Review TRENDS in Microbiology Vol.10 No.
5 May 2002 227
How do extracellular pathogens cross
the blood–brain barrier?
Xavier Nassif, Sandrine Bourdoulous, Emmanuel Eugène and Pierre-Olivier Couraud
Bacterial invasion of the meninges can occur as a consequence of bloodstream
invasion by some bacterial pathogens. Bacteria enter the central nervous
system following a direct interaction with the luminal side of the cerebral
endothelium, which constitutes the blood–brain barrier. To breach the barriers
protecting the brain, extracellular pathogens must cross a monolayer of tight
junction-expressing endothelial or epithelial cells. The limited number of
pathogens capable of crossing these tight barriers and invading the meninges
suggests that they display very specific attributes. For Neisseria meningitidis,
type IV pili have been identified as being essential for meningeal invasion and it
is believed other, as-yet-unidentified factors are also involved.
Published online: 10 April 2002
Few bacterial pathogens are capable of invading
the brain owing to the protective effect of the
physiological barriers between the bloodstream
and the central nervous system (CNS). In some
circumstances, bacteria can invade the brain
not by crossing one of these barriers but by the
dissemination of a localized infection (e.g. in
pneumococcal otitis media [1]) or dissemination via
a neural route [2]. However, in most cases, invasion
of the meninges is a complication of bloodstream
invasion [3] that occurs secondary to nasopharyngeal
(Haemophilus influenzae, Streptococcus pneumoniae
and Neisseria meningitidis) or gastrointestinal
(Escherichia coli K1, Listeria monocytogenes and
group B Streptococcus) colonization.
Bacteria enter the CNS following direct
interaction with the luminal side of the cerebral
endothelia, which constitute the blood–brain barrier
(BBB). The limited number of pathogens capable
of crossing these tight barriers and invading the
meninges suggests that they have very specific
attributes. Bacteria presenting a brain tropism are
listed in Table 1; they are invasive pathogens that
either multiply extracellularly (S. pneumoniae,
Xavier Nassif* N. meningitidis, H. influenzae and, in newborns,
Emmanuel Eugène
INSERM U570,
E. coli K1 and Group B Streptococcus) or
Faculté de Médecine intracellularly, within macrophages (Mycobacterium
Necker-Enfants Malades, tuberculosis and L. monocytogenes).
Université Paris V,
It is likely that the intracellular pathogens
156 Rue de Vaugirard,
75015 Paris, France. invade the meninges using infected leukocytes as
*e-mail: nassif@necker.fr ‘Trojan horses’ [4]. This invasion could be a direct
Sandrine Bourdoulous
consequence of the bacteria infecting a pool of
Pierre-Olivier Couraud leukocytes that recirculate through the brain;
Institut Cochin, alternatively, the bacteria could alter leukocyte gene
Département de Biologie
expression such that the infected cells become capable
Cellulaire,
22 rue Méchain, of crossing the endothelium that lines the brain
75014 Paris, France. capillaries. However, these hypotheses have yet to
http://tim.trends.com 0966-842X/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S0966-842X(02)02349-1
be investigated. The extracellular pathogens raise
different issues. Our knowledge of the mechanisms
by which these pathogens invade the brain remains
incomplete, essentially for two reasons: (1) most of
these pathogens are restricted to humans, and animal
models closely mimicking the human disease are not
available; and (2) there is no good in vitro model of
the human BBB. However, the availability of the
complete genome sequence for some of these bacteria,
as well as observations obtained from various models
and clinical cases, allows reasonable hypotheses to be
made regarding the mechanisms these pathogens use
to invade the meninges. Bacteria with a meningeal
tropism are not only capable of interacting with and
crossing a monolayer of tight-junction-forming cells
such as those forming the BBB, they can also induce
a high-grade bacteremia. The importance of the
intensity of the bacteremia in meningeal invasion
occurring after the neonatal period has recently
been reviewed [5–7]. This review will concentrate
specifically on the mechanisms of extravasation
from the cerebral capillaries by the extracellular
pathogens responsible for meningeal invasion after
the neonatal period.
The structure of the barriers between the bloodstream
and the CNS
Two different structures separate the bloodstream
from the CNS: the BBB and the blood–cerebrospinal
fluid barrier (BCSFB), which is present at the
choroid plexuses.
The blood–brain barrier
The BBB is formed by the endothelium lining the
brain capillaries. This endothelium differs from
that lining other capillaries by the presence of
intercellular tight junctions, the presence of pericytes
within the capillary basement membrane and the fact
that astrocyte foot processes ensheath the capillaries
[8,9]. The BBB is responsible for low-rate fluid-phase
endocytosis and the electrical resistance of the tight
junctions is extremely high (transendothelial
resistances >1000 ohm.cm2); these factors constitute
the principal barrier to the passive movement of
fluid, electrolytes, macromolecules and cells via the
paracellular pathway. The permeability of the BBB
in physiological situations is regulated mainly by
astrocyte-derived factors that control both the
transport properties of the endothelium and the
organization of the tight junction architecture.
228 Review TRENDS in Microbiology Vol.10 No.5 May 2002
Table 1. Main pathogens capable of invading the meninges
Pathogen Type of pathogenesis Period when meningitis
is usually observed
Streptococcus pneumoniae Extracellular After 3 months of age
Neisseria meningitidis Extracellular After 3 months of age
Haemophilus influenzae Extracellular After 3 months of age
Escherichia coli K1 Extracellular Neonatal meningitis
Streptococcus agalactiae Extracellular Neonatal meningitis
Mycobacterium tuberculosis Intracellular After 3 months of age
Listeria monocytogenes Intracellular All ages
The blood–cerebrospinal fluid barrier
The choroid plexuses contain permeable (fenestrated)
capillaries close to the basal layer of epithelial cells
that secrete cerebrospinal fluid (CSF) and control
transport between the bloodstream and the CSF.
Choroid plexus epithelial cells closely resemble a
transporting epithelium. An occluding band of tight
junctions close to the CSF side of the cells restricts
movement via the paracellular pathway, although
these junctions are slightly more permeable than
those of the BBB [10,11].
The tight junctions
The molecular composition of tight junctions has been
the subject of intensive investigations over the past few
years [9,12,13]. To date, multiple integral membrane
proteins localized in both endothelial and epithelial
tight junctions have been identified: occludin, claudins
and junction-associated membrane (JAM) proteins
[14,15]. Occludin and claudins (an expanding family of
16 members) contain four transmembrane domains
and a carboxy-terminal cytoplasmic domain that
mediates the interaction of these proteins with
peripheral membrane proteins. Claudin-1 and -5 are
expressed at the BBB and play a crucial role in the
organization and function of the tight junctions,
whereas occludin appears to be dispensable.
JAMs, which are also expressed at the BBB, have
a single transmembrane domain and belong to the
immunoglobulin superfamily. However, it remains
unclear whether JAMs are components of the tight
junction strand or are localized between the strands.
Several peripheral membrane proteins underlying
tight junction strands have been identified, including
three proteins belonging to the membrane-associated
guanylate kinase homolog (MAGUK) family, ZO-1,
ZO-2 and ZO-3, which have been localized to the
cytoplasmic surface of tight junctions.
Interaction of bacterial pathogens with the
blood–brain barriers
For most bacteria, the precise site of entry into the
CSF remains unknown. In the case of H. influenzae,
experimental data obtained in monkeys reveal a
higher density of bacteria in the ventricular CSF
compared with the lumbar space, suggesting that
bacteria enter the CSF at the choroid plexuses [16].
Proof that extracellular pathogens can interact
http://tim.trends.com
directly with the brain barriers has been obtained
for N. meningitidis. Postmortem examination
of the brain of a patient who died of fulminant
meningococcemia showed numerous bacteria
interacting with the endothelium of both the
choroid plexuses and the meninges [17]. The
fact that bacteria were found interacting with both
components of the brain/CSF barriers does not allow
conclusions to be drawn regarding the exact site of
meningeal invasion. However, it does demonstrate
that a direct interaction between the bacteria and the
components of the barriers protecting the brain can
take place. In addition, meningococcal interaction
with endothelial cells is not restricted to the brain
barrier endothelia. Indeed, N. meningitidis can
adhere to most vessels of the body (X. Nassif,
unpublished), thus demonstrating that crossing the
brain barriers is not a consequence of a specific
interaction with brain endothelial cells.
To breach the brain barriers an extracellular
pathogen must cross a monolayer of endothelial
or epithelial cells expressing tight junctions. An
important question to be addressed is whether
bacteria use the trans- or paracellular pathway. The
former would require transcytosis, and the latter
opening of the tight junctions. The lack of suitable
animal models and of an in vitro model mimicking the
different components of the BCSFB in humans has
hampered the precise analysis of this step. However,
indirect evidence suggests that both N. meningitidis
and S. pneumoniae can use the transcellular route
to cross tight-junction-forming monolayers. First,
in vivo, N. meningitidis has been localized inside
endothelial cells of the brain of an infected patient
(Fig. 1), and second, N. meningitidis was observed
to cross a monolayer of tight-junction-forming cells
in vitro using the transcellular route and without
altering the organization of tight junctions [18,19]. In
the case of S. pneumoniae, it has been demonstrated
that this pathogen can cross a monolayer of brain
endothelial cells, clearly showing it is capable of
transcytosis [20]. If these extracellular pathogens
cross host cell monolayers using the transcellular
route this implies that they are internalized and then
cross the cytoplasm to be exocytosed.
In the case of S. pneumoniae, all bacteria are
capable of adhering to brain endothelial cells in vitro,
although only those expressing a specific adhesin,
the choline-binding protein, can be efficiently
internalized and transcytosed through the
cell monolayer [20]. Furthermore, efficient
internalization requires expression of the platelet-
activating factor (PAF) receptor by the endothelial
cells [21]. Following this internalization, bacteria are
either killed intracellularly and recycled in vacuoles
back to the luminal surface, or traffic through the
cells to exit at the basolateral surface. S. pneumoniae
can form opaque or transparent colonies. Transparent
colonies harbor less capsular polysaccharide and
more cell wall phosphorylcholine than opaque
 Review TRENDS in Microbiology Vol.10 No.5 May 2002
(a) (b)
(c) (d)
Fig. 1. Interaction of Neisseria meningitidis with endothelial cells both in vitro with human umbilical
vein endothelial cells (HUVEC) (a) and (b), and in vivo (c) and (d). (a) Scanning electron micrograph
of HUVEC monolayers. (b) Transmission electron micrograph of a HUVEC monolayer infected
by N. meningitidis for 4 h (localized adhesion). (a) and (b) show cellular protrusions engulfing
meningococci inside endothelial cells. (c) and (d) show a section of a choroid plexus capillary from
a postmortem examination of a pediatric patient who died of fulminant meningococcemia [(d) is a
higher magnification]. In (c), arrows point to diplococci (right) located inside a capillary endothelial
cell and one bacterium (left) adhering to the endothelial apical cell surface, probably in the process of
being engulfed. In (d) the arrows point to structures reminiscent of microvilli-like cellular protrusions
which are limited to the vicinity of the bacteria. These images demonstrate that N. meningitidis is
capable of interacting with the endothelial cells of brain microvessels. Scale bars = 1 µm.
colonies. Bacteria that produce opaque colonies,
which survive better in the bloodstream than
transparent colonies, are largely killed when
intracellular. Consistent with the in vivo observation
that transparent forms of S. pneumoniae cause
invasive diseases, most of the bacteria capable of
transcytosis form transparent colonies.
Meningococcal invasion
Two virulence factors are essential for meningeal
invasion by N. meningitidis: the capsular
polysaccharide, which is required for bacterial
survival in extracellular fluids; and type IV pili,
multimeric structures essential for the adhesion of
virulent capsulated N. meningitidis to host cells [22,23].
This latter process appears to be caused by a tip-
located adhesin designated PilC [24]. The role of the
type IV pili was established by demonstrating that
meningeal invasion was associated with an increase
in the expression of PilC [17]. Non-capsulated
meningococci, as well as the closely related pathogen
Neisseria gonorrhoeae, use several mechanisms for
host cell invasion [25,26]. These involve interactions
of host cellular receptors with bacterial surface
components, including the Opa and Opc outer
membrane proteins, the porins and the
lipooligosaccharide. Our understanding of these
mechanisms has benefited from numerous
http://tim.trends.com
229
investigations [27–34]. The role of these components
in meningeal invasion by virulent capsulated strains
of N. meningitidis remains to be identified precisely,
given that, as already mentioned, non-piliated
capsulated isolates are unable to interact efficiently
with the host cells. Type IV pili are thus the only
components that appear able to initiate the
interaction of virulent capsulated N. meningitidis
with the cellular components of the brain barriers.
However, following the initial interaction, the pili
retract and the meningococci appear non-piliated.
[35,36]. This suggests that other bacterial
components are involved in the interaction of piliated
capsulated meningococci with host cells. A reduction
of the production of the polysaccharide capsule
could, at this stage, allow the aforementioned outer
membrane components to mediate these interactions.
A proposed mechanism of internalization of
capsulated meningococci following pilus-mediated
adhesion is summarized in Fig. 2. Capsulated piliated
bacteria adhere to the apical surface of the cells
and only a small proportion of the bacteria are
internalized. This adhesion is associated with the
formation of ‘cortical plaques’ beneath bacterial
colonies on the apical surface [37]. These structures
result from the localized polymerization of cortical
actin associated with the clustering of integral
membrane proteins, such as intercellular adhesion
molecule-1 (ICAM-1), CD44 and the epithelial growth
factor (EGF) receptor, as well as ezrin, a member of
the ezrin-radixin-moesin (ERM) protein family [37,38].
These cytoskeletal modifications can lead to the
formation of microvilli-like structures, as shown in
Fig. 1, which could then be responsible for bacterial
internalization. Transmission electron microscopy
analysis of sections of the choroid plexus capillaries of
a brain tissue obtained from a fulminant case shows
the presence of extracellular structures, reminiscent
of epithelial microvilli, in the proximity of bacteria
in the process of being internalized and not at the
surface of non-infected cells (Fig. 1). The use of a
dominant-negative form of ezrin inhibits actin
polymerization, thus demonstrating the essential
role of ezrin recruitment in these cytoskeletal
modifications [38]. This is consistent with the known
function of the ERM proteins, which play a major
regulatory role in many of the morphogenic changes
of the plasma membrane, including the formation of
microvilli [39,40], by acting as linkers between the
plasma membrane and the actin cytoskeleton.
ERM proteins interact with the cytoplasmic
domain of transmembrane proteins (so-called ERM-
binding proteins), such as CD44 or ICAM-1, via their
amino-terminal domain, and interact with F-actin
by their carboxy-terminal domain. The cytoskeletal
changes induced by pilus-mediated adhesion occur
in a two-step process: (1) formation of the organizing
center of the cortical plaque, that is, recruitment of
ezrin, moesin and ERM-binding transmembrane
proteins; and (2) polymerization of cortical actin. The
230 Review TRENDS in Microbiology Vol.10 No.5 May 2002
?
ICAM-1
CD44
ErbB2 CD46
P P
PP PP
Ezrin
Cortactin
Rho
F-actin
Cdc42
Fig. 2. Internalization of Neisseria meningitidis by endothelial cells.
The first step requires interaction between bacterial pili and a cellular
membrane receptor, which is likely to be an ezrin-radixin-moesin
(ERM)-binding protein, the best candidate being CD46. This interaction
leads to the subsequent recruitment of ezrin, cortical actin and cortactin
and to the formation of microvilli. This signaling requires the activation
of two small GTPases, Rho and Cdc42. Another signaling event is linked
to the activation, at the site of bacterial interaction, of the tyrosine
kinase receptor ErbB2. The phosphorylation of ErbB2 provides a
docking site for the cytosolic tyrosine kinase Src, which phosphorylates
cortactin and subsequently promotes bacterial internalization.
Recently, the cytosolic domain of CD46, a pilus receptor, has been
shown to be phosphorylated by an Src kinase; whether this is a
consequence of ErbB2 activation is unknown [51]. It should be pointed
out that the mechanism of activation of ezrin is unknown. Abbreviation:
ICAM-1, intercellular adhesion molecule 1. Modified from [45] by
copyright permission of The Rockefeller University Press.
first step is the consequence of the direct interaction
between the pili and a cellular receptor. It is tempting
to speculate that this latter component might be an
ERM-binding transmembrane protein, thus being
responsible for ERM protein activation and the
subsequent recruitment of other transmembrane
proteins. Such a mechanism has been demonstrated
for Group A Streptococcus, and is responsible for ezrin
activation following the interaction of hyaluronic acid
with CD44 [41]. In the case of N. meningitidis, CD46
has been proposed as a receptor for type IV pili [42].
Future work should further document the role of
CD46 and its phosphorylation [51] in the recruitment
of ezrin, with particular attention being given to the
ERM-binding site present in its cytosolic domain.
The second step, actin polymerization, requires
the activation of the GTPase Rho and its downstream
effector Rho kinase; Cdc42 is also involved in this
process, as expected for the formation of microvilli-
like structures, which, unlike ruffles, do not require
Rac activation. Whereas other bacterial pathogens
responsible for dramatic cytoskeletal modifications
inject bacterial proteins with exchange-factor activity
into the cytosol via a type III secretion system [43],
http://tim.trends.com
ErbB2-P
P P P P
Cortactin-P
Src
TRENDS in Microbiology
the analysis of the genome content of N. meningitidis
did not reveal a type III secretion system and the
mechanism of activation of these GTPases by this
pathogen is unknown. The role of signaling following
the recognition of CD46 by pili remains to be explored
[44]. Another signaling event is associated with the
internalization of N. meningitidis by endothelial cells.
Indeed, ErbB2, a receptor tyrosine kinase of the EGF
receptor family, is specifically recruited beneath
bacterial colonies in the cortical plaques following
pilus-mediated adhesion [45]. ErbB2 is activated
by homodimerization and is required for efficient
meningococcal internalization.
The internalization of N. meningitidis is likely to
promote its subsequent transcytosis through cellular
barriers in vivo. As previously mentioned, type IV pili
are the only bacterial components identified so far that
are involved in the formation of the microvilli-like
membrane protrusions induced by N. meningitidis.
However, the fact that pili retract and disappear
following the initial pilus-mediated interaction
strongly argues in favor of the involvement of
additional meningococcal components, such as Opa
proteins, in both entry and transcytosis [46]. Recently,
the importance of the IgA protease in cleaving
Lamp-1 and thus increasing intracellular bacterial
survival was demonstrated [47]. The recent
availability of the meningococcal genome sequence
[48,49] as well as genetic tools allowing the
construction of libraries of random mutations in
bacterial genes should allow the precise identification
of the factors required for these steps and facilitate
the understanding of the cellular mechanisms
involved in bacterial survival and transcytosis.
The importance of bacteremia
A specific aspect of the pathogenesis of extracellular
bacteria capable of invading the meninges is the
 Review TRENDS in Microbiology Vol.10 No.5 May 2002 231

presence of high-level bacteremia. The importance has been made in our understanding of the
of bacteremia in meningeal invasion was initially mechanisms of pathogenesis, numerous questions
demonstrated for H. influenzae. Using animal remain to be answered, such as the adaptation
models, it was shown that meningeal invasion to extracellular growth and the mechanism of
directly correlated with the intensity and/or length of transcytosis. A major remaining question is why
the bacteremia [3,50,16]. The bacterial factors known do some bacteria invade the meninges? Without
to participate in the in vivo multiplication of bacterial antibiotic treatment, bacterial meningitis is a lethal
pathogens include (1) the capsule, which is essential; disease; the factors responsible for invasion of the
(2) the lipooligosaccharide; and (3) the iron-chelation meninges thus cannot participate in bacterial
systems. However, these attributes are not restricted dissemination, especially when one considers that at
to extracellular neuroinvasive bacteria, but are least some of the pathogens responsible for bacterial
shared by most extracellular pathogens. One might meningitis are restricted to humans. This suggests
Acknowledgements
X.N.’s and E.E.’s laboratory speculate that these attributes are more efficient that the factors responsible for breaching the
is supported by INSERM, in allowing extracellular growth of neuroinvasive blood–brain barriers serve an important purpose
the Université Paris V bacteria compared with other extracellular other than killing the host. For Haemophilus,
René Descartes, the
Ministère de la Recherche
pathogens. Alternatively, these bacteria could express meningococci and pneumococci, an important step
(PRFMMIP) and the some as-yet-unknown virulence factors or metabolic in their life cycle is nasopharyngeal colonization.
Direction Générale des pathways that could be responsible for the adaptation A popular hypothesis is to consider that the
armées. S.B.’s and P-O.C.’s to extracellular growth. characteristics required to interact with the brain
laboratory is supported
by CNRS, INSERM, the
barriers might have been selected initially to interact
Université Paris V René Conclusions with nasopharyngeal cells; if this is the case,
Descartes, the Ministère Invasion of the meninges by extracellular pathogens meningeal invasion might simply be an unlucky
de la Recherche
appears to be a consequence of their ability to induce a consequence of the expression of virulence factors in
(PRFMMIP) and the
Direction Générale des high level of bacteremia and to interact directly with the bloodstream that were initially designed to allow
Armées the brain barriers. Even though significant progress mucosal colonization.
References 12 Tsukita, S. and Furuse, M. (1999) Occludin and 23 Nassif, X. et al. (1994) Roles of pilin and PilC in
1 Marra, A. and Brigham, D. (2001) Streptococcus claudins in tight-junction strands: leading or adhesion of Neisseria meningitidis to human
pneumoniae causes experimental meningitis supporting players? Trends Cell Biol. 9, 268–273 epithelial and endothelial cells. Proc. Natl. Acad.
following intranasal and otitis media infections 13 Zahraoui, A. et al. (2000) Tight junction, a Sci. U. S. A. 91, 3769–3773
via a nonhematogenous route. Infect. Immun. platform for trafficking and signaling protein 24 Rudel, T. et al. (1995) Neisseria PilC protein
69, 7318–7325 complexes. J. Cell Biol. 151, F31–36 identified as type-4 pilus tip-located adhesin.
2 Jin, Y. et al. (2001) Neural route of cerebral Listeria 14 Martin-Padura, I. et al. (1998) Junctional Nature 373, 357–359
monocytogenes murine infection: role of immune adhesion molecule, a novel member of the 25 Nassif, X. et al. (1999) Interactions of pathogenic
response mechanisms in controlling bacterial immunoglobulin superfamily that distributes at Neisseria with host cells. Is it possible to
neuroinvasion. Infect. Immun. 69, 1093–1100 intercellular junctions and modulates monocyte assemble the puzzle? Mol. Microbiol.
3 Moxon, E.R. and Ostrow, P.T. (1977) Haemophilus transmigration. J. Cell Biol. 142, 117–127 32, 1124–1132
influenzae meningitis in infant rats: role of 15 Tsukita, S. et al. (1999) Structural and signalling 26 Dehio, C. et al. (2000) Host cell invasion by
bacteremia in pathogenesis of age-dependent molecules come together at tight junctions. pathogenic Neisseriae. Subcell. Biochem.
inflammatory responses in cerebrospinal fluid. Curr. Opin. Cell Biol. 11, 628–633 33, 61–96
J. Infect. Dis. 135, 303–307 16 Smith, A.L. (1987) Pathogenesis of Haemophilus 27 Virji, M. et al. (1992) Expression of the Opc
4 Drevets, D.A. and Leenen, P.J. (2000) Leukocyte- influenzae meningitis. Pediatr. Infect. Dis. J. protein correlates with invasion of epithelial and
facilitated entry of intracellular pathogens into 6, 783–786 endothelial cells by Neisseria meningitidis.
the central nervous system. Microbes Infect.
17 Pron, B. et al. (1997) Interaction of Neisseria Mol. Microbiol. 6, 2785–2795
2, 1609–1618
meningitidis with the components of the 28 Virji, M. et al. (1993) Meningococcal Opa and Opc
5 Kim, K.S. (2000) Escherichia coli invasion of brain
blood–brain barrier correlates with an increased proteins: their role in colonization and invasion of
microvascular endothelial cells as a pathogenetic
expression of PilC. J. Infect. Dis. human epithelial and endothelial cells.
basis of meningitis. Subcell. Biochem. 33, 47–59
176, 1285–1292 Mol. Microbiol. 10, 499–510
6 Huang, S.H. and Jong, A.Y. (2001) Cellular
18 Merz, A.J. et al. (1996) Traversal of a polarized 29 Meyer, T.F. (1998) Pathogenic Neisseria –
mechanisms of microbial proteins contributing
epithelium by pathogenic Neisseriae: facilitation interplay between pro- and eukaryotic worlds.
to invasion of the blood–brain barrier.
by type IV pili and maintenance of epithelial Folia Microbiol. (Praha) 43, 311–319
Cell. Microbiol. 3, 277–287
barrier function. Mol. Med. 2, 745–754 30 van Putten, J.P. et al. (1998) Gonococcal invasion
7 Kim, K.S. (2001) Escherichia coli translocation at
19 Pujol, C. et al. (1997) Interaction of Neisseria of epithelial cells driven by P.IA, a bacterial ion
the blood–brain barrier. Infect. Immun.
meningitidis with a polarized monolayer of channel with GTP binding properties. J. Exp.
69, 5217–5222
epithelial cells. Infect. Immun. 65, 4836–4842 Med. 188, 941–952
8 Goldstein, G.W. and Betz, A.L. (1986) The
blood–brain barrier. Sci. Am. 255, 74–83 20 Ring, A. et al. (1998) Pneumococcal trafficking 31 Bauer, F.J. et al. (1999) Mutagenesis of the
9 Kniesel, U. and Wolburg, H. (2000) Tight across the blood–brain barrier. Molecular analysis Neisseria gonorrhoeae porin reduces invasion in
junctions of the blood–brain barrier. Cell. Mol. of a novel bidirectional pathway. J. Clin. Invest. epithelial cells and enhances phagocyte
Neurobiol. 20, 57–76 102, 347–360 responsiveness. Mol. Microbiol. 31, 903–913
10 Segal, M. (1998) The Blood–CSF barrier and the 21 Cundell, D.R. et al. (1995) Streptococcus 32 Naumann, M. et al. (1999) Host cell interactions
choroid plexus. In Introduction to the Blood–Brain pneumoniae anchor to activated human cells by and signalling with Neisseria gonorrhoeae. Curr.
Barrier (Pardridge, W.M., ed.), pp. 251–258, the receptor for platelet-activating factor. Nature Opin. Microbiol. 2, 62–70
Cambridge University Press 377, 435–438 33 Virji, M. et al. (1999) Critical determinants of host
11 Strazielle, N. and Ghersi-Egea, J.F. (2000) 22 Virji, M. et al. (1991) The role of pili in the receptor targeting by Neisseria meningitidis and
Choroid plexus in the central nervous system: interactions of pathogenic Neisseria with cultured Neisseria gonorrhoeae: identification of Opa
biology and physiopathology. J. Neuropathol. Exp. human endothelial cells. Mol. Microbiol. adhesiotopes on the N-domain of CD66 molecules.
Neurol. 59, 561–574 5, 1831–1841 Mol. Microbiol. 34, 538–551

http://tim.trends.com
232 Review TRENDS in Microbiology Vol.10 No.5 May 2002

34 Edwards, J.L. et al. (2000) Neisseria gonorrhoeae
elicits membrane ruffling and cytoskeletal
rearrangements upon infection of primary human
endocervical and ectocervical cells. Infect. Immun.
68, 5354–5363
35 Pujol, C. et al. (1999) The meningococcal PilT
protein is required for induction of intimate
attachment to epithelial cells following pilus-
mediated adhesion. Proc. Natl. Acad. Sci. U. S. A.
96, 4017–4022
36 Merz, A.J. et al. (2000) Pilus retraction powers
bacterial twitching motility. Nature 407, 98–102
37 Merz, A.J. et al. (1999) Type IV pili of pathogenic
Neisseriae elicit cortical plaque formation in
epithelial cells. Mol. Microbiol. 32, 1316–1332
38 Eugene, E. et al. (2002) Microvilli-like structures
are associated with the internalization of
virulent capsulated Neisseria meningitidis into
vascular endothelial cells. J. Cell Sci.
115, 1231–1241
39 Mangeat, P. et al. (1999) ERM proteins in cell
adhesion and membrane dynamics. Trends Cell
Biol. 9, 187–192
40 Yonemura, S. and Tsukita, S. (1999) Direct
involvement of ezrin/radixin/moesin (ERM)-
binding membrane proteins in the organization of
microvilli in collaboration with activated ERM
proteins. J. Cell Biol. 145, 1497–1509
41 Cywes, C. and Wessels, M.R. (2001) Group A
Streptococcus tissue invasion by CD44-mediated
cell signalling. Nature 414, 648–652
42 Kallstrom, H. et al. (1997) Membrane cofactor
protein (MCP or CD46) is a cellular pilus receptor for
pathogenic Neisseria. Mol. Microbiol. 25, 639–647
43 Sansonetti, P.J. (2001) Microbes and microbial
toxins: paradigms for microbial–mucosal
interactions – III. Shigellosis: from symptoms to
molecular pathogenesis. Am. J. Physiol.
Gastrointest. Liver Physiol. 280, G319–G323
44 Kallstrom, H. et al. (1998) Cell signaling by the
type IV pili of pathogenic Neisseria. J. Biol. Chem.
273, 21777–21782
45 Hoffmann, I. et al. (2001) Activation of ErbB2
receptor tyrosine kinase supports invasion of
endothelial cells by Neisseria meningitidis. J. Cell
Biol. 155, 133–143
46 Wang, J. et al. (1998) Opa binding to cellular
CD66 receptors mediates the transcellular
traversal of Neisseria gonorrhoeae across
polarized T84 epithelial cell monolayers.
Mol. Microbiol. 30, 657–671
47 Hopper, S. et al. (2000) Effects of the
immunoglobulin A1 protease on Neisseria
gonorrhoeae trafficking across polarized T84
epithelial monolayers. Infect. Immun.
68, 906–911
48 Parkhill, J. et al. (2000) Complete genome
sequence of a serogroup A strain of Neisseria
meningitidis Z2491. Nature 404, 502–506
49 Tettelin, H. et al. (2000) Complete genome
sequence of Neisseria meningitidis serogroup B
strain MC58. Science 287, 1809–1815
50 Moxon, E.R. et al. (1984) Pathogenesis of
meningitis: experimental studies on the
molecular basis of Haemophilus influenzae
infection. Infection 12, 23–28
51 Lee, S.W. et al. (2002) CD46 is phosphorylated at
tyrosine 354 upon infection of epithelial cells by
Neisseria gonorrhoeae. J. Cell Biol. 156, 951–957

Bacterial avoidance of phagocytosis

Jean Celli and B. Brett Finlay

Phagocytosis constitutes the primary line of host innate and adaptive defence Phagocytosis can be viewed as either an opportunity
against incoming microbial pathogens, providing an efficient means for their or an obstacle for microbial pathogens. Although
removal and destruction. However, several virulent bacteria that do not pathogens with an intracellular ‘lifestyle’ have
function as intracellular pathogens have evolved mechanisms to avoid and developed many sophisticated strategies to enter
prevent phagocytosis that constitute an essential part of their pathogenic and survive within phagocytic cells, other bacterial
capacity. Some of these mechanisms include preventing recognition by pathogens have evolved mechanisms to prevent their
phagocytic receptors or blocking uptake by professional phagocytes. Recently, phagocytosis as part of their pathogenic profile. Such
the molecular mechanisms of such antiphagocytic properties have been mechanisms allow the organism to avoid destruction
elucidated for some pathogens. Such mechanisms illustrate the diversity of via degradative endocytic pathways and in some
mechanisms bacterial pathogens use to avoid phagocytic uptake. instances block phagocytic responses, thereby
impairing the development of cellular immunity and
Published online: 21 March 2002 enhancing extracellular survival. In this review, we
focus on some bacterial pathogens, the antiphagocytic
Phagocytosis is a normal receptor-mediated process properties of which have been characterized at the
that uses an actin-based mechanism to internalize molecular level, and discuss the strategies and
particles such as bacteria and yeasts (~0.5–5 µm mechanisms these bacteria use to escape phagocytosis.
in diameter) into phagocytic cells. Professional
Jean Celli phagocytes, including neutrophils and macrophages, Diversity of phagocytosis mechanisms
B. Brett Finlay* are characterized by their ability to express a set of A fundamental step in phagocytosis is the receptor-
Biotechnology
Laboratory, University
phagocytic receptors designed to recognize, bind mediated recognition of an extracellular particle
of British Columbia, to and trigger internalization of foreign objects, destined for phagocytosis. Different phagocytic
Room 237, 6174 including pathogens, debris and apoptotic cells. receptors can be involved in particle recognition,
University Boulevard,
Following uptake, internalized particles are depending on whether or not opsonins are involved
Vancouver, BC,
Canada V6T 1Z3. destroyed as they progress along the degradative in the phagocytic process. Opsonin-dependent
*e-mail: bfinlay@ endocytic pathway, cumulating in mature phagocytosis involves either Fcγ receptors (FcγR) or
interchange.ubc.ca phagolysosomes. Resident macrophages, which complement receptors (CR1, CR3 and CR4), which
Jean Celli reside in tissues, and neutrophils, which migrate bind particles that have either IgG or complement
Centre d’Immunologie, from the blood to the site of infection, constitute the bound to their surface, respectively [1]. Opsonin-
INSERM-CNRS de
primary line of innate defence against most bacterial independent phagocytosis is triggered by engagement
Marseille-Luminy,
Case 906, 13288 Marseille pathogens by providing a means for their removal and of a variety of cellular receptors capable of recognizing
Cedex 9, France. destruction [1]. and binding molecular motifs directly on the surface

http://tim.trends.com 0966-842X/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S0966-842X(02)02343-0