may remain unchanged due to a net replacement of developmentally generated granule cells (B). Changes in neurogenesis can
result in increased representation of interneurons (C), a larger pool of adult generated immature neurons (D) or the generation of
mature neurons with distinct physiological and biochemical properties (E). Conceivably, neurogenesis may be altered in any one of
these ways in MDD. Conversely, AD drugs may influence DG function in more than one way to exert their behavioral effects. (For
B/W version, see page 700 in the volume.)
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 39

The dentate gyrus in Alzheimer’s disease

Thomas G. Ohm

Institute of Integrative Neuroanatomy, Department of Clinical Cell and Neurobiology, Charité CCM, 10098 Berlin,
Germany

Abstract: As part of the hippocampus, the dentate gyrus is considered to play a crucial role in associative
memory. The reviewed data suggest that the dentate gyrus withstands the formation of plaques, tangles and
neuronal death until late stages of Alzheimer’s disease (AD). However, changes related to a disconnecting
process, and more subtle intrinsic alterations, may contribute to disturbances in memory and learning
observed in early stages of AD.

Keywords: hippocampus; granule cells; interneuron; Alzheimer’s disease; dentate gyrus; review

Basic anatomy originating from 650,000 to 1,100,000 (Gomez Isla

The human dentate gyrus belongs to the heteroge-
neously composed allocortex, which is located
mainly in the antero-medial temporal lobe. As part
of the hippocampus, the dentate gyrus is considered
to play a crucial role in associative memory,
especially with respect to events (‘what happens’)
(Morris, 2006). To perform this task, the dentate
gyrus is highly organized and numerous synaptic
interactions take place involving at least 11 types of
interneurons, which form an intrinsic network
(Freund and Buzsaki, 1996). In cognitively normal
humans,
9–22 million (mean: 14.6) (Seress, 1988;
West and Gundersen, 1990; West et al., 1994;
Bobinski et al., 1996a, 1997; Simic et al., 1997;
Harding et al., 1998; Korbo et al., 2004) densely
packed principal (projection) neurons, the granule
cells, form the narrow granule cell layer.
The major input of the dentate gyrus comes
through the perforant path representing axons
Corresponding author. Tel.: +49 (0)30 450 528202;
Fax: +49 (0)30 450 528913; E-mail: thomas_georg.ohm@
charite.de
et al., 1996; West and Slomianka, 1998; von
Gunten et al., 2006) excitatory entorhinal projec-
tion neurons located in the pre-a layer or — in a
different nomenclature — layer II. Pre-a belongs to
the lamina principalis externa, the superimposed
concept for the superficial layers, which is sepa-
rated by the stratum dissecans from the deeper
layers, together called lamina principalis interna.
Perforant path axons end preferentially within the
outer two-thirds of the superficial molecular layer
mainly on the (apical) dendrites of the granule
cells, but also on dendrites of interneurons such as
the basket cells. Origins of other important affer-
ents are cholinergic neurons of the basal forebrain
mainly located in the medial septal and diagonal
band nucleus and association fibres from the mossy
cells of sector CA4. These afferents are also at-
tracted to the inner molecular layer where they
terminate together with axons stemming from the
reunion nucleus of the thalamus. In humans, the
relative proportion of the respective input is not
exactly known. Furthermore, the wide-spreading
projection from noradrenergic and serotoninergic
nuclei of the brain stem (nucleus coeruleus and

DOI: 10.1016/S0079-6123(07)63039-8 723

724
oral raphe nuclei), along with some hypothalamic
(supra- and/or tuberomamillary) projections form a
profuse input to the molecular layer of the dentate
gyrus. Neurons from the pre-b layer (also termed
layer III) uses the perforant path for a direct pro-
jection to the sector CA1 of the Ammon’s horn
(probably encoding information used for spatial
memory (‘where is it happening’) (Morris, 2006).
Here the NMDA receptor-mediated synaptic plas-
ticity integrates the spatial- and event-related asso-
ciative memory components. Several findings suggest
that the CA3 pyramidal cells, which receive the gran-
ule cell output, aid at ‘pattern completion’ if recall of
spatial information is challenged under situations of
incomplete data about the environment (Nakazawa
et al., 2002). A hippocampal commissural system
(psalterium) is virtually absent (Amaral et al., 1984;
Demeter et al., 1985) and only a scanty callosal input
arrives from the contralateral hippocampus. This is in
contrast to a rodent brain and the classical descrip-
tions made by, e.g., Kölliker or de Nó.
The granule cells convey the processed informa-
tion via the mossy fibers, i.e., the granule cells’
axons, to the cornu ammonis (CA). Here they ter-
minate with unusually large and zinc-rich synaptic
boutons on proximal dendrites of pyramidal cells
of the sector CA3 of the hippocampus, thereby
forming the stratum lucidum. Mossy fiber collat-
erals are abound, terminating on inhibitory local
circuit neurons of the dentate gyrus’s polymorphic
layer and on the glutamatergic mossy cells
( ¼ modified hilar/CA4 pyramidal cells with osten-
tatious dendritic excrescences). In monkey, mossy
fiber collaterals terminate on basal dendrites of
granule cells, which may be the morphological ba-
sis for a recurrent excitatory feedback loop (Seress
and Frotscher, 1990). Others, however, when per-
forming biocytin-tracing on human hippocampal
tissue obtained from surgery, did not find evidence
for this (Lim et al., 1997).
The local interneuron’s axon terminates with
inhibitory synapses on the dendrites of granule
cells or on other interneurons. The granule cells
mostly direct their dendrites into the molecular
layer (apical dendrites) but in one-third (Lim et al.,
1997), — much more abundant than in rodents —
also into the polymorphic or plexiform layer (basal
dendrites) (Seress and Mrzljak, 1987). The axon of
the mossy cell synapses with the dendrite of
the basket cell, a pyramid-shaped GABAergic,
parvalbumin-containing interneuron. Their short
axons profusely give rise of collaterals in the gran-
ule cell layer where these branches form inhibitory
axosomatic synapses with the granule cells.
Common to both projection neurons and inter-
neurons is their richness in calcium-binding
proteins. Granule cells contain calbindin and
calmyrin, whereas calbindin, parvalbumin or
calretinin is found in interneurons.
Alzheimer’s disease-related histopathology
The formation and deposition of certain aggregated
proteins is most striking among the several histo-
logical changes tied up to Alzheimer’s disease (AD).
Normally, aggregation of these proteins does not
occur, and even more surprisingly, this material is
only rarely found in animals. Two proteins seem to
be central: tau and Ab.
Tau is a microtubule associated, primarily
neuronal-axonal protein involved in microtubule
assembly and stabilisation, processes controlled
by the degree of tau’s phosphorylation (normal
2–3 mol phosphate per mol tau) and O-
GlcNAcetylation. In humans, alternative splicing
of a single gene generates six isoforms, all of which
are normally among the most soluble proteins
known so far, and all of which are involved in AD
(Spillantini and Goedert, 1998). After hyper-
phosphorylation of some of the
30 potential
phosphorylation sites tau tends to self-polymerise.
Moderate hyperphosphorylation (4–6 mol phos-
phate per mol tau) results in sequestration of nor-
mal tau whereas higher phosphorylation (10 mol
phosphate per mol tau) is associated with filament
formation. In AD tau is hyperphosphorylated and
forms highly characteristic paired helical filaments
(PHF; interval between two twists 80 nm, closest
width 10 nm and broadest width 20 nm) and — to
a lesser amount — also straight filaments. These
filaments represent the protein backbone of in-
traneuronal histopathological features, i.e., tangles
and neuropil threads (AD-related ‘tau-pathology’).
These structures can be stained selectively with
highly sensitive silver stains such as the Gallyas

stain that was proven to be as sensitive and selec-
tive as immunostains (Schönheit et al., 2004).
Immunostains, however, allow the determination
of the hierarchically occurring phosphorylation
and thereby the staging of tangles (Uchihara et al.,
2001; Augustinack et al., 2002) unveiling that
changes in tau phosphorylation precede tangle
formation (Bancher et al., 1989; Braak et al., 1994).
Tangles occur in somata of neurons, which may
live for years (Bobinski et al., 1998) or even decades
with a tangle (Morsch et al., 1999). When the
tangle-bearing neuron has deceased, inert remnants
remain for quite some time, termed ‘ghost tangles’.
Eventually, they become cleared by astrocytes.
Ghost tangles are usually less-densely twisted, less
argyrophilic, affine for acid dyes and Congo red,
ubiquitinated and even immunoreactive against
anti-Ab-antibodies. Neuropil threads are likewise
silver-stainable filaments but primarily located
within the dendrites of neurons, often before
the respective soma has formed a tangle (Braak
et al., 1986).
The Alzheimer-related tau pathology shows a
regular, highly area-specific, lamina-specific and
cell type-specific spread through the cortex allow-
ing their classification into six stages. Neither neu-
ritic plaques nor any other plaque subtype shows a
similar degree of reliability in the distribution and
spread. The Braak-staging [for more details see:
(Braak and Braak, 1991)], developed on Gallyas-
stained tissue, distinguishes two ‘entorhinal’ (stages
I and II), two ‘limbic’ (stages III and IV) and two
‘isocortical stages’ (stages V and VI). The stages
I–II do not show clinical or insipient signs of AD
(Bancher et al., 1993; Braak et al., 1993). The stage
III displays a severe involvement of layers pre-a of
the transentorhinal and entorhinal cortices includ-
ing first occurrence of ghost tangles as sign of be-
ginning neuronal loss and deafferentiation of the
dentate gyrus. Stage IV is marked off by an abun-
dance of ghost tangles in layer pre-a, severer tangle
formation in layer pri-a and now commencing tan-
gle formation in layer pre-b. Stages V and VI meet
the classical neuropathological criteria of AD
(Khatchaturian, 1985; Jellinger, 1997; Jellinger
and Bancher., 1997) and generally are associated
with dementia. At stage VI the entorhinal cortex
may have an almost denuded layer pre-a, due to
725
the death of more than 90% of the neurons
(Gomez Isla et al., 1996). Ghost tangles are
numerous and may even be cleared in considera-
ble numbers from the neuropil by astrocytes. This
may feign a non-tangle related neuron loss (Kril
et al., 2002). In terms of time, the development of
these six stages was estimated to take almost five
decades (Ohm et al., 1995). Although many neu-
rons of the cortex may have developed a tangle in
these late stages, it is to note that only few of the
many nerve cell types are vulnerable (Braak and
Del Tredici, 2004).
Ab is physiologically formed by proteolytic
processing of a receptor-like transmembrane pro-
tein (APP, amyloid precursor protein) (Dyrks
et al., 1988) through a sequential action of so-
called beta- and gamma-secretases (Haass, 2004).
APP is member of a larger superfamily including
several splice-variants of APP and APP-like pro-
teins (APPLP1 and 2) and which may share com-
mon functions (Bush et al., 1994). Apparently the
APP, APPLP1 and APPLP2 showed a similar
pattern of expression in the hippocampus with
both mRNA and protein in granule cells and
pyramidal cells of all sectors of the Ammon’s horn
(McNamara et al., 1998). Ab is extracellularly
deposited in a variety of states and shapes, collec-
tively termed ‘plaques’. When Ab adopts a b-sheet
state, it can be stained as an amyloid-like material
(‘b-amyloid-plaque’) with classical amyloid stains
such as Congo red or thioflavine. In many cases, a
core of b-amyloid is surrounded by a surface of
dystrophic neurites and swollen glial processes,
and forms by this the so-called ‘neuritic plaque’.
b-pleated Ab, however, represents only a limited
amount of all extracellularly deposited Ab-peptide.
This other Ab forms the ‘diffuse’ or ‘pre-amyloid’
plaques and stains either with immunocyto-
chemical tools or a few specific silver stains. The
often-used term ‘senile plaque’ refers frequently to
an undefined melange of Ab-states and shapes. The
distribution pattern of diffuse and amyloid plaques
is not congruent. Ab deposits in phases that are
characterized by a hierarchical neuroanatomical
pattern (Thal et al., 2002). The relationship
between plaques and tangles is a matter of debate
since long. For many years the so-called amyloid
cascade hypothesis has dominated. The hypothesis
726
postulates that Ab-plaques are formed before tan-
gles and may even cause their formation. Only re-
cently, work using the well-defined circuitry of the
hippocampal formation provided strong evidence
that tangles develop either before or independent
from plaques (Schönheit et al., 2004) and follows
an anterograde pattern (Thal et al., 2002; Schön-
heit et al., 2004).
Apart from the tau-pathology and the Ab-
pathology, there is an inconspicuous pathology,
which, not surprisingly, correlates best with the de-
gree of cognitive impairment. This is the loss of
synapses reported by Davies and colleagues (Davies
et al., 1987), later confirmed in many subsequent
electron-microscopical studies and (indirectly) by
semi-quantitative immunolabelling of synaptic mar-
ker proteins (Scheff and Price, 2003). Synapse loss
takes place in terminal zones of neurons prone to
develop tangles. Thus, it is not surprising that
number of tangles also correlate strongly with the
cognitive decline. Tangle-bearing neurons express
baser levels of the message for the synaptic marker
protein synaptophysin (Callahan and Coleman,
1995), suggesting that synapse dysfunction and loss
is not only the self-evident consequence of neuronal
loss but may occur in still living neurons. This is
further indicated by a reduction in the synapse to
neuron ratio in dentate gyrus (48%) (Bertoni-
Freddari et al., 1996). Interestingly, dendrites which
lay within plaques show similar numbers of spines
compared to adjacent dendrites or dendritic seg-
ments which lay in the inter-plaque region of the
neuropil (Einstein et al., 1994). Apart from numeric
loss of synapses there are morphological changes,
which may differ depending on the nature of the
participating fibers and cells. From mossy fiber ter-
minals decreases in number and area of spines and
post-synaptic thickenings are reported, whereas non-
mossy fiber terminals seem to behave in an opposite
way, i.e., displaying more synapses, spines and post-
synaptic thickenings (Kiktenko et al., 1997).
AD related histopathology in the dentate gyrus
Tau-pathology
Densely packed tangles with a globose shape (in
contrast to the torch-like shaped tangles, e.g., in
pyramidal cells of the Ammon’s horn) built-up by
PHFs occur only in late stages of AD, i.e., stage VI
of Braaks’ classification (Braak and Braak, 1991).
The percentage of tangle-bearing neurons among
all neurons of the granule cell layer is 1.7–4.2 in
severe AD (Bobinski et al., 1997). Spherical tau-
aggregates consisting of straight filaments
(18–25 nm in diameter), however, are seen in many
granule cells when probing severe Alzheimer cases
with TG3 (Wakabayashi et al., 1997), a mono-
clonal antibody binding a phosphate-dependent
epitope in PHFs. Interneurons seem to be very
resistant, perhaps because of their high content of
calcium-binding proteins such as parvalbumin or
calretinin (Braak et al., 1991; Nitsch and Ohm,
1995; Freund and Buzsaki, 1996). However, an
obligatory absence cannot be presumed because in
frontal cortex 4–6% parvalbumin-containing neu-
rons also reacted with an antibody raised against
tau from tangles (Iwamoto and Emson, 1991).
Others have reported only a 0–0.7% subpopula-
tion of parvalbumin-containing neurons of the
superior frontal gyrus which may have tangles
(Sampson et al., 1997). In the hippocampus, how-
ever, only two cells out of 1950 examined parval-
bumin-containing interneurons showed — an even
questionable — tangle formation (Ohm et al.,
2002). Also for NOS-containing neurons, it is
shown that they resist tangle formation. The rel-
ative scarceness of NOS neurons in the dentate
gyrus does not allow forming a definite opinion
about this region. However, double-labelling of
other hippocampal neurons has revealed that more
than 3300 tangle-bearing neurons are negative for
NOS. Starting with Braak-stage IV, some large
multipolar neurons of the plexiform layer and ad-
jacent rim of the sector CA4 develop tangles ex-
tending into proximal dendrites. In stage V, a
small number of closely arranged tangles are seen
in modified pyramidal cells, thereby easily distin-
guishable from the ones in the multipolar neurons
described above. In stage VI, no new type accrues
but the number of afflicted neurons increases con-
siderably. Neuropil threads are mainly located in
dendrites of neurons that undergo hyper-
phosphorylation and, subsequently, formation of
argyrophilic aggregation of tau (Braak and Braak,
1991). This may indicate that a neuron’s main

receptive structure, the dendrite, has lost normal
structure and function. Beginning with stage III of
Braaks’ classification, hyperphosphorylated tau
diffusely stains the outer molecular layer of the
dentate gyrus. At stages V and particularly at stage
VI, argyrophilic neuropil threads are seen (Thal et
al., 2000). This corresponds to the beginning of
tangle formation (coarse granular material) in the
granule cells at stage V, and marked numbers of
tangles exhibiting a globose shape in stage VI
(Braak and Braak, 1991).
Plaques
Ab-deposits in CA1 occur in phase 2 of Ab depo-
sition, the dentate gyrus is involved later, namely in
phase 3 (Thal et al., 2002). A correlative analysis
suggests that phases 1 and 2 are associated with
CDR 0, i.e., no signs of cognitive deterioration
(Thal et al., 2002). Neuritic plaques are seen in a
zone between the outer two and the inner third of
the molecular layer, i.e., within the terminal zone of
the perforant path (Crain and Burger, 1988). They
form later than the tangles in the layer pre-a of the
entorhinal cortex, namely not before Braak-stage
IV (Braak and Braak, 1991). The caveat that spot
check-like examinations might overlook changes
is ruled out in studies using close-meshed serial
sections throughout the whole hippocampal
formation (Schönheit et al., 2004). The limbic
stage III with hyperphosphorylated tau highlight-
ing the outer molecular layer and neuritic plaques
formed at stage IV probably represents neuro-
pathological signs of insipient AD (Bancher et al.,
1993; Braak et al., 1993). At stage V, the numbers
increase and a row-like appearance becomes clear.
A dense ribbon of neuritic plaques (Braak and
Braak, 1991) characterizes the stage VI. Interest-
ingly, the terminal zone of the mossy fibers is
almost devoid of neuritic plaques even in late and
severe stages of AD.
Synaptic changes
In 1988, the first report on AD-related synaptic
loss quantitatively assessed by electron microscopy
was published (Bertoni-Freddari et al., 1988).
727
Bertoni-Freddari and colleagues reported highly
significant reductions in synaptic length, surface
density and numeric density determined in the
supragranular band (innermost part of the inner
molecular layer). Later ultrastructural examina-
tions have confirmed this (Scheff and Price, 1998),
and also assigned it to the outer molecular layer
(Dekosky et al., 1996; Scheff et al., 1996). The
reduction has been calculated to a 21% decrease in
numeric synaptic densities and a 26% decrease in
the thickness of the outer molecular layer, and a
27% decrease in numeric synaptic densities and
26% reduction in thickness of the inner molecular
layer vs. control, respectively. None of the studies,
however, has analysed Braak-staged material,
which would have allowed making estimates
about the dynamic of the process. Given that
synaptophysin represents synapses, a significant
reduction was visualized correlating to the cogni-
tive decline. Early, mild and severe Alzheimer cases
are accompanied by a 23, 49 and 67% reduction in
the outer third of the molecular layer, and a 30, 39
and 62% loss in the middle third, respectively. No
significant changes are seen in the inner third of the
molecular layer (Masliah et al., 1994). At the time
being, we have no information about the relative
changes in synapse subtypes, e.g., excitatory vs.
inhibitory forms. Many transmitters coexist within
the same terminal (Torrealba and Carrasco, 2004).
It is, therefore, not insentient to consider the
possibility of selective defects associated with
only one of the transmitters leaving the other
transmitter and the ‘synapse’ intact.
Because synapses often terminate on dendritic
spines, changes in spine number, spine density and
spine shape have been determined as well. Spine
numbers of the dentate gyrus are reduced in AD
cases (Einstein et al., 1994) by 35–44%, depending
on the topographical relationship between plaques
and dendrites. Dendrites crossing plaques when
drawing through the molecular layer show 9%
fewer spines than those traversing in plaque-free
regions. This difference, however, is statistically
not significant. As there are no data concerning the
numbers of axons within plaque-free and plaque
regions, the putative role of plaques (be it Ab-
peptide or Ab-amyloid) on spine maintenance or
plasticity in AD remains penumbral. Others have
728
found a reduction in granule cell spine density by
60% (Williams and Matthysse, 1986), or 40% (De
Ruiter and Uylings, 1987) in Golgi-stained mate-
rial and of 23% in DiO-labelled neurons (Ji et al.,
2003). One should, in this context, keep in mind
that determinations of spine densities not only de-
pend on post- and peri-mortem factors more than
other morphometrical measures (Uylings et al.,
1986), but also on shrinkage of the dendrite or the
selection of an appropriate or representative den-
dritic intercept. In addition, Golgi stains do show
only a small proportion of neurons and they may
not be representative. Determinations of the
post-synaptic density protein SAP97 has unveiled
no significant decrease in AD dentate gyrus
(Wakabayashi et al., 1999).
Detailed data on intrinsic connectivity in AD is
apparently lacking. It is, however, conceivable, that
certain subtypes of dentate interneurons react with
adaptive changes analogous to those seen under
other pathological circumstances [e.g., epilepsy
(Magloczky et al., 2000)]. It is to note in this con-
text, however, that epilepsy may be a symptom of
advanced AD (Lozsadi and Larner, 2006).
With respect to output synapses, mossy fibre
terminals are the sole ones. They are marked by
zinc and dynorphin. Densitometrical analysis of
Timm-stained zinc has found an increase in AD
(Goldsmith and Joyce, 1995). This may reflect
either an increase in zinc levels per terminal or an
increase in terminal with unchanged levels of zinc,
or higher shrinkage in Alzheimer tissue with more
zinc-rich terminal per tissue volume. The copper–
zinc superoxide dismutase, expressed in granule
cells, might be unaltered on protein and message
level (Ceballos et al., 1991), if the low numbers of
investigated cases were representative. However, a
35% reduction is found in the stratum lucidum of
CA3 for hMTH1, another neuronal protein sug-
gested to protect from zinc-mediated oxidative
damage (Furuta et al., 2001). Decreases in number
and area of spines and post-synaptic thickenings
are reported with respect to mossy fiber terminals
on CA3 dendrites (Kiktenko et al., 1997).
Chromogranin B, a large dense core vesicle pro-
tein and within the hippocampus particularly rich
in mossy fibres, is found reduced (46%) in
immunocytostaining intensity (Marksteiner et al.,
2000). 3[H] kainate, which labels binding sites on
mossy fibre terminals, is found reduced in autora-
diography of the stratum lucidum (34%)
(Represa et al., 1988). The dendritic tree (basal as
well as apical) of the CA3 pyramids, however, is
unchanged (Flood et al., 1987a, b) suggesting no
overall loss of synaptic contacts. In line herewith,
non-mossy fibre terminals (of unknown origin)
form more synapses and are associated with
a larger area of spines and form post-synaptic
thickenings (Kiktenko et al., 1997).
Neuronal loss
In cognitively normal individuals,
9–22 million
granule cells are counted with stereological tech-
niques (Seress, 1988; West and Gundersen, 1990;
West et al., 1994; Bobinski et al., 1996a, 1997;
Simic et al., 1997; Harding et al., 1998; Korbo
et al., 2004). The broad range (>factor 2) is prob-
ably not only due to a large inter-individual var-
iation but also to differences in the sampling
scheme or anatomical delineation in regions where
the dentate gyrus does not display a C-shaped
band. Thus, relative changes seem to be more in-
formative. Unfortunately, the picture remains
somewhat blurred. The reported neuronal cell loss
in the granule cell layer ranges from 0% (Korbo et
al., 2004), 8% (n.s.) (Bobinski et al., 1997), around
11% (n.s.) (West et al., 1994, 2004), 25% (n.s)
(Simic et al., 1997) to 43% (po0.05) (Bobinski
et al., 1996b). Curiously, researchers who found
also the 43% loss described the 8% loss. However,
the fact that five out of six studies have failed to
show a statistically significant change suggests that
putative dysfunction of the dentate gyrus is not
caused by simple loss of granule cells. Interestingly
to note is a transient tendency for more granule
cells in pre-clinical Alzheimer (+13%) of Braak
stages II–IV (West et al., 2004). This may be the
result of an increase in neurogenesis, as it appears
to occur in some Alzheimer brains (Jin et al., 2004).
However, granule cell dispersion, occurring in
epilepsy, and suggestive for substantial reactive
neurogenesis, does apparently not take place in
AD. Because most granule cells normally express
calbindin, calbindin expression and protein levels

are of interest. A
25% reduction (not statistically
significant) is found in a radioactive in situ hybrid-
isation analysis (Maguire-Zeiss et al., 1995) and a
85% reduction mRNA level is seen in hippocampal
homogenate determinations (Iacopino and
Christakos, 1990). The calbindin protein level in
whole hippocampal homogenates is reported to be
67% (Iacopino and Christakos, 1990). The reduc-
tions may reflect a loss of calbindin-mRNA and
subsequently also calbindin occurring before gran-
ule cells fade away. Others, pondered that loss of
calbindin in many of the granule cells might relate
to the cognitive decline in AD despite of the fact
that the total number of granule cells might be un-
altered (Greene et al., 2001; Palop et al., 2003).
Calbindin loss, however, is a well-known feature of
dentate granule cells in epilepsy (Magloczky et al.,
1997) where it not necessarily associates with
cognitive decline or even dementia. In epilepsy, a
group with severe hippocampal cell loss (mainly in
CA1), displaying the so-called Ammon’s horn
sclerosis, shows a marked reduction in granule cell
calbindin levels. A second group with medial tem-
poral lobe epilepsy (but not with Ammon’s horn
sclerosis) displays normal calbindin levels in their
granule cells. Thus, alternatively to a putative
causal role in cognitive decline, depletion of cal-
bindin may increase the resistance of granule cells
(Nagerl et al., 2000). Interestingly, immunocyto-
chemical demonstration of calmyrin, another
granule cell-associated calcium-binding protein,
and which is lacking in interneurons, showed no
differences between controls and Alzheimer cases
in stages IV–V (Bernstein et al., 2005). Also
neuron-specific enolase is found unaltered
(Wakabayashi et al., 1999). Together this further
supports the notion that granule cells withstand
neuronal death in AD.
Apart from loss of granule cells, local interneu-
rons have to be considered, too. Many of these are
characterized by the presence of certain calcium-
binding protein: parvalbumin, calretinin and cal-
bindin. It seems that at least some of the inter-
neuron types (Freund and Buzsaki, 1996) have lost
the parvalbumin-phenotype or are even numeri-
cally reduced. Data reported, however, are not
unanimous. A 60% reduction in neuronal density
of parvalbumin-containing neurons is found for
729
the dentate gyrus and sector CA4 together (Brady
and Mufson, 1997), whereas others have found a
slight increase in the dentate gyrus (Satoh et al.,
1991). The latter study is flawed by the fact that
the total number of cases and immunoreactive cells
is low. In other brain areas, a reduction is found,
ranging between 20 and 72%, depending on brain
region and layer (Arai et al., 1987; Solodkin et al.,
1996; Mikkonen et al., 1999). Again, others differ
when reporting no loss (Hof et al., 1991; Fonseca
et al., 1993; Sampson et al., 1997; Leuba et al.,
1998). In general, however, the loss of parvalbu-
min-containing neurons is seen only in late
Alzheimer and especially in those regions, which
undergo massive differentiation or loss of principal
neurons. Thus, it is likely that at least those
GABAergic interneurons, which contain parval-
bumin, are primarily resistant against the tau
pathology and untimely cell death. Another inter-
neuron-associated calcium-binding protein is
calretinin, which is found in differentially shaped
neurons of the human dentate gyrus (Nitsch and
Ohm, 1995). With respect to the AD hippocampus
no indication for a loss in the numbers of immuno-
stained neurons is reported (Brion and Resibois,
1994). In line herewith, also other brain areas
display an unchanged overall neuronal density of
calretinin-positive neurons (Hof et al., 1993;
Fonseca and Soriano, 1995; Leuba et al., 1998;
Mikkonen et al., 1999). The observed layer-specific
increase of small neurons that is concomitantly
found with a decrease in large calretinin-contain-
ing neurons (Sampson et al., 1997; Leuba et al.,
1998; Mikkonen et al., 1999) supports the notion
that more subtle plastic–adaptive changes occur.
In this context it is also to note that the intensity of
the calretinin immunoreaction with a staining
gradient (apical dendrites stronger, basal weaker)
is inverted in the entorhinal cortex of AD patients
(Mikkonen et al., 1999). This suggests a plastic and
adaptive process. Coincided with calretinin, the
innermost part of the molecular layer is normally
strongly immunolabelled by secretoneurin, a mem-
ber of the chromogranin family. In AD, this band
is significantly reduced (52%), whereas the in-
tensity of the calretinin immunostaining is only
slightly diminished (16%) (Kaufmann et al.,
1998). The rest of the inner molecular layer is
730
unchanged, but a 40% decrease in secretoneurin is
found in the outer molecular layer. The third cal-
cium binding protein of various interneurons is
calbindin, which, however, is also typical for gran-
ule cells (Sloviter et al., 1991; Seress et al., 1993).
This implies that determination of protein or mes-
sage levels represent the contribution of both neu-
ronal classes. The immunocytochemical analysis of
the dentate gyrus reveals only few calbindin-con-
taining granule cells and interneurons in Al-
zheimer brains (Iritani et al., 2001). No changes
in the number and distribution of GABAergic
neurons are seen in the hippocampus between
controls and AD (Yew et al., 1999). Some inter-
neurons of the dentate gyrus, mostly located just
underneath the granule cell layer, contain neuro-
peptide Y (NPY). These peptidergic large poly-
morphic neurons are reduced in numbers, to only
a mild degree in the rostral and middle part of the
hippocampus, but evident caudally. Although
there are neurons with intense immunostaining
and a preserved dendritic tree, most others, how-
ever, show a loss or regression of their dendritic
tree (Chan-Palay et al., 1986). A further interneu-
ron subgroup is identified by its content of som-
atostatin. These neurons are evenly spread
throughout the polymorphic layer, lacking in the
granule cell layer, and are only occasionally found
in the molecular layer. In AD, only vestiges remain
with greatly diminished dendritic processes
(Chan-Palay, 1987). This is in line with biochemi-
cal findings, which have determined a 70% loss of
somatostatin in hippocampal tissue (Davies and
Terry, 1981). Also in other cortex areas,
somatostatin-containing interneurons show a
decreased numeric density (>70%) (Kumar,
2005) and/or cytoskeletal changes reminiscent of
tau-pathology (van de Nes et al., 2002). Together
this suggests that this population is more vulner-
able than other interneurons. Multipolar CCK-8-
containing neurons are occasionally found in the
molecular layer and numerous polymorphic ones
in the anterior part of the dentate gyrus, mainly in
the subgranular zone of the polymorphic layer
(Lotstra and Vanderhaeghen, 1987). In contrast to
animals, CCK-8-positive fibers seem to emanate
almost exclusively from intrahippocampal neu-
rons, because fimbria and alveus were found
devoid of CCK-8-positive profiles and only rarely
they were seen in the angular bundle. The only
exception seems to be a projection via the perfo-
rant path probably originating from pre-a neu-
rons. Changes in CCK are not reported for AD
hippocampus. Some larger enkephalin-containing
neurons are found in the molecular layer of the
dentate gyrus (Kulmala, 1985). No changes are
seen in AD (Kulmala, 1985) when examining the
dentate gyrus or the whole hippocampus (Yew
et al., 1999). However, a Leu- and Met-encephalin-
receptor binding loss is found in hippocampal ho-
mogenates (Rinne et al., 1993). Others have found
a statistically significant 48% loss of m- and a 36%
loss of k-opiod-binding sites whereas no changes
have been seen for d-binding sites in quantitative
autoradiography (Mathieu-Kia et al., 2001).
Prominent terminal-like Substance P staining is
seen in the dentate gyrus and in CA1–4 fields, and
multipolar immunolabelled neurons are located in
the polymorphic layer of the dentate gyrus as well
as in CA1–4 (Kowall et al., 1993). These neurons
are NADPH-negative, in contrast to most of the
Substance P-containing neurons of the isocortex.
Substance P-receptor immunocytochemistry
shows a light marking of CA2 pyramidal cells
and occasional labelling of basket cells in CA1–4.
In AD, staining intensity is reduced in the dentate
gyrus but sparing the hilar neurons (Kowall et al.,
1993). Nitric oxide is generated in neurons with
NADPH-diaphorase activity (NOS-neurons). The
distribution and morphological types does not
differ significantly between controls and Alzheimer’
disease patients in all sectors of the Ammon’s horn
and subiculum (Hyman et al., 1992). Also the total
number of NOS neurons is unaltered. However,
whereas in CA4 and CA3 a considerable decline
(though not statistically significant) in the numbers
is found, the subiculum shows a light increase.
Only few NOS-neurons are occasionally found in
the dentate gyrus, i.e., in 2 of 20 examined hippo-
campi. Double labelling has revealed that more
than 3300 tangle-bearing neurons are negative for
NOS. The planimetrically determined size of the
neurons does not show changes (exception CA1)
but regressive changes are found in the form
of foreshortened dendrites and distorted and
beaded axons. NADPH-diaphorase activity is

histochemically visualized and displays a striking
loss in the terminal zone of the perforant path in
AD (Rebeck et al., 1993). In controls, this zone
stands out clear due to the high level of reaction
product and gives the molecular layer a sublam-
inated aspect. As with NOS antibody staining,
only few neurons are stained with NADPH-
diaphorase activity in the granule cell layer. CA4
appeared to be unaltered.
The excitatory–inhibitory neurochemistry
Because the dentate gyrus receives a glutamatergic
excitatory input from the layer pre-a and transmits
processed information to CA3, the glutamatergic
system has attracted researchers and numerous ne-
urochemical studies have been performed. Early
studies using microdissected specimens of the
termination zone of the perforant path, and subse-
quent determinations of the glutamate content
therein, have demonstrated an 83% reduction in
AD (Hyman et al., 1987). More recent studies have
determined possible changes at a higher definition
in terms of receptor subtypes and (sub)regional
distribution. In contrast to significant decreases in
other parts of the hippocampus (or the hippocam-
pus as a whole), the dentate gyrus shows no sig-
nificant changes for NMDA-R1, 2A and 2B
message or protein (despite of trends toward de-
creased levels) as determined by Western blotting of
micropunches (Wakabayashi et al., 1999; Mishizen-
Eberz et al., 2004), radioactive in situ hybridisation
(Ulas and Cotman, 1997) and immunocytochemis-
try (Aronica et al., 1998). An immunocytochemical
study with Braak-staged tissue has found no
changes in early stages but an increase in staining
intensity at stages IV and later in all CA fields
concomittant with a decrease in staining of the
outer molecular layer (Ikonomovic et al., 1999).
NMDA-receptor message and protein is located on
neurons, i.e., granule cells and pyramidal cells,
showing expression differences between the hippo-
campal subfields. NMDA-R1, -2A and -2B message
is highest in granule cells, followed by CA2 and 3,
protein highest in CA1 for NMDA-R1 and -2A and
granule cells for 2B (Mishizen-Eberz et al., 2004).
Likewise AMPA receptors have been examined and
731
no changes are seen in the dentate gyrus in quan-
titative autoradiography (Dewar et al., 1991) or a
reduction in the outer molecular layer is found,
together with an increase in the polymorphic layer
(Geddes et al., 1992). However, Kd and Bmax may
have changes, making an interpretation difficult.
Again, more recent studies give a more defined
picture. Glu-R1 is found unaltered in an in situ
hybridisation study (Pellegrini-Giampietro et al.,
1994) and in immunocytochemical investigations
(Aronica et al., 1998). Others, however, find a
subtly modified picture. Glu-R1 immunoreactive
neurons are apparently unaltered in number
(Ikonomovic et al., 1995; Aronica et al., 1998;
Wakabayashi et al., 1999) but show an increase in
staining intensity in the molecular layer (Hyman
et al., 1994; Ikonomovic et al., 1995; Wakabayashi
et al., 1999) especially in the inner (Hyman et al.,
1994) and supragranular part (Ikonomovic et al.,
1995), and in the plexiform layer (Ikonomovic
et al., 1995). This increase is on one hand attributed
to an increase in the number of immunoreactive
fibers but on the other hand due to an increase of
immunoreactivity. The latter may also explain why
the proportion of Alzheimer cases displaying a sig-
nal (specific immunoreactivity over background) is
about twice as high as of controls (Hyman et al.,
1994). With respect to CA3 (and CA4), i.e., the
terminal zone of the granule cells, an increase is
found (Aronica et al., 1998), which seems to follow
the progression of the Braak-stages (though not
statistically significant) (Carter et al., 2004). Curi-
ously, Western blotting shows a significant 70%
reduction in the dentate gyrus of AD patients,
which is accompanied by virtually identical levels of
neuron-specific enolase (Wakabayashi et al., 1999).
The latter suggests that no major granule cell loss
has taken place and is in line with most of the cell
counting studies. The apparent contradiction be-
tween the Western blot and immunocytochemical
data obtained from identical cases (Wakabayashi et
al., 1999) may be due to unmasking of epitopes or
increasing of affinity, which is not observed under
denaturation conditions in Western blotting.
Glu-R2 and Glu-R2/3 immunoreactivity is found
to be unaltered (Hyman et al., 1994) or increased in
the (inner) molecular layer and plexiform layer of
the dentate gyrus (Ikonomovic et al., 1995; Aronica
732
et al., 1998). The number of Glu-R2 immunoreac-
tive neurons in the dentate gyrus is reduced
(Ikonomovic et al., 1995). Microdissected punches
of hippocampi, grouped according to the Braak-
classification into stages 0–II, stages III–IV and
stages V–VI, and subjected to Western blotting, has
contained comparable amounts of immunoreactive
protein (Carter et al., 2004). Interestingly, however,
double-labelling of Glu-R2 and tangles in the
entorhinal cortex and subiculum/CA1 by the
MC1 antibody has resulted in a complementary
pattern: number of Glu-R2 immunoreactive cells
diminished when the number of MC1 positive (i.e.,
tangle-bearing cells) increased (Ikonomovic et al.,
1997). Others, however, report similar immunore-
activity in both tangle-bearing and tangle-free neu-
rons for any of the studied Glu-R subtypes (Hyman
et al., 1994). Glu-R4, though relatively sparse in the
hippocampus, is found unaltered in terms of stain-
ing pattern and intensity. In sharp contrast to
staining with Glu-R1 and Glu-R2 antibodies, the
immunoreaction is not localised to dendrites or ax-
ons but highlighted neuronal somata including
CA3 pyramidal cells and granule cells (Hyman
et al., 1994). Quantitative 3[H] kainate autoradiog-
raphy reveals a statistically significant 54% reduc-
tion in the supragranular layer (Represa et al.,
1988). Kainate receptors can be visualised with an-
tibodies raised against Glu-R5/6/7. Immunoreac-
tive neurons are granule cells, and mossy cells as
well as pyramidal cells of the CA sectors. The som-
atodendritic compartment is labelled. Other neu-
rons marked are non-pyramidal cells of the stratum
oriens and some glial cells in the alveus and fimbria.
No significant changes are seen in Alzheimer brains
(Aronica et al., 1998).
With respect to the inhibitory GABAergic sys-
tem autoradiographic studies show a similar Kd for
3
[H] GABA hippocampal binding sites between
controls and Alzheimer cases (Chu et al., 1987).
GABAA receptors are found unaltered, whereas
GABAB binding sites are significantly reduced in
the molecular layer of the dentate gyrus, the stra-
tum pyramidale and lacunosum-moleculare of the
sector CA1 (Chu et al., 1987). A similar picture is
seen with 3[H] flunitrazepame except for the reduc-
tion in the molecular layer of the gyrus dentatus
(Jansen et al., 1990; Penney et al., 1990), which is
found unaltered or even displaying a slight in-
crease. Micropunches of hippocampi subjected to
Western blotting of b1, b2, a1 and a5 subunits of
GABAA receptors show no differences between
controls and Alzheimer individuals. A significant
decrease is seen only for a5 in CA3 (20%) (Ri-
ssman et al., 2003), whereas a5-ligand binding
show only in CA1 a significant 27% decrease (Ho-
well et al., 2000). Immunocytochemical analyses
give similar results for GABAA-R-b2/3 (Mizukami
et al., 1997) and GABAA-R-a1 (Singer et al., 2006)
when comparing cases with Braak-stages I–VI. In
severe stages, the GABAA-R-a1 is found decreased
locally on soma and dendritic surface of mossy
cells. Labelling of interneurons with GABAA-R-a1
or GABAA-R-b2/3 remains unaltered though they
display shrunken processes. GABAB-R1, however,
show a transient but significant increase in stages II
and IV in sectors CA4 and CA3, whereas a 64%
loss of immunoreactive neurons of sector CA1
takes place between stages I–II and II–IV which
further proceeds to 70% in stages V–VI (Iwakiri et
al., 2005). The finding that between stages I–II and
III—IV, only 12% of Nissl-stained neurons are lost
(and 39% in stages V–VI) suggests that receptor
loss precedes neuronal death. In situ hybridisation
for GABAA-R-b2 in cases with Braak stages I–VI
shows — in line with the protein findings — no
changes. In contrast, GABAA-R-b3 message is
reduced in all hippocampal sectors but not in the
CA4 (Mizukami et al., 1998). In a parenthetical
note GABAA-R-a1 message localisation and signal
intensity is reported to be apparently unaltered
in Alzheimer hippocampus (Pellegrini-Giampietro
et al., 1994).
Disconnected dentate gyrus
From the above it becomes conceivable that the
classical histopathological signs of AD, which cor-
relate strongly with the degree of dementia (tan-
gles), are late events. This also holds true for
neuronal cell loss. In contrast, clinical signs such as
dementia may develop much earlier. In terms of the
Braak classification tangles and neuropil threads
evolve in stage V or later, whereas loss of cognitive
and memory function as assessed by the mini

mental state examination (MMSE) test may occur
already in stage III (Bancher et al., 1993; Braak
et al., 1993). The dentate gyrus is considered to
play a crucial role in associative memory, especially
with respect to events (‘what happens’) (Morris,
2006). This raises the possibility that changes which
lead to interruption or perturbation of the granule
cells’ input isolates the dentate gyrus and causes
impairment of learning and memory (Hyman et al.,
1984). To discharge one single granule cell,
400
axons of the perforant path have to convey the exci-
tatory input from the pre-a neurons (McNaughton
et al., 1991). This suggests that loss of entorhinal
pre-a neurons may be sufficient to cause learning
and memory defects. Support for this concept
comes from several lines of evidence. Correlative
clinico-neuropathological studies are, unfortu-
nately, not unambiguous. In general, the numbers
of tangles in entorhinal neurons correlate with both
Braak staging and deterioration of cognition
(Garcia-Sierra et al., 2000, 2001). An early study
in 1996 indicated that a 57% loss of layer pre-a
neurons (
650,000 in CDR 0 control cases vs.
285,000 in CDR 0.5 individuals) is associated with
a Clinical Dementia Rating (CDR) score of 0.5,
which represents beginning dementia (Gomez Isla
et al., 1996). At that time, the total entorhinal
neuron loss is only
30%. This is confirmed by a

35 and 50% pre-a layer neuron loss in CDR 0.5
cases (Kordower et al., 2001; Price et al., 2001),
respectively, but not in a newer study (only 1–2%
loss) (Hof et al., 2003). Although there is no direct
translating between CDR and Braak scores, it is
suggested that a CDR 0.5 may represent Braak-
stage III and a MMSE range of 26–29 (Hof et al.,
2003). Severe AD (CDR 3–5, MMSE 0 or Braak
stage VI) is found associated with a
90% loss of
pre-a neurons (Gomez Isla et al., 1996) and a
laminar-specific spongiosis in the perforant path
termination zone (Duyckaerts et al., 1998) as well
as ALZ50 immunoreactivity highlights the perfo-
rant path terminal zone (Hyman et al., 1988). The
immunocytochemical detection of the intrinsic pre-
synaptic vesicle protein synaptophysin, suggests a
functional or terminal loss in the termination zone
of the perforant path (Cabalka et al., 1992;
Masliah et al., 1994). Demonstration of the my-
elinated fibres by the Weil-stain has indicated a
733
marked rarefaction of white matter along the
course of the perforant tract together with a loss
of oligodendrocytes thereabout (Morys et al.,
1994). In vivo imaging by MRI scans in psycho-
metrically assessed individuals supports this. A
decrease in white matter volume in the parahippo-
campal gyrus that includes the perforant path is
found already in amnesic mild cognitive impaired
individuals (the MMSE scores suggest that this
might represent very mild AD) (Petersen et al.,
2006; Stoub et al., 2006).
Post-synaptically, the loss of perforant path in-
put seems to induce changes in granule cells. Gran-
ule cells remodel their dendritic tree in AD as
determined in quantitative morphometrical analy-
ses on Golgi stained (Flood et al., 1985, 1987a, b)
and with Lucifer yellow intracellularly filled (Ein-
stein et al., 1994) neurons. In the Golgi studies, the
granule cells’ apical dendrites of severe Alzheimer
cases are compared to controls. The controls con-
sist of three differentially aged groups of non-
demented individuals. The middle-aged group
(mean of age 52.2972.60 SEM), the old-aged
group (comparable to the Alzheimer group)
(73.4072.36 years) and the very old-aged group
(90.2071.85 years), however, have various degrees
of AD-related histopathology. Flood and col-
leagues have found a significantly reduced total
dendritic length (37%) and a significantly reduced
average segment length (dendritic length/number of
segments; 34%) but no significant changes in
total segment numbers or width of the fascia den-
tata (granule cell layer + molecular layer) when
comparing the old-aged group with the Alzheimer
group. Interestingly, the Alzheimer group has only
a 16% shorter total dendritic length than the mid-
dle-aged group, which has the lowest degree of
AD-related neuropathology (no changes or only
‘very slight or slight changes’). This difference as
well that between the very old group and the
Alzheimer group is not statistically significant in
post-hoc tests. It is tempting to speculate that the
difference in the dendritic tree measures between
the middle-aged group (with no or only very few
AD-related neuropathological changes) and the old
group (with some more changes but no clinical
signs of AD) might represent an early stage in the
evolution of AD where regenerative plasticity is
734
dominant (Flood et al., 1985). Later, when differ-
entiation proceeds the balance between regenerative
(progressive) and degenerative (regressive) plasticity
should shift towards the formation of a reduced
dendritic arborisation. Similar to the Golgi studies,
the analysis of Lucifer yellow-filled apical granule
cell dendrites unveils a statistically significant re-
duction in dendritic length (by 50%) and also a
shorter average segment length (19%) (Einstein et
al., 1994). Proliferative signs, however, such as in-
creased number of dendritic twigs or unusual ex-
crescences are not observed.
Spine numbers are reduced in AD cases by
35–60% (Williams and Matthysse, 1986; De Ruiter
and Uylings, 1987; Einstein et al., 1994; Ji et al.,
2003). However, it is likely that spine plasticity will
also occur. In early stages, first loss of afferents may
induce signals, which allow the local sprouting in
order to attract ‘new’ synaptic input from
neighboured afferents. Given that
400 axons
(synapses) may be necessary to discharge a granule
cell (McNaughton et al., 1991), this may be an at-
tempt to keep the neuronal net properly working.
Additionally, the existing terminals are statistically
significantly enlarged in terms of apposition length
(+18% in outer molecular layer and +12% in the
inner molecular layer) and — as a mild trend — as
well as total synaptic contact area [in outer and in-
ner molecular layer +4% (n.s.)] (Scheff et al., 1996;
Scheff and Price, 1998) probably to serve the same
purpose. Later, when differentiation proceeds, the
balance between regenerative (progressive) and de-
generative (regressive) plasticity shifts towards the
formation of synaptically stripped dendrites (Scheff
and Price, 2003). AP180, a protein considered to be
crucially involved in neuronal clathrin-mediated
synaptic recycling, is expressed in granule cells and
stains the molecular layer of cognitive normal
humans homogeneously. In AD, a loss of granule
cell body staining but an increase in the staining
intensity and width of the inner molecular layer is
reported (Yao et al., 1999). This suggests that gran-
ule cells show plastic adaptation probably respond-
ing to challenges on their dendritic input.
Apart from disconnection of principal cells, in-
terneurons may also become disconnected from the
entorhinal input. This may alter the balance be-
tween excitatory and inhibitory effects. A
comparative morphometrical analysis of apical
and basal dendrites of parvalbumin-containing
GABAergic interneurons located within or directly
at the border of the granule cell layer has been
made in Braak-staged cases (Ohm et al., 2002). A
non-significant and transient increase in dendritic
length, branch order and number of segments of
the apical dendrites is reported for an early Braak-
stage (stage II) followed by a statistically significant
decline of the respective measures in stages IV and
V. Basal dendrites, in contrast, remained stable.
Because apical dendrites receive entorhinal input
from pre-a neurons whereas basal dendrites do not
(Zipp et al., 1989), the dendrite-specific change
suggests an input-specific plastic change. A
35–50% pre-a layer neuron loss is reported for
mild cognitive decline (Kordower et al., 2001; Price
et al., 2001), which may represent Braak-stage III
(Hof et al., 2003). It is therefore likely that a sub-
stantial loss of input (>50%) is required to induce
the regressive changes which may eventually lead
to the interneuron’s own death. It may also play an
important role how fast this input decreases. In-
terestingly, the transient increase in dendritic
length, segment numbers and branches (Ohm et
al., 2002) parallels a transient tendency for more
granule cells in pre-clinical Alzheimer (+13%) of
Braak stages II–IV (West et al., 2004). The latter
may be the result of an increase in neurogenesis, as
it appears to occur in some Alzheimer brain (Jin et
al., 2004), the former as an attempt to sprout.
Closing remark
The available data suggest that the dentate gyrus
withstands the formation of plaques, tangles and
neuronal death until late stages of AD. However,
changes related to a disconnecting process, mainly
from loss of entorhinal input, and more subtle in-
trinsic alterations may contribute to disturbances
in memory and learning observed already in early
stages of AD.
References
Amaral, D.G., Insausti, R. and Cowan, W.M. (1984) The com-
missural connections of the monkey hippocampal formation.
J. Comp. Neurol., 224: 307–336.

Arai, H., Emson, P.C., Mountjoy, C.Q., Carassco, L.H. and
Heizmann, C.W. (1987) Loss of parvalbumin-immunoreac-
tive neurones from cortex in Alzheimer-type dementia. Brain
Res., 418: 164–169.
Aronica, E., Dickson, D.W., Kress, Y., Morrison, J.H. and
Zukin, R.S. (1998) Non-plaque dystrophic dendrites in
Alzheimer hippocampus: a new pathological structure
revealed by glutamate receptor immunocytochemistry.
Neuroscience, 82: 979–991.
Augustinack, J.C., Schneider, A., Mandelkow, E.M. and
Hyman, B.T. (2002) Specific tau phosphorylation sites cor-
relate with severity of neuronal cytopathology in Alzheimer’s
disease. Acta Neuropathol. (Berl.), 103: 26–35.
Bancher, C., Braak, H., Fischer, P. and Jellinger, K.A. (1993)
Neuropathological staging of Alzheimer lesions and intellec-
tual status in Alzheimer’s and Parkinson’s disease patients.
Neurosci. Lett., 162: 179–182.
Bancher, C., Brunner, C., Lassmann, H., Budka, H.,
Jellinger, K., Wiche, G., Seitelberger, F., Grundke Iqbal, I.,
Iqbal, K. and Wisniewski, H.M. (1989) Accumulation of
abnormally phosphorylated tau precedes the formation of
neurofibrillary tangles in Alzheimer’s disease. Brain Res.,
477: 90–99.
Bernstein, H.G., Blazejczyk, M., Rudka, T., Gundelfinger,
E.D., Dobrowolny, H., Bogerts, B., Kreutz, M.R., Kuznicki,
J. and Wojda, U. (2005) The Alzheimer disease-related cal-
cium-binding protein Calmyrin is present in human forebrain
with an altered distribution in Alzheimer’s as compared to
normal ageing brains. Neuropathol. Appl. Neurobiol., 31:
314–324.
Bertoni-Freddari, C., Fattoretti, P., Casoli, T., Caselli, U. and
Meier-Ruge, W. (1996) Deterioration threshold of synaptic
morphology in aging and senile dementia of Alzheimer’s
type. Anal. Quant. Cytol. Histol., 18: 209–213.
Bertoni-Freddari, C., Meier-Ruge, W. and Ulrich, J. (1988)
Quantitative morphology of synaptic plasticity in the aging
brain. Scanning Microsc., 2: 1027–1034.
Bobinski, M., Wegiel, J., Tarnawski, M., de Leon, M.J.,
Reisberg, B., Miller, D.C. and Wisniewski, H.M. (1998) Du-
ration of neurofibrillary changes in the hippocampal pyrami-
dal neurons. Brain Res., 799: 156–158.
Bobinski, M., Wegiel, J., Tarnawski, M., Reisberg, B., de Leon,
M.J., Miller, D.C. and Wisniewski, H.M. (1997) Relation-
ships between regional neuronal loss and neurofibrillary
changes in the hippocampal formation and duration and
severity of Alzheimer disease. J. Neuropathol. Exp. Neurol.,
56: 414–420.
Bobinski, M., Wegiel, J., Wisniewski, H.M., Tarnawski, M.,
Reisberg, B., de Leon, M.J. and Miller, D.C. (1996a)
Neurofibrillary pathology — correlation with hippocampal
formation atrophy in Alzheimer disease. Neurobiol. Aging,
17: 909–919.
Bobinski, M., Wegiel, J., Wisniewski, H.M., Tarnawski, M.,
Reisberg, B., de Leon, M.J. and Miller, D.C. (1996b)
Neurofibrillary pathology — correlation with hippocampal
formation atrophy in Alzheimer disease. Neurobiol. Aging,
17: 909–919.
735
Braak, E., Braak, H. and Mandelkow, E.M. (1994) A sequence
of cytoskeleton changes related to the formation of neuro-
fibrillary tangles and neuropil threads. Acta Neuropathol.
Berl., 87: 554–567.
Braak, E., Strotkamp, B. and Braak, H. (1991) Parvalbumin-
immunoreactive structures in the hippocampus of the human
adult. Cell Tissue Res., 264: 33–48.
Braak, H. and Braak, E. (1991) Neuropathological stageing of
Alzheimer-related changes. Acta Neuropathol. (Berl.), 82:
239–259.
Braak, H., Braak, E., Grundke-Iqbal, I. and Iqbal, K. (1986)
Occurrence of neuropil threads in the senile human brain and
in Alzheimer’s disease: a third location of paired helical
filaments outside of neurofibrillary tangles and neuritic
plaques. Neurosci. Lett., 65: 351–355.
Braak, H. and Del Tredici, K. (2004) Poor and protracted
myelination as a contributory factor to neurodegenerative
disorders. Neurobiol. Aging, 25: 19–23.
Braak, H., Duyckaerts, C., Braak, E. and Piette, F. (1993)
Neuropathological staging of Alzheimer-related changes
correlates with psychometrically assessed intellectual status.
In: Corain B., Iqbal K., Nicolini M., Winblad B., Wisniewski
H. and Zatta P. (Eds.), Alzheimer’s Disease: Advances in
Clinical and Basic Research. Wiley, Chicester, pp. 131–137.
Brady, D.R. and Mufson, E.J. (1997) Parvalbumin-immunore-
active neurons in the hippocampal formation of Alzheimer’s
diseased brain. Neuroscience, 80: 1113–1125.
Brion, J.P. and Resibois, A. (1994) A subset of calretinin-
positive neurons are abnormal in Alzheimer’s disease. Acta
Neuropathol. (Berl.), 88: 33–43.
Bush, A.I., Pettingell, W.-H.J., de Paradis, M., Tanzi, R.E. and
Wasco, W. (1994) The amyloid beta-protein precursor and its
mammalian homologues. Evidence for a zinc-modulated
heparin-binding superfamily. J. Biol. Chem., 269: 26618–26621.
Cabalka, L.M., Hyman, B.T., Goodlett, C.R., Ritchie, T.C.
and Vanhoesen, G.W. (1992) Alteration in the pattern of
nerve terminal protein immunoreactivity in the perforant
pathway in Alzheimer’s disease and in rats after entorhinal
lesions. Neurobiol. Aging, 13: 283–291.
Callahan, L.M. and Coleman, P.D. (1995) Neurons bearing
neurofibrillary tangles are responsible for selected synaptic
deficits in Alzheimer’s disease. Neurobiol. Aging, 16:
311–314.
Carter, T.L., Rissman, R.A., Mishizen-Eberz, A.J., Wolfe,
B.B., Hamilton, R.L., Gandy, S. and Armstrong, D.M.
(2004) Differential preservation of AMPA receptor subunits
in the hippocampi of Alzheimer’s disease patients according
to Braak stage. Exp. Neurol., 187: 299–309.
Ceballos, I., Javoy-Agid, F., Delacourte, A., Defossez, A.,
Lafon, M., Hirsch, E., Nicole, A., Sinet, P.M. and Agid, Y.
(1991) Neuronal localization of copper-zinc superoxide di-
smutase protein and mRNA within the human hippocampus
from control and Alzheimer’s disease brains. Free Radic.
Res. Commun., 12–13(Pt 2): 571–580.
Chan-Palay, V. (1987) Somatostatin immunoreactive neurons
in the human hippocampus and cortex shown by immuno-
gold/silver intensification on vibratome sections: coexistence
736
with neuropeptide Y neurons, and effects in Alzheimer-type
dementia. J. Comp. Neurol., 260: 201–223.
Chan-Palay, V., Lang, W., Haesler, U., Köhler, C. and
Yasargil, G. (1986) Distribution of altered hippocampal
neurons and axons immunoreactive with antisera against
neuropeptide Y in Alzheimer’s-type dementia. J. Comp.
Neurol., 248: 376–394.
Chu, D.C.M., Penney Jr., J.B. and Young, A.B. (1987) Quan-
titative autoradiography of hippocampal GABAB and
GABAA receptor changes in Alzheimer’s disease. Neurosci.
Lett., 82: 246–252.
Crain, B.J. and Burger, P.C. (1988) The laminar distribution of
neuritic plaques in the fascia dentata of patients with
Alzheimer’s disease. Acta Neuropathol. (Berl.), 76: 87–93.
Davies, C.A., Mann, D.M.A., Sumpter, P.Q. and Yates, P.O.
(1987) A quantitative morphometric analysis of the neuronal
and synaptic content of the frontal and temporal cortex in
patients with Alzheimer’s disease. J. Neurol. Sci., 78:
151–164.
Davies, P. and Terry, R.D. (1981) Cortical somatostatin-like
immunoreactivity in cases of Alzheimer’s disease and senile
dementia of the Alzheimer type. Neurobiol. Aging, 2: 9–14.
De Ruiter, J.P. and Uylings, H.B.M. (1987) Morphometric and
dendritic analysis of fascia dentata granule cells in human
aging and senile dementia. Brain Res., 402: 217–229.
Dekosky, S.T., Scheff, S.W. and Styren, S.D. (1996) Structural
correlates of cognition in dementia: quantification and as-
sessment of synapse change. Neurodegeneration, 5: 417–421.
Demeter, S., Rosene, D.L. and van Hoesen, G.W. (1985) In-
terhemispheric pathways of the hippocampal formation, pre-
subiculum, and entorhinal and posterior parahippocampal
cortex in the rhesus monkey: the structure and organization
of the hippocampal commissures. J. Comp. Neurol., 233:
30–47.
Dewar, D., Chalmers, D.T., Graham, D.I. and McCulloch, J.
(1991) Glutamate metabotropic and AMPA binding sites are
reduced in Alzheimer’s disease — an autoradiographic study
of the hippocampus. Brain Res., 553: 58–64.
Duyckaerts, C., Colle, M.A., Seilhean, D. and Hauw, J.J.
(1998) Laminar spongiosis of the dentate gyrus: a sign of
disconnection, present in cases of severe Alzheimer’s disease.
Acta Neuropathol. (Berl.), 95: 413–420.
Dyrks, T., Weidemann, A., Multhaup, G., Salbaum, J.M.,
Lemaire, H.G., Kang, J., Müller-Hill, B., Masters, C.L. and
Beyreuther, K. (1988) Identification, transmembrane orien-
tation and biogenesis of the amyloid A4 precursor of
Alzheimer’s disease. EMBO J., 7: 949–957.
Einstein, G., Buranosky, R. and Crain, B.J. (1994) Dendritic
pathology of granule cells in Alzheimer’s disease is unrelated
to neuritic plaques. J. Neurosci., 14: 5077–5088.
Flood, D.G., Buell, S.J., Defiore, C.H., Horwitz, G.J. and
Coleman, P.D. (1985) Age-related dendritic growth in
dentate gyrus of human brain is followed by regression in
the ‘oldest old’. Brain Res., 345: 366–368.
Flood, D.G., Buell, S.J., Horwitz, G.J. and Coleman, P.D.
(1987a) Dendritic extent in human dentate gyrus granule cells
in normal aging and senile dementia. Brain Res., 402: 205–216.
Flood, D.G., Guarnaccia, M. and Coleman, P.D. (1987b) Den-
dritic extent in human CA2-3 hippocampal pyramidal neu-
rons in normal aging and senile dementia. Brain Res., 409:
88–96.
Fonseca, M. and Soriano, E. (1995) Calretinin-immunoreactive
neurons in the normal human temporal cortex and in Al-
zheimer’s disease. Brain Res., 691: 83–91.
Fonseca, M., Soriano, E., Ferrer, I., Martinez, A. and Tunon,
T. (1993) Chandelier cell axons identified by parvalbumin-
immunoreactivity in the normal human temporal cortex and
in Alzheimer’s disease. Neuroscience, 55: 1107–1116.
Freund, T.F. and Buzsaki, G. (1996) Interneurons of the
hippocampus. Hippocampus, 6: 347–470.
Furuta, A., Iida, T., Nakabeppu, Y. and Iwaki, T. (2001) Ex-
pression of hMTH1 in the hippocampi of control and
Alzheimer’s disease. Neuroreport, 12: 2895–2899.
Garcia-Sierra, F., Hauw, J.J., Duyckaerts, C., Wischik, C.M.,
Munoz, J. and Mena, R. (2000) The extent of neurofibrillary
pathology in perforant pathway neurons is the key determi-
nant of dementia in the very old. Acta Neuropathol. (Berl.),
100: 29–35.
Garcia-Sierra, F., Wischik, C.M., Harrington, C.R., Luna-Mu-
noz, J. and Mena, R. (2001) Accumulation of C-terminally
truncated tau protein associated with vulnerability of the
perforant pathway in early stages of neurofibrillary pathol-
ogy in Alzheimer’s disease. J. Chem. Neuroanat., 22: 65–77.
Geddes, J.W., Ulas, J., Brunner, L.C., Choe, W. and Cotman,
C.W. (1992) Hippocampal excitatory amino acid receptors in
elderly, normal individuals and those with Alzheimer’s dis-
ease: non-N-methyl-D-aspartate receptors. Neuroscience, 50:
23–34.
Goldsmith, S.K. and Joyce, J.N. (1995) Alterations in hippo-
campal mossy fiber pathway in schizophrenia and Al-
zheimer’s disease. Biol. Psychiatry, 37: 122–126.
Gomez Isla, T., Price, J.L., McKeel Jr., D.W., Morris, J.C.,
Growdon, J.H. and Hyman, B.T. (1996) Profound loss of
layer II entorhinal cortex neurons occurs in very mild Al-
zheimer’s disease. J. Neurosci., 16: 4491–4500.
Greene, J.R., Radenahmad, N., Wilcock, G.K., Neal, J.W. and
Pearson, R.C. (2001) Accumulation of calbindin in cortical
pyramidal cells with ageing; a putative protective mechanism
which fails in Alzheimer’s disease. Neuropathol. Appl.
Neurobiol., 27: 339–342.
von Gunten, A., Kovari, E., Bussiere, T., Rivara, C.B., Gold,
G., Bouras, C., Hof, P.R. and Giannakopoulos, P. (2006)
Cognitive impact of neuronal pathology in the entorhinal
cortex and CA1 field in Alzheimer’s disease. Neurobiol.
Aging, 27: 270–277.
Haass, C. (2004) Take five — BACE and the gamma-secretase
quartet conduct Alzheimer’s amyloid beta-peptide genera-
tion. EMBO J., 23: 483–488.
Harding, A.J., Halliday, G.M. and Kril, J.J. (1998) Variation in
hippocampal neuron number with age and brain volume.
Cereb. Cortex, 8: 710–718.
Hof, P.R., Bussiere, T., Gold, G., Kovari, E., Giannakopoulos,
P., Bouras, C., Perl, D.P. and Morrison, J.H. (2003) Stereo-
logic evidence for persistence of viable neurons in layer II of

the entorhinal cortex and the CA1 field in Alzheimer disease.
J. Neuropathol. Exp. Neurol., 62: 55–67.
Hof, P.R., Cox, K., Young, W.G., Celio, M.R., Rogers, J. and
Morrison, J.H. (1991) Parvalbumin-immunoreactive neurons
in the neocortex are resistant to degeneration in Alzheimer’s
disease. J. Neuropathol. Exp. Neurol., 50: 451–462.
Hof, P.R., Nimchinsky, E.A., Celio, M.R., Bouras, C. and
Morrison, J.H. (1993) Calretinin-immunoreactive neocortical
interneurons are unaffected in Alzheimer’s disease. Neurosci.
Lett., 152: 145–148.
Howell, O., Atack, J.R., Dewar, D., Mckernan, R.M. and Sur,
C. (2000) Density and pharmacology of alpha5 subunit-con-
taining GABA(A) receptors are preserved in hippocampus of
Alzheimer’s disease patients. Neuroscience, 98: 669–675.
Hyman, B.T., van Hoesen, G.W. and Damasio, A.R. (1987)
Alzheimer’s disease: glutamate depletion in the hippocampal
perforanth pathway zone. Ann. Neurol., 22: 37–40.
Hyman, B.T., Van Horsen, G.W., Damasio, A.R. and Barnes,
C.L. (1984) Alzheimer’s disease: cell-specific pathology iso-
lates the hippocampal formation. Science, 225: 1168–1170.
Hyman, B.T., Kromer, L.J. and van Hoesen, G.W. (1988) A
direct demonstration of the perforath pathway terminal zone
in Alzheimer’s disease using the monoclonal antibody Alz-50.
Brain Res., 450: 392–397.
Hyman, B.T., Marzloff, K., Wenniger, J.J., Dawson, T.M.,
Bredt, D.S. and Snyder, S. (1992) Relative sparing of nitric
oxide synthase-containing neurons in the hippocampal for-
mation in Alzheimer’s disease. Ann. Neurol., 32: 818–820.
Hyman, B.T., Penney, J.-B.J., Blackstone, C.D. and Young,
A.B. (1994) Localization of non-N-methyl-D-aspartate gluta-
mate receptors in normal and Alzheimer hippocampal for-
mation. Ann. Neurol., 35: 31–37.
Iacopino, A.M. and Christakos, S. (1990) Specific reduction of
calcium-binding protein (28-kilodalton calbindin-D) gene ex-
pression in aging and neurodegenerative diseases. Proc. Natl.
Acad. Sci. U.S.A., 87: 4078–4082.
Ikonomovic, M.D., Mizukami, K., Davies, P., Hamilton, R.,
Sheffield, R. and Armstrong, D.M. (1997) The loss of
GluR2(3) immunoreactivity precedes neurofibrillary tangle
formation in the entorhinal cortex and hippocampus of Al-
zheimer brains. J. Neuropathol. Exp. Neurol., 56: 1018–1027.
Ikonomovic, M.D., Mizukami, K., Warde, D., Sheffield, R.,
Hamilton, R., Wenthold, R.J. and Armstrong, D.M. (1999)
Distribution of glutamate receptor subunit NMDAR1 in the
hippocampus of normal elderly and patients with Al-
zheimer’s disease. Exp. Neurol., 160: 194–204.
Ikonomovic, M.D., Sheffield, R. and Armstrong, D.M. (1995)
AMPA-selective glutamate receptor subtype immunoreactiv-
ity in the hippocampal formation of patients with Al-
zheimer’s disease. Hippocampus, 5: 469–486.
Iritani, S., Niizato, K. and Emson, P.C. (2001) Relationship of
calbindin D28K-immunoreactive cells and neuropathological
changes in the hippocampal formation of Alzheimer’s dis-
ease. Neuropathology, 21: 162–167.
Iwakiri, M., Mizukami, K., Ikonomovic, M.D., Ishikawa, M.,
Hidaka, S., Abrahamson, E.E., Dekosky, S.T. and Asada, T.
(2005) Changes in hippocampal GABABR1 subunit
737
expression in Alzheimer’s patients: association with Braak
staging. Acta Neuropathol. (Berl.), 109: 467–474.
Iwamoto, N. and Emson, P.C. (1991) Demonstration of ne-
urofibrillary tangles in parvalbumin-immunoreactive inter-
neurones in the cerebral cortex of Alzheimer-type dementia
brain. Neurosci. Lett., 128: 81–84.
Jansen, K.L.R., Faull, R.L.M., Dragunow, M. and Synek, B.L.
(1990) Alzheimer’s disease — changes in hippocampal N-
methyl-D-aspartate, quisqualate, neurotensin, adenosine,
benzodiazepine, serotonin and opioid receptors — an auto-
radiographic study. Neuroscience, 39: 613–627.
Jellinger, K.A. (1997) Neuropathological staging of Alzheimer-
related lesions: the challenge of establishing relations to age.
Neurobiol. Aging, 18: 369–375.
Jellinger, K.A. and Bancher, C. (1997) Proposals for re-eval-
uation of current autopsy criteria for the diagnosis of Al-
zheimer’s disease. Neurobiol. Aging, 18: S55–S65.
Ji, Y., Gong, Y., Gan, W., Beach, T., Holtzman, D.W. and
Wisniewski, T. (2003) Apolipoprotein E isoform-specific reg-
ulation of dendritic spine morphology in apolipoprotein E
transgenic mice and Alzheimer’s disease patients. Neurosci-
ence, 122: 305–315.
Jin, K., Peel, A.L., Mao, X.O., Xie, L., Cottrell, B.A., Henshall,
D.C. and Greenberg, D.A. (2004) Increased hippocampal
neurogenesis in Alzheimer’s disease. Proc. Natl. Acad. Sci.
U.S.A., 101: 343–347.
Kaufmann, W.A., Barnas, U., Humpel, C., Nowakowski, K.,
DeCol, C., Gurka, P., Ransmayr, G., Hinterhuber, H., Win-
kler, H. and Marksteiner, J. (1998) Synaptic loss reflected by
secretoneurin-like immunoreactivity in the human hippo-
campus in Alzheimer’s disease. Eur. J. Neurosci., 10:
1084–1094.
Khatchaturian, Z.S. (1985) Diagnosis of Alzheimer’s disease.
Arch. Neurol., 31: 545–548.
Kiktenko, A.I., Uranova, N.A. and Denisov, D.V. (1997)
Quantitative characteristics of changes in synaptic contacts in
the hippocampus in Alzheimer’s disease. Neurosci. Behav.
Physiol., 27: 681–682.
Korbo, L., Amrein, I., Lipp, H.P., Wolfer, D., Regeur, L., Os-
ter, S. and Pakkenberg, B. (2004) No evidence for loss of
hippocampal neurons in non-Alzheimer dementia patients.
Acta Neurol. Scand., 109: 132–139.
Kordower, J.H., Chu, Y., Stebbins, G.T., Dekosky, S.T., Co-
chran, E.J., Bennett, D. and Mufson, E.J. (2001) Loss and
atrophy of layer II entorhinal cortex neurons in elderly peo-
ple with mild cognitive impairment. Ann. Neurol., 49:
202–213.
Kowall, N.W., Quigley, B.-J.J., Krause, J.E., Lu, F., Kosofsky,
B.E. and Ferrante, R.J. (1993) Substance P and substance P
receptor histochemistry in human neurodegenerative dis-
eases. Regul. Pept., 46: 174–185.
Kril, J.J., Patel, S., Harding, A.J. and Halliday, G.M. (2002)
Neuron loss from the hippocampus of Alzheimer’s disease
exceeds extracellular neurofibrillary tangle formation. Acta
Neuropathol. (Berl.), 103: 370–376.
Kulmala, H.K. (1985) Immunocytochemical localization of
enkephalin-like immunoreactivity in neurons of human
738
hippocampal formation: effects of ageing and Alzheimer’s
disease. Neuropathol. Appl. Neurobiol., 11: 105–115.
Kumar, U. (2005) Expression of somatostatin receptor subtypes
(SSTR1-5) in Alzheimer’s disease brain: an immunohisto-
chemical analysis. Neuroscience, 134: 525–538.
Leuba, G., Kraftsik, R. and Saini, K. (1998) Quantitative dis-
tribution of parvalbumin, calretinin, and calbindin D-28k
immunoreactive neurons in the visual cortex of normal and
Alzheimer cases. Exp. Neurol., 152: 278–291.
Lim, C., Blume, H.W., Madsen, J.R. and Saper, C.B. (1997)
Connections of the hippocampal formation in humans: I. The
mossy fiber pathway. J. Comp. Neurol., 385: 325–351.
Lotstra, F. and Vanderhaeghen, J.J. (1987) Distribution of
immunoreactive cholecystokinin in the human hippocampus.
Peptides, 8: 911–920.
Lozsadi, D.A. and Larner, A.J. (2006) Prevalence and causes of
seizures at the time of diagnosis of probable Alzheimer’s
disease. Dement. Geriatr. Cogn. Disord., 22: 121–124.
Magloczky, Z., Halasz, P., Vajda, J., Czirjak, S. and Freund,
T.F. (1997) Loss of Calbindin-D28K immunoreactivity from
dentate granule cells in human temporal lobe epilepsy.
Neuroscience, 76: 377–385.
Magloczky, Z., Wittner, L., Borhegyi, Z., Halasz, P., Vajda, J.,
Czirjak, S. and Freund, T.F. (2000) Changes in the distribu-
tion and connectivity of interneurons in the epileptic human
dentate gyrus. Neuroscience, 96: 7–25.
Maguire-Zeiss, K.A., Li, Z.W., Shimoda, L.M. and Hamill,
R.W. (1995) Calbindin D28k mRNA in hippocampus, supe-
rior temporal gyrus and cerebellum: comparison between
control and Alzheimer disease subjects. Brain Res. Mol.
Brain Res., 30: 362–366.
Marksteiner, J., Lechner, T., Kaufmann, W.A., Gurka, P.,
Humpel, C., Nowakowski, C., Maier, H. and Jellinger, K.A.
(2000) Distribution of chromogranin B-like immunoreactiv-
ity in the human hippocampus and its changes in Alzheimer’s
disease. Acta Neuropathol. (Berl.), 100: 205–212.
Masliah, E., Mallory, M., Hansen, L., Deteresa, R., Alford, M.
and Terry, R. (1994) Synaptic and neuritic alterations during
the progression of Alzheimer’s disease. Neurosci. Lett., 174:
67–72.
Mathieu-Kia, A.M., Fan, L.Q., Kreek, M.J., Simon, E.J. and
Hiller, J.M. (2001) Mu-, delta- and kappa-opioid receptor
populations are differentially altered in distinct areas of
postmortem brains of Alzheimer’s disease patients. Brain
Res., 893: 121–134.
McNamara, M.J., Ruff, C.T., Wasco, W., Tanzi, R.E.,
Thinakaran, G. and Hyman, B.T. (1998) Immunohistochem-
ical and in situ analysis of amyloid precursor-like protein-1
and amyloid precursor-like protein-2 expression in Alzheimer
disease and aged control brains. Brain Res., 804: 45–51.
McNaughton, B.L., Barnes, C.A., Mizumori, S.J.Y., Green,
E.J. and Sharp, P.E. (1991) In: Morell F. (Ed.), Kindling and
Synaptic Plasticity. Birkhäuser, Basel, pp. 110–123.
Mikkonen, M., Alafuzoff, I., Tapiola, T., Soininen, H. and Mi-
ettinen, R. (1999) Subfield- and layer-specific changes in parv-
albumin, calretinin and calbindin-D28K immunoreactivity in
the entorhinal cortex in Alzheimer’s disease. Neuroscience, 92:
515–532.
Mishizen-Eberz, A.J., Rissman, R.A., Carter, T.L.,
Ikonomovic, M.D., Wolfe, B.B. and Armstrong, D.M.
(2004) Biochemical and molecular studies of NMDA recep-
tor subunits NR1/2A/2B in hippocampal subregions
throughout progression of Alzheimer’s disease pathology.
Neurobiol. Dis., 15: 80–92.
Mizukami, K., Grayson, D.R., Ikonomovic, M.D., Sheffield,
R. and Armstrong, D.M. (1998) GABAA receptor beta 2 and
beta 3 subunits mRNA in the hippocampal formation of
aged human brain with Alzheimer-related neuropathology.
Brain Res. Mol. Brain Res., 56: 268–272.
Mizukami, K., Ikonomovic, M.D., Grayson, D.R., Rubin,
R.T., Warde, D., Sheffield, R., Hamilton, R.L., Davies, P.
and Armstrong, D.M. (1997) Immunohistochemical study of
GABA(A) receptor beta2/3 subunits in the hippocampal for-
mation of aged brains with Alzheimer-related neuropatho-
logic changes. Exp. Neurol., 147: 333–345.
Morris, R.G.M. (2006) Elements of a neurobiological theory of
hippocampal function: the role of synaptic plasticity, synap-
tic tagging and schemas. Eur. J. Neurosci., 23: 2829–2846.
Morsch, R., Simon, W. and Coleman, P.D. (1999) Neurons
may live for decades with neurofibrillary tangles. J.
Neuropathol. Exp. Neurol., 58: 188–197.
Morys, J., Sadowski, M., Barcikowska, M., Maciejewska, B.
and Narkiewicz, O. (1994) The second layer neurones of the
entorhinal cortex and the perforant path in physiological
ageing and Alzheimer’s disease. Acta Neurobiol. Exp.
(Wars.), 54: 47–53.
Nagerl, U.V., Mody, I., Jeub, M., Lie, A.A., Elger, C.E. and
Beck, H. (2000) Surviving granule cells of the sclerotic human
hippocampus have reduced Ca(2+) influx because of a loss
of calbindin-D(28k) in temporal lobe epilepsy. J. Neurosci.,
20: 1831–1836.
Nakazawa, K., Quirk, M.C., Chitwood, R.A., Watanabe, M.,
Yeckel, M.F., Sun, L.D., Kato, A., Carr, C.A., Johnston, D.,
Wilson, M.A. and Tonegawa, S. (2002) Requirement for
hippocampal CA3 NMDA receptors in associative memory
recall. Science, 297: 211–218.
van de Nes, J.A., Sandmann-Keil, D. and Braak, H. (2002)
Interstitial cells subjacent to the entorhinal region expressing
somatostatin-28 immunoreactivity are susceptible to devel-
opment of Alzheimer’s disease-related cytoskeletal changes.
Acta Neuropathol. (Berl.), 104: 351–356.
Nitsch, R. and Ohm, T.G. (1995) Calretinin immunoreactive
structures in the human hippocampal formation. J. Comp.
Neurol., 360: 475–487.
Ohm, T.G., Müller, H., Braak, H. and Bohl, J. (1995) Close-
meshed prevalence rates of different stages as a tool to un-
cover the rate of Alzheimer’s disease-related neurofibrillary
changes. Neuroscience, 64: 209–217.
Ohm, T.G., Münch, S., Schönheit, B., Zarski, R. and Nitsch, R.
(2002) Transneuronally altered dendritic processing of tan-
gle-free neurons in Alzheimer’s disease. Acta Neuropathol.
(Berl.), 103: 437–443.

Palop, J.J., Jones, B., Kekonius, L., Chin, J., Yu, G.Q., Raber,
J., Masliah, E. and Mucke, L. (2003) Neuronal depletion of
calcium-dependent proteins in the dentate gyrus is tightly
linked to Alzheimer’s disease-related cognitive deficits. Proc.
Natl. Acad. Sci. U.S.A., 100: 9572–9577.
Pellegrini-Giampietro, D.E., Bennett, M.V. and Zukin, R.S.
(1994) AMPA/kainate receptor gene expression in normal
and Alzheimer’s disease hippocampus. Neuroscience, 61:
41–49.
Penney, J.B., Maragos, W.F., Greenamyre, J., Debowey, D.L.,
Hollingsworth, Z. and Young, A.B. (1990) Excitatory amino
acid binding sites in the hippocampal region of Alzheimer’s
disease and other dementias. J. Neurol. Neurosurg. Psychi-
atry, 53: 314–320.
Petersen, R.C., Parisi, J.E., Dickson, D.W., Johnson, K.A.,
Knopman, D.S., Boeve, B.F., Jicha, G.A., Ivnik, R.J., Smith,
G.E., Tangalso, E.G., Braak, H. and Kokmen, E. (2006)
Neuropathologic features of amnestic mild cognitive impair-
ment. Arch. Neurol., 63: 645–646.
Price, J.L., Ko, A.I., Wade, M.J., Tsou, S.K., Mckeel, D.W.
and Morris, J.C. (2001) Neuron number in the entorhinal
cortex and CA1 in preclinical Alzheimer disease. Arch. Ne-
urol., 58: 1395–1402.
Rebeck, G.W., Marzloff, K. and Hyman, B.T. (1993) The pat-
tern of NADPH-diaphorase staining, a marker of nitric oxide
synthase activity, is altered in the perforant pathway terminal
zone in Alzheimer’s disease. Neurosci. Lett., 152: 165–168.
Represa, A., Duyckaerts, C., Tremblay, E., Hauw, J.J. and Ben
Ari, Y. (1988) Is senile dementia of the Alzheimer type as-
sociated with hippocampal plasticity? Brain Res., 457:
355–359.
Rinne, J.O., Lonnberg, P., Marjamaki, P., Molsa, P., Sako, E.
and Paljarvi, L. (1993) Brain methionine- and leucine-
enkephalin receptors in patients with dementia. Neurosci.
Lett., 161: 77–80.
Rissman, R.A., Mishizen-Eberz, A.J., Carter, .L., Wolfe, B.B.,
De Blas, A.L., Miralles, C.P., Ikonomovic, M.D. and Arm-
strong, D.M. (2003) Biochemical analysis of GABA(A) re-
ceptor subunits alpha 1, alpha 5, beta 1, beta 2 in the
hippocampus of patients with Alzheimer’s disease neuropa-
thology. Neuroscience, 120: 695–704.
Sampson, V.L., Morrison, J.H. and Vickers, J.C. (1997) The
cellular basis for the relative resistance of parvalbumin and
calretinin immunoreactive neocortical neurons to the pathol-
ogy of Alzheimer’s disease. Exp. Neurol., 145: 295–302.
Satoh, J., Tabira, T., Sano, M., Nakayama, H. and Tateishi, J.
(1991) Parvalbumin-immunoreactive neurons in the human
central nervous system are decreased in Alzheimer’s disease.
Acta Neuropathol. (Berl.), 81: 388–395.
Scheff, S.W. and Price, D.A. (1998) Synaptic density in the
inner molecular layer of the hippocampal dentate gyrus in
Alzheimer disease. J. Neuropathol. Exp. Neurol., 57:
1146–1153.
Scheff, S.W. and Price, D.A. (2003) Synaptic pathology in Al-
zheimer’s disease: a review of ultrastructural studies. Ne-
urobiol. Aging, 24: 1029–1046.
739
Scheff, S.W., Sparks, D.L. and Price, D.A. (1996) Quantitative
assessment of synaptic density in the outer molecular layer of
the hippocampal dentate gyrus in Alzheimer’s disease.
Dementia, 7: 226–232.
Schönheit, B., Zarski, R. and Ohm, T.G. (2004) Spatial and
temporal relationships between plaques and tangles in Al-
zheimer-pathology. Neurobiol. Aging, 25: 697–711.
Seress, L. (1988) Interspecies comparison of the hippocampal
formation shows increased on the regio superior in the Am-
mon’s horn of the human brain. J. Hirnforsch., 29: 35–340.
Seress, L. and Frotscher, M. (1990) Morphological variability is
a characteristic feature of granule cells in the primate fascia
dentata: a combined Golgi/electron microscope study. J.
Comp. Neurol., 293: 253.
Seress, L., Gulyas, A.I., Ferrer, I., Tunon, T., Soriano, E. and
Freund, T. (1993) Distribution, morphological features, and
synaptic connections of parvalbumin- and calbindin D28k-
immunoreactive neurons in the human hippocampal forma-
tion. J. Comp. Neurol., 337: 208–230.
Seress, L. and Mrzljak, L. (1987) Basal dendrites of granule
cells are normal features of the fetal and adult dentate gyrus
of both monkey and human hippocampal formations. Brain
Res., 405: 169–174.
Simic, G., Kostovic, I., Winblad, B. and Bogdanovic, N. (1997)
Volume and number of neurons of the human hippocampal
formation in normal aging and Alzheimer’s disease. J. Comp.
Neurol., 379: 482–494.
Singer, D., Lehmann, J., Hanisch, K., Hartig, W. and
Hoffmann, R. (2006) Neighbored phosphorylation sites as
PHF-tau specific markers in Alzheimer’s disease. Biochem.
Biophys. Res. Commun., 346: 819–828.
Sloviter, R.S., Sollas, A.L., Barbaro, N.M. and Laxer, K.D.
(1991) Calcium-binding protein (Calbindin-D28K) and parv-
albumin immunocytochemistry in the normal and epileptic
human hippocampus. J. Comp. Neurol., 308: 381–396.
Solodkin, A., Veldhuizen, S.D. and van Hoesen, G.W. (1996)
Contingent vulnerability of entorhinal parvalbumin-containing
neurons in Alzheimer’s disease. J. Neurosci., 16: 3311–3321.
Spillantini, M.G. and Goedert, M. (1998) Tau protein pathology
in neurodegenerative diseases. Trends Neurosci., 21: 428–433.
Stoub, T.R., Detoledo-Morrell, L., Stebbins, G.T., Leurgans,
S., Bennett, D.A. and Shah, R.C. (2006) Hippocampal dis-
connection contributes to memory dysfunction in individuals
at risk for Alzheimer’s disease. Proc. Natl. Acad. Sci. U.S.A.,
103: 10041–10045.
Thal, D.R., Holzer, M., Rub, U., Waldmann, G., Gunzel, S.,
Zedlick, D. and Schober, R. (2000) Alzheimer-related tau-
pathology in the perforant path target zone and in the hip-
pocampal stratum oriens and radiatum correlates with onset
and degree of dementia. Exp. Neurol., 163: 98–110.
Thal, D.R., Rub, U., Orantes, M. and Braak, H. (2002) Phases
of A beta-deposition in the human brain and its relevance for
the development of AD. Neurology, 58: 1791–1800.
Torrealba, F. and Carrasco, M.A. (2004) A review on electron
microscopy and neurotransmitter systems. Brain Res. Brain
Res. Rev., 47: 5–17.
740
Uchihara, T., Nakamura, A., Yamazaki, M. and Mori, O.
(2001) Evolution from pretangle neurons to neurofibrillary
tangles monitored by thiazin red combined with Gallyas
method and double immunofluorescence. Acta Neuropathol.
(Berl.), 101: 535–539.
Ulas, J. and Cotman, C.W. (1997) Decreased expression of
N-methyl-D-aspartate receptor 1 messenger RNA in select
regions of Alzheimer brain. Neuroscience, 79: 973–982.
Uylings, H.B.M., Ruiz-Marcos, A. and van Pelt, J. (1986) The
metric analysis of three-dimensional dendritic tree patterns: a
methodological review. J. Neurosci. Methods, 18: 127–151.
Wakabayashi, K., Hansen, L.A., Vincent, I., Mallory, M. and
Masliah, E. (1997) Neurofibrillary tangles in the dentate
granule cells of patients with Alzheimer’s disease, Lewy body
disease and progressive supranuclear palsy. Acta
Neuropathol. (Berl.), 93: 7–12.
Wakabayashi, K., Narisawa-Saito, M., Iwakura, Y., Arai, T.,
Ikeda, K., Takahashi, H. and Nawa, H. (1999) Phenotypic
down-regulation of glutamate receptor subunit GluR1 in
Alzheimer’s disease. Neurobiol. Aging, 20: 287–295.
West, M.J., Coleman, P.D., Flood, D.G. and Troncoso, J.C.
(1994) Differences in the pattern of hippocampal neuronal
loss in normal ageing and Alzheimer’s disease. Lancet, 344:
769–772.
West, M.J. and Gundersen, H.J.G. (1990) Unbiased stereolog-
ical estimation of the number of neurons in the human
hippocampus. J. Comp. Neurol., 296: 1–22.
West, M.J., Kawas, C.H., Stewart, W.F., Rudow, G.L. and
Troncoso, J.C. (2004) Hippocampal neurons in pre-clinical
Alzheimer’s disease. Neurobiol. Aging, 25: 1205–1212.
West, M.J. and Slomianka, L. (1998) Total number of neurons
in the layers of the human entorhinal cortex. Hippocampus,
8: 69–82.
Williams, R.S. and Matthysse, S. (1986) Age-related changes
in Down syndrome brain and the cellular pathology of
Alzheimer disease. Prog. Brain Res., 70: 49–67.
Yao, P.J., Morsch, R., Callahan, L.M. and Coleman, P.D.
(1999) Changes in synaptic expression of clathrin assembly
protein AP180 in Alzheimer’s disease analysed by
immunohistochemistry. Neuroscience, 94: 389–394.
Yew, D.T., Li, W.P., Webb, S.E., Lai, H.W. and Zhang, L.
(1999) Neurotransmitters, peptides, and neural cell adhesion
molecules in the cortices of normal elderly humans and Al-
zheimer patients: a comparison. Exp. Gerontol., 34: 117–133.
Zipp, F., Nitsch, R., Soriano, E. and Frotscher, M. (1989)
Entorhinal fibers form synaptic contacts on parvalbumin-
immunoreactive neurons in the rat fascia dentata. Brain Res.,
495: 161–166.
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 40

Hippocampal atrophy and disconnection in incipient

and mild Alzheimer’s disease

Leyla deToledo-Morrell, Travis R. Stoub and Changsheng Wang

Department of Neurological Sciences, Rush University Medical Center, Chicago, IL 60612, USA

Abstract: Quantitative imaging techniques allow the in vivo investigation of age and disease related
changes in the brain and their relation to cognitive function. In this chapter we review imaging evidence
indicating that the entorhinal cortex and hippocampus show atrophy very early in Alzheimer’s disease
(AD) and in individuals who are at risk of developing AD compared to age appropriate controls. Fur-
thermore, the extent and rate of atrophy of the entorhinal cortex, a brain region pathologically involved
very early in the disease process, can predict who among the elderly will develop AD. Techniques that
assess the integrity of white matter further demonstrate that alterations in the parahippocampal white
matter in the region that includes the perforant path could partially disconnect the dentate gyrus and other
hippocampal subfields from incoming sensory information. Such partial disconnection and degradation in
transmission of sensory information in people at risk of AD and in patients with very mild AD could
contribute to the memory dysfunction associated with the early stages of the disease.

Keywords: entorhinal cortex; perforant path; imaging; aging; memory

Introduction The hippocampal dentate gyrus has been impli-

High-resolution structural magnetic resonance im-
aging (MRI) techniques provide a unique tool for
examining alterations in brain anatomy in vivo
during healthy aging and in age-related diseases.
In addition, such techniques allow us to (a) exam-
ine the relation between alterations in given
brain regions and the sequential development of
behavioral symptoms in degenerative diseases, and
(b) delineate the specific role of certain mesial
temporal lobe structures, such as the entorhinal
cortex and hippocampus in human memory
function, because of the age or disease-related
degeneration in those structures.
Corresponding author. Tel.: +1 (312) 942 5399;
Fax: +1 (312) 563 3570; E-mail: ldetoled@rush.edu
cated as the sub-region most sensitive to the effects
of advancing age (Small et al., 2004). Although
synaptic loss in the dentate molecular layer has
been demonstrated in patients with Alzheimer’s
disease (AD) (Scheff and Price, 2003; Scheff et al.,
2006), cell loss associated with the disease seems
to be more prominent in the CA1 region of the
hippocampus (West et al., 1994, 2000). In addi-
tion, AD-related pathological changes such as
neurofibrillary tangles are most notable in the
subiculum and the CA1 region (Van Hoesen and
Hyman, 1990). Unfortunately, until recently, in
vivo structural imaging techniques did not have
the resolution to differentiate sub-regions of the
hippocampal formation (hippocampus proper and
dentate). However, a recently published structural
MRI protocol acquired with a high-field (4 Tesla)

DOI: 10.1016/S0079-6123(07)63040-4 741

742
magnet holds promise for differentiating hippo-
campal sub-fields in vivo in future investigations
(Mueller et al., 2007).
In this chapter, we will review the evidence
indicating that there is atrophy of both the
entorhinal cortex and hippocampus in the very
early stages of AD, as well as in individuals with
amnestic mild cognitive impairment (MCI) who
are at high risk of developing AD (Petersen et al.,
1999; Petersen, 2000). These elderly individuals
have impairment in memory function, but do not
meet criteria for dementia. Furthermore, we will
demonstrate that white matter changes in the
region of the parahippocampal gyrus that includes
the perforant path in people with amnestic MCI
and mild AD could exacerbate the memory dys-
function characteristic of the initial stages of AD,
by partially disconnecting the hippocampus from
incoming sensory information.
Since the MRI techniques in the experiments
from our laboratory to be described were similar,
they will be detailed first.
Acquisition and quantitation of MRI data
All MR images were acquired on a 1.5 Tesla Gen-
eral Electric Signa scanner with the manufacturer’s
three-dimensional (3D) Fourier transform spoiled
gradient recalled (SPGR) pulse sequence. The ac-
quisition parameters for the T1 weighted sequence
were as follows: 124 contiguous images acquired in
the coronal plane, 1.6 mm-thick sections, matrix ¼
256
192, field of view ¼ 22 cm, TR/TE ¼ 34/7,
flip angle ¼ 351, signals averaged ¼ 1.
The analyze software package (Mayo Clinic
Foundation, Rochester, MN, USA) was used for
determining the volume of regions of interest and
for co-registering sequential scans. Both the
entorhinal cortex and hippocampal formation
were manually segmented as described below.
To correct for individual differences in brain size,
entorhinal and hippocampal volumes were divided
by total intracranial volume (an accepted measure
of pre-morbid brain size) derived from sagittaly
reformatted 5 mm slices. To compute intracranial
volume, the inner table of the cranium was traced
in consecutive sagittal sections spanning the whole
brain. At the level of the foramen magnum, a
straight line was drawn from the inner surface of
the clivus to the occipital bone.
Entorhinal cortex volume was quantified with
the use of a protocol developed and validated
in our laboratory, technical details of which
are presented in Goncharova et al. (2001). The
advantage of this protocol is that entorhinal vol-
ume is measured from the same oblique coronal
sections most commonly used for hippocampal
volumetry, so that one of these two adjacent
structures is not over estimated at the expense of
the other.
Briefly, both entorhinal and hippocampal
volumes were computed separately for the right
and left hemispheres from coronal slices reformat-
ted to be perpendicular to the long axis of the
hippocampus. For the entorhinal cortex, tracing
began with the first section in which the gyrus
ambiens, amygdala and the white matter of the
parahippocampal gyrus first appeared visible.
The dorsomedial border in rostral sections was
the sulcus semiannularis and in caudal sections the
subiculum. The shoulder of the collateral sulcus
was used as the lateral border, a somewhat
conservative criterion that allowed consistency in
tracings and avoided the use of different lateral
borders, depending on individual differences in the
depth of the collateral sulcus (Insausti et al., 1998).
The last section traced was three 1.6 mm slices
rostral to the image in which the lateral geniculate
first appeared visible.
The protocol and validation procedures used for
quantifying hippocampal volume were published
previously (Wilson et al., 1996; deToledo-Morrell
et al., 1997). For hippocampal volumetry, tracings
started with the first section where the hippocam-
pus could be clearly differentiated from the
amygdala by the alveus and included the fimbria,
the dentate gyrus, the hippocampus proper and the
subiculum. Tracings continued on all consecutive
slices until the slice before the full appearance of
the fornix. Investigators involved in MRI analyses
were blinded to clinical information until all such
analyses were completed.
Figure 1 shows sample tracings of the entorhinal
cortex and hippocampal formation. Both regions
of interest were normalized using the following

Fig. 1. A sample coronal image illustrating the segmentation of the entorhinal cortex (outlined, right side) and hippocampus (outlined,
left side).
formula: (absolute volume in mm3/intracranial
volume in mm3)
1000.
Hippocampal and entorhinal cortex atrophy in
incipient and mild AD
The entorhinal cortex and hippocampus are part
of the mesial temporal lobe memory system
(Squire and Zola-Morgan, 1991; Young et al.,
1997). The entorhinal cortex receives sensory input
from the neocortex and provides the hippocampus
with multi-modal sensory information via the per-
forant path. These brain regions have received
special attention in both post mortem and in vivo
investigations on the pathophysiology of AD,
since memory dysfunction is one of the earliest
hallmarks of the disease.
Post mortem pathological studies have impli-
cated the entorhinal cortex and the transentorhinal
region as sites that are involved in the early stages
of AD and in individuals with MCI (Braak and
Braak, 1991, 1995; Gomez-Isla et al., 1996; Braak
et al., 1998; Kordower et al., 2001). In vivo MRI
investigations have also demonstrated atrophy of
both the entorhinal cortex and hippocampus not
only in patients diagnosed with mild AD, but also
in individuals with amnestic MCI and those with
subjective cognitive complaints (Convit et al.,
1997; Jack et al., 1997, 1999; de Leon et al.,
1996; deToledo-Morrell et al., 1997, 2000a, b;
Killiany et al., 2000, 2002; Xu et al., 2000;
Dickerson et al., 2001; Du et al., 2001; Jessen
et al., 2006; Saykin et al., 2006).
In an earlier study from our laboratory (Di-
ckerson et al., 2001) we compared MRI-derived
743
entorhinal and hippocampal volumes in healthy
elderly controls with no cognitive impairment, in
people who presented at the clinic with cognitive
complaints, but who did not meet criteria for
dementia and in patients with very mild AD.
Interestingly, the two patient groups differed sig-
nificantly from controls in entorhinal volume, but
not from each other; in contrast, they differed
from each other as well as from controls in hip-
pocampal volume, with the very mild AD cases
showing the greatest atrophy.
Recently, interest has increased in directly
comparing the accuracy with which the volumes
of these two mesial temporal lobe structures
can differentiate those who convert from a non-
demented status to AD (i.e., those who are at risk
of developing AD). The study described below was
undertaken to examine this question using an
initial baseline MRI in a group of participants
diagnosed with amnestic MCI (deToledo-Morrell
et al., 2004).
The participants consisted of 27 elderly individ-
uals (Z65 years of age) who received a clinical
diagnosis of amnestic MCI during their baseline
evaluation. They had a significant deficit in the
memory domain, but did not meet criteria for de-
mentia. After the baseline diagnosis of MCI, all
participants were followed with yearly clinical
evaluations. During a 36-month follow-up period,
10 of the 27 participants originally diagnosed with
amnestic MCI converted to AD.
All clinical evaluations were carried out at
the Rush Alzheimer’s Disease Center (RADC,
Chicago, IL) as previously described (Wilson et al.,
1996; deToledo-Morrell et al., 1997). Briefly, the
744
evaluation incorporated the Consortium to Estab-
lish a Registry for Alzheimer’s Disease (CERAD,
Morris et al., 1989) procedures and included a
medical history, neurological examination, neuro-
psychological testing, informant interview and
blood tests.
Exclusion criteria for entry into the study con-
sisted of evidence of other neurologic, psychiatric
or systemic conditions that could cause cognitive
impairment (e.g., stroke, alcoholism, major de-
pression, a history of temporal lobe seizures). The
clinical diagnosis of probable AD in the longitu-
dinal clinical evaluations followed NINCDS/AD-
RDA guidelines (McKhann et al., 1984); it
required a history of cognitive decline and neuro-
psychological test evidence of impairment in at
least two cognitive domains, one of which had to
be memory.
Demographic data and Mini Mental State
Examination (MMSE, Folstein et al., 1975) scores
for the MCI-converters and non-converters are
presented in Table 1. The maximum score on the
MMSE is 30, with scores Z27 being considered
within the normal range. The individuals who
developed AD (converters) did not differ from the
non-converters in age, but there was a significant
difference between them in MMSE scores and in
years of education.
Total (right+left) normalized entorhinal and
hippocampal volumes derived from the baseline
MRI for MCI converters and non-converters
are shown in Fig. 2. Converters differed signifi-
cantly from non-converters in total entorhinal
[t(25) ¼ 3.090, p ¼ 0.005] and hippocampal
[t(25) ¼ 2.442, p ¼ 0.022] volume. Multivariate lo-
gistic regression analyses adjusted for education
were carried out to examine how well baseline
entorhinal and hippocampal volumes could
Table 1. Demographic characteristics of participants
MCI converters (N ¼ 10)
Age (years)
Education (years)
Female/male
MMSE score
 Significantly differently from each other (po0.01).
predict conversion to AD. In these analyses, total
normalized entorhinal and hippocampal volumes
were used as predictors to determine the extent to
which each region of interest contributed sepa-
rately to describing group differences. The results
demonstrated that although both total entorhinal
and total hippocampal volumes were independent
predictors of conversion [w2(1) ¼ 7.3, p ¼ 0.007,
odds ratio per 0.1 unit volume change ¼ 2.024 for
the entorhinal cortex and w2(1) ¼ 6.2, p ¼ 0.013,
odds ratio ¼ 1.318 for the hippocampus], ent-
orhinal volume was a better predictor with a con-
cordance rate of 90.6%. Thus, for every 0.1 unit
decrease in entorhinal cortex volume, the odds of
conversion doubled, while the same unit decrease
in hippocampal volume increased the chances of
conversion by only 30%.
Similar analyses were carried out on entorhinal
and hippocampal volumes separated by hemi-
sphere. In this case, the right hemisphere ent-
orhinal volume best predicted conversion with a
concordance rate of 93.5%. Unlike these findings,
hemispheric differences in hippocampal volume
did not lead to differential predictions regarding
conversion.
The in vivo anatomical results presented here
are in agreement with post mortem pathological
findings and underscore the early involvement of
the entorhinal cortex in AD. Taken together,
the results just described indicate that entorhinal
volume is a good predictor of conversion from
MCI to AD. In agreement with these results,
others have also reported that baseline entorhinal
volume differentiated those at risk for developing
AD better than hippocampal volume (Killiany
et al., 2002).
Our results also demonstrated that the right
hemisphere entorhinal volume was the best
MCI non-converters (N ¼ 17)
82.7 (74.5) 81.1 (78.1)
(75.6–89.2) (66.1–98.0)
18.4 (72.1) 15.2 (73.1)
6/4 9/8
26.1 (71.4) 28.0 (71.8)
 745

Fig. 2. Box plot comparing MRI-derived normalized entorhinal (left hand side) and hippocampal (right hand side) volumes in
participants with MCI who converted to a diagnosis of Alzheimer’s disease, in contrast to those who did not. The central box shows
the data between the upper and lower quartiles, with the median represented by the line. The height of the line is the interquartile range
(IQR); the ‘‘whiskers’’ extend from the upper and lower quartiles to a distance of 1.5 IQR away or to the most extreme data point
within that range, whichever is closer (adapted with permission from deToledo-Morrell et al., 2004).

predictor of risk of AD. Very similar findings have
been recently reported by another laboratory
(Tapiola et al., 2006). The reason for this hemi-
spheric difference in entorhinal volume in those
who are at risk of developing AD is not clear since,
in general, post mortem pathological investigations
do not carry out hemispheric comparisons.
It should be pointed out that changes in the
entorhinal cortex in elderly individuals seem to be
predictive of conversion not only from MCI to
AD, but also of the transition from normal cog-
nition to MCI or AD. In an interesting report, de
Leon and his colleagues (de Leon et al., 2001) used
MRI guided resting state [2-18F]-2-deoxy-D-glu-
cose/positron-emission tomography to assess met-
abolic changes in given brain regions of interest.
They found that a reduction in entorhinal cortex
metabolism was predictive of cognitive decline
among cognitively normal elderly subjects. Al-
though obtained with a different imaging modal-
ity, these results are very much in line with our
findings and provide further evidence for the very
early involvement of the entorhinal cortex in the
development of AD. Similarly, Rodrigue and Raz
(2004) found that the rate of entorhinal atrophy in
healthy elderly people was associated with greater
memory decline.
MRI Predictors of risk of incident AD: a
longitudinal study
As discussed above, a large number of cross-sec-
tional MRI investigations have demonstrated at-
rophy of both the hippocampus and entorhinal
cortex in patients with mild to moderate AD and
in people with cognitive impairment who are at
risk of developing AD. Longitudinal studies that
measured the volumes of these two mesial tempo-
ral lobe structures have shown the annual rate of
atrophy of both structures to be greater in patients
with AD and in those with MCI compared to eld-
erly control subjects (Jack et al., 1998, 2000, 2004;
Cardenas et al., 2003; Du et al., 2003, 2004).
However, most of these longitudinal investigations
used scans from two-time points separated by 1–5
746

years for determining rate of atrophy, rather than significant predictor of event time, all models were
multiple yearly scans. education adjusted.
The experiment from our laboratory described The results of the proportional odds models
above examined MRI markers of AD using a demonstrated that initial diagnosis of MCI was an
baseline scan and longitudinal clinical evaluations. important predictor of incident AD [w2(1) ¼
In the next study, we used proportional odds 16.569, po0.0001; Fig. 3A]. Furthermore, both
models to assess the relationship between rate of baseline entorhinal volume and its rate of atrophy
atrophy of mesial temporal lobe structures were independent predictors of incident AD; their
and risk of incident AD among non-demented addition to initial diagnosis improved the model
elderly individuals (Stoub et al., 2005). Rate of [w2(2) ¼ 10.904, po0.0043]. Lower baseline ent-
atrophy of regions of interest was derived from orhinal volume was associated with greater risk of
yearly scans. AD (see Fig. 3B); for every 0.25 unit of decrease in
Data reported here were obtained from 58 non- baseline entorhinal volume, the odds of incident
demented elderly participants (age 65 or older) AD increased by a factor of 4.98. Similarly, in-
who were followed with annual clinical evaluations creased rate of decline in entorhinal volume was
as well as high-resolution MRI scans for up to 5 associated with greater risk of AD (see Fig. 3C);
years (baseline and 5 years of follow-up). The for every 0.02 unit of increase in rate of entorhinal
clinical evaluations and MRI acquisition and anal- atrophy, the likelihood of developing AD in-
ysis protocols were as described above. Twenty- creased by a factor of 2.18.
three of the 58 participants received a diagnosis of Surprisingly, hippocampal volume was not an
amnestic MCI during the baseline evaluation independent predictor of risk of AD; the addition
and 35 were healthy controls with no cognitive of baseline hippocampal volume and its slope of
impairment. decline to initial diagnosis did not improve the
The number of years until a diagnosis of AD model based on initial diagnosis alone
was modeled using proportional odds models [w2(2) ¼ 2.701, p ¼ 0.26].
(Kalbfleish and Prentice, 1980). For this purpose, APOE e4 allele status, a risk factor for AD, was
AD was treated as an all absorbing state, with available for 53 of 58 participants (see Table 2).
onset at the first evaluation at which NINCDS/ When the analyses described above were carried
ADRDA criteria (McKhann et al., 1984) were out on data from these 53 participants, results
met. For participants who did not develop AD, the
number of years to last clinical diagnosis was the Table 2. Baseline demographic characteristics of participants
event time, which was coded as censored. Covari-
Non-converters Converters to AD
ates considered were demographic variables (age,
sex, years of education), baseline diagnosis (MCI N 44 14
or control), and volumetric measures (normalized Sex
total hippocampal and entorhinal volumes). Male 16 4
Female 28 10
Volumes of regions of interest derived from scans Age (years) 80 (76) 81 (76)
following a diagnosis of AD were not included in MMSE 28.5 (71.5) 26.9 (71.8)
the analyses (for further analytic details, see Stoub Education (years) 15 (73) 18 (73)
et al., 2005). APOE status
Fourteen of the 58 non-demented participants 4/4 0 0
3/4 8 5
developed AD during the follow-up period; three 3/3 25 6
of these 14 participants entered the study as con- 2/4 0 2
trol subjects with no cognitive impairment and the 2/3 7 0
rest with a diagnosis of amnestic MCI. Demo- 2/2 0 0
graphic characteristics for participants who did Not available 4 1
and did not develop AD are presented in Table 2. Significantly different p ¼ 0.002.
 Significantly different p ¼ 0.019.
Since education, but not age, was found to be a

remained similar to those already detailed. How-
ever, APOE e4 allele status (i.e., the presence or
absence of any e4 allele) was not found to be an
important predictor of incident AD when added to
747
models based on initial diagnosis and entorhinal
and hippocampal volume [w2(2) ¼ 1.972, p ¼ 0.16],
although the presence or absence of any APOE e4
allele by itself was a predictor of event time
[w2(1) ¼ 5.00, p ¼ 0.025]. These results are in line
with those from a prior report (Bennett et al.,
2003) which demonstrated that the effect of the e4
allele, which was strongly associated with the like-
lihood of clinical AD, was no longer significant
after controlling for the effects of AD pathology,
suggesting that such pathology mediates the effect
of allele status on clinical disease. In our experi-
ments, MRI-derived hippocampal and entorhinal
atrophy could be viewed as surrogate markers of
the underlying pathology.
The major finding of this study is that among
non-demented individuals, the initial size of the
entorhinal cortex and its rate of atrophy are
significant risk factors for AD, with initial small
size and steep declines in volume being associated
with incident AD. In contrast, baseline size and
rate of hippocampal atrophy were not found to be
independent risk factors. Other investigators have
reported that hippocampal atrophy at baseline in
patients with MCI is associated with risk of AD
(Jack et al., 1999). Although our results may,
at first sight, seem contrary to this report, it is
important to point out that in this paper the in-
vestigators assessed only MRI-based hippocampal
atrophy and did not directly compare hippocam-
pal and entorhinal volumes in their models.
Fig. 3. Estimates of the probability of not yet having been di-
agnosed with Alzheimer’s disease as a function of years of fol-
low-up for two different populations. The dotted line represents
the reference population and the solid line depicts the hypo-
thetical population. The reference population in all panels con-
sists of people who have no cognitive impairment (NCI) at
baseline, have an average education level of 16 years, whose
normalized entorhinal cortex volume is 1.00 (chosen to be close
to the overall mean of 0.97027 at baseline), and for whom the
slope of the rate of entorhinal atrophy is 0.03 (which is close
to the mean slope of 0.03074). For the model shown in A, the
hypothetical population is different from the reference only in
that the initial diagnosis is MCI, not NCI. In B, the hypothet-
ical population is different from the reference in terms of base-
line entorhinal cortex volume, which was 25% smaller than 1 or
0.75. The hypothetical population shown in C is different from
the reference population only in that entorhinal cortex volume
declines faster with a slope of 0.05 rather than 0.03 (adapted
with permission from Stoub et al., 2005).
748
As discussed earlier in this chapter, post mortem
studies have suggested that AD-related pathology
may start in the transentorhinal and entorhinal
regions and then spread to the hippocampus
(Braak and Braak, 1991, 1995; Braak et al.,
1998). The fact that shrinkage of the entorhinal
cortex, but not of the hippocampus was a predic-
tor of AD in our study is in support of these post
mortem findings and suggests that the extent of
entorhinal involvement in the disease process may
precede hippocampal involvement.
Hippocampal disconnection due to parahippocampal
white matter changes in incipient AD: relation to
memory dysfunction
As discussed up to this point, previous histopath-
ological findings (Hyman et al., 1984; Braak and
Braak, 1991, 1995; Gomez-Isla et al., 1996; Braak
et al., 1998; Mufson et al., 1999; Kordower et al.,
2001), as well as in vivo MRI results indicate that
the entorhinal cortex and hippocampus are path-
ologically involved very early in patients with AD
and in those with MCI who are at high risk for
developing AD. More specifically, two studies
(Gomez-Isla et al., 1996; Kordower et al., 2001)
found a loss of entorhinal cortex layer II neurons
in patients with AD and in those with MCI, com-
pared to elderly controls. These neurons receive
multimodal sensory input from primary sensory
and association cortices and project this informa-
tion to the hippocampus via the perforant path,
a white matter tract located in the anterior medial
portion of the parahippocampal gyrus (Van
Hoesen and Pandya, 1975; Van Hoesen et al.,
1975; Amaral et al., 1987).
The loss of layer II neurons in the entorhinal
cortex and their axons is important since it would
result in alterations in information flow to the
hippocampus. In addition, damage to the white
matter of the parahippocampal gyrus could dis-
rupt afferent connections to the entorhinal cortex
and ultimately disconnect multi-modal sensory in-
put to the hippocampus, information vital to the
formation of new memories. These changes in
white matter may be detectable in vivo with im-
aging techniques, but have received less attention
in investigations on the pathophysiology of incip-
ient or clinically diagnosed AD and the anatomical
underpinnings of the memory dysfunction associ-
ated with the disease.
Up to this point we have concentrated on a
description of structural alterations detected in
vivo in the entorhinal cortex and hippocampus. In
the next experiment (Stoub et al., 2006) we turned
our attention to assessing the contribution of al-
terations in the parahippocampal white matter, in
addition to entorhinal and hippocampal atrophy,
to the memory dysfunction in people with amnes-
tic MCI who are at risk of developing AD.
The participants in the study consisted of 40
older individuals (mean age 77.977.5 years) who
met criteria for amnestic MCI and 50 healthy con-
trols (mean age 78.176.0 years) with no cognitive
impairment. Clinical evaluations and acquision of
MRI scans were as described before. White matter
volume changes throughout the brains of individ-
uals with amnestic MCI, compared to age-
matched healthy controls were assessed with vox-
el-based morphometry (VBM, Good et al., 2001).
Automated techniques such as VBM that provide
quick, unbiased means of comparing changes in
white matter at the voxel level are especially useful
when there is no a priori knowledge of where in the
brain such white matter changes may be taking
place. In addition to white matter changes, we as-
sessed differences between the groups in entorhinal
cortex and hippocampal volume using manual
segmentation and volumetric measurement of the
two structures since, in our hands, these measures
are better at detecting subtle changes in small gray
matter regions.
Group differences in whole-brain white matter
volumes were assessed with the two-sample t sta-
tistic within the Statistical Parametric Mapping
software (SPM 99); the significance threshold was
set at 0.001. Regions of statistically significant
difference were identified according to the Mon-
treal Neurological Institute template and then
converted to Talariach coordinates (Talairach
and Turnoux, 1988) with the use of MNITOTAL
software. The Talairach coordinates were trans-
formed to actual brain areas by using the Talai-
rach Daemon system (Lancaster et al., 2000).
These coordinates were used to construct regions

of interest that were then applied to individual
white matter density maps to extract individual
volume values.
In order to relate structural changes to memory
function, individual memory Z scores were calcu-
lated based on combined performance on declar-
ative memory tasks using seven declarative
memory scores. These consisted of immediate
and delayed recall of the East Boston Story (Al-
bert et al., 1991) and of Story A from the Logical
Memory of the Wechsler memory scale — Revised
(Wechsler, 1987). An additional test involved the
learning and retention of a 10-word list from the
CERAD battery (Morris et al., 1989). The three
scores for this test included Word List Memory
(the total number of words immediately recalled
after each of three consecutive presentations of the
list), Word List Recall (the number of words re-
called after a delay) and Word List Recognition
(the number of words correctly recognized in a
four-alternative, forced-choice format, adminis-
tered after Word List Recall).
The demographic characteristics of all partici-
pants are listed in Table 3. The amnestic MCI
group did not differ from healthy controls in age
or education, but had significantly lower [t(88) ¼
6.479, po0.001] MMSE scores and memory Z
scores [t(88) ¼ 9.921, po0.001], which is not
surprising.
As shown in Fig. 4, there was a significant de-
crease (po0.001) in white matter volume in par-
ticipants with MCI compared to controls in the
anterior-medial aspect of both parahippocampal
gyri, in the region that includes the perforant path.
Table 3. Demographic characteristics and memory Z scores of
participants
Healthy MCI
controls
Number of participants 50 40
Age (years) 78.1 (76.0) 77.9 (77.5)
Education (years) 15.0 (72.8) 16.2 (73.1)
Male/female 15/35 16/24
MMSE score 29.0 (70.9) 27.2 (71.6)
Memory Z score 0.5337 0.5948
(70.4681) (70.6111)
 po0.001.
749
In addition to the parahippocampal white matter
change, the group of MCI participants also
showed significant entorhinal and hippocampal
atrophy (po0.001 for both) similar to what our
laboratory and others have reported. The atrophy
in both structures was bilateral.
For predicting memory function from MRI
measures, each of the three MRI indices (i.e., total
parahippocampal white matter volume, entorhinal
cortex volume and hippocampal volume) was en-
tered singly into a regression analysis that used as
the dependent variable individual memory Z
scores. Each of the three anatomical measures
was found to be a significant predictor of memory
function [F(1,88) ¼ 24.08, po0.001 for hippocam-
pal volume; F(1,88) ¼ 10.35, po0.002 for
entorhinal cortex volume; and F(1,88) ¼ 11.11,
po0.001 for parahippocampal white matter
volume]. However, when all three MRI measures
were entered simultaneously into a multiple
regression model, only hippocampal and white
matter volume measures were found to be signifi-
cant predictors of memory function [t(86) ¼ 3.01,
p ¼ 0.003 and t(86) ¼ 2.22, po0.029 respectively].
The major contribution of these data is that,
in addition to a reduction in the size of the hip-
pocampus, a decrease in white matter volume in
the region of the parahippocampal gyrus that
includes the perforant path significantly contrib-
utes to the memory dysfunction in people with
MCI. The perforant path that supplies the hip-
pocampus with multi-modal information is path-
ologically involved very early in AD (Hyman et
al., 1984). The investigation by Hyman and his
collaborators was carried out on post mortem
tissue and demonstrated that in patients with
very mild AD, there is disconnection of the hip-
pocampus from sensory cortical inputs as a re-
sult of loss of layer II cells in the entorhinal
cortex. The authors hypothesized that such
disconnection could contribute to memory dys-
function during the very early stages of AD;
however, the report did not include cognitive
data proximal to death. Our results provide an
in vivo demonstration of hippocampal discon-
nection resulting from white matter changes in
the region of the perforant path that impacts
memory function.
750

Fig. 4. Color map showing significant (p ¼ 0.001) voxels of decreased white matter density in participants with amnestic MCI
compared with controls, superimposed on coronal and sagittal slices of a template based on data for all subjects in the study. The
image is masked to include white matter regions. The colors correspond to the t values shown on the color bar. Note the bilaterality of
significant differences in the white matter of the parahippocampal gyrus (adapted with permission from Stoub et al., 2006). (See Color
Plate 40.4 in color plate section.)

White matter volume change may reflect not
only loss of afferent and efferent fibers in the re-
gion of the parahippocampal gyrus, but also par-
tial demyelination in remaining fibers. Recently
developed diffusion tensor imaging (DTI) that de-
tects microstructural alterations in normal appear-
ing white matter could aid in determining the
microstructural changes in remaining fibers that
may further degrade impulse transmission. DTI
detects microstructural alterations in white matter
by measuring the directionality of molecular diffu-
sion (fractional anisotropy, FA). Highly organized
and well myelinated white matter tracts have high
FA because diffusion is constrained by the tract’s
cellular organization. As white matter is damaged,
FA decreases due to decreased anisotropic diffu-
sion. In fact, in the most recent study from our
laboratory (Wang et al., 2006), using a high res-
olution DTI protocol, we demonstrated that in
very mild AD cases there is not only loss of white
matter volume in the region of the parahippocam-
pal gyrus, but that remaining fibers are not
normal, as indicated by lower FA values com-
pared to age appropriate controls.
Conclusions
In this chapter, we reviewed evidence from in vivo
MRI investigations that demonstrate significant
entorhinal cortex and hippocampal atrophy not
only in patients with very mild AD, but also in
individuals who are at risk of developing AD.
Atrophy in these mesial temporal lobe structures
important for memory function can predict who
will develop AD a number of years prior to a
clinical diagnosis. Recently, there has been in-
creased interest in investigating alterations in white
matter that could lead to degradation in informa-
tion flow due to partial disconnection of different
brain regions. We have shown that white matter
volume loss in the region of the parahippocampal
gyrus contributes significantly to memory dysfunc-
tion in individuals with incipient AD. The exact

underlying mechanism of the white matter volume
loss in the parahippocampal region cannot be de-
termined in vivo with the tools currently available
to us. One may speculate that a decrease in white
matter fibers in this region, reflected in white mat-
ter volume change, may cause a disruption of in-
put to the dentate gyrus and hippocampus proper
from the entorhinal cortex. Because these incom-
ing fibers arise from cells in the entorhinal cortex,
entorhinal atrophy may, in part, be the origin of
the decrease in white matter volume. Future in
vivo imaging studies using high field scanners are
necessary to fully elaborate the changes post-
synaptic to the entorhinal input; such changes in
response to entorhinal atrophy are likely to in-
clude the dentate gyrus.
Acknowledgments
This research was supported by grants P01
AG09466, P30 AG10161 and R01 AG17917 from
the National Institute on Aging, National Insti-
tutes of Health.
References
Albert, M., Smith, L., Scherr, P., Taylor, J., Evans, D. and
Funkenstein, H. (1991) Use of brief cognitive tests to identify
individuals in the community with clinically diagnosed
Alzheimer’s disease. Int. J. Neurosci., 57(3–4): 67–178.
Amaral, D.G., Insausti, R. and Cowan, W.M. (1987) The ent-
orhinal cortex of the monkey. I. Cytoarchitectonic organiza-
tion. J. Comp. Neurol., 264(3): 326–355.
Bennett, D.A., Wilson, R.S., Schneider, J.A., Evans, D.A.,
Aggarwal, N.T., Arnold, S.E., Cochran, E.J., Berry-Kravis,
E. and Bienias, J.L. (2003) Apolipoprotein E epsilon4 allele,
AD pathology, and the clinical expression of Alzheimer’s
disease. Neurology, 60(2): 246–252.
Braak, H. and Braak, E. (1991) Neuropathological staging of
Alzheimer-related changes. Acta Neuropathol., 82(4):
239–259.
Braak, H. and Braak, E. (1995) Staging of Alzheimer’s disease-
related neurofibrillary changes. Neurobiol. Aging, 16(3):
271–288.
Braak, H., Braak, E., Bohl, J. and Bratzke, H. (1998) Evolution
of Alzheimer’s disease related cortical lesions. J. Neural
Trans. Suppl., 54: 97–106.
Cardenas, V.A., Du, A.T., Hardin, D., Ezekiel, F., Weber, P.,
Jagust, W.J., Chui, H.C., Schuff, N. and Weiner, M.W.
(2003) Comparison of methods for measuring longitudinal
751
brain change in cognitive impairment and dementia. Ne-
urobiol. Aging, 24(4): 537–544.
Convit, A., de Leon, M.J., Tarshish, C., DeSanti, S., Tsui, W.,
Rusinek, H. and George, A. (1997) Specific hippocampal
volume reductions in individuals at risk for Alzheimer’s
disease. Neurobiol. Aging, 18(2): 131–138.
Dickerson, B.C., Goncharova, I., Sullivan, M.P., Forchetti, C.,
Wilson, R.S., Bennett, D.A., Beckett, L.A. and deToledo-
Morrell, L. (2001) MRI-derived entorhinal and hippocampal
atrophy in incipient and very early Alzheimer’s disease.
Neurobiol. Aging, 22(5): 747–754.
Du, A.T., Schuff, N., Amend, D., Laakso, M.P., Hsu, Y.Y.,
Jagust, W.J., Yaffe, K., Kramer, J.H., Reed, B., Norman, D.,
Chui, H.C. and Weiner, M.W. (2001) MRI of entorhinal
cortex and hippocampus in mild cognitive impairment and
Alzheimer’s disease. J. Neurol. Neurosurg. Psychiatry, 71(4):
441–447.
Du, A.T., Schuff, N., Kramer, J.H., Ganzer, S., Zhu, X.P.,
Jagust, W.J., Miller, B.L., Reed, B.R., Mungas, D., Yaffe,
K., Chui, H.C. and Weiner, M.W. (2004) Higher atrophy rate
of entorhinal cortex than hippocampus in AD. Neurology,
62(3): 422–427.
Du, A.T., Schuff, N., Zhu, X.P., Jagust, W.J., Miller, B.L.,
Reed, B.R., Kramer, J.H., Mungas, D., Yaffe, K.,
Chui, H.C. and Weiner, M.W. (2003) Atrophy rates of ent-
orhinal cortex in AD and normal aging. Neurology, 60(3):
481–486.
Folstein, M.F., Folstein, S.E. and McHugh, P.R. (1975) ‘‘Mini-
mental state.’’ A practical method for grading the mental
status of patients for the clinician. J. Psychiatr. Res., 12(3):
189–192.
Gomez-Isla, T., Price, J.L., McKeel, D.W., Morris, J.C., Grow-
don, J.H. and Hyman, B.T. (1996) Profound loss of layer. II.
Entorhinal cortex neurons occurs in very mild Alzheimer’s
disease. J. Neurosci., 16(14): 4491–4500.
Good, C.D., Johnstrude, I.S., Ashburner, J., Henson, R.N.,
Friston, K.J. and Frakoviak, R.S. (2001) A voxel-based
morphometric study of ageing in 465 normal adult human
brains. Neuroimage, 14(1): 21–36.
Goncharova, I.I., Dickerson, B.C., Stoub, T.R. and deToledo-
Morrell, L. (2001) MRI of entorhinal cortex: a reliable pro-
tocol for volumetric measurement. Neurobiol. Aging, 22(5):
737–745.
Hyman, B.T., Van Hoesen, G.W., Damasio, A.R. and Barnes,
C.L. (1984) Alzheimer’s disease: cell specific pathology iso-
lates the hippocampal formation. Science, 225(4667):
1168–1170.
Insausti, R., Juottonen, K., Soininen, H., Insausti, A.M., Part-
anen, K., Vainio, P., Laakso, M.P. and Pitkanen, A. (1998)
MR volumetric analysis of the human entorhinal, perirhinal
and temporopolar cortices. Am. J. Neuroradiol., 19(4):
659–671.
Jack Jr., C.R., Petersen, R.C., Xu, Y., O’Brien, P.C., Smith,
G.E., Ivnik, R.J., Boeve, B.F., Tangalos, E.G. and Kokmen,
E. (2000) Rates of hippocampal atrophy correlate with
change in clinical status in aging and AD. Neurology, 55(4):
484–489.
752
Jack Jr., C.R., Petersen, R.C., Xu, Y., O’Brien, P.C., Smith,
G.E., Ivnik, R.J., Tangalos, E.G. and Kokmen, E. (1998)
Rate of medial temporal lobe atrophy in typical aging and
Alzheimer’s disease. Neurology, 51(4): 993–999.
Jack Jr., C.R., Petersen, R.C., Xu, Y.C., O’Brien, P.C., Smith,
G.E., Ivnik, R.J., Boeve, B.F., Waring, S.C., Tangalos, E.G.
and Kokmen, E. (1999) Prediction of AD with MRI-based
hippocampal volume in mild cognitive impairment. Neurol-
ogy, 52(7): 1397–1403.
Jack Jr., C.R., Petersen, R.C., Xu, Y.C., Waring, S.C., O’Brien,
P.C., Tangalos, E.G., Smith, G.E., Ivnik, R.J. and Kokmen,
E. (1997) Medial temporal atrophy on MRI in normal aging
and very mild Alzheimer’s disease. Neurology, 49(3): 786–794.
Jack Jr., C.R., Schlung, M.M., Gunter, J.L., O’Brien, P.C.,
Weigand, S.D., Knopman, D.S., Boeve, B.F., Ivnik, R.J.,
Smith, G.E., Cha, R.H., Tangalos, E.G. and Petersen, R.C.
(2004) Comparison of different MRI brain atrophy measures
with clinical disease progression in AD. Neurology, 62(4):
591–600.
Jessen, F., Feyen, L., Freymann, K., Tepest, R., Maier, W.,
Heun, R., Schild, H.H. and Scheef, L. (2006) Volume reduc-
tion of the entorhinal cortex in subjective memory impair-
ment. Neurobiol. Aging, 27(12): 1751–1756.
Kalbfleish, J.D. and Prentice, R.L. (1980) The Statistical Anal-
ysis of Failure Time Data. Wiley, New York.
Killiany, R.J., Gomez-Isla, T., Moss, M., Kikinis, R., Sandor,
T., Jolesz, F., Tanzi, R., Jones, K., Hyman, B.T. and Albert,
M.S. (2000) Use of structural magnetic resonance imaging to
predict who will get Alzheimer’s disease. Ann. Neurol., 47(4):
430–439.
Killiany, R.J., Hyman, B.T., Gomez-Isla, T., Moss, M.B.,
Kikinis, R., Jolesz, F., Tanzi, R., Jones, K. and Albert, M.S.
(2002) MRI measures of entorhinal cortex vs. hippocampus
in preclinical AD. Neurology, 58(8): 1188–1196.
Kordower, J.H., Chu, Y., Stebbins, G.T., DeKosky, S.T., Co-
chran, E.J., Bennett, D. and Mufson, E.J. (2001) Loss and
atrophy of layer. II. Entorhinal cortex neurons in elderly
people with mild cognitive impairment. Ann. Neurol., 49(2):
202–213.
Lancaster, J.L., Woldorff, M.G., Parsons, L.M., Liotti, M.,
Freitas, C.S., Rainey, L., Kochunov, P.V., Nickerson, D.,
Mikiten, S.A. and Fox, P.T. (2000) Automated Talairach
atlas labels for functional brain mapping. Hum. Brain
Mapp., 10(3): 120–131.
de Leon, M.J., Convit, A., George, A.E., Golomb, J., DeSanti,
S., Tarshish, C., Rusinek, H., Bobinski, M., Ince, C., Miller,
D. and Wisniewski, H. (1996) In vivo structural studies of the
hippocampus in normal aging and in incipient Alzheimer’s
disease. Ann. N.Y. Acad. Sci., 777: 1–13.
de Leon, M.J., Convit, A., Wolf, O.T., Tarshish, C.Y., DeSanti,
S., Rusinek, H., Tsui, W., Kandil, E., Sherer, A.J., Roche,
A., et al. (2001) Prediction of cognitive decline in normal
elderly subjects with 2-[(18)F]fluoro-2-deoxy-D-glucose/posi-
tron emission tomography (FDG/PET). Proc. Natl. Acad.
Sci. U.S.A., 98(19): 10966–10971.
McKhann, G., Drachman, D., Folstein, M., Katzman, R.,
Price, D. and Stadlan, E.M. (1984) Clinical diagnosis of
Alzheimer’s disease: report of the NINCDS-ADRDA work
group under the auspices of department of health and human
services task force on Alzheimer’s disease. Neurology, 34(7):
939–944.
Morris, J.C., Heyman, A., Mohs, R.C., van Belle, G., Fill-
enbaum, G., Mellits, E.D. and Clark, C. (1989) The Con-
sortium to Establish a Registry for Alzheimer’s Disease
(CERAD). Part I. Clinical and neuropsychological assess-
ment of Alzheimer’s disease. Neurology, 39(9): 1159–1165.
Mueller, S.G., Stables, L., Du, A.T., Schuff, N., Truran, D.,
Cashdollar, N. and Weiner, M.W. (2007) Measurement of
hippocampal subfields and age-related changes with high
resolution MRI at 4T. Neurobiol. Aging, 28(5): 719–726.
Mufson, E.J., Chin, E.Y., Cochran, E.J., Beckett, L.A., Ben-
nett, D.A. and Kordower, J.H. (1999) Entorhinal cortex
beta-amyloid load in individual with mild cognitive impair-
ment. Exp. Neurol., 158(2): 469–490.
Petersen, R.C. (2000) Aging, mild cognitive impairment and
Alzheimer’s disease. Dementia, 18: 789–805.
Petersen, R.C., Smith, G.E., Waring, S.C., Ivnik, R.J., Tanga-
los, E.G. and Kokmen, E. (1999) Mild cognitive impairment:
clinical characterization and outcome. Arch. Neurol., 56(3):
303–308.
Rodrigue, K.M. and Raz, N. (2004) Shrinkage of the entorhinal
cortex over five years predicts memory performance in
healthy adults. J. Neurosci., 24(4): 956–963.
Saykin, A.J., Wishart, H.A., Rabin, L.A., Flashman, L.A.,
West, J.D., McHugh, T.L. and Mamourian, A.C. (2006)
Older adults with cognitive complaints show brain atrophy
similar to that of amnestic MCI. Neurology, 67(5): 834–842.
Scheff, S.W. and Price, D.A. (2003) Synaptic pathology in
Alzheimer’s disease: a review of ultrastructural studies.
Neurobiol. Aging, 24(8): 1029–1046.
Scheff, S.W., Price, D.A., Schmitt, F.A. and Mufson, E.J.
(2006) Hippocampal synaptic loss in early Alzheimer’s dis-
ease and mild cognitive impairment. Neurobiol. Aging,
27(10): 1372–1384.
Small, S.A., Chawla, M.K., Buonocore, M., Rapp, P.R. and
Barnes, C.A. (2004) Imaging correlates of brain function in
monkeys and rats isolates a hippocampal subregion differ-
entially vulnerable to aging. Proc. Natl. Acad. Sci. U.S.A.,
101(18): 7181–7186.
Stoub, T.R., Bulgakova, M., Leurgans, S., Bennett, D.A.,
Fleischman, D., Turner, D.A. and deToledo-Morrell, L.
(2005) MRI predictors of risk of Alzheimer’s disease: a lon-
gitudinal study. Neurology, 64(9): 1520–1524.
Stoub, T.R., deToledo-Morrell, L., Stebbins, G.T., Leurgans,
S., Bennett, D.A. and Shah, R.C. (2006) Hippocampal dis-
connection contributes to memory dysfunction in individuals
at risk for Alzheimer’s disease. Proc. Natl. Acad. Sci. U.S.A.,
103(26): 10041–10045.
Squire, L.R. and Zola-Morgan, S. (1991) The medial temporal
lobe memory system. Science, 253(5026): 1380–1386.
Talairach, J. and Turnoux, P. (1988) Co-Planar Stereotaxic
Atlas of the Human Brain. Thieme, New York.
Tapiola, T., Pennanen, C., Tapiola, M., Tervo, S., Kivipelto,
M., Hänninen, T. and Pihlajamäki, M., et al. (2006) MRI of

hippocampus and entorhinal cortex in mild cognitive im-
pairment: a follow-up study. Neurobiol. Aging, (in press).
deToledo-Morrell, L., Dickerson, B., Sullivan, M.P., Spanovic,
C., Wilson, R.S. and Bennett, D. (2000a) Hemispheric differ-
ences in hippocampal volume predict verbal and spatial
memory performance in patients with Alzheimer’s disease.
Hippocampus, 10(2): 136–142.
deToledo-Morrell, L., Goncharova, I., Dickerson, B., Wilson,
R.S. and Bennett, D.A. (2000b) From healthy aging to
Alzheimer’s disease: in vivo detection of entorhinal cortex
atrophy. Ann. N.Y. Acad. Sci., 911: 240–253.
deToledo-Morrell, L., Sullivan, M.P., Morrell, F., Wilson,
R.S., Bennett, D.A. and Spencer, S. (1997) Alzheimer’s dis-
ease: in vivo detection of differential vulnerability of brain
regions. Neurobiol. Aging, 18(5): 463–468.
deToledo-Morrell, L., Stoub, T.R., Bulgakova, M., Wilson,
R.S., Bennett, D.A., Leurgans, S., Wuu, J. and Turner, D.A.
(2004) MRI-derived entorhinal volume is a good predictor of
conversion from MCI to AD. Neurobiol. Aging, 25(9):
1197–1203.
Van Hoesen, G.W. and Hyman, B.T. (1990) Hippocampal for-
mation: anatomy and the patterns of pathology in Al-
zheimer’s disease. Prog. Brain Res., 83: 445–457.
Van Hoesen, G.W. and Pandya, D.N. (1975) Some connections
of the entorhinal (area 28) and perirhinal (area 35) cortices of
the rhesus monkey. III. Efferent connections. Brain Res.,
95(1): 39–59.
Van Hoesen, G.W., Pandya, D.N. and Butters, N. (1975)
Some connections of the entorhinal (area 28) and perirhinal
753
(area 35) cortices of the rhesus monkey. II. Frontal lobe
afferents. Brain Res., 95(1): 25–38.
Wang, C., Medina, D., Stebbins, G., Shah, R., Bammer, R.,
Moseley, M. and deToledo-Morrell, L. (2006) Changes in the
microstructural integrity of parahippocampal white matter
tracts in Alzheimer’s disease. Soc. Neurosci. Abstr., on line.
Wechsler, D. (1987) Wechsler Memory Scale: Revised Manual.
Psychological Corporation, San Antonio, TX.
West, M.J., Coleman, P.D., Flood, D.G. and Troncoso, J.C.
(1994) Differences in the pattern of hippocampal neuronal
loss in normal aging and Alzheimer’s disease. Lancet,
344(8925): 769–772.
West, M.J., Kawas, C.H., Martin, L.J. and Troncoso, J.C.
(2000) The CA1 region of the human hippocampus is a hot
spot in Alzheimer’s disease. Ann. N.Y. Acad. Sci., 908:
255–259.
Wilson, R.S., Sullivan, M.P., deToledo-Morrell, L., Stebbins,
G.T., Bennett, D.A. and Morrell, F. (1996) Association of
memory and cognition in Alzheimer’s disease with volumetric
estimates of temporal lobe structures. Neuropsychology,
10(4): 459–463.
Xu, Y., Jack, C.R., O’Brien, P.C., Kokmen, E., Smith,
G.E., Ivnik, R.J., Boeve, B.F., Tangalos, R.G. and Peter-
sen, R.C. (2000) Usefulness of MRI measures of entorhinal
cortex versus hippocampus in AD. Neurology, 54(9):
1760–1767.
Young, B., Otto, T., Fox, G.D. and Eichenbaum, H. (1997)
Memory representation within the parahippocampal region.
J. Neurosci., 17(13): 5183–5195.
Plate 40.4. Color map showing significant (p ¼ 0.001) voxels of decreased white matter density in participants with amnestic MCI
compared with controls, superimposed on coronal and sagittal slices of a template based on data for all subjects in the study. The
image is masked to include white matter regions. The colors correspond to the t values shown on the color bar. Note the bilaterality of
significant differences in the white matter of the parahippocampal gyrus (adapted with permission from Stoub et al., 2006). (For B/W
version, see page 750 in the volume.)
H.E. Scharfman (Ed.)
Progress in Brain Research, Vol. 163
ISSN 0079-6123
Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 41

Epileptogenesis in the dentate gyrus:

a critical perspective

F. Edward Dudek1, and Thomas P. Sutula2

1
Department of Physiology, University of Utah School of Medicine, 420 Chipeta Way, Suite 1700, Salt Lake City,
UT 84108, USA
2
Department of Neurology, University of Wisconsin, Madison, WI, USA

Abstract: The dentate gyrus has long been a focal point for studies on the molecular, cellular, and
network mechanisms responsible for epileptogenesis in temporal lobe epilepsy (TLE). Although several
hypothetical mechanisms are considered in this chapter, two that have garnered particular interest and
experimental support are: (1) the selective loss of vulnerable interneurons in the region of the hilus and
(2) the formation of new recurrent excitatory circuits after mossy fiber sprouting. Histopathological data
show that specific GABAergic interneurons in the hilus are lost in animal models of TLE, and several
lines of electrophysiological evidence, including intracellular analyses of postsynaptic currents, support
this hypothesis. In particular, whole-cell recordings have demonstrated a reduction in the frequency of
miniature inhibitory postsynaptic currents in the dentate gyrus and other areas (e.g., CA1 pyramidal
cells), which provides relatively specific evidence for a reduction in GABAergic input to granule cells.
These studies support the viewpoint that modest alterations in GABAergic inhibition can have significant
functional impact in the dentate gyrus, and suggest that dynamic activity-dependent mechanisms of
GABAergic regulation add complexity to this local synaptic circuitry and to analyses of epileptogenesis.
In regard to mossy fiber sprouting, a wide variety of experiments involving intracellular or whole-cell
recordings during electrical stimulation of the hilus, glutamate microstimulation, and dual recordings
from granule cells support the hypothesis that mossy fiber sprouting forms new recurrent excitatory
circuits in the dentate gyrus in animal models of TLE. Similar to previous studies on recurrent excitation
in the CA3 area, GABA-mediated inhibition and the intrinsic high threshold of granule cells in the
dentate gyrus tends to mask the presence of the new recurrent excitatory circuits and reduce the likelihood
that reorganized circuits will generate seizure-like activity. How cellular alterations such as neuron loss in
the hilus and mossy fiber sprouting influence functional properties is potentially important for under-
standing fundamental aspects of epileptogenesis, such as the consequences of primary initial injuries,
mechanisms underlying network synchronization, and progression of intractability. The continuous
nature of the axonal sprouting and formation of recurrent excitation could account for aspects of the
latent period and the progressive nature of the epileptogenesis. Future studies will need to identify

Corresponding author. Tel.: +1 (801)587-5880;

Fax: +1 (801)581-8075; E-mail: ed.dudek@hsc.utah.edu

DOI: 10.1016/S0079-6123(07)63041-6 755

756

precisely how these hypothetical mechanisms and others contribute to the process whereby epileptic
seizures are initiated or propagated through an area such as the dentate gyrus. Finally, in addition to
its unique features and potential importance in epileptogenesis, the dentate gyrus may also serve as a
model for other cortical structures in acquired epilepsy.

Keywords: epilepsy; seizure; granule cell; hilus; interneuron; axon sprouting

Introduction A definition of epileptogenesis

Overview Epileptogenesis is broadly considered to be the

For decades, the dentate gyrus has attracted the
attention of epilepsy researchers, and this interest
has led to many publications that focus on this
particular brain structure. The high level of
interest stems from the development of several
concepts that have been derived from traditional
brain science and from both basic and clinical
epilepsy research. The goal of this chapter is to
consider some of the key hypotheses, concepts
and controversies in the field of epilepsy that are
centered on the role of the dentate gyrus in
‘‘epileptogenesis’’ and on how different possible
mechanisms in the dentate gyrus may be altered
in a manner that might lead to temporal lobe ep-
ilepsy (TLE). Epileptogenesis is broadly defined as
the biological process responsible for the devel-
opment of spontaneous recurrent seizures after an
acquired brain insult or a primary abnormality
such as a mutation or developmental defect.
Even focusing on epileptogensis alone, the field
is too extensive and complicated to review in this
brief chapter. Thus, our aim is to target a few
basic concepts in epilepsy research, and provide
commentary in those areas where we believe
there are particular problems in the field and/or
where a different or modified perspective would
be useful. We will provide only minimal discus-
sion of neurogenesis and GABA-A receptors as
they relate to epileptogenesis in the dentate
gyrus because these topics are amply covered by
Parent (this volume) and Coulter and Carlson
(this volume), respectively. Similarly, a detailed
discussion of the electrophysiological studies of
the human dentate gyrus in relation to epilepto-
genesis is provided by Williamson and Patrylo
(this volume).
collective molecular-, cellular-, and systems-level
mechanisms whereby spontaneous recurrent sei-
zures develop after a brain insult in a manner that
is time-dependent. For purposes of this review, we
will consider epileptogenesis in a more specific and
restrictive sense (Stables et al., 2003), rather than
the loose and more general meaning often associ-
ated with the term. The more restrictive definition
of epileptogenesis does not include: (1) acute sei-
zures associated with brain injury, (2) pharmaco-
logical treatments in animals or isolated tissue that
induce seizure-like activity, or (3) genetic abnor-
malities that are associated with increased seizure
susceptibility. While those processes are of interest
for understanding the acute mechanisms of net-
work synchronization and are sometimes included
in the broader less restrictive definition of epile-
ptogenesis, we believe for purposes of this critical
analysis that focusing on the more restrictive defi-
nition has advantages for understanding how
chronic epilepsy develops and progresses. We be-
lieve the term epileptogenesis also applies to the
progressive worsening of the epilepsy — in terms
of the increased frequency and/or severity of sei-
zures — after the first of the spontaneous recurrent
seizures. A critical debate in the field is whether
acquired epilepsy (versus genetic epilepsy) requires
actual brain damage and neuronal loss; some
researchers believe, particularly in regard to
acquired pediatric epilepsy (e.g., after hypoxia,
prolonged febrile seizures, or status epilepticus
(SE) in children and immature animals), that frank
injury with neuronal death does not need to be
present for the development of chronic epilepsy
(i.e., epileptogenesis) to occur. This issue has been
controversial, particularly in research on the dent-
ate gyrus. Many studies have regarded increased

seizure susceptibility as a surrogate marker
for epileptogenesis, and so another critical issue
involves the question: how good are the data that
an animal shows epileptogenesis (i.e., becomes
‘‘epileptic’’ with spontaneous recurrent seizures)?
Concepts of epileptogenesis pertinent to the dentate
gyrus
Several questions about epileptogenesis in the
dentate gyrus have captured the attention of
the epilepsy research community during the last
decade; these concepts will serve as targets for
discussion in terms of interpretation of the pres-
ently available literature and suggestions for future
research. The hippocampus has long been seen as
a critical structure in TLE, particularly as an
epileptic focus (i.e., presence of interictal spikes),
as an epileptogenic zone (site of seizure initiation),
and thus, as a primary target for resections aimed
at surgical treatment of intractable TLE (see
Engel, 1989 for review of the earlier literature).
This emphasis on the hippocampus in TLE, and by
extrapolation of the dentate gyrus, however, has
been challenged from many perspectives (e.g., see
Gloor, 1992; Bertram, 1997; Bertram et al., 1998).
Because the dentate gyrus has traditionally been
viewed as the first stage or ‘‘gate’’ into the classical
trisynaptic circuit of the hippocampus, a hippo-
campo-centric view has led to a dentate-centric
focus in the minds of many researchers. This is not
to dismiss the concept that the hippocampus and
dentate gyrus are likely important in TLE; rather,
it is to acknowledge that other areas may be as
important (or even more important), and that the
mechanisms that occur during epileptogenesis in
the dentate gyrus may also be present in other
brain regions. The observation of ‘‘maximal dent-
ate activation’’ strengthened the dentate ‘‘gate’’
concept in epilepsy research, and provided a
hypothetical electrophysiological basis for how
the dentate could serve a ‘‘closed gate’’ versus an
‘‘open gate’’ function (Heinemann et al., 1992;
Lothman et al., 1992; Stringer and Lothman,
1992).
Another basis for interest in the dentate gyrus
by epilepsy researchers has been the hypothesis
757
(e.g., Sloviter, 1991; Sloviter et al., 2003) that the
loss of hilar neurons (or ‘‘endfolium sclerosis’’ —
considered to be a subset of the more general
histopathological condition of mesial temporal
sclerosis; Margerison and Corsellis, 1966) is the
central feature or minimum essential substrate of
TLE. Several studies on animal models of trau-
matic brain injury (e.g., Lowenstein et al., 1992;
Toth et al., 1997) and ischemia (e.g., Crain et al.,
1988; Hsu and Buzsaki, 1993; Williams et al.,
2004), which may lead to epilepsy, have found that
the hilus is particularly prone to neuronal injury;
therefore, interest has focused on the vulnerability
of the different types of hilar neurons, and how
this might alter the electrophysiological properties
of the dentate network. At least two separate
concepts or hypotheses have emerged concerning
the consequences of hilar neuron loss and the
possible development of increased network excit-
ability: (1) a reduction in the number of interneu-
rons (e.g., Obenaus et al., 1993; Houser and
Esclapez, 1996) and/or (2) a loss of excitatory
input to interneurons (i.e., the ‘‘dormant basket
cell hypothesis’’; e.g., Sloviter, 1987; Sloviter et al.,
2003).
In a separate but related line of research, at least
four groups (de Lanerolle et al., 1989; Sutula et al.,
1989; Houser et al., 1990; Babb et al., 1991) found
that the presence of mesial temporal sclerosis in
specimens from surgical resections for intractable
TLE was correlated with ‘‘mossy fiber sprouting,’’
where the Timm stain method densely stains the
inner molecular layer, which is not present in
normal tissue (e.g., Tauck and Nadler, 1985; see
Sutula and Dudek, this volume). Although the
dentate gyrus is strategically located, and both
hilar neuron loss and mossy fiber sprouting are
common features of TLE, the dentate gyrus and its
granule cells are clearly not required for epilepto-
genesis and both hilar neuron loss and mossy fiber
sprouting in the dentate gyrus do not appear to be
necessary or sufficient for spontaneous recurrent
seizures to occur. For reasons that we will discuss
and that are also developed in another chapter
(Sutula and Dudek, this volume), these observa-
tions do not diminish the potential importance of
understanding how the dentate gyrus contributes
to hippocampal epileptogenesis and the syndromes
758
of TLE. In the sections that follow, we will con-
sider how some of the features of the organization
and electrophysiology of the dentate gyrus are
specifically pertinent to the role of the dentate
gyrus in epileptogenesis.
Epileptogenesis in the dentate gyrus: the laminar
organization of the hippocampus leads to large field
potentials and thus an increased ability to detect
epileptiform events
While the dentate gyrus and hippocampus are
commonly involved in TLE, which is the most
common type of intractable epilepsy, the high
interest and potential importance of this region of
limbic circuitry derives from the frequent obser-
vation of interictal discharges and apparent onset
of seizures during intrahippocampal depth record-
ings both clinically and in animal models of TLE,
in addition to the characteristic histopathological
(and clinical imaging) abnormalities (see Engel,
1989 for classical overview). An important point is
that because of the highly organized and laminar
structure of the hippocampus, including the dent-
ate gyrus, synchronous activity in any particular
part of the hippocampus generates especially large
field potentials in the extracellular space. There-
fore, compared to other structures with less
packed and layered cell bodies, such as the am-
ygdala and entorhinal cortex, epileptiform events
(including interictal spikes and seizures) would be
expected to be much more readily observed in the
hippocampus and dentate gyrus than other parts
of the cerebral cortex. Thus, a seizure may start in
a nearby structure (e.g., amygdala or entorhinal
cortex) with less lamination that generates a
smaller and more-difficult-to-detect field potential,
and then spread to the dentate gyrus and other
hippocampal areas, where the relatively large-
amplitude field potentials of the seizure might be
first detected. For example, considerable interest
has focused on the occurrence of ‘‘fast ripples,’’
which is vernacular that refers to high-frequency
electrographic oscillations (i.e., approximately
200–500 Hz), as markers of epileptic seizure onset
in the dentate gyrus and elsewhere in the hippo-
campus and parahippocampal areas. Based on
their time course, these events are almost certainly
synchronous action potentials (i.e., small popula-
tion spikes) that are likely to be much larger in
the dentate gyrus and hippocampus than other
seizure-prone areas because of their laminar struc-
ture (Bragin et al., 2002). As a result of the laminar
organization, it is and will continue to be difficult
to assess where electrographic seizures actually
start, and whether an area in the hippocampus,
or in another structure, is the actual initiation
site. Nonetheless, the ability to analyze different
components of the field potentials in the dentate
gyrus (and hippocampus proper) in response to
electrical stimulation of synaptic inputs and during
spontaneous seizures provides important potential
avenues for research.
Epileptogenesis in the dentate gyrus: the normal
dentate gyrus has a high threshold for seizure
generation, so it may not be particularly prone to
chronic epileptogenesis
It is widely believed that the most important
epileptogenic structures are particularly suscepti-
ble to seizure generation in the normal brain, and
these structures are even more prone to ‘‘hyperex-
citability’’ in the epileptic brain, where the term
hyperexcitability describes a general increase in
response to a particular stimulus or enhanced
tendency to generate repetitive synchronous neu-
ronal discharges manifesting as a burst of popu-
lation spikes. If this is the case, one could argue
that the dentate gyrus — as studied in the normal
brain — should be relatively low on the list of
candidate structures for epileptogenesis. Although
the dentate gyrus is a unique structure with a
high degree of neuronal homogeneity among
the granule cells and heterogeneity in the hilus, it
has long been known that the dentate granule
cells are highly resistant to epileptiform activity
after treatment with convulsive agents, such as
GABA-A receptor antagonists (e.g., Fricke and
Prince, 1984). In the normal brain, the dentate
gyrus appears to be the least susceptible part of the
hippocampus to generate either brief interictal-like
bursts or prolonged seizure-like activity, at least
when compared to the CA1 and CA3 areas. This

resistance to epileptiform activity is widely
believed to be due to strong GABAergic inhibi-
tion, but the relatively negative resting membrane
potential, high threshold for action potential
generation, and high degree of spike accommoda-
tion to maintained depolarization (e.g., see Staley
et al., 1992) are probably equally as important
reasons why the dentate granule cells are relatively
seizure-resistant under normal conditions. Many
studies in epilepsy research have not considered
these fundamental properties of the normal dent-
ate gyrus; and on these grounds, the dentate gyrus
may be an unlikely area for seizure onset during
epileptogenesis (i.e., be an epileptogenic zone),
because seizure threshold of the dentate gyrus
often appears high compared to other structures
even when tested in chronically epileptic animals.
The transformation of the synaptic interactions in
the dentate during epileptogenesis could, however,
be more complex and important than is apparent
from the previous discussion. Indeed, if properties
such as a more negative resting membrane poten-
tial, higher action potential threshold, lower
spontaneous firing rate, and resistance to repeti-
tive firing endow the dentate gyrus with resistance
to synchronous network discharge and epilepto-
genesis, then processes such as hilar neuronal
loss and formation of recurrent excitatory circuits
by mossy fiber sprouting could be especially
important in the transformation of the dentate
gyrus into a more epileptogenic structure.
Models of epileptogenesis in the dentate gyrus:
evidence for filtering and gating properties in the
dentate gyrus
The perforant path projection from the entorhinal
cortex to the dentate gyrus is the first synapse of
the classic tri-synaptic hippocampal circuit,
and this synapse to the granule cells is generally
believed to be relatively resistant to transmission
of activity into CA3 (i.e., the dentate can be
considered to be a gate that is usually closed; see
Hsu, this volume). The ‘‘gate’’ property of the
dentate gyrus is often thought to impede the prop-
agation of normal electrical activity and seizures
into hippocampus, and when this gating property
759
fails, the dentate may allow propagation of activity
into other structures that are projection targets
from the hippocampus (Heinemann et al., 1992;
Lothman et al., 1992; Stringer and Lothman,
1992). There is evidence that the ‘‘gate’’ function
of the dentate gyrus may operate in an ‘‘all or
none’’ fashion, as implied by the observation of
‘‘maximal dentate activation’’ associated with
propagation of epileptiform activity. One basis
for the ‘‘gate’’ concept is the relatively high thresh-
old for excitation of the dentate, and according to
this perspective, reduction in the ‘‘gate’’ function
of the dentate gyrus could be an epileptogenic
mechanism that would promote increased excita-
bility (i.e., hyperexcitability) to perforant path
stimuli, transforming synchronous excitatory
postsynaptic potentials (EPSPs) with superim-
posed action potentials into a large burst of ac-
tion potentials. This mechanism could apply to
synaptic inputs that include interictal spikes and
actual seizures; thus, it has been considered that
the ‘‘gate’’ to the hippocampus normally restricts
or blocks epileptiform activity. While this ‘‘gating’’
property may restrict epileptiform activity, the
parallel pathway from the entorhinal cortex pro-
jecting directly to the CA3 and CA1 areas may still
transmit discharges into the hippocampus regard-
less of the properties of the dentate gyrus. Other
studies suggest that the dentate gyrus can serve as
a filter whereby activity is blocked from entering
the dentate under some conditions but not others
(e.g., see Iijima et al., 1996; Behr et al., 1998; ver-
sus Ang et al., 2006). Although the dentate gyrus
may not be necessary for transmission of signals
into the hippocampus, it is not fully understood
how the dentate gyrus processes electrical infor-
mation in relation to normal hippocampal inte-
gration (see Kesner, this volume) or how the
synaptically reorganized dentate gyrus affects the
transmission of seizures into the hippocampus in
animal models of epilepsy or patients with TLE
(see Sutula and Dudek, this volume). Thus, the
dentate gyrus presumably plays an important role
in epileptogenesis, even though the entorhinal cor-
tex appears to project ‘‘downstream’’ to the CA3
and CA1 areas of the hippocampus. The concept
that the dentate simply functions as a ‘‘gate’’ that
normally blocks seizure activity in the entorhinal
760
cortex from propagating into the hippocampus
may, however, be an oversimplification and
require a more sophisticated assessment in the
complex network of serial and parallel pathways
projecting into the hippocampus and to parahip-
pocampal structures, such as the subiculum, which
has also been implicated in TLE (Cohen et al.,
1998).
Epileptogenesis in the dentate gyrus: repeated
activation of the dentate gyrus can promote
propagation of seizures into the hippocampus
The high seizure threshold of the normal dentate
gyrus becomes dramatically reduced after it has
experienced a few electrically induced afterdis-
charges (i.e., maximal dentate activation), and so
the dentate appears to be highly sensitive to pre-
vious electrical activity (Heinemann et al., 1992;
Lothman et al., 1992; Stringer and Lothman,
1992). Even a single afterdischarge in the dentate
gyrus increases synaptic transmission for periods
of as long as 3 months, and induction of long-term
potentiation increases susceptibility to evoked net-
work activity and afterdischarges (Sutula and
Steward, 1986, 1987; Sayin et al., 1999). These
forms of network plasticity occur over a relatively
short time frame and are therefore unlikely to ac-
count for chronic epileptogenesis, but relatively
long-lasting alterations could promote network
synchronization in the dentate gyrus and hippo-
campal pathways. The potential role of the dentate
in modulating, reducing, or filtering some forms of
electrical activity and possibly single seizures
would be degraded by activity-dependent enhance-
ment of synaptic transmission in granule cells, by
allowing more propagation of clusters of seizures
into the hippocampus proper. Therefore, although
some forms of entorhinal cortical activity can nor-
mally bypass the dentate gyrus during propagation
into the hippocampus, it remains possible that the
altered ability of repetitive seizures or seizure clus-
ters to spread into the hippocampus after maximal
dentate activation or kindling may be a critical
feature of epileptogenesis during TLE. The present
approaches with hippocampal slices, or even intact
anaesthetized preparations, might fail to identify
potentially important functions regarding sponta-
neous interictal spikes and seizures versus electri-
cally stimulated events.
Animal models of epileptogenesis during TLE:
alterations of the dentate gyrus
Models and hypothetical mechanisms of
epileptogenesis
Animal models simulating histopathological
features of human TLE in the dentate gyrus have
been used for the study of chronic epileptogenesis.
The two most widely used models of epileptogen-
esis in adult animals, SE and kindling, have been
used to examine processes of epileptogeneis fol-
lowing an initial injury (e.g., SE) and repeated
brief seizures (i.e., kindling). Both of these induc-
tion scenarios mimic common but distinctive clin-
ical features of human TLE, namely syndromes
that follow an initial precipitating injury and
cryptogenic cases that gradually progress to in-
tractability. These models have provided experi-
mental opportunities to assess two general types of
hypothetical mechanisms derived from histopath-
ological alterations in the dentate gyrus commonly
associated with human TLE that would change the
balance of excitation and inhibition, specifically
reducing GABAergic inhibition and/or increasing
glutamatergic excitation. The hypothetical reduc-
tion in inhibition has been proposed to be due
to a decrease in inhibitory interneurons and/or a
reduction of excitatory synaptic input to inhibitory
interneurons (i.e., the ‘‘dormant basket cell’’
hypothesis). Both of these effects could occur as
a result of the histopathological observation of a
loss of hilar neurons. The hypothesis of enhanced
excitation in the dentate gyrus has focused on new
recurrent excitatory circuits associated with mossy
fiber sprouting, as detected by increased Timm
stain product in the inner molecular layer of the
dentate gyrus. Although there are many differ-
ences between the kindling and SE models, both
show some of the histopathological alterations in
the dentate gyrus associated with TLE, and have
at least some utility to address these hypothetical
mechanisms of epileptogenesis.

Models of epileptogenesis based on kindling
Kindling is a progressive decrease in the threshold
for induction of afterdischarges to daily electrical
stimulations of the amygdala or other limbic struc-
tures. This animal model progressively develops a
chronic irreversible state where low-intensity stim-
uli trigger prolonged afterdischarges and seizures.
Kindling initially induces low-level apoptosis in
the hilus (Bengzon et al., 1997) and cumulatively
results in progressive loss of neurons not only
in the hilus, but in CA3 and CA1 in a pattern
resembling human hippocampal sclerosis (Cavazos
et al., 1994). Kindling also induces mossy fiber
sprouting, and the Timm stain in the molecular
layer progressively increases over many months
with prolonged kindling (Cavazos et al., 1991),
although the density of staining in the inner
molecular layer is generally less than is typically
seen in SE models where epileptogenesis occurs
over several weeks and months.
Models of epileptogenesis based on SE
The two main approaches that have been invoked
to induce SE and subsequent epileptogenesis
involve injection of chemotoxins or electrical
stimulation. Kainic acid (i.e., kainate; Ben-Ari,
1985; Nadler, 1991) and pilocarpine (Turski et al.,
1983; with or without lithium pretreatment)
are the most commonly used chemotoxins, but
repetitive electrical stimulation of certain limbic
structures can also induce SE (e.g., Lothman et al.,
1990). The repetitive seizures during SE cause
neuronal loss in the hilus of the dentate gyrus and
in the CA3 and CA1 areas of the hippocampus,
plus other brain areas. The neuronal death in the
hilus is thought to contribute to mossy fiber
sprouting (see Sutula and Dudek, this volume).
Specific neuronal injury in the hilus and
hippocampus versus more extensive brain damage
associated with human TLE
The extrahippocampal damage associated with the
models based on SE is often severe and extensive,
and thus it has been argued that the SE models do
761
not recapitulate the key histopathological features
of TLE (Sloviter, 1991). Based on an interpreta-
tion of Margerison and Corsellis (1966) (see also
Meldrum and Bruton, 1992), some workers have
argued that the SE models have too little damage
in areas that are completely damaged in TLE
(e.g., CA1 and hilus), but too much damage in
extrahippocampal areas. Although the hilus
was seen to be the most common area for severe
neuronal loss (about 65% of patients) in a histo-
pathological study on autopsy tissue from patients
institutionalized due to epilepsy and other poten-
tial disorders, Margerison and Corsellis (1966)
also emphasized that extrahippocampal damage is
commonly observed in many patients with classic
features of mesial temporal scerosis. For example,
Margerison and Corsellis (1966) state that their
study of the whole brain ‘‘has emphasized the
need to see beyond the temporal lobes and to take
into account the possible importance of damage
in other parts of the brain as well.’’ More recent
imaging studies have emphasized the more wide-
spread injury to areas that communicate with the
hippocampus, such as the amygdala and the ent-
orhinal and perirhinal cortex (see Cascino, 2005),
although other brain areas are damaged in some
patients (Margerison and Corsellis, 1966). Thus,
the hippocampus in general and the dentate gyrus
in particular are brain regions of considerable
interest in relation to human TLE, but other areas
are also involved. Furthermore, it is not clear that
the SE models are actually such a poor reflection
of TLE, as some would suggest.
Histopathological alterations in the dentate gyrus of
animal models of epileptogenesis
The models of TLE based on SE often have
more damage to other brain areas than those TLE
patients with rather focal hippocampal and para-
hippocampal lesions, but the SE models show many
histopathological similarities to human TLE in the
dentate gyrus. The SE models typically show exten-
sive histopathological damage to the hilus of the
dentate gyrus (Fig. 1) and to the CA3 and/or CA1
areas of the hippocampus that is roughly similar to
human TLE. These models have been employed
762

Fig. 1. Nissl staining to show hilar neurons in a control rat versus a rat with kainate-induced epilepsy. (A and B) Sections in the septal
hippocampus show some loss of hilar neurons in a kainate-treated rat (B) compared to control rat (A). (C and D) Near the temporal
end of the hippocampus, the loss of hilar neurons in the kainate-treated rat was severe (D). m: molecular layer; g: granule layer; h:
hilus; CA3: CA3 pyramidal cell layer; bar ¼ 500 mm. Adapted with permission from Buckmaster and Dudek (1997) Fig. 1, p. 388.

extensively, because they are relatively easy to use
(although mortality can be a problem), and they
show histopathological alterations in the hilus of the
dentate gyrus that broadly reflect some of the fea-
tures of human TLE. In the SE models, spontane-
ous seizures develop with an apparent latent period
and progressively increase in seizure severity and
frequency. The SE models provide a contrast to
kindling, where neuronal injury is more gradually
induced in the hilus and other regions, mossy fiber
sprouting gradually progresses as a function of the
number of evoked seizures, and spontaneous sei-
zures develop more slowly, typically after 90–100
seizures evoked by daily stimulation (Sayin et al.,
2003). When spontaneous seizures are present in
more extensively kindled animals, mossy fiber
sprouting is more prominent and interneurons are
lost in the hilus (Sayin et al., 2003). Thus, the kin-
dling and SE models recapitulate certain features of
the histopathological changes and seizure suscepti-
bility characteristic of TLE, but with rates of in-
duction and progression that distinctively mimic
clinical features of some TLE syndromes, including
onset with a delay after initial injury and progres-
sion to intractability. It should be emphasized that
not all cases of human TLE develop after an ob-
vious initial injury and not all cases progress to in-
tractability, so neither the SE models nor kindling
are universally applicable to the broad range of
syndromes of human TLE. It should be further em-
phasized that most histopathological studies on hu-
mans with TLE are based on patients who have
suffered from intractable epilepsy for many years or
even decades, while research on animals (e.g., the SE
models) does not usually involve periods greater
than a few weeks or months.
Damage to the hilus is a common characteristic of
TLE
The dormant basket cell hypothesis
Many studies over the last two decades have
aimed to address the issue of which types of

neurons (see Houser, this volume) are lost in the
hilus in human tissue samples and in animal mod-
els of TLE (particularly the SE models). Although
certain vulnerable GABAergic interneurons are
lost, many other GABAergic neurons remain in-
tact after a repetitive stimulation protocol to the
perforant path that causes ‘‘hyperexcitability’’
of dentate field potentials (Sloviter, 1987, 1991).
Because many inhibitory basket cells were still
present within the dentate gyrus in human TLE
and in animal models, Sloviter (1987, 1991) pro-
posed the ‘‘dormant basket cell’’ hypothesis, which
essentially states that TLE is associated with loss
of excitatory mossy cells that normally project to
GABAergic basket cell interneurons; therefore; the
basket cells in the hilus of TLE patients and an-
imal models of TLE have lost excitatory synaptic
input (i.e., become ‘‘dormant’’), thus leading to the
hyperexcitability. This hypothesis, however, is
based on the relatively acute effects of repetitive
extracellular stimulation (i.e., after hours and a
few days), when chronic epileptic seizures either do
not occur or are quite rare. Furthermore, nearly
all of the electrophysiological data presented in
support of this hypothesis involve in vivo paired-
pulse experiments with extracellular stimulation
and recording techniques using a range of inter-
pulse intervals and repetitive stimulation frequen-
cies (e.g., Sloviter, 1987; Sloviter et al., 2003) that
are difficult to interpret and are generated by
complicated physiological processes at many
other cellular sites and levels (e.g., the perforant
path-to-granule cell synapse; GABA-A receptors,
chloride ion homeostasis, etc). A reduction of
afferent input to GABAergic neurons has been
reported in the CA1 area (Bekenstein and Loth-
man, 1993), but this study did not directly assess
excitation to interneurons and alterations of a va-
riety of physiological processes are likely to con-
tribute to a net defect in inhibitory mechanisms.
For example, Doherty and Dingledine (2001)
found a reduced excitatory drive onto interneu-
rons in the dentate gyrus after SE, and this effect
was due to a use-dependent mechanism involving
group II metabotropic glutamate receptors. Sev-
eral other studies have addressed the dormant
basket cell hypothesis. Based on a series of exper-
iments using hippocampal slices from animals with
763
chronic epilepsy (i.e., SE models), Bernard et al.
(1998) (see also Esclapez et al., 1997) have pro-
vided several lines of evidence that basket cells are
not ‘‘dormant’’ in TLE. In addition, studies on
miniature and spontaneous inhibitory postsynap-
tic currents (mIPSCs and sIPSCs) of dentate gran-
ule cells from the kainate model provide evidence
that interneurons are spontaneously active even
when iontopic gluatamatergic receptors are
blocked. When in vitro electrophysiological exper-
iments were conducted on hippocampal slices after
the pharmacological blockade of ionotropic glut-
amatergic receptors to isolate mIPSCs and sIPSCs
from glutamatergic synaptic circuits (Shao and
Dudek, 2005), the frequency of sIPSCs in dentate
granule cells at o1 week and >3 months (i.e.,
from chronically epileptic rats) after kainate-
induced SE was not significantly different from
controls, and was much higher than when the
slices were bathed in tetrodotoxin to block so-
dium-mediated action potentials and isolate mI-
PSCs. The observation that blocking action
potentials with tetrodotoxin greatly reduced the
frequency of IPSCs (i.e., the frequency of sIPSCs
was much higher than mIPSCs) in all groups
indicates that the interneurons that project to
granule cells in kainate-treated rats (and controls)
are clearly not ‘‘dormant’’ (i.e., they are sponta-
neously active) either o1 week or >3 months (i.e.,
from chronically epileptic rats) after kainate-in-
duced SE, even when recorded in isolated slices
where GABAergic inhibitory circuits are intact and
bathed in pharmacological agents that block glut-
amatergic EPSCs. Although this approach relies
on mIPSCs and sIPSCs in granule cells as a meas-
ure of the upstream GABAergic inhibitory net-
work, it is not affected by potential differences in
the responses to extracellular stimulation, nor is it
influenced by simultaneous alterations in glut-
amatergic synapses or circuits. This analysis is
likely also somewhat of a simplification, since it
has focused on somatic inhibition (i.e., the primary
target of basket cells), but other studies with com-
bined dendritic and somatic recording in the CA1
area of slices from pilocarpine-treated rats have
provided evidence for differential effects of epile-
ptogenesis on dendritic and somatic inhibition
(Cossart et al., 2001). It remains possible, however,
764
that basket cells and some other remaining inter-
neurons may experience a significant reduction of
excitatory synaptic input due to the loss of input
from glutamatergic neurons in the hilus and else-
where, and this loss of excitatory input could
importantly affect the ‘‘balance’’ of synaptic
excitation and inhibition. Experiments with
paired-pulse stimulation, particularly in intact
animals, lack the mechanistic resolution to pro-
vide evidence supporting the dormant basket cell
hypothesis. Given the numerous studies at the cel-
lular and synaptic levels indicating that surviving
interneurons have significant spontaneous and
evoked activity in SE models, and the interpretive
ambiguity and limitations of paired-pulse meas-
urements using a range of interpulse intervals and
repetitive stimulation, overall the evidence for dor-
mancy of basket cells as a mechanism of epilepto-
genesis is weak and indirect. Quantitative
electrophysiological studies with whole-cell
recordings of EPSCs to assess changes in the ex-
citatory synaptic inputs to the different types of
GABAergic interneurons during epileptogenesis,
preferably using several different animal models of
TLE (i.e., including but not limited to kindling and
SE models), could be useful to further evaluate
this hypothesis.
Loss of interneurons
Whereas the dormant basket cell hypothesis has
not garnered support from recent electrophysio-
logical experiments using modern recording tech-
niques, data with a wide range of morphological
and physiological techniques from many labora-
tories in human tissue and animal models directly
support the hypothesis that GABAergic inhibition
undergoes alterations in association with modest
loss of specific interneurons in TLE. Early studies
in a variety of hippocampal and cortical areas
using various immunohistochemical and in situ
hybridization techniques have shown that some
GABAergic interneurons (i.e., a relatively small
number, and only specific types) are lost in TLE
(Fig. 2; e.g., Sloviter, 1987; Obenaus et al., 1993;
Buckmaster and Dudek, 1997; Gorter et al., 2001).
Immunohistochemical data and silver stains after
forebrain ischemia have shown that selectively
vulnerable neurons, including cells that are
likely GABAergic interneurons, are damaged.
Similar studies, focusing on immunohistochemical
techniques, identified several different types of
GABAergic interneurons that were damaged after
traumatic brain injury, and many of the injured
interneuron types corresponded to those lost in
TLE. These histopathological alterations in inter-
neuron populations may contribute to the hyper-
excitability often seen shortly after experimental
SE and other brain insults (e.g., Lowenstein et al.,
1992). Several other mechanisms, however, can
hypothetically be responsible for or contribute to
the hyperactive responses to extracellular stimula-
tion within days after experimental SE, including
alterations arising from direct SE-induced changes
in chloride distribution and other GABA-A
receptor-mediated mechanisms (e.g., Kapur et al.,
1999).
Several laboratories have used high-resolution
whole-cell recordings of mIPSCs to test more di-
rectly the hypothesis of a loss of interneuron in-
put to granule cells (Kobayashi and Buckmaster,
2003; Shao and Dudek, 2005) and CA1 pyramidal
cells (Wierenga and Wadman, 1999), and they
have found that the frequency of mIPSCs is re-
duced in these models, which is consistent with a
loss of inhibitory GABAergic terminals (i.e.,
death of GABAergic interneurons) in TLE. Sim-
ilar data showing a reduction in the frequency of
mIPSCs have been reported after lateral fluid
percussion injury (Toth et al., 1997). An impor-
tant advantage of this approach is that it is
independent of the complexities and uncertainties
of extracellular stimulation within a multisynaptic
circuit. Because all of the experimental record-
ings were conducted in tetrodotoxin to block
action potential-mediated activity (and in some
cases, ionotropic glutamatergic antagonists to
block fast EPSPs), the number of other potential
confounding mechanisms and difficult-to-control
variables was reduced. In contrast to the lack of
evidence for dormancy of basket cells, the evi-
dence from direct experiments on several animal
models from several laboratories that epilepto-
genesis is associated with a modest but measur-
able reduction in inhibition to dentate granule
 765

Fig. 2. GAD mRNA-containing neurons in the dentate gyrus of a control (A) and pilocarpine-treated (B) rat. (A) In the normal
dentate gyrus, numerous GAD mRNA-labeled neurons are present in the hilus (H). Many labeled neurons are also located along the
inner border of the granule cell layer (G) and in the molecular layer (M). (B) In a pilocarpine-treated animal, the number of labeled
neurons throughout the hilus (H) is substantially reduced. Some GAD mRNA-containing neurons persist along the base of the granule
cell layer (G) and in the molecular layer (M); scale bar ¼ 200 mm. Adapted with permission from Obenaus et al. (1993) Fig. 4, p. 4476.

cells associated with a loss of specific types of
vulnerable interneurons in the hilus of the dentate
gyrus (and also in CA1) is quite strong. Because
this reduction in the number of specific interneu-
rons ultimately would be expected to enhance the
effects of excitatory synaptic input to the granule
cell population, and because assessments of GAB-
Aergic inhibition using extracellular stimulation
depend heavily on the intensity of the stimulation,
the types of analyses that employ electrical stim-
ulation of an afferent input such as the perforant
path are problematic. Furthermore, the presence
of fewer interneurons will lead to complex effects
during the repetitive activation expected to occur
as multiple interictal spikes and seizures begin
to occur in the entorhinal cortex and dentate
gyrus.
Mossy fiber sprouting is also a common feature of
TLE
Axon sprouting and increased recurrent excitation
Mossy fiber sprouting has attracted considerable
attention because the Timm stain shows a dra-
matic and reproducible change in the distribution
of the mossy fiber axons in the inner molecular
layer of the dentate gyrus associated with TLE
(Fig. 3). The ease and relative simplicity of this
staining technique and the apparent clarity of the
results have drawn attention to the dentate gyrus
and this particular form of synaptic reorganization
of a local-circuit pathway. The fact that numerous
laboratories have independently observed Timm
stain in the inner molecular layer of human tissue
766

Fig. 3. The Timm stain showed dark labeling in the granule cell layer and inner molecular layer in rats with kainate-induced epilepsy
versus control rats. (A) Hippocampal sections from a control rat had relatively little Timm stain (with light cresyl violet counter-
staining) in the granule cell and inner molecular layers. (B–E) Sections of the middle region along the septotemporal axis of the
hippocampus from rats with kainate-induced epilepsy displayed different degrees of abnormal Timm staining in the granule cell layer
(B shows the least and E the most). Arrows indicate regions of Timm staining in the inner molecular layer. M: molecular layer; g:
granule layer; h: hilus; CA3: CA3 pyramidal cell layer; bar ¼ 500 mm. Adapted with permission from Buckmaster and Dudek (1997)
Fig. 7, p. 395.

from patients with intractable epilepsy (de Lane-
rolle et al., 1989; Sutula et al., 1989; Houser et al.,
1990; Babb et al., 1991) supports the concept that
it is one of the changes likely to be associated with
the broadly defined concept of synaptic reorgan-
ization. Animal models, ranging from pilocarpine-
to kainate-treated rats to the kindling model, have
all revealed a time-dependent increase in Timm
stain in the inner molecular layer of the dentate
gyrus. As reviewed elsewhere in this volume
(Sutula and Dudek, this volume), several lines of
electrophysiological and ultrastructural data sug-
gest that most if not all of the new synapses in the
‘‘reorganized’’ dentate gyrus are excitatory. For
example, several laboratories have shown that hi-
lar (and perforant path) stimulation can lead to
delayed or long-latency EPSPs and prolonged
spike bursts several weeks or months after kain-
ate treatment, particularly when slices from rats
with chronic epilepsy are bathed in GABA-A
receptor antagonists and/or high potassium,
versus similar experiments on slices from control
animals where hilar stimulation evokes one anti-
dromic action potential and sometimes an EPSP
(Tauck and Nadler, 1985; Cronin et al., 1992;
Wuarin and Dudek, 1996; Patrylo and Dudek,
1998; Hardison et al., 2000, Lynch and Sutula,
2000). The advantage of hilar over perforant path
stimulation in these earlier studies was that elec-
trical stimulation of the hilus would be expected
to activate the mossy fiber axons of the granule
cells themselves versus the powerful monosynaptic
inputs of the perforant path, but in either case,
the key result is that the EPSPs in slices from rats

with robust mossy fiber sprouting occurred with a
longer and more variable latency than would be
expected from a direct monosynaptic input. These
experiments were founded conceptually on the
studies of recurrent excitation in the CA3 area
(Traub and Wong, 1982, 1983; Miles and Wong,
1987), where low-intensity stimulation triggered
all-or-none synaptic bursts with a long–and vari-
able latency and paired intracellular recordings
showed multisynaptic interactions, only when
GABA-A mediated inhibition was blocked. The
advantage of performing these experiments during
blockade of GABA-A mediate inhibition is based
not only on the previous work showing that
GABA-mediated inhibition has a ‘‘masking’’ effect
on recurrent excitation (Christian and Dudek,
1988; Traub and Wong, 1982, 1983; Miles and
Wong, 1987), but also because it controls at least
partially for epileptogenesis-associated differences
in GABA-mediated inhibition (see above), partic-
ularly since the experimental design in these stud-
ies compared slices from the animal model of
epilepsy with exactly equivalent slices from control
animals. Electrical stimulation, however, activates
axons of passage and can thus be nonspecific.
More direct experiments using glutamate micro-
stimulation techniques that do not activate fibers
of passage, including glutamate microdrops
(Wuarin and Dudek, 1996; Lynch and Sutula,
2000) and focal photoactivation of caged gluta-
mate (Molnar and Nadler, 1999; Wuarin and
Dudek, 2001), showed that (1) relatively selective
stimulation of dentate granule cells could cause
EPSCs in other granule cells in normal medium,
(2) these recurrent excitatory circuits increased in
density and effectiveness over the ensuing months
after experimental SE, and (3) when the slices were
bathed in bicuculline, the new local excitatory cir-
cuits of the granule cells could generate novel net-
work bursting to focal uncaging of glutamate in
the granule cell layer (but not elsewhere) several
months after SE, but not at earlier times and not in
slices from control preparations. Although these
experiments with glutamate microdrops and focal
uncaging of glutamate used more specific tech-
niques than extracellular electrical stimulation to
activate granule cells, and showed that the forma-
tion of new recurrent excitatory circuits and the
767
consequent generation and propagation of
epileptiform activity in the dentate gyrus was time
dependent (Wuarin and Dudek, 2001), they did
not directly address the issue of monosynaptic
connections between individual granule cells. Dual
intracellular recordings provided direct evidence
that mossy fiber sprouting in rats with pilocarpine-
induced epilepsy is associated with monosynaptic
recurrent excitatory connections among dentate
granule cells (Scharfman et al., 2003). The most
direct and quantitative ultrastructural studies also
point to a high preponderance of new excitatory
synapses to dentate granule cells versus interneu-
rons (Zhang and Houser, 1999; Buckmaster et al.,
2002). Although synaptic reorganization may have
a unique importance in the dentate gyrus, a more
likely scenario is that axon sprouting and the for-
mation of new recurrent excitatory circuits is a
widespread phenomenon after injury that occurs
throughout many areas of the hippocampus and
neocortex during epileptogenesis. For example,
numerous laboratories have provided morpholog-
ical and electrophysiological evidence that a sim-
ilar mechanism of axonal sprouting and enhanced
recurrent excitation occurs in the CA1 area during
epileptogenesis (Meier and Dudek, 1996; Perez et
al., 1996; Esclapez et al., 1999; Smith and Dudek,
2001, 2002; Shao and Dudek, 2004). As expected
from previous experiments performed on recurrent
excitatory circuits in the CA3 area (Traub
and Wong, 1982, 1983; Miles and Wong, 1987;
Christian and Dudek, 1988), several laboratories
have shown that reduction of GABA-A mediated
inhibition or an increase in extracellular potassium
has an unmasking effect on the new local recurrent
excitatory circuits in the dentate gyrus. The un-
derlying concept is that GABA-A mediated inhi-
bition, along with the membrane properties of
the granule cells (e.g., the highly negative resting
potentials, high threshold, and resistance to repet-
itive firing), will tend to impede the likelihood of
multisynaptic interactions, which is the prime
mechanism by which increased recurrent excita-
tion would be expected to lead to seizure activity.
This concept may provide a hypothetical explana-
tion for how seizures might not occur under some
circumstances, but would emerge under others;
that is, seizures would be generated and/or
768
propagated when synaptic inhibition is slightly
and transiently depressed and/or extracellular
potassium increased in concentration, as this par-
ticular set of changes would allow multisynaptic
interactions to generate a seizure.
Increased recurrent excitation versus increased
inhibition as a consequence of mossy fiber sprouting
Although most experiments have focused on and
supported the hypothesis that mossy fiber sprout-
ing during chronic epileptogenesis leads to in-
creased recurrent excitation among granule cells,
some data suggest that the sprouted mossy fibers
preferentially synapse on interneurons (e.g., Slovi-
ter, 1992; Kotti et al., 1997; Harvey and Sloviter,
2005). The hypothesis is that the ‘‘dormant’’ bas-
ket cells, which have presumably lost excitatory
synaptic input during repetitive seizures, become
re-innervated by the sprouted mossy fibers. The
histopathological data include light micrographs
of basket cells surrounded by Timm stain, but as
pointed out previously by Ribak and Peterson
(1991), similar micrographs can be obtained from
the temporal pole of the hippocampus of normal
animals, and the light-microscopic techniques do
not show whether mossy fibers near interneurons
actually synapse with them. Ultrastructural obser-
vations have reported mossy fiber synapses on in-
terneurons (Cavazos et al., 2003; see Sutula and
Dudek, this volume), but the number of synaptic
contacts from mossy fibers to interneurons has not
been rigorously evaluated and there is no quanti-
tative evidence that there is an increase over nor-
mal levels. Indeed, contacts by sprouted mossy
fibers on interneurons appear to be infrequent if
not rare (see Sutula and Dudek, this volume).
Most of the electrophysiological evidence for
the ‘‘hyper-inhibition’’ hypothesis is essentially
based on variations of the paired-pulse technique
(Sloviter, 1992; Buckmaster and Dudek, 1997;
Wilson et al., 1998; Harvey and Sloviter, 2005),
which is potentially confounded by technical issues
(e.g., see Waldbaum and Dudek, 2005, 2006).
Thus, although mossy fiber sprouting in the dent-
ate gyrus- and sprouting-based synaptic
reorganization in other areas of the hippocampus
and temporal lobe may involve new inhibitory cir-
cuits from sprouted mossy fibers to interneurons,
the preponderance of evidence from many labora-
tories with a wide range of techniques using several
different animal models point to the conclusion
that mossy fiber sprouting leads predominantly to
new recurrent excitatory circuits.
Mossy fiber sprouting and increased seizure
susceptibility
Many studies have noted the close association
between the presence of mossy fiber sprouting and
decreased seizure threshold (i.e., in the kindling
model) or the presence of spontaneous recurrent
seizures (i.e., in the models of epileptogenesis based
on SE). Nonetheless, many tissue specimens from
patients with intractable TLE lack sprouting, and
seizure frequency does not appear to be related to
the degree of mossy fiber sprouting in animal mod-
els (e.g., Buckmaster and Dudek, 1997). This issue
is complex, however, because the distribution of
mossy fiber sprouting is not homogeneous (e.g.,
more robust sprouting generally occurs in the tem-
poral versus septal hippocampus in animal mod-
els). Plus, axonal sprouting and the formation of
new recurrent excitatory circuits likely occur in
many other structures in addition to the dentate
gyrus, and synaptic reorganization in these other
areas is difficult to detect. Gorter et al. (2001) re-
ported that rats subjected to amygdala-stimulated
SE could be separated into two groups: (1) animals
that showed a progressive increase in seizure fre-
quency over time and that had robust mossy fiber
sprouting, and (2) rats with a low seizure frequency
and little or no mossy fiber sprouting. Longo and
Mello (1997) have reported that pretreatment with
the protein-synthesis inhibitor cycloheximide
blocks mossy fiber sprouting but does not prevent
the development of spontaneous recurrent seizures.
Williams et al. (2002), however, found that cyclo-
heximide pretreatment had no detectable effect on
mossy fiber sprouting. In this study, some rats did
not experience SE (i.e., had few if any convulsive
seizures during pilocarpine treatment), but still had

at least a few subsequent spontaneous seizures with
little or no Timm stain in the inner molecular layer,
consistent with the data of Gorter et al. (2001)
suggesting that mossy fiber sprouting is not neces-
sary for the development of occasional spontane-
ous seizures. It should not be surprising that the
dentate gyrus and mossy fiber are not required for
limbic seizures (for more extensive discussion on
this point, see Sutula and Dudek, this volume).
More recently, Toyoda and Buckmaster (2005)
also attempted to determine if cycloheximide
blocked mossy fiber sprouting after pilocarpine-in-
duced SE; they infused cycloheximide directly into
the area around the hippocampus for prolonged
periods from an osmotic minipump, provided
evidence that the cycloheximide was present in
the area around the dentate gyrus, and also ob-
served robust Timm stain in the inner molecular
layer (i.e., cycloheximide had no detectable effect
on mossy fiber sprouting). It is not clear how one
pretreatment injection of cycloheximide would be
expected to block the weeks/months-long continu-
ous process that is responsible for sprouting of the
mossy fibers and the formation of new synaptic
contacts with dentate granule cells. Presently,
therefore, the proposal that cycloheximide blocks
mossy fiber sprouting is problematic. The obser-
vation of spontaneous seizures before little or any
mossy fiber sprouting has occurred (i.e., between 5
and 7 days after kainate-induced SE in the study of
Hellier et al., 1999), and other data where animals
have shown spontaneous seizures with little or no
evidence of mossy fiber sprouting weeks after pi-
locarpine treatment (Williams et al., 2002), suggest
that other mechanisms (e.g., loss of interneurons)
besides mossy fiber sprouting in the dentate gyrus
can lead to the development of spontaneous
seizures after experimental SE. Although axonal
sprouting after neuronal injury in other brain areas
could account for the occurrence of spontaneous
seizures, particularly in the cases of spontaneous
seizures within a few weeks after SE, a more likely
explanation is that the death of vulnerable inter-
neurons during the SE (e.g., see Houser and
Esclapez, 1996) decreased seizure threshold (i.e.,
increased seizure susceptibility) and led to occa-
sional seizures.
769
Does mossy fiber sprouting inhibit seizures?
After spontaneous seizures in pilocarpine-treated
rats, c-fos expression has been reported to occur
preferentially in interneurons versus dentate
granule cells (Harvey and Sloviter, 2005), which
has been interpreted to mean that mossy fiber
sprouting leads to new synapses onto inhibitory
interneurons rather than to granule cells (i.e.,
hyper-inhibition), and that mossy fiber sprouting
suppresses rather than promotes seizures. Harvey
and Sloviter (2005) evaluated c-fos staining 1 h
after a spontaneous seizure, while in similar stud-
ies, Peng and Houser (2005) euthanized animals
15 min after a seizure and found Fos expression
preferentially in granule cells. Peng and Houser
(2005) found that staining was observed in inter-
neurons when animals were euthanized after 1 h,
the Fos staining in interneurons persisted longer
than in granule cells, and the enhanced Fos ex-
pression in the granule cells was not observed
when a second seizure occurred shortly after a
previous seizure. The results of Harvey and Slovi-
ter (2005) could have arisen if they studied animals
that had had more than one seizure in close
temporal proximity. In addition, progressive acti-
vation of granule cells and interneurons during
spontaneous seizures would be expected to occur
over hundreds of milliseconds, and seizures typi-
cally last up to 3 min; therefore, c-fos expression
lacks the temporal resolution to differentiate
whether granule cells or interneurons are sequenti-
ally or preferentially activated during spontaneous
seizures. Harvey and Sloviter (2005) also proposed
that population spikes are highly prevalent in the
dentate gyrus during spontaneous seizures that
occur a few days after SE, but are rare or nonex-
istent when spontaneous seizures occur weeks
later, as might be expected if inhibition increased
in the dentate gyrus over time after SE. Alternative
mechanisms besides increased inhibition are
possible, including depolarization inactivation, as
occurs during paroxysmal depolarization shifts af-
ter blockade of GABA-A receptors. Thus, it is not
clear whether these data apply to the question of
whether mossy fiber sprouting leads to hyper-in-
hibition and is potentially compensatory.
770
Reorganization of GABAergic neurons
Another hypothesis, based on increased
immunohistochemical staining of GABAergic ter-
minals in human tissue and in animal models of
TLE is whether other remaining GABAergic ter-
minals might sprout new synaptic connections to
partly (or totally) restore GABA-mediated inhibi-
tion (e.g., see Davenport et al., 1990; Bausch,
2005). If this occurred, one would expect an in-
crease in the frequency of mIPSCs with time after
experimental SE, but this was not seen in rats with
kainate-induced epilepsy (Shao and Dudek, 2005);
however, other techniques may reveal axon
sprouting, synaptic reorganization and recovery
of inhibition during epileptogenesis.
The dentate gyrus as a model for studies of
epileptogenesis associated with TLE
Although the dentate gyrus and the hilus clearly
have an important and probably have a unique
role in hipocampal integration and epileptogene-
sis, they may also serve as a model for other
cortical structures where epileptogenesis involves
neuronal death and synaptic reorganization in ad-
dition to possible alterations in neurotransmitter
receptors and voltage-gated channels. Although
epilepsy is caused by a great variety of age-depend-
ent etiologies and the mechanisms appear to be
equally heterogeneous, the cellular and network
mechanisms of (1) a loss of vulnerable interneu-
rons, and (2) axon sprouting with progressive for-
mation of new recurrent excitatory circuits have
the most support from independent experiments
in numerous laboratories and multiple animal
models. Although other molecular, cellular and
network mechanisms have been implicated in the
dentate gyrus, these two cellular mechanisms that
have been extensively documented in the dentate
gyrus may be generally applicable in many other
temporal and neocortical structures. Precisely how
these mechanisms and others in the dentate
gyrus and elsewhere lead to the epileptic seizures
associated with TLE will require further studies.
While the heterogeneity of TLE etiologies and the
mechanisms underlying neuronal and network
synchronization demand continuing attention to
a diverse range of potential mechanisms and in-
teractions spanning molecular, cellular, and sys-
tems levels, the insights gained from study of
neuronal loss and mossy fiber sprouting during the
last two decades have been a substantial advance
and a major accomplishment for epilepsy research.
References
Ang, C.W., Carlson, G.C. and Coulter, D.A. (2006) Massive
and specific dysregulation of direct cortical input to the hip-
pocampus in temporal lobe epilepsy. J. Neurosci., 26:
11850–11856.
Babb, T.L., Kupfer, W.R., Pretorius, J.K., Crandall, P.H. and
Levesque, M.F. (1991) Synaptic reorganization by mossy
fibers in human epileptic fascia dentata. Neuroscience, 42:
351–363.
Bausch, S.B. (2005) Axonal sprouting of GABAergic interneu-
rons in temporal lobe epilepsy. Epilepsy Behav., 7: 390–400.
Behr, J., Lyson, K.J. and Mody, I. (1998) Enhanced propaga-
tion of epileptiform activity through the kindled dentate
gyrus. J. Neurophysiol., 79: 1726–1732.
Bekenstein, J.W. and Lothman, E.W. (1993) Dormancy of in-
hibitory interneurons in a model of temporal lobe epilepsy.
Science, 259: 97–100.
Ben-Ari, Y. (1985) Limbic seizure and brain damage produced
by kainic acid: mechanisms and relevance to human temporal
lobe epilepsy. Neuroscience, 14: 375–404.
Bengzon, J.K.Z., Elmér, E., Nanobashvili, A., Kokaia, M. and
Lindvall, O. (1997) Apoptosis and proliferation of dentate
gyrus neurons after single and intermittent limbic seizures.
Proc. Natl. Acad. Sci., 94: 10432–10437.
Bernard, C., Esclapez, M., Hirsch, J.C. and Bernard, C. (1998)
Interneurones are not so dormant in temporal lobe epilepsy:
a critical reappraisal of the dormant basket cell hypothesis.
Epilepsy Res., 32: 93–103.
Bertram, E.H. (1997) Functional anatomy of spontaneous sei-
zures in a rat model of limbic epilepsy. Epilepsia, 38: 95–105.
Bertram, E.H., Zhang, D.X., Mangan, P., Fountain, N. and
Rempe, D. (1998) Functional anatomy of limbic epilepsy: a
proposal for central synchronization of a diffusely hyperex-
citable network. Epilepsy Res., 32: 194–205.
Bragin, A., Mody, I., Wilson, C.L. and Engel, J. (2002) Local
generation of fast ripples in epileptic brain. J. Neurosci., 22:
2012–2021.
Buckmaster, P.S. and Dudek, F.E. (1997) Neuron loss, granule
cell axon reorganization, and functional changes in the dent-
ate gyrus of epileptic kainate-treated rats. J. Comp. Neurol.,
385: 385–404.
Buckmaster, P.S., Zhang, G.F. and Yamawaki, R. (2002) Axon
sprouting in a model of temporal lobe epilepsy creates a pre-
dominantly excitatory feedback circuit. J. Neurosci., 22:
6650–6658.

Cascino, G.D. (2005) Temporal lobe epilepsy: more than hip-
pocampal pathology. Epilepsy Curr., 5: 187–189.
Cavazos, J.E., Das, I. and Sutula, T.P. (1994) Neuronal loss
induced in limbic pathways by kindling: evidence for induc-
tion of hippocampal sclerosis by repeated brief seizures. J.
Neurosci., 14: 3106–3121.
Cavazos, J.E., Golarai, G. and Sutula, T.P. (1991) Mossy fiber
synaptic reorganization induced by kindling: time course of
development, progression, and permanence. J. Neurosci., 11:
2795–2803.
Cavazos, J.E., Zhang, P., Qazi, R. and Sutula, T.P. (2003)
Ultrastructural features of sprouted mossy fiber synapses in
kindled and kainic acid-treated rats. J. Comp. Neurol., 458:
272–292.
Christian, E.P. and Dudek, F.E. (1988) Characteristics of local
excitatory circuits studied with glutamate microapplication in
the CA3 area of rat hippocampal slices. J. Neurophysiol., 59:
90–109.
Cohen, I., Navarrao, V., Clemenceau, S., Baulac, M. and Miles,
R. (1998) On the origin of interictal activity in human tem-
poral lobe epilepsy in vitro. Science, 298: 1418–1421.
Cossart, R., Dinocourt, C., Hirsch, J.C., Merchan-Perez, A.,
De Felipe, J., Ben-Air, Y., Esclapez, M. and Bernard, C.
(2001) Dendritic but not somatic GABAergic inhibition
is decreased in experimental epilepsy. Nat. Neurosci., 4:
52–62.
Coulter, D.A. and Carlson, G.C. (this volume) Functional reg-
ulation of the dentate gyrus by GABA-mediated inhibition.
In: Scharfman, H. (Ed.), The Dentate Gyrus. Progress in
Brain Research. Elsevier, Amsterdam.
Crain, B.J., Westerkam, W.D., Harrison, A.H. and Nadler, J.V.
(1988) Selective neuronal death after transient forebrain is-
chemia in the Mongolian gerbil: a silver impregnation study.
Neuroscience, 27: 387–402.
Cronin, J., Obenaus, A., Houser, C.R. and Dudek, F.E. (1992)
Electrophysiology of dentate granule cells after kainate-in-
duced synaptic reorganization of the mossy fibers. Brain
Res., 573: 305–310.
Davenport, D.J., Brown, W.J. and Babb, T.L. (1990) Sprouting
of GABAergic and mossy fiber axons in dentate gyrus fol-
lowing intrahippocampal kainate in the rat. Exp. Neurol.,
109: 180–190.
Doherty, J. and Dingledine, R. (2001) Reduced excitatory drive
onto interneurons in the dentate gyrus after status epileptic-
us. J. Neurosci., 21: 2048–2057.
Engel, J. (1989) Seizures and Epilepsy (Contemporary Neurol-
ogy Series). F.A. Davis Company, Philadelphia, PA.
Esclapez, M., Hirsch, J.C., Ben Ari, Y. and Bernard, C. (1999)
Newly formed excitatory pathways provide a substrate for
hyperexcitability in experimental temporal lobe epilepsy. J.
Comp. Neurol., 408: 449–460.
Esclapez, M., Hirsch, J.C., Khazipov, R., Ben Ari, Y. and
Bernard, C. (1997) Operative GABAergic inhibition in hip-
pocampal CA1 pyramidal neurons in experimental epilepsy.
Proc. Natl. Acad. Sci. U.S.A., 94: 12151–12156.
Fricke, R.A. and Prince, D.A. (1984) Electrophysiology of
dentate gyrus granule cells. J. Neurophysiol., 51: 195–209.
771
Gloor, P. (1992) Role of the amygdala in temporal lobe epi-
lepsy. In: Aggleton J.P. (Ed.), The Amygdala: Neurobiolog-
ical Aspects of Emotion, Memory, and Mental Dysfunction.
Wiley-Liss, Inc., New York, pp. 505–538.
Gorter, J.A., van Vliet, E.A., Aronica, E. and Lopes da Silva,
F.H. (2001) Progression of spontaneous seizures after status
epilepticus is associated with mossy fibre sprouting and ex-
tensive bilateral loss of hilar parvalbumin and somatostatin-
immunoreactive neurons. Eur. J. Neurosci., 13: 657–669.
Hardison, J.L., Okazaki, M.M. and Nadler, J.V. (2000) Modest
increase in extracellular potassium unmasks effect of recur-
rent mossy fiber growth. J. Neurophysiol., 84: 2380–2389.
Harvey, B.D. and Sloviter, R.S. (2005) Hippocampal granule
cell activity and c-Fos expression during spontaneous sei-
zures in awake, chronically epileptic, pilocarpine-treated rats:
Implications for hippocampal epileptogenesis. J. Comp. Ne-
urol., 488: 442–463.
Heinemann, U., Beck, H., Dreier, J.P., Ficker, E., Stabel, J. and
Zhang, C.L. (1992) The dentate gyrus as a regulated gate
for the propagation of epileptiform activity. Epilepsy Res.
Suppl., 7: 273–280.
Hellier, J.L., Patrylo, P.R., Dou, P., Nett, M., Rose, G.M. and
Dudek, F.E. (1999) Assessment of inhibition and epilepti-
form activity in the septal dentate gyrus of freely behaving
rats during the first week after kainate treatment. J. Neuro-
sci., 19: 10053–10064.
Houser, C.R. (this volume) Interneurons of the dentate gyrus:
An overview of cell types, terminal fields and neurochemical
identity. In: Scharfman, H. (Ed.), The Dentate Gyrus.
Progress in Brain Research. Elsevier, Amsterdam.
Houser, C.R. and Esclapez, M. (1996) Vulnerability and plas-
ticity of the GABA system in the pilocarpine model of spon-
taneous recurrent seizures. Epilepsy Res., 26: 207–218.
Houser, C.R., Miyashiro, J.E., Swartz, B.E., Walsh, G.O.,
Rich, J.R. and Delgado-Escueta, A.V. (1990) Altered pat-
terns of dynorphin immunoreactivity suggest mossy fiber
reorganization in human hippocampal epilepsy. J. Neurosci.,
10: 267–282.
Hsu, M. and Buzsaki, G. (1993) Vulnerability of mossy fiber
targets in the rat hippocampus to forebrain ischemia. J. Ne-
urosci., 13: 3964–3979.
Iijima, T., Witter, M.P., Ichikawa, M., Tominage, T., Kajiwara,
R. and Matsumoto, G. (1996) Entorhinal-hippocampal in-
teractions revealed by real-time imaging. Science, 272:
1176–1179.
Kapur, J., Haas, K.F. and Macdonald, R.L. (1999) Physiolog-
ical properties of GABAA receptors from acutely dissociated
rat dentate granule cells. J. Neurophysiol., 81: 2464–2471.
Kesner, R.P. (this volume) A behavioral analysis of dentate
gyrus function. In: Scharfman, H. (Ed.), The Dentate Gyrus.
Progress in Brain Research. Elsevier, Amsterdam.
Kobayashi, M. and Buckmaster, P.S. (2003) Reduced inhibition
of dentate granule cells in a model of temporal lobe epilepsy.
J. Neurosci., 23: 2440–2452.
Kotti, T., Riekkinen, P.J. and Miettinen, R. (1997) Character-
ization of target cells for aberrant mossy fiber collaterals in
the dentate gyrus of epileptic rat. Exp. Neurol., 146: 323–330.
772
de Lanerolle, N.C., Kim, J.H., Robbins, R.J. and Spencer,
D.D. (1989) Hippocampal interneuron loss and plasticity in
human temporal lobe epilepsy. Brain Res., 495: 387–395.
Longo, B.M. and Mello, L.E. (1997) Blockade of pilocarpine-
or kainate-induced mossy fiber sprouting by cycloheximide
does not prevent subsequent epileptogenesis in rats. Neuro-
sci. Lett., 226: 163–166.
Lothman, E.W., Bertram, E.H., Kapur, J. and Stringer, J.L.
(1990) Recurrent spontaneous hippocampal seizures in the
rat as a chronic sequela to limbic status epilepticus. Epilepsy
Res., 6: 110–118.
Lothman, E.W., Stringer, J.L. and Bertram, E.H. (1992) The
dentate gyrus as a control point for seizures in the hippo-
campus and beyond. Epilepsy Res. Suppl., 7: 301–313.
Lowenstein, D.H., Thomas, M.J., Smith, D.H. and McIntosh,
T.K. (1992) Selective vulnerability of dentate hilar neurons
following traumatic brain injury: a potential mechanistic link
between head trauma and disorders of the hippocampus. J.
Neurosci., 12: 4846–4853.
Lynch, M. and Sutula, T. (2000) Recurrent excitatory connec-
tivity in the dentate gyrus of kindled and kainic acid-treated
rats. J. Neurophysiol., 83: 693–704.
Margerison, J.H. and Corsellis, J.A. (1966) Epilepsy and the
temporal lobes. A clinical, electroencephalographic and ne-
uropathological study of the brain in epilepsy, with particular
reference to the temporal lobes. Brain, 89: 499–530.
Meier, C.L. and Dudek, F.E. (1996) Spontaneous and stimu-
lation-induced synchronized burst afterdischarges in the iso-
lated CA1 of kainate-treated rats. J. Neurophysiol., 76:
2231–2239.
Meldrum, B.S. and Bruton, C.J. (1992) Epilepsy. In: Adams
J.H. and Duchen L.W. (Eds.), Greenfield’s Neuropathology.
Oxford, New York, pp. 1246–1283.
Miles, R. and Wong, R.K. (1987) Inhibitory control of local
excitatory circuits in the guinea-pig hippocampus. J. Physiol.,
388: 611–629.
Molnar, P. and Nadler, J.V. (1999) Mossy fiber-granule cell
synapses in the normal and epileptic rat dentate gyrus studied
with minimal laser photostimulation. J. Neurophysiol., 82:
1883–1894.
Nadler, J.V. (1991) Kainic acid as a tool for the study of tem-
poral lobe epilepsy. Life Sci., 29: 2031–2042.
Obenaus, A., Esclapez, M. and Houser, C.R. (1993) Loss of
glutamate decarboxylase mRNA-containing neurons in the
rat dentate gyrus following pilocarpine-induced seizures. J.
Neurosci., 13: 4470–4485.
Parent, J.M. (this volume) Adult neurogenesis in the intact and
epileptic dentate gyrus. In: Scharfman, H. (Ed.), The Dentate
Gyrus. Progress in Brain Research. Elsevier, Amsterdam.
Patrylo, P.R. and Dudek, F.E. (1998) Physiological unmasking
of new glutamatergic pathways in the dentate gyrus of hip-
pocampal slices from kainate-induced epileptic rats. J. Ne-
urophysiol., 79: 418–429.
Peng, Z. and Houser, C.R. (2005) Temporal patterns of Fos
expression in the dentate gyrus after spontaneous seizures in
a mouse model of temporal lobe epilepsy. Epilepsy Curr., 6:
57–58.
Perez, Y., Morin, F., Beaulieu, C. and Lacaille, J.C. (1996)
Axonal sprouting of CA1 pyramidal cells in hyperexcitable
hippocampal slices of kainate-treated rats. Eur. J. Neurosci.,
8: 736–748.
Ribak, C.E. and Peterson, G.M. (1991) Intragranular mossy
fibers in rats and gerbils form synapses with the somata
and proximal dendrites of basket cells in the dentate gyrus.
Hippocampus, 1: 355–364.
Sayin, U., Osting, S., Hagen, J. and Rutecki, P.S.T. (2003)
Spontaneous seizures and loss of axo-axonic and axo-somatic
inhibition induced by repeated brief seizures in kindled rats.
J. Neurosci., 23: 2759–2768.
Sayin, U., Rutecki, P. and Sutula, T. (1999) Alterations in
nmda dependent currents in granule cells of the dentate gyrus
contribute to the induction but not the permanence of
kindling. J. Neurophysiol., 81: 564–574.
Scharfman, H.E., Sollas, A.L., Berger, R.E. and Goodman,
J.H. (2003) Electrophysiological evidence of monosynaptic
excitatory transmission between granule cells after seizure-
induced mossy fiber sprouting. J. Neurophysiol., 90:
2536–2547.
Shao, L.R. and Dudek, F.E. (2004) Increased excitatory synap-
tic activity and local connectivity of hippocampal CA1
pyramidal cells in rats with kainate-induced epilepsy. J.
Neurophysiol., 92: 1366–1373.
Shao, L.-R. and Dudek, F.E. (2005) Changes in mIPSCs and
sIPSCs after kainate treatment: evidence for loss of inhibitory
input to dentate granule cells and possible compensatory re-
sponses. J. Neurophysiol., 94: 952–960.
Sloviter, R.S. (1987) Decreased hippocampal inhibition and
a selective loss of interneurons in experimental epilepsy.
Science, 235: 73–76.
Sloviter, R.S. (1991) Permanently altered hippocampal struc-
ture, excitability, and inhibition after experimental status
epilepticus in the rat: the ‘‘dormant basket cell’’ hypothesis
and its possible relevance to temporal lobe epilepsy. Hippo-
campus, 1: 41–66.
Sloviter, R.S. (1992) Possible functional consequences of synap-
tic reorganization in the dentate gyrus of kainate-treated rats.
Neurosci. Lett., 137: 91–96.
Sloviter, R.S., Zappone, C.A., Harvey, B.D., Bumanglag, A.V.,
Bender, R.A. and Frotscher, M. (2003) ‘‘Dormant basket cell’’
hypothesis revisited: relative vulnerabilities of dentate gyrus
mossy cells and inhibitory interneurons after hippocampal
status epilepticus in the rat. J. Comp. Neurol., 459: 44–76.
Smith, B.N. and Dudek, F.E. (2001) Short- and long-term
changes in CA1 network excitability after kainate treatment
in rats. J. Neurophysiol., 85: 1–9.
Smith, B.N. and Dudek, F.E. (2002) Network interactions me-
diated by new excitatory connections between CA1 pyram-
idal cells in rats with kainate-induced epilepsy. J.
Neurophysiol., 87: 1655–1658.
Stables, J.P., Bertram, E., Dudek, F.E., Holmes, G., Mathern,
G., Pitkanen, A. and White, H.S. (2003) Therapy discovery
for pharmacoresistant epilepsy and for disease-modifying
therapeutics: summary of the NIH/NINDS/AES Models II
Workshop. Epilepsia, 44: 1472–1478.

Staley, K.J., Otis, T.S. and Mody, I. (1992) Membrane prop-
erties of dentate gyrus granule cells: comparison of sharp
microelectrode and while-cell recordings. J. Neurophysiol.,
67: 1346–1358.
Stringer, J.L. and Lothman, E.W. (1992) Bilateral maxi-
mal dentate activation is critical for the appearance of an
afterdischarge in the dentate gyrus. Neuroscience, 46:
309–314.
Sutula, T., Cascino, G., Cavazos, J., Parada, I. and Ramirez, L.
(1989) Mossy fiber synaptic reorganization in the epileptic
human temporal lobe. Ann. Neurol., 26: 321–330.
Sutula, T.P. and Dudek, F.E. (this volume) Unmasking recur-
rent excitation generated by mossy fiber sprouting in the ep-
ileptic dentate gyrus: an emergent property of a complex
system. In: Scharfman, H. (Ed.), The Dentate Gyrus.
Progress in Brain Research. Elsevier, Amsterdam.
Sutula, T.P. and Steward, O. (1986) Quantitative analysis of
synaptic potentiation during kindling of the perforant path.
J. Neurophysiol., 56: 732–746.
Sutula, T.P. and Steward, O. (1987) Facilitation of kindling by
prior induction of long-term potentiation in the perforant
pathway. Brain Res., 420: 109–117.
Tauck, D.L. and Nadler, J.V. (1985) Evidence of functional
mossy fiber sprouting in hippocampal formation of kainic
acid-treated rats. J. Neurosci., 5: 1016–1022.
Toth, Z., Hollrigel, G.S., Gorcs, T. and Solesz, I. (1997) In-
stantaneous perturbation of dentate interneuronal networks
by a pressure wave-transient delivered to the neocortex. J.
Neurosci., 17: 8106–8117.
Toyoda, I. and Buckmaster, P.S. (2005) Prolonged infu-
sion of cycloheximide does not block mossy fiber sprouting
in a model of temporal lobe epilepsy. Epilepsia, 46:
1017–1020.
Traub, R.D. and Wong, R.K. (1982) Cellular mechanism of
neuronal synchronization in epilepsy. Science, 216: 745–747.
Traub, R.D. and Wong, R.K. (1983) Synchronized burst dis-
charge in disinhibited hippocampal slice. II. Model of cellular
mechanism. J. Neurophysiol., 49: 459–471.
Turski, W.A., Cavalheiro, E.A., Schwarz, M., Czuczwar, S.J.,
Kleinrok, Z. and Turski, L. (1983) Limbic seizures produced
by pilocarpine in rats: behavioural, electroencephalographic
and neuropathological study. Behav. Brain Res., 9: 315–335
773
In: H. Scharfman (Ed.), The Dentate Gyrus. Progress in
Brain Research. Elsevier, Amsterdam.
Waldbaum, S. and Dudek, F. (2005) Parametric assessment of
the consistency of the paired-pulse protocol for studying
hippocampal inhibition and hyperexcitability. Soc. Neurosci.
Abstr., 31: 276.12.
Waldbaum, S. and Dudek, F.E. (2006) Increased paired-pulse
suppression after low-dose application of the selective
GABAA receptor antagonist SR-95531, GABAzine: The
paired-pulse technique can provide misleading evidence for
‘‘hyperinhibition’’ when GABAA-mediated inhibition is de-
creased. Soc. Neurosci. Abstr., 32: 794.18.
Wierenga, C.J. and Wadman, W.J. (1999) Miniature inhibitory
postsynaptic currents in CA1 pyramidal neurons after kin-
dling epileptogenesis. J. Neurophysiol., 82: 1352–1362.
Williams, P.A., Pou, P. and Dudek, F.E. (2004) Epilepsy and
synaptic reorganization in a model of perinatal hypoxia-is-
chemia. Epilepsia, 45: 1210–1218.
Williams, P.A., Wuarin, J.-P., Dou, P., Ferraro, D.J. and
Dudek, F.E. (2002) Reassessment of the effects of cyclohexi-
mide on mossy fiber sprouting and epileptogenesis in the pi-
locarpine model of temporal lobe epilepsy. J. Neurophysiol.,
88: 2075–2087.
Williamson, A. and Patrylo, P.R. (this volume) Physiological
studies of human dentate granule cells. In: Scharfman, H.
(Ed.), The Dentate Gyrus. Progress in Brain Research. Else-
vier, Amsterdam.
Wilson, D.L., Kahn, S.U., Engel, J., Isokawa, M., Babb, T.L.
and Behnke, E.J. (1998) Paired pulse suppression and
facilitation in human epileptogenic hippocampal formation.
Epilepsy Res., 31: 211–230.
Wuarin, J.P. and Dudek, F.E. (1996) Electrographic seizures
and new recurrent excitatory circuits in the dentate gyrus of
hippocampal slices from kainate-treated epileptic rats. J. Ne-
urosci., 16: 4438–4448.
Wuarin, J.P. and Dudek, F.E. (2001) Excitatory synaptic input
to granule cells increases with time after kainate treatment. J.
Neurophysiol., 85: 1067–1077.
Zhang, N. and Houser, C.R. (1999) Ultrastructural localization
of dynorphin in the dentate gyrus in human temporal lobe
epilepsy: a study of reorganized mossy fiber synapses. J.
Comp. Neurol., 405: 472–490.
 Subject Index

a-Adrenoceptors 302 Seizure susceptibility 689–690

a-Amino-3-hydroxy-5- methylisoxazole-4-
propionic acid (AMPA) 401
Aberrant neurogenesis hypothesis 536
Acetylcholine 63–78, 461
Acetylcholinesterase (AChE)
Reorganization after entorhinal denervation
509–510
Activity-regulated cytoskeleton (Arc) 671
Arc-associated protein (Arc) as synapse
consolidation mediator 453, 455–457
Adrenalectomy (ADX) 355, 358–361
Adrenoceptors 300–302
a-Adrenoceptors 302
b-Adrenoceptors 301
Adult neurogenesis in the intact and epileptic DG
529–537
Hippocampal neurogenesis
Modulation 532
Seizure-induced 533–534
Integration and function of adult born DGCs
530–531
Neonatal and adult dentate granule cell
neurogenesis 530
Regulation 531–533
Afferent and efferent connections 17, 32–34
Afferent fiber lamination formation 136–139
Cholinergic and GABAergic afferents from
medial septum 138–139
Commissural/associational (C/A) fibers
137–138
Entorhinal afferents 136–137
Non-hippocampal afferent connections 33
Afterhyperpolarization (AHP) 186
Aging effects
GABAergic circuitry 682–687
Glutamatergic circuitry 687–689
on dentate circuitry and function 679–690
on dentate filter function 679–681
Spatial learning and memory 689
Allocentric coordinate 620
Allopregnanolone 405
Alzheimer’s disease (AD) 463–464
Dentate gyrus 723–734
Disconnected DG 732–734
Excitatory–inhibitory systems 731–732
Histopathology 724–726
Entorhinal cortex atrophy 743–745
Hippocampal atrophy 741–751
MRI 741–743
Neuropeptide Y (NPY) and its receptors in
292–293
Somatostatin (SST) levels in 268–269
b-Amyloid degradation by 268
White matter changes 748–750
Amygdalo-entorhinal cortex (AE) 46
Amyloid b deposition 725–726
Androgen receptors (AR) 399, 407
Animal models of epilepsy 139, 183–195
Antidepressants (ADs)
Neurogenic effects, effects on neurogenesis
707–708
Serotonin-dependence 708–710
Aplysia 309
Apolipoprotein E receptor 2 (ApoER2) 135
Apolipoprotein E receptor 3 (ApoER3) 148
Apoptosis 355, 357, 362
Chronic stress and 364
Arachidonoylethanolamide (AEA) 321,
331
Archicortex 23
Astrocytes 513
Astrocytic b-adrenoceptors 301
Astroglial cells 404, 513–514
Autoassociative memory function 579
Autoradiograms 671
Axo-axonic cells 644–645

775
776
Axon sprouting 542
Axonal degeneration techniques 44
Axonal reorganization, after entorhinal
denervation 505–510
Basal forebrain inputs 18–19
Basket cells 155, 157–159, 644, 762–763, 762–764
Basolateral amygdala 622
Behavioral analysis of dentate gyrus 567–575
Conjunctive encoding 567–568
Encoding vs. retrieval of spatial information
572–574
Spatial pattern separation 568–572
Biotinylated dextran amine (BDA) 69
Blood-derived cells 513
Braak-staging 725, 733–734
Brain-derived neurotrophic factor (BDNF) 115,
373–378, 400
Arc 456
Consolidation 454–456, 460–461
Effect on neurogenesis 378
Effects on learning and memory 378
Effects on synaptic transmission 377–378
Gene regulation 375–376
in Huntington’s disease 385
in neurodegenerative diseases 385
in neuropsychiatric disease 385–386
in pain 384–385
in Rett syndrome 385
Localization, transport and release 374–375
Physiologic regulation of 377
Role in GABAergic synapses 377
Roles during development 376–377
Seizure 376
Translation control in dendrites 457–460
Brain slice methods 111
Brain-specific Na+-dependent inorganic
phosphate cotransporter (BNPI) 75–76
BNPI/VGLUT1 75–76
DNPI/VGLUT2 76–77
VGLUT3 75, 77–78
Brainstem inputs 19
Brodman’s areas 28a and 28b 46
CA1 (regio superior) 4–5, 418–419
CA1 interneurons 207
CA2 4–5, 418–419
CA3 (regio inferior) 30, 123, 207, 418–419, 422
Anatomical connectivity from DG to 10,
109–111
CA3 backprojection to DG 627–636 (see also
separate entry)
Lesion-induced mossy fiber (MF) sprouting
99–100
Lesions 572–574
MF trajectory and termination in 86–97
Pyramidal cells 577–588
Role in spatial pattern separation 568–572
CA3 backprojection to DG 627–636
Associative networks and reverberatory circuits
635
Backprojection 629–633
Anatomical evidence 629–630
Physiological evidence 630–633
Functional implications 633–635
Hippocampal information processing 633–634
Pathological conditions 634–635
Ca2+ channels
Ca2+/calmodulin-dependent protein kinase
(CaMKII) 478
Ca2+/phospholipid-dependent protein kinase
(PKC) 477
in human DGCs 186
in mossy fiber (MF) transmission 112
Interleukin 1-b (IL-1-b) and 348–349
Cajal-Retzius cells 137, 140, 144
Calbindin 728–729
Calcitonin generelated peptide (CGRP)-
immunoreactive neurons 66, 323
Calretinin 226–227, 508
Labeling comparison in mouse and rat 227
cAMP response element binding protein (CREB)
267, 427
cAMP-dependent protein kinase (PKA) 478–479
cAMP-dependent signaling pathways 267
Carbachol (CCh) 329
CART-immunoreactive mossy cells 31–32
Catecholaminergic brainstem-dentate connections
71–74
Serotonergic afferents 71–73
CB1 receptors 319–321
Endogenous ligands
(arachidonoylethanolamide) 321
Expression and distribution 322, 324–326
Coexpression with other receptor subtypes
326

Immunohistochemical studies 323
Role in GABAergic neurons 322–324
CB2 receptors 319–321
C-fos protein 604, 769
cGMP-dependent protein kinase (PKG) 479
Chandelier cells 155, 163
Chemokines 515–516
Cholecystokinin (CCK) 30, 101, 115, 201, 225
Choline acetyltransferase (ChAT) 65–66,
138–139
Cholinergic neurons 18–19, 63–78, 133–140
Chronic stress effects in dentate gyrus 362–367
Code conversion 582–585
Commissural/association (C/A) fibers 63–78
Conjunctive encoding 567–568
Cornu ammonis 594, 724
Cortical hem
in DG and hippocampus development 144
Corticosteroids, in DG 355–367
Dentate function after chronic stress 362–367
Dentate function in the absence of 357–361
Cell turnover and morphology 357–359
Physiology 359–361
Functional relevance of corticosteroid effects in
DG 365–367
Dose dependence 365–366
in health and disease 366–367
Physiological variations in corticosteroid levels
361–362
Cell turnover 361
Physiology 361–362
Prolonged exposure, damage due to 355
Systems activated by stress 355–357
Corticotrophin 311
Cortistatin (CST) 265
Crossed entorhinodentate fibers 507, 510
Cycloheximide 769
Cytokines 339 (see also Pro-inflammatory
cytokines)
Dehydroepiandrosteronesulfate (DHEAS) 401
Dendritic changes, following entorhinal
denervation 514–516 (see also
Transneuronal changes)
Dense-core vesicles 251
Densitometry 671
Dentate gyrus (DG) (see also Hippocampal DG)
Aging-related changes in 683–685
777
Anatomy and architecture 3–22, 418–421
as a filter or gate 601–611
Dentate gate vs. filter 605–607
Dentate filter function 607–608
Behavioral analysis
CA3 backprojection to (see CA3)
Cell and fiber layers development in 133–140
Cortical signals transformation in
Extrinsic afferent systems to 63–78 (see also
Extrinsic afferent systems to DG)
Interneurons of 217–229
Modeling 639–655
Neurotrophins in 371–387
Norepinephrine and 299–312
Opioid systems in 245–258
Plastic processes in 417–442 (see also Plasticity)
Pro-inflammatory cytokines and their effects in
339–350 (see also Pro-inflammatory
cytokines)
Depolarization-induced suppression of inhibition
(DSI) 319, 327–329
Depression 464–465
NPY and its receptors in 293–294
Detonator synapses 111, 422, 433
Development of dentate gyrus 133–140 (see also
Reelin)
Afferent fiber 136–139
Functional considerations 139–140
Granule cell generation 134–136
Diacylglycerol lipase (DAGL) 321
Diffusion tensor imaging (DTI) 750
DiI-labeled granule neurons 170
Dopamine 63–78, 461
Dormant basket cell hypothesis of epilepsy 204
Dorsal raphe nucleus (DRN) 703
Dorsolateral (DLE) entorhinal cortex 46
Doublecortin (DCX) 157
Dual-labeling immunohistochemical techniques
323
Dynorphins 113, 245–258
Anatomical distribution of 249–250
Immunoreactivity 256
Location in hippocampal formation 248
Prodynorphin-derived opioids 246
a-Neo-endorphin 246
b-Neo-endorphin 246
Dynorphin A (1–8) 246
Dynorphin A (1–17) 246
778
Dynorphin B (1–13) 246
[Leu5]-enkephalin 246
Leumorphin 246
Electroconvulsive therapy (ECT) 293
Embryonic marginal zone 593
Emx2 gene 146
Endocannabinoids (ECs) in DG 319–332
CB1 receptors 320–321 (see also separate entry)
CB2 receptor 320–321
Markers of EC system 322–327
Modulating glutamatergic transmission in DG
329–330
Non-CB receptor targets of 330–331
Physiological role in dentate 327–331
Role in neurogenesis 330
Enkephalins 113, 245–258
Anatomical distribution of 247–249
Location in hippocampal formation 248
Proenkephalin-derived opioids 246
C-terminally extended forms of [Met5]-
enkephalin 246
[Leu5]-enkephalin 246
[Met5]-enkephalin 246
Entorhinal afferents 136–137
Entorhinal cortex 3, 45–46
Entorhinal cortex lesions (ECL) 503–519
Projection to DG 17–18
Subdivisions 46
Entorhinal denervation, structural reorganization
of DG after 501–519
Acetylcholinesterase (AChE) reorganization
509–510
Axonal reorganization 505–510
Commissural/associational (C/A) mossy cell
axons 505, 508
Crossed entorhinodentate fibers, reorganization
510
Dendritic changes 514–516 (see also
Transneuronal changes)
Fiber systems reorganization 505–510
GABAergic C/A projection 509
Glial changes 511–514
Immune response 518
Neuroanatomical species differences 503–505
Entorhinal-dentate projection 43–58 (see also
Perforant path)
Epilepsy 371 (see also Epileptic DG)
Adult neurogenesis in 529–537
Epilepsy models, neurotrophins roles in 379–381
NPY and its receptors in 292–293
Epileptic DG
Human dentate characteristics of 185
Mossy fiber sprouting in 541–558 (see also
Mossy fiber sprouting)
Epileptogenesis 650–652, 755–770
Definition 756–757
in DG 757–760
Filtering and gating properties 759–760
Propagation of seizures 760
TLE, animal models 760–762 (see also
Temporal lobe epilepsy)
Histopathological alterations 761–762
Kindling 761
Status epilepticus 761
Epileptic animals
Pharmacology and subunit expression
alterations 240
Synaptic GABAA receptor expression
Changes in 240
Upregulation in 240
Tonic GABAA current 241
Zinc sensitivity 240
Estradiol 402
17b-Estradiol 401, 405
Estradiol treatment 256
Estrogen receptor (ER) 399, 402
Effect on seizures 405
ERa distribution in DG 406–408
ERb distribution in DG 408
Subcellular localization 407
Estrous cycle 402
Eukaryotic elongation factor 2 (eEF2) 458–459
Experimental autoimmune encephalomyelitis
(EAE) 342
Extended preparation 44
Extracellular signal-regulated kinase (ERK) 480
Extracellular unit recording 304–305
Extrasynaptic GABAA receptors 235
Alterations in animal models of epilepsy 240
(see also Epileptic animals)
Altered zinc sensitivity 240
Composition and function of 236
Expressed by dentate granule cells 236–237
Upregulation in granule cells of epileptic
animals 240
 779

Extrinsic afferent systems to DG 17–19, 63–78 GABAergic commissural projection 74–75

Afferent projections 17 GABAergic inhibition 548–551
Basal forebrain inputs 18–19 in human DGCs 190
BNPI/VGLUT1 75–76 GABAergic interneurons 14–16, 63–78, 164
Brainstem inputs 19 GABAergic neurons 322–324, 582, 587, 770
CA1–CA3 subfields 64 Septohippocampal cholinergic systems and
Catecholaminergic brainstem-dentate 68–69
connections 71–74 Transporters, co-localization 118–119
Commissural connections 74–75 GABA-mediated inhibition (see also
DNPI/VGLUT2 76–77 Extrasynaptic GABAA receptors)
Entorhinal cortex projection 17–18 Functional regulation of DG 235–241
Glutamatergic innervation 75 Gatekeeper function 237–240
Noradrenergic and dopaminergic afferents Tonic and phasic inhibition in dentate granule
73–74 cells 237
Septo-hippocampal connections 65–69 Genetic regulation of DG development 143–150
Supramammillary input 19, 69–71 Cortical hem 144
VGLUT3 75, 77–78 Developmental signaling systems in adult DG
150
Fadrozole 405 Emx2 146
Fatty acid amide hydrolase (FAAH) 321 Lhx5 146
Feedforward interneurons 304–305 Mutants with defects in development 148
Field excitatory postsynaptic potential (fEPSP) Neurogenic niche 147–148
359–360 Transcription factors 144–146
Filopodia 167–180 Glia
Filtering function, dentate-hilar 607–608 Glia-guided neuronal migration 134
Fimbria stimulation 630 Role of glia in synaptic remodeling 404
Flinder’s Sensitive Line (FSL) 293 Glial changes, following entorhinal denervation
Fluoxetine 293, 708 511–514
Fornix 618 Astroglia 513–514
Fractional anisotropy (FA) 750 Blood-derived cells 513
Functional cellular connectivity 202–203 Microglia 511–513
Functional significance of laminated organization Oligodendroglia and NG2-positive cells 514
139–140 Glial fibrillary acid protein (GFAP) 35, 136, 404
Fusiform cells (spiny and aspiny) in hilus Global remapping 299, 308, 312, 592
155–164 Glucocorticoid receptor (GR) 355, 357
Glutamate 113
g-EEG 299–312 Glutamatergic innervation of DG 63–78
g-Frequency oscillation 207 Glutamatergic transmission in DG
g-Oscillations 621 Transport, co-localization 118–119
GABA (g-aminobutyric acid) neurons/GABAergic Glutamic acid decarboxylase (GAD) 65, 77,
neurons 217, 603–604 217–229, 641, 765
Axonal arborizations and postsynaptic targets Double labeling 220
of 221–223 GAD expression 116–117
Distribution of 218–221 GAD mRNA-labeled neurons distribution
GABAergic afferents from medial septum 218
138–139 GAD65 219
GABAergic C/A projection after entorhinal Glycogen metabolism 299–312
denervation 509 Gonadal hormones 256, 399
780

Granule cells, dentate (DGCs) 8–11, 643–644, Connectivity of 202–203

664–666 Firing patterns 205
Apical dendrites of 156 Functional identification and activity in vivo
Biophysical properties of 664–665 199–213
Cell layer 5 Slow oscillations of 210–213
Comparative anatomy 24–28 Hilar somatotatin (SST) interneurons 269–270
Co-release of glutamate and GABA 120–121 Excitotoxicity 273–274
Density, regulation 148–150 Hippocampal information processing 633–634
Description 155–157 Hippocampal interneurons 14
Glutamate as transmitter for 11 Hippocampal mossy fibers and terminals
Golgi-impregnated granule cells 27 Supra- and infrapyramidal bands of 87–90
Granule cell association hypothesis 203 Human DG 23–28
Granule cell generation 133–140 Granule cells, physiological studies of
Compact granule cell layer formation 183–195
134–136 Huntington’s disease 385
Sites of 134 Hypothalamic-pituitary-adenocortical (HPA)
in primates 26 system 705–707
Mossy cell 13–14
Mossy fibers 9–10 IL-1 receptor-associated kinase (IRAK) 346
NPY actions 289 Infrapyramidal 134, 87–90, 134
Pyramidal basket cell 11–13 Inhibitory postsynaptic potential (IPSP) 111
Single unit recordings 665–666 Interleukin 18 (IL-18) 349–350
Tonic and phasic inhibition 237 Interleukin 1-b (IL-1-b)
Granule neuron dendrites Action in DG 345–350
Development of 168–173 Ca2+ channels 348–349
Maturation 173–176 IL-1 signalling 346
Morphological development and maturation of Synaptic plasticity 346–348
167–180 Interneurons of DG 14, 110, 157, 217–229
Primary period of 167 Classification 217
Temporal and spatial gradients of 168 GABA as primary neurotransmitter 217–229
(see also GABA)
HCN current 186–187 GABAergic interneurons 14–16
Hebbian plasticity 434–435 Intracellular labeling of 217–229
Hebb-Williams maze 571–574 Neurochemical identity 223–228
Heterosynaptic LTD 437 Intracellular labeling techniques 174, 217–229
Hilar cells with axonal projections to the perforant
path (HIPP) cells 15–16, 222, 583–584, 645 k-Opioid receptors (KORs) 246–247
Hilar commissural-association pathway-related Immunoreactivity 254
(HICAP) cells 56, 164, 223, 645–646 Kainate model 405
Hilar interneurons 303–304 Kainic acid 118, 761, 405
NPY effects on 289–290 Lateral perforant path (LPP) 299–300, 307–308,
Hilar mossy cells 199–213, 630 426, 462–463, 493–494
Anatomy 200–202 Learning and memory, BDNF effect on 378
Basic properties 200 Lef1 mutants 147
b/g-Oscillations 208–210 Lesioning techniques 503
y-Oscillations 207 Lesion-induced sprouting 85–102
g-Frequency oscillation 207 Leumorphin 246
Cellular properties of 203–204 Lhx5 gene 146
 781

LIM domain kinase (LIMK) gene 455 Microglia 511–513

Locus coeruleus (LC) 300, 305–306
Activation 305
Glutamatergic activation of 304
Modulation of NE 306–311
Stimulation 304–307
Long-term depression (LTD) 417, 436–438, 453,
473–488
Associative LTD 437, 438
Depotentiation related to 437–438
Heterosynaptic LTD 429, 437, 493–494
Homosynaptic LTD 438
Induction, cellular mechanisms, comparison
473–494
in sparse coding 438
mGluR-induced LTD 489–491
Long-term potentiation (LTP) 245–258, 417,
423–436, 578, 590
Long-term memory (LTM) 616
Low-frequency stimulation (LFS) 111
LRP family (LRP6) 146
Lucifer yellow 733–734
Major depressive disorder (MDD) 697 (see also
Depression)
Genes, environment 703–707
5-HT system and hippocampal neurogenesis
703–705
Neurogenesis, stress effects on 705–707
Hippocampal dysfunction and atrophy 701–702
Mass-associated TLE (MaTLE) 184
Maximal dentate activation (MDA) 603
Medial (ME) entorhinal cortex 46, 591–592, 619
Medial perforant path (MPP) 299–301, 305,
307–308, 426
Medial septum/diagonal band of Broca (MSDB)
65
MSDB cholinergic innervation 65–67
MSDB GABAergic innervation 67–68
Medial temporal lobe sclerosis (MTS) 183–184,
189
Median raphe 63–78
Metabotropic glutamate receptors (mGluRs)
193
Metaplasticity 438–441
Methylazoxymethanolacetate (MAM) 89, 712
mGluR5 receptor 345, 349
mGluR-dependent LTD 482, 489–491
MicroRNAs (miRNAs) 459–460
Mild cognitive impairment (MCI) 742, 748–749
Mineralocorticoid receptor (MR) 355–356
Mini mental state examination (MMSE) 732–733,
744
Modeling DG 639–655
Molecular layer perforant path-associated cells
(MOPP) 15, 56, 223
Monoglyceride lipase (MGL) 321
Mossy cells 13–14, 159–161, 644 (see also Hilar
mossy cells)
Mossy fiber sprouting in epileptic DG 541–558
(see also Seizure-induced mossy fiber
sprouting)
Mossy fibers (MFs), dentate 9–10, 58–102, 568
Anatomy 110–111
Development and cognitive function
Development 97–98
Genetics and breeding experiments 99
in CA3 (regio inferior)
Intra- and supragranular mossy fiber collaterals
94–95
Mossy fiber LTP 432
Mossy fiber sprouting 183–195
Plasticity of 431–436
Projections to CA1 (regio superior) 91–92
Projections to CA3 10
Projections to the hilus 9–10, 92, 95–96
Sprouting 99–100, 765–769
Structural plasticity 85–102
Supra- and infrapyramidal 87–90
Synaptic transmission 109–123
Terminals 287–289, 581–582
Tetanic stimulation 590
Timm staining 93–94
Trajectory and termination 86–97
Transverse and longitudinal trajectories of
90–91
Visualization 86
Zinc in 33
a-Neo-endorphin 246
b-Neo-endorphin 246
Na+ currents 186
Naloxone 568
NE (see Norepinephrine)
Neonates 167–180
782

Neprilysin 268 Neuronal cell death, TNF-a and 341

Nerve growth factor (NGF) 115, 371–373 Neuronal migration 134
Neuroanatomical organization, DG 3–22 Neuronal network, connectivity matrix for 642
Cell types and their connectivity 8–19 (see also Neurons of DG 14–17
Granule cells; Mossy cells; Pyramidal Axo-axonic cell 15
basket cells) Long-spined cell 16
Comparative neuroanatomy 6–8 of the molecular layer 15
among rat, monkey, humans 6–8 Neurons of polymorphic cell layer 15–17
Extrinsic inputs 17–19 Neuropeptide Y (NPY) in
Layers of, 5 DG 115, 227–228, 285–294, 686
Granule cell layer 5 (see also granule cells, (see also Y1 receptors; Y2 receptors;
dentate) Y4 receptors; Y5 receptors)
Molecular layer 5 Effects on hilar interneurons 289–290
Polymorphic layer (hilus) 5, 9–10 Effects on LTP 290
Long-spined cell 16 Effects on neurogenesis 291–292
Nissl and Timm’s staining 6 Effects on synaptosomes 290–291
Septotemporal axis 5 Electrophysiological effects of 287–294
Neurochemical identity, of interneurons in the fascia dentata 286–287
223–228 in disease 292–294
Calretinin 226–227 Alzheimer’s disease (AD) 292–293
Cholecystokinin (CCK) 225 Depression 293–294
Neuropeptide Y (NPY) 227–228 Epilepsy 292–293
Parvalbumin (PV)-labeled neurons 224–225 Neurons containing NPY and their synaptic
Somatostatin 225–226 contacts 286
NeuroD mutant 149–150 Neuropil 702
Neurodegenerative diseases, BDNF role in 385 Neuroprotection 399
Neurogenesis 143–144, 355, 464, 662, 697–715 Neuropsychiatric disease, BDNF role in 385–386
and antidepressants 707 Neurotransmitter transport of human dentate
and MDD (see Major depressive disorder) granule cells 191–193
and sex steroids 402–403 Neurotrophins (NT) in DG 371–387
Antidepressant treatments, neurogenic effects of Brain-derived neurotrophic factor (BDNF)
707–708 (see also separate entry)
BDNF effect on 378 Nerve growth factor (NGF) 372–373
Endocannabinoid (EC) role in 330 Neurotrophin-3 (NT-3) 371, 378–379
NPY effects on 291–292 Neurotrophin-4/5 371, 379
Opioids and 256–257 Role in epilepsy models 379–381 (see also
Preclinical studies 710–713 Epilepsy)
Antidepressant (AD) drugs, behavioral effects Seizure regulation of 379
of 712–713 Signal transduction 372
Learning, hippocampal neurogenesis in Structure 371–372
711–712 NF-kB 340, 342
Neurogenesis blockade and depression Nissl staining 762
symptoms 711 Nissl and Timm’s staining 6
Stress based depression paradigms NMDAR/mGluR-dependent LTD 474–481
710–711 Ca2+/calmodulin-dependent protein kinase
Serotonin-dependent ADs and hippocampal (CaMKII) 478
neurogenesis 708–710 Ca2+/phospholipid-dependent protein kinase
Neurokinin B 115 (PKC) 477
 783

cAMP-dependent protein kinase (PKA) 478–479 Granule cell neurogenesis 662–663

Cascades involved with 475 Granule cell numbers 662
cGMP-dependent protein kinase (PKG) 479 Granule cell synaptic contacts, changes in
Expression 476 663–664
Extracellular signal-regulated kinase (ERK) 480 Immediate early genes 670–672
Induction 475 In situ hybridization 672
Kinase pathways 477 Perforant path-granule cell synapse 666–670
Phosphatase pathways 476 Novel environment 311
Phospholipase A2 480 Nuclear translocation 134
Presynaptic vesicular release 479–480
NMDAR-dependent LTD 481–482 d-Opioid receptors (DORs) 246–247, 251–252
mGluR-dependent LTD 482 (see also separate Opioid physiology 254–255
entry) Subcellular locations of 249
N-methyl-D-aspartate receptors (NMDAR) 119, m-Opioid receptors (MORs) 246–247, 252–253
175, 401, 474–488, 567 (see also NMDAR- Offline behavioral states 579
dependent LTD; NMDAR/mGluR- Oligodendrocytes 158, 514
dependent LTD) Opioid systems in DG 245–258
Non-hippocampal afferent connections 33 b-Endorphin 246
Non-human primates, 23–38 Anatomical distribution of 247–251
Non-neuronal cells 511 Endomorphin 246
Non-self/self duality 619–620 Gonadal steroids and 256
Norepinephrine (NE) 299–312, 461 Neurogenesis and 256–257
Adrenoceptor distribution 300–302 (see also Opioid peptides 246 (see also Dynorphins;
Adrenoceptors) Enkephalins)
Hypotheses 299 Release 251
Locus coeruleus(LC)-NE modulation and Role in mossy fiber LTP 432–433
environmental events 310–312 Opioid physiology 254–255
NE innervation 300 Opioid receptors 246–247
NE-induced plasticity 306–310 ORL1 247
LC firing and NE release patterns 309–310 Prenatal morphine and 257
Pathway selectivity 307–308 Seizures and 255–256
Physiology 302–306 Subcellular locations of 249
EEG recording 305
Evoked potential recording 305–306 p55 TNFR 340, 342
Extracellular unit recording 304–305 Pain, BDNF role in 384–385
Glycogen metabolism 302–303 Parachlorophenylalanine (PCPA) 703–704
Hilar interneurons 303–304 Paradoxical TLE (PTLE) 184
Intracellular recording 303 Parahippocampal region 43–58
Release modulation by glutamate and vice-versa Parvalbumin (PV)-labeled neurons 224–225,
310 514–515
Normal aging, hippocampal granule cells in Parvalbumin-immunostained basket cells in
661–674 human DG 37
Aged granule cells Pathway selectivity, in NE-induced plasticity
Biophysical properties of 664–665 307–308
Cholinergic responses in vitro in 668 Pediatric epilepsy 756
Single unit recordings in granule cells 665–666 Perforant path 17, 43–58
DG, vulnerability of 672–673 Contralateral entorhinal-dentate projection
Granule cell dendrites 664 43–58
784

Longitudinal organization 43, 54–58 Subcellular localization of 407

Nomenclature 45–46 Pro-inflammatory cytokines, and their effects in
Radial organization in 47–54 DG 339–350 (see also Interleukin 1-b;
Synaptic organization of 56–57 Interleukin 18; Tumour necrosis factor-a)
Transection 503–519 in CNS, distribution 340
in mice 503 Projection cells
Transverse organization in 51, 53 Ultrastructure and synaptic connectivity of
Phaseolus vulgaris-leucoagglutinin (PHAL) 155–164
tracing 508, 510 Pyramidal basket cell, dentate 11–13, 158, 160
Phenylchloromethamphetamine (PCA) 704
Phosphatase pathways 476 Radial glial cells in the adult DG 157
Phospholipase A2 480 Radial glial scaffold 135
Physiological studies of human dentate granule Radial organization in perforant path 47–52,
cells 183–195 52–54
GABAergic inhibition 190 Rate remapping 592
Membrane properties 184–187 Receptor tyrosine kinases (RTKs) 372
Neuromodulation 193–195 (see also Recurrent excitation, in DG 551–558
Neuromodulation) Circuit formation 554
Synaptic properties 187–193 Emergent property 557
Picrotoxin 238 Functional effect 558
Pilocarpine 761 Inhibition state 551–553
Plaques 727 Reelin 133–140, 148, 536
Plasticity 555–556 Effect on granule cell migration 135
at mossy fiber (MF) CA3 synapse 121–122 Effect on granule cell orientation 135
Excitatory and inhibitory circuits 555–556 Effect on radial glial scaffold 135
in somatostain (SST) receptors expression Reinnervation 507
272–273 REM (rapid eye movement) 602
in the DG 417–442 Rett syndrome, BDNF role in 385
Metaplasticity 438–441
of MFs dentate mossy fibers 85–102 Schaffer collateral axons 473–494, 628
Neural grafting, testing MF growth and Seizure disorders 689–690
100–101 Hippocampal-dependent seizures 690
Sex steroids and 402–405 Seizure regulation
Synaptic plasticity 423–436 of BDNF expression 376
Polymorphic layer (hilus) 5, 9–10 of NGF expression 373
Mossy fibers 9–10 of trk receptor expression 374
Neurons of 15–17 trk receptor activation following 381–384
Postmitotic neurons 530 Seizures and hippocampal neurogenesis 533–534
Postsynaptic targets of GABA neurons 221–223 Caudal subventricular zone (SVZ) 535–536
Pregnenolone 405 Functional significance 534–536
Prenatal morphine 257 Mechanisms 536–537
Presynaptic mechanisms in synaptic consolidation Temporal lobe epilepsy (TLE) models 533
462 Seizure-induced mossy fiber sprouting
Presynaptic vesicular release 479–480 541–558
Progenitor cells 529–537 Selective serotonin reuptake inhibitor (SSRI)
Progesterone (PR) 399, 402 708–709
3a,5a-tetrahyroprogesterone 405–406 Septal fibers 18–19
Distribution in DG 409 Septo-hippocampal (SH) connections 65–69

GABAergic systems and, interactions between
68–69
Medial septum/diagonal band (MSDB)
cholinergic innervation of DG 65–67
MSDB GABAergic innervation of DG 67–68
Septotemporal axis 5
Serotonin 63–78, 403, 461, 532
Serotonin-dependent antidepressants 708
Sex steroids
and dentate physiology 400–402
and dentate plasticity 402–405 (see also
Plasticity)
and neurogenesis 402–403
and synaptic remodeling 403–405
and the DG 399–410
Effect, mediation of 406–410
ERa distribution in DG 406–408
ERb distribution in DG 408
Local steroid synthesis, role of 409–410
PR distribution in DG 409
Electrophysiological studies 401
in adulthood 401
Neuroprotective effects 405–406
Subcortical mediation of sex steroid effects
410
Sharp waves-ripple complexes (SPW) 207
Slow oscillations, of hilar mossy cells 210–213
Somatostatin (SST) in DG 225–226, 265–278, 686
Activity-dependent expression of 270–272
as a subset of hilar interneurons 269–270
Cortistatin (CST) and 265
Distribution 266
Electrophysiological effects of 274–277
in hippocampus 270
Antiepileptic properties in 271
Inhibiting LTP 275
K+ channels activated by 266
Levels in Alzheimer’s disease 268–269
SST receptor expression
in DG 272
Plasticity in 272–273
Synaptic potentiation reduction by 265
Spatial information encoding vs. retrieval
572–574
Spatial learning 689
Spikes, dentate 586
Potentiation
and pairing requirements 308
785
in anesthetized rats 306–307
In vitro 307
Spinophilin 403–404
Spoiled gradient recalled (SPGR) pulse 742
Sprouting 507, 534, 541–558 (see also Mossy fiber
sprouting)
Cholinergic 509
Commissural/associational fibers 507–508
Fiber systems 507
Reinnervation process 507
Mossy fibers (MFs), lesion-induced 99–100
Species difference in 517–518
Status epilepticus (SE) 533
Stem cell 529–530
Stimulus/response couplet 616–617
Stimulus-induced mGluR-dependent LTD 488
Stress
Exposure of rat to 356
Systems activated by 355–357
Subcortical and commissural afferents 63–78
Subiculum 43–58
Supragranular mossy fibers 94–95, 156
Supramamillo-dentate connections (SUM) 69–71
Supramammillary area (SUM) 63–78
Suprapyramidal bands
of DG 134
of hippocampal mossy fibers and
terminals 87–90
Synapse 505
Synaptic consolidation 453–465
(see also Brain-derived neurotrophic factor)
Extrinsic modulation of 461–462
Functions and clinical implications of 463–465
Alzheimer’s disease 463–464
Depression 464–465
Neurogenesis 464
Lateral perforant path (LPP) 462–463
Mechanisms, functions, and therapeutic
implications 453–465
Presynaptic mechanisms 462
Synaptic consolidation 453
Translocation 460
Synaptic plasticity in DG 299, 311–312, 423–436
in hippocampus, forms
LTD 344
LTP 344
IL-1b and 346–348
TNF-a and 343–345
786

Synaptic potentiation Tianeptine 708

and spike potentiation in vitro 307 Timm staining 18, 760, 765–766
Synaptic remodeling 399 and species differences, in mossy fibers 93–94
Glia in 404 of DG 25
Sex steroids and 403–405 Nissl and Timm’s staining 6
Synaptic transmission Timm stained hippocampal sections
and TNF-a 343 of cat 88
BDNF effects on 377–378 of dog 88
Synaptic transmission, MFs of European hedgehog 88
CA3-interneuron synapses 123 of guinea pig 88
Communication from DG to CA3 area of man 88
109–123 of rat 88
‘Detonator synapse’ 111 T-maze 622
GABA transporters, co-localization 118–119 TNF receptors (TNFR) 340
Glutamate transporters, co-localization kB (NF-kB) 340
118–119 p55 TNFR 340
Long-term plasticity at 121–122 Topographical organization 43–58
Multiple neurotransmitters/modulators in Tracing techniques 508
113–117 Transcription factors,
Pre- and post-synaptic constituents 114, patterning the cortex and
119–120 hippocampus 144–146
Presynaptic modulation 117–118 Transneuronal changes
Transmission at MF-CA3 pyramidal neuron Cellular mechanisms in 515
synapse, quantal nature of 113 Chemokines role in 515–516
Transmitter release, properties of 112–113 Degeneration 512–516
Voltage clamp methods to the study 111 in granule cells 514–515
Synaptodendrosome (SD) 458 in parvalbumin-positive neurons 514–515
Synaptophysin immunohistochemistry 687 Transplants 85–102
Tri-synaptic circuit/pathway 63, 109, 417
y-EEG 303, 311 Tropomyosinrelated kinase (trk) receptors 372,
y-Oscillations 621 381–384, 460
Tau 724 Tumour necrosis factor (TNF) 339–350 (see also
Temporal lobe epilepsy (TLE) 183, 204, 240–241, TNF receptors)
533–537, 755
Animal models of 183–195 Vasopressin 311
Dormant basket cell hypothesis 762–764 Very low-density lipoprotein receptor (VLDLR)
GABAergic neurons, reorganization of 770 135, 148
Hilar damage 762 Vesicular glutamate transporters
Interneurons, loss of 764–765 (VGLUT), 75–76
Mass-associated TLE (MaTLE) 184 BNPI/VGLUT1 75–76
Mossy fiber sprouting 765–769 DNPI/VGLUT2 76–77
Axon sprouting 765–768 VGLUT1, 75–76, 325
Recurrent excitation vs inhibition 768 VGLUT2, 75
Seizure susceptibility 768–769 VGLUT3, 75, 77–78
Paradoxical TLE (PTLE) 184 Voltage sensitive dye imaging
Pilocarpine model 534–536 techniques 238
Tetrodotoxin (TTX) 454–455
Thorny excrescences 28 Wnt signaling pathway 146–148
 787

Y1 receptors 285–294 Z scores 749

Y2 receptors 285–294 Zinc (Zn2+)
Y4 receptors 286 in mossy fibers 33
Y5 receptors 286 Zinc-sensitive synaptic GABAA receptors 240–241