Cell Tissue Res (1988) 252:367-375
a n d 3"tssue
Basement-membrane components associated
with the extracellular matrix of the lymph node
Randall H. Kramer 1' 2, Steven D. Rosen ~, and Kevin A. McDonald 2
Departments of 1 Anatomy and 2 Stomatology, Schools of Medicine and Dentistry, University of California, San Francisco, USA
Summary. Lymph nodes contain an extensive array of ex-
tracellular matrix fibers frequently referred to as "reticular
fibers" because of their reticular pattern and positive reac-
tion with silver stains. These fibers are known to contain
primarily type-III collagen. In the present study, frozen and
plastic-embedded sections of mouse and human lymph
nodes were subjected to immunostaining with a panel of
monospecific antibodies directed against type-IV collagen,
type-III collagen, laminin, entactin, and heparan sulfate
proteoglycan. Immunofluorescent staining revealed that, in
addition to being uniformly stained with antibodies to type-
III collagen, these fibers also stained positively with anti-
bodies to type-IV collagen and to other basement-mem-
brane-specific components. Furthermore, the basement-
membrane-specific antibodies stained the outer surface of
individual fibers. These same type-III collagen-rich fibers
were distinct from blood vascular basement membranes
since they did not react with antibodies to factor VIII-re-
lated antigen, an endothelial-cell-specific marker. The role
of these basement-membrane-specific components asso-
ciated with the reticular fibers of lymphoid tissue is un-
known. However, it is possible that the ligands promote
attachment of reticular fibroblasts as well as macrophages
and lymphocytes to the extracellular matrix fibers.
Key words: Lymph node - Extracellular matrix - Reticular
fibers - Basement membrane - Mouse - Human
Traditionally, the extracellular matrix of the lymph node
has been defined as consisting of a fibrous capsule and
associated trabeculae and a network of "reticular fibers"
that permeate the tissue and provide a continuous tridimen-
sional network that intimately supports the lymphoid cells
as well as the blood vessels. Special histochemical tech-
niques are required for visualizing these fibers, and the
silver impregnation method remains the usual means for
their staining. The reticular cells surrounding these fibers
are believed to be responsible for forming and maintaining
this connective tissue framework of the lymph node.
Early work (Clark 1962) as well as more recent studies
Send offprint requests to: R.H. Kramer, Box 0512, HSW-604, De-
partments of Anatomy and Stomatology, University of California,
San Francisco, CA 94143, USA
coen
Research
9 Springer-Verlag 1988
(Becker et al. 1976; D'Ardenne et al. 1983; Junqueira et al.
1978; Linder et al. 1978; von der Mark 1981; Unsworth
et al. 1984; Macarak et al. 1986) demonstrated that type
III collagen is a principal constituent of the reticular fibers,
but that smaller amounts of type I collagen are present
as well. In addition, Linder et al. (1978), Stenman and Va-
heri (1978), and D'Ardenne et al. (1983) established that
fibronectin is also associated with reticular fibers in lymph
nodes. D'Ardenne et al. (1983) also demonstrated the co-
distribution of type III collagen and fibronectin in struc-
tures resembling reticular fibers.
The identity of the component(s) responsible for the
argyrophilia of reticular fibers following staining with the
silver impregnation method remains unknown (Unsworth
et al. 1984). One complication in resolving this point is that
these fibers may vary in their macromolecular composition
from tissue to tissue, as recently suggested by the work
of Macarak et al. (1986). For example, in contrast to the
lymph node, the human placental interstitium is composed
of large-diameter (30-35 nm), cross-banded fibers contain-
ing type I collagen, together with small-diameter (10-
15 nm) fibers containing type III collagen that often sur-
round the type I collagen fibers (Amenta et al. 1986). How-
ever, it is generally believed the argyrophilia in fibers of
lymph nodes is not caused by a specific biochemical entity,
but rather represents a diverse set of glycoproteins that
form silver deposit following histochemical staining. The
silver precipitate is believed to be formed by the reaction
of silver diamine ion with aldehydes following the oxidation
of vicinal hydroxyl groups in glycoproteins (Stern 1979).
Some evidence suggests that carbohydrate groups asso-
ciated with the fibers, perhaps heparan sulfate proteogly-
cans (Montes et al. 1980), are involved in the affinity for
the silver stain. It is noteworthy that basal lamina also pro-
duces a positive reaction with these stains, presumably due
to the presence of heavily glycosylated type IV collagen,
fibronectin, laminin, and heparan sulfate proteoglycans
(Burgeson 1982).
Ultrastructural studies of the reticular fibers of the
lymph node and spleen indicate that at their core are bun-
dles of collagen fibrils, frequently surrounded by an in-
distinct and discontinuous layer of amorphous ground sub-
stance. This amorphous material is believed to correspond
to the silver staining seen at the level of the light microscope
(Clark 1962; Fresen and Wellensiek 1959; Sorenson 1960).
Several authors have suggested that in some respects, this
368
amorphous material resembles a basal lamina that is local-
ized at the interface between the reticular cell and the core
of collagen fiber (Clark 1962; Forkert et al. 1977; Lennert
1978; Nopajaroonsri et al. 1971; Sorenson 1960).
These ultrastructural observations were the first indica-
tion that basement-membrane-like material may be asso-
ciated with reticular fibers. Of course, the most distinct
characteristic of the basement-membrane is its sheet-like
ultrastructure with uniformity of thickness. Since the base-
ment-membrane-like material observed by electron micros-
copy in association with lymph node reticular fibers con-
sisted of short segments or patches of basal-lamina-like ma-
terial (Sorenson 1960; Clark 1962; Nopajaroonsri etal.
1971; Forkert et al. 1977; Lennert 1978), these structures
do not rigidly satisfy the original definition that character-
izes authentic basement-membranes.
However, basement-membranes are also distinguished
by their unique biochemical composition (Heathcote and
Grant 1981 ; Kleinman et al. 1981, 1982; Laurie et al. 1982;
Abrahamson 1986). All basement-membranes contain the
following major macromolecules, which represent a diverse
group of basement-membrane-specific components: type IV
collagen, laminin, heparan sulfate proteoglycan, and entac-
tin. In the present study we have reported evidence that
reticular fibers in lymph node are positive for all of these
components. These components are seen in addition to the
well-known interstitial proteins, type III collagen and fibro-
nectin. Furthermore, our results indicate that this base-
ment-membrane-like material is located on the outer sur-
face of the reticular fibers. The presence of this additional
extracellular matrix material on the reticular fibers raises
some interesting questions about the origin and function
of the fibers.
Materials and methods
Preparation o f tissue sections
Healthy adult mice (C57B1/6 and Balb/c) and rats (Sprague-
Dawley) were anesthetized and sacrificed by cervical dislo-
cation or decapitation. The peripheral and mesenteric
lymph nodes were removed. Normal para-aortic human
lymph nodes were obtained from kidneys donated for trans-
plant surgery. For frozen sections, nodes were immediately
snap-frozen in 2-methyl-butane cooled in liquid nitrogen.
Sections (4-16 lam thick) were cut using an IEC cryostat
at - 2 0 ~ and placed on acid-cleaned glass slides. If not
used immediately, the slides were stored at - 7 0 ~ C for up
to one month. For immunostaining, the slides were im-
mersed in acetone ( - 2 0 ~ C) for 10 min and rehydrated in
Dulbecco's phosphate-buffered saline (PBS).'
For plastic embedding, fresh lymph nodes were fixed
and infiltrated with G M A plastic resin by a modification
of the method of Beckstead et al. (1985). Briefly, the tissue
was fixed with 4% paraformaldehyde in 0.1 M phosphate
buffer, pH 7.4, for 2-6 h at 4 ~ C. After washing in a solu-
tion of 3% sucrose in 0.1 M phosphate buffer overnight,
the tissue was dehydrated in cold acetone and infiltrated
with the G M A plastic resin at 4 ~ C. Sections 1-4 ktm thick
were cut with glass knives on a Sorvall microtome.
1 Abbreviations used in this paper: FITC fluorescein isothiocyanate;
PBS phosphate-buffered saline; RITC tetraethyl rhodamine iso-
thiocyanate
Antibodies
The various antisera and antibodies were individually tested
for the optimal dilution. Affinity-purified secondary anti-
bodies specific to mouse or rabbit IgG (purchased from
Boehringer-Mannheim Biochemicals) were conjugated with
horseradish peroxidase, fluorescein isothiocyanate (FITC)
or rhodamine isothiocyanate (RITC).
For studies of mouse and rat lymph nodes, affinity-
purified rabbit antibodies against mouse laminin and type
IV collagen isolated from the EHS tumor matrix were gifts
of Dr. Hynda Kleinman and Dr. George Martin (NIH).
In addition, rabbit antisera prepared against mouse laminin
(Sakashita et al. 1980) and type IV collagen (Bachinger
et al. 1982) purified from the PF HR9 endodermal mouse
cell line were obtained from Drs. Eva Engvall and Erkki
Ruoslahti (La Jolla Cancer Foundation) and Dr. Hans-
Peter Bachinger (University of South Carolina), respective-
ly, and produced results similar to those obtained with the
antibodies prepared against the EHS-derived antigens. Af-
finity-purified rabbit antibodies against the heparan sulfate
proteoglycan isolated from the EHS tumor were obtained
from Dr. J.R. Hassel (NIH) and have been shown to react
specifically with the protein core of the proteoglycan (Has-
sel et al. 1980). Finally, antiserum against mouse entactin
was a gift of Dr. A. Chung (University of Pittsburgh).
For studies of human lymph nodes, affinity-purified
rabbit antibodies to human type IV collagen were obtained
from Dr. Heinz Furthmayr (Yale University). Mouse
monoclonal antibody against human type IV collagen was
obtained from Miles Laboratories. Mouse monoclonal anti-
body against human type III collagen (clone no. III-53),
originally obtained from Dr. Kazushi Iwata (Fuji Chemical
Co, Japan), was a gift from Dr. R. Bhatnager (University
of California, San Francisco). Rabbit antisera to human
factor VIII-related antigen was from Calbiochem. Finally,
rabbit antisera to dinitrophenyl-bovine serum albumin
(BSA), fl-galactosidase, and human fibrinogen was from
Cooper Biomedical.
Immunohistochemistry
We used standard frozen-section immunohistochemical
procedures. The sections were fixed by an exposure to cold
acetone for 10 min, and nonspecific background staining
was then blocked by incubation for 30 min with 30% nor-
mal goat serum. Subsequently the sections were incubated
for 1 h with the primary antiserum or antibodies (diluted
in 1% normal serum in PBS), washed 5 times, then incu-
bated for 1 h with the appropriate secondary antiserum
coupled to FITC, RITC or horseradish peroxidase (also
diluted in 1% normal serum in PBS). After washing, the
sections were overlaid with a glycerol-PBS medium (for im-
munofluorescence) or Permount (for peroxidase). For dou-
ble immunofluorescence with human lymph node tissue,
sections were first incubated with mouse monoclonal anti-
body against human type III collagen, then with rabbit
anti-human type IV collagen (or with rabbit anti-human
factor VIII-related antigen). FITC-conjugated goat anti-
mouse and RITC-conjugated goat anti-rabbit antibodies
were then used to stain the respective primary antibody.
Slides were examined with a Zeiss photomicroscope
equipped with epiluminescence and appropriate filter com-
binations that permitted selective viewing of either fluores-
cein- or rhodamine-labeled antibodies.

The plastic-embedded tissue sections were stained using
the avidin-biotin-peroxidase complex system according to
the method described by Beckstead et al. (1985). Briefly,
the sections were digested with 0.25% trypsin in PBS for
45-60 min at 37 ~ C, then washed. Background staining was
blocked with 3% normal goat serum. The sections were
then incubated with the primary antibody for 1-12 h and
washed again. Next, the sections were exposed to biotiny-
lated secondary antibody (usually goat anti-rabbit IgG) for
1-2 h, then washed a third time. Finally, sections were cov-
ered with the avidin-biotin-peroxidase complex and incu-
bated for 1-2 h. Staining was developed using 3,3-'diamino-
benzidine as chromogen substrate and hydrogen peroxide
as described (Beckstead et al. 1985).
Some plastic-embedded sections were stained by the
silver impregnation technique according to the modification
of Gridley (1951). To permit comparison of the silver-
stained structures with those that reacted positively with
antibodies to basement-membrane components, serial plas-
tic-embedded sections were cut and adjacent sections at-
tached to glass slides and stained for anti-laminin or for
reticular fibers by the silver impregnation technique.
To confirm the specificity of the antiserum, control
blocking experiments were performed on mouse lymph
node sections with mouse laminin and type IV collagen.
These antigens were isolated from the EHS tumor according
to the procedures of Kleinman et al. (1982). In these experi-
369
Fig. l A-F. Localization of
basement-membrane-specific
components in reticular fibers of
mouse lymph nodes. Lymph
nodes were embedded in plastic
(A and B) or frozen (C-F) and
sections were prepared. Silver-
stained reticular fibers (arrows)
are present throughout paracortex
(A). Adjacent section (B) stained
for laminin yields strong reaction
of basement-membrane
surrounding a high endothelial
venule (*) and moderate reaction
of reticular fibers (arrowheads).
Laminin (C), heparan sulfate
proteoglycan (D), and type IV
collagen (E and F) were localized
by immunofluorescent staining in
neighboring sections of lymph
node. All 3 antigens were present
in blood vessels as well as in
fibers randomly distributed in
paracortex. Usually walls of blood
vessels (*) stained more intensely
than neighboring fibers. Note
tube-like staining patterns that
frequently outline network of
fibers in (C) (arrowheads). Low-
magnification view of paracortex
stained with affinity-purified
antibodies to mouse type IV
collagen demonstrates extensive
interconnection of fiber network
(E). Preincubation of antibodies
to type IV collagen with excess
purified mouse type IV collagen
completely abolishes staining (F).
Bars: A, B 100 ~tm; C, D 20 Ixm;
E, F 40 lam
ments, the appropriate dilution of antibodies (anti-mouse
laminin or anti-mouse type IV collagen) that yielded a
strong but specific immunofluorescence for the reticular
fibers and the basement-membrane of blood vessels was
preincubated with 10 to 100 pg of the respective purified
antigen for 1 h at 37~ C. The solution was then subjected
to microcentrifugation (9000 g) for 10 min, and the super-
natant was applied to the tissue section for immunofluores-
cent staining. Other control experiments used normal (non-
immune) or hyperimmune sera to irrelevant antigens (rabbit
anti-dinitrophenyl-BSA, anti-fl-galactosidase, and anti-hu-
man fibrinogen [Cooper Biomedical]).
Results
Reticular fibers present in lymph nodes have traditionally
been detected by silver impregnation methods, which reveal
an interlacing network of fibrous material that extends
throughout the tissue. We examined the staining pattern
of these fibers in plastic-embedded sections of mouse lymph
nodes treated by the silver impregnation method (Fig. 1 A).
As expected, argyrophilic fibers of various diameters were
present throughout the entire parenchyma of the node, but
were sparser in the germinal centers, which are associated
with B-cell proliferation. Adjacent plastic-embedded sec-
tions were incubated with antibodies to laminin, a base-
ment-membrane-specific glycoprotein (Kleinman et al.
370

1981, 1982), and stained by the avidin-biotin-peroxidase

method. A network of fibers similar in distribution to the
reticular fibers demonstrated a positive reaction, although
the intensity of the reaction product was generally less than
that observed for basement-membrane of adjacent blood
vessels (Fig. 1 B). Similar staining of the fibers in adjacent
plastic-embedded sections was detected with antibodies to
type IV collagen (not shown).
In further studies we employed cryostat sections of
mouse lymph nodes. Sections were stained with a panel
of antibodies to several known constituents of basement-
membranes. As expected, intense staining was observed in
the basement-membrane region of blood vessels. This pat-
tern was noted for all antibodies to the basement-mem-
brane-specific components tested: anti-laminin (Fig. 1 C),
anti-heparan sulfate proteoglycan (Fig. 1 D), and anti-type
IV collagen (Fig. 1 E). In addition, however, thin but dis-
tinct streaks of moderately stained material were present
throughout the tissue that closely resembled the pattern
of reticular fibers seen with silver staining.
The co-distribution of collagen types III and IV in and
around the reticular fibers was clearly established by dou-
ble-immunofluorescent staining of human lymph node tis-
sue sections. In these experiments, sections were stained
sequentially with mouse monoclonal antibodies to human
type III collagen (FITC) and affinity-purified rabbit poly-
clonal antibodies to human type IV collagen (RITC). As
expected from previous studies, the reticular fibers were
intensely and uniformly stained with the anti-type III anti-
bodies (Fig. 2 A and B). The staining with anti-type IV col-
lagen antibodies co-distributed with that of anti-type III
collagen (Fig. 2A'), except that the former pattern was fre-
quently more granular and in some cases was discontinuous
(Fig. 2 A ' and B'). Careful comparisons indicated that
nearly all fibers that stained with anti-type III collagen also
stained with anti-type IV collagen. Fig. 2 A-B. Distribution of type III collagen and type IV collagen
In some o f the larger fibers, identified by staining with in extracellular matrix fibers of para-aortic human lymph node.
anti-type III collagen, a remarkable pattern for type IV For double immunofluorescence, cryostat sections of lymph node
collagen was observed. Whereas the fibers were uniformly were incubated with mouse monoclonal antibody to human type
stained with antibodies to type III collagen (Fig. 2 B), stain- Ill collagen (A and B) followed by rabbit polyclonal antiserum
ing with anti-type IV collagen antibodies revealed only a to human type IV collagen (A' and B'). Lastly sections were stained
discontinuous outline of the fibers (Fig. 2 B'). Fine, but dis- with mixture of FITC-conjugated goat anti-mouse IgG and RITC-
crete, bands of staining sometimes formed a lattice-like pat- conjugated goat antirabbit IgG. At low magnification, antibody
tern over the surface of the fiber, and at other times a to type III collagen stains network of fibers in paracortical region
of lymph node (A). Antibodies to type IV collagen outline same
periodic ribbed pattern was detected (Fig. 2 B'). On occa- structures, although in many areas stain is irregular or discontin-
sions when the fiber was cut in cross-section, distinct stain- uous (A'). In high-magnification view of adjacent area, larger fibers
ing of the outer perimeter of the fiber was observed are strongly and uniformly stained throughout their cores with
(Fig. 2B'). A similar immunofluorescent staining pattern antibody to type III collagen (B). In contrast, antibodies to type
was observed when sections were reacted with antibodies IV collagen localize primarily to outer portion of fibers (B'). Where
to laminin (Fig. 3). A schematic diagram summarizing the fibers are apparently cut in cross-section (*), staining for this base-
proposed organization of these fibers is shown in Fig. 4. ment-membrane-specific component is clearly restricted to surface
The use of antiserum to factor VIII-associated antigen of structure. Unusual staining patterns resembling "ringlets" are
(a vascular endothelial cell-specific marker) in double-im- present (arrowheads). Bar: 100 lam
munofluorescence-stained sections clearly identified the
positively stained vascular endothelial cells (Fig. 5). Blood
vessels that were identified with antibodies to Factor VIII ules in the lymph node (Clark 1962). Consistent with our
(Fig. 5 B) were also diffusely stained with anti-type III colla- findings, Becker et al. (1976) showed that antibodies to type
gen (Fig. 5A). However, the population of type III-positive III collagen stain both reticular fibers and blood vessels
fibers that were negative for factor VIII, and thus were in the spleen. The relative staining intensities for reticular
not associated with blood vessels, was by far in the majority. fibers and blood vessels in lymph node are summarized
The weak, but positive, staining of blood vessels for type in Table 1. Various controls were performed in these experi-
III collagen may be attributable to the almost continuous ments. Normal control serum or antibody always yielded
layer of reticular cells and associated reticular fibers known negligible staining. Hyperimmune sera to irrelevant anti-
to surround arterioles, capillaries, and high endothelial ven- gens (dinitrophenyl-BSA, fl-galactosidase, or fibrinogen)

also gave negative results. In another set o f control experi-
ments with cryostat sections o f mouse l y m p h nodes, anti-
bodies to mouse EHS type IV collagen were incubated (neu-
tralized) with an excess o f purified antigen before reacting
with tissue sections. Staining was completely abolished,
demonstrating the specificity of the antisera (Fig. 1 F). In
parallel experiments, purified mouse laminin was also effec-
tive in neutralizing the activity o f rabbit anti-laminin anti-
bodies (not shown).
Discussion
The identification of basement-membrane-specific c o m p o -
nents in association with fibers in lymph nodes was unex-
pected. However, a n u m b e r o f early ultrastructural studies
(Clark 1962; F o r k e r t et al. 1977; Sorenson 1960)described
reticular fibers in which centrally located collagen fibers
were frequently encased in an outer lamina of a m o r p h o u s
matrix that interfaced with reticular cells. The ultrastructure
o f these short segments or patches o f laminae was suggested
to be similar to that of b a s e m e n t - m e m b r a n e (Clark 1962;
F o r k e r t e t a l . 1977; Lennert 1978; N o p a j a r o o n s r i e t a l .
1971 ; Sorenson 1960). Until now, the a n a t o m i c distribution
o f basement-membranes was thought to be limited to those
tissues where the b a s e m e n t - m e m b r a n e formed the interface
between p a r e n c h y m a l cells and the connective tissue. In
most cases, a sheet of parenchymal cells, such as epithelium
or endothelium, produces a subjacent layer o f basement-
membrane. Alternatively, the individual cell is enclosed
within a sheath o f b a s e m e n t - m e m b r a n e (fat cells, muscle
fibers) that separates it from the interstitium. By definition,
authentic b a s e m e n t - m e m b r a n e s are a composite o f 2 contin-
uous sheets o f uniform thickness, the basal lamina and the
reticular lamina which together form an inseparable unit
(Laurie and L e b l o n d 1983; Timpl et al. 1984; T o d d and
B o w m a n 1857; Vracko 1975).
371
Fig. 3. Low-magnification view of
immunofluorescent staining for
laminin on surface of extracellular
matrix fibers of lymph node.
Same area of lymph node is
shown in Fig. 2B, but is here
stained with antibodies to
laminin. Note "ringlet-like"
structures and interconnecting
lattice of laminin-positive material
on surface of fibers (arrows).
Cores of fibers, as in Fig. 2B,
stain for type III collagen (not
shown). Bar: 100 tam
type III collagen fibrils
Fig. 4. Schematic diagram of proposed model for organization of
extracellular matrix macromolecules associated with reticular fibers
of lymph node. Based on present immunofluorescence observations
it is proposed that reticular fibers are composed of a core of type
III collagen-rich fibrils surrounded by a discontinuous, but inter-
connecting, lattice of extracellular matrix containing basement
membrane-specific components
While the ultrastructure observed in ultrathin sections
with a p p r o p r i a t e stains is still an i m p o r t a n t criterion in
the identification o f basement-membranes, the overall
three-dimensional arrangement o f the extracellular matrix
is also critical. Typically, the reticular lamina forms the
connecting interface between the basal lamina and the adja-
cent connective tissue. High-resolution electron microscopy
of the basal lamina permits detection o f 2 distinct layers,
the lamina densa and lamina rara. The lamina rara is less
electron-dense than the lamina densa, while the lamina
densa can be resolved into a c o m p a c t network o f fibrillar
arrays embedded in a granular material. While all base-
372

Fig. 5. Localization of extracellular matrix fibers and factor VIII-related antigen in para-aortic human lymph nodes. Cryostat sections
of lymph node were subjected to double immunofluorescence staining by first incubating sections with mouse monoclonal antibody
to human type III collagen, followed by rabbit polyclonal antibody to human factor VIII-related antigen. Next, sections were incubated
with a mixture of FITC-conjugated goat anti-mouse IgG and R1TC-conjugated goat anti-rabbit IgG. Fibers are stained (arrowheads)
with anti-type III collagen (A) as are walls of neighboring blood vessels (,). However, anti-factor VIII antiserum, while clearly staining
endothelial cells of blood vessels, fails to stain fibers (B). Bar: 10 gm

ment-membranes have a similar ultrastructure, there is con-
siderable heterogeneity a m o n g various tissues. For example,
some basement-membranes are extremely thick and repre-
sent the fusion o f basal lamina derived from two cell types
as in the glomerulus and lung (Farquhar 1978).
In recent years, it has become apparent that basement-
membranes can also be identified according to their bio-
chemical composition, which is highly specific for these
structures (Timpl et al. 1984). In the red pulp of the spleen,
the endothelial cells that line the sinusoids are encircled
by ring fibers, which consist of basement-membrane enclos-
ing a sparse core of collagen fibers (Chen and Weiss 1972).
This finding, in early ultrastructural studies, has recently
been supported by immunohistochemical work demonstrat-
ing that these ring fibers appear to have a core of types
I and III collagen and a peripheral deposit of laminin
(Drenckhahn and Wagner 1986). It would thus appear that
reticular fibers of the lymph node and the ring fibers of
the spleen may both have a similar structure, consisting
of a core of interstitial collagens with an outer layer of
material containing basement-membrane-specific macro-
molecules. In this regard, it is interesting to note that Ka-
limo et al. (1985) have reported that reticular fibers sur-
rounding a primary lymphoma of the brain react with silver
stains and are also positive for collagen type III as well
as type IV collagen and laminin by immunostaining.
Whether the basement-membrane-specific components de-
tected in the present study correspond to these previously
identified amorphous zones in reticular fibers awaits confir-
mation by immunoelectron microscopy.
Silver staining, the method classically employed to dem-
onstrate the reticular fibers of lymph nodes, is relatively
nonspecific. In fact, it labels carbohydrates commonly
found in glycoproteins. Many extracellular matrix proteins,
including fibronectin, collagen types III, IV, and V, and
laminin, are heavily glycosylated. Thus, the positive reac-
tion of these lymph node fibers with this staining technique
is presumed to be due to the presence of these glycoproteins
and possibly other as yet unidentified components. Al-
though the fibers identified in the present study were inten-
sely stained with type III collagen, it is difficult to prove
by light-microscopic methods that the argyrophilic fibers
detected with silver stain are one and the same.
The association of basement-membrane-specific macro-
molecules with matrix fibers was not due to the proximity
of these fibers to blood vessels. This was established by
a number of observations. First, the reticular fibers exhib-
ited a characteristic morphology easily distinguished from
that of vascular basement-membranes. Second, while vascu-
lar basement-membranes always produced intense and con-
tinuous immunofluorescent staining with basement-mem-
brane-specific antisera (i.e. anti-laminin, anti-type IV colla-

Table 1. Summary of distribution of antigens specific for basement
membrane and interstitial connective tissue in the human lymph
node*
Antisera Reticular fibers Blood vessels
Laminin + ++ ++ ++
Type IV collagen + + ++ +
HS proteoglycan + ++
Entactin + + ++ ++
Fibronectin + ++ ++ ++
Type I collagen + + + Our finding that the specific staining of basement-mem-
Type III collagen + +++ ++ + brane components is usually weaker in the reticular fibers
Factor VIII-Ag - ++ ++
Fibrinogen - +__
* Data summarized from observations of frozen sections of para-
aortic human lymph node incubated with monospecific antibodies
followed by staining with fluorescein-labeled secondary antibodies.
Staining intensity of reticular fibers and blood vessels in paracorti-
cal region of lymph nodes was scored as follows: intense
(+ + + +), strong ( + + +), moderate (+ + ), weak (+), equivocal
(+), and negative (-). All blood vessels examined, including arter-
ioles, capillaries, and high endothelial venules, demonstrated a sim-
ilar staining pattern
gen), the corresponding staining of reticular fibers was less
intense and discontinuous. Finally, double immunostaining
with antibodies to factor VIII-related antigen and with anti-
serum to either type III or IV collagen indicated that all
blood vessels down to the smallest capillaries were clearly
distinct from positive reticular fibers. Factor VIII-related
antigen is associated with yon Willebrand factor, a large
glycoprotein known to be synthesized by vascular endothe-
lium and megakaryocytes, that is important in the coagu-
lation pathway (Hoyer 1981). Immunohistochemical stain-
ing of cells for this antigen is a definitive marker for vascu-
lar endothelium (Jaffe et al. 1973; Kramer etal. 1987a).
Weibel-Palade bodies in the cytoplasm of vascular endothe-
lial cells have been shown to be a storage reservoir for
von Willebrand protein (Ewenstein et al. 1987). These re-
sults argue strongly that the basement-membrane-positive
fibers are not simply collapsed blood vessels.
It is also unlikely that the basement-membrane compo-
nents were exclusively associated with the lymphatic en-
dothelium that lines the lymphatic sinuses of the lymph
node. The ubiquitous distribution of the basement-mem-
brane-specific material on the fibers was too extensive to
represent associations with lymphatic endothelium. In vir-
tually every fiber that stained positively with anti-type III
collagen antibodies, basement-membrane-specific compo-
nents were also detected. Positive fibers were found in every
compartment of the node except for germinal centers, where
reticular fibers are known to be sparse. In contrast, lym-
phatic endothelium is restricted to the linings of nodal sin-
uses (Fossum 1980). More important, recent work by Liotta
and collaborators (Barsky et al. 1983) indicates that lym-
phatic endothelium does not deposit detectable amounts
of basement-membrane materials such as laminin or type
IV collagen.
The possibility that basement-membrane macromole-
cules present in the structures associated with blood vessels
could solubilize and then associate with the reticular fibers,
thereby generating a positive staining reaction, also seems
unlikely since identical results were obtained when tissues
were fixed in formaldehyde, dehydrated in acetone, and
373
embedded in plastic. Under these conditions tissue proteins
would exhibit poor solubility, because they are first immo-
bilized by aldehyde fixation and then dehydrated and infil-
trated with a nonpolar medium. Furthermore, it is well-
established that basement-membrane-specific proteins, in-
cluding laminin and type IV collagen, are extremely insolu-
ble because they are stabilized by noncovalent and covalent
intramolecular interactions, including extensive cross-link-
ing by disulfide- and aldehyde-derived bonds (Heathcote
and Grant 1981).
than in the neighboring blood vessels probably explains
why basement-membrane staining of reticular fibers has not
been previously reported. This weak staining pattern in
fibers could result from a lower antigen density or poor
access of the antibody probe to the respective matrix mole-
cule. Alternatively, the staining could have been overlooked
because of the relatively small diameters of reticular fibers
(0.5-2.0 gm) and their tortuous arrangement in the tissue,
as compared with the relatively large cylinders of blood
vessel basement-membranes (10-50 gin).
The function of the identified extracellular matrix fibers
coated with basement-membrane-specific components is
unknown. The predominant cell type associated with stro-
real fibers of the lymph node is the reticular fibroblast.
These cells are characterized by their extensive array of
branching cell processes that closely follow the pattern of
reticular fiber network. Studies with monoclonal antibodies
specific to these cells indicate that they are found through-
out the parenchyma of the node, including the paracortex,
medullary cords, sinuses and follicles (Van-Vliet etal.
1986). Other studies based on enzyme histochemical reac-
tions indicate that the reticular fibroblast is distinct from
the capsule/trabecular fibroblast and the dendritic reticu-
lure cell (Beckstead 1983). Since the reticular fibroblast is
always found in close association with the reticular fibers,
it is believed that these cells synthesize and maintain the
fibers. Electron microscopy has demonstrated the intimate
relationship between the cell processes and the surface of
the reticular fiber (Clark 1962; Nopajaroonsri et al. 1971),
thus suggesting that the reticular cell forms adhesions with
these fibers. The results of the present study suggest that
basement-membrane-specific material is located at the sur-
face of the fiber and could be important in these adhesions.
Of course, fibronectin and type III collagen may also be
important in the reticular fibroblast/extracellular matrix in-
teraction. Most extracellular matrix macromolecules, in-
cluding laminin, fibronectin, and types I, III, and IV colla-
gens, promote cell adhesion; numerous cell-surface recep-
tors for these ligands have been identified (Reviewed in
Hynes 1987; Rouslahti et al. 1986).
Alternatively, other cell types in the lymph node may
interact with reticular fibers and associated basement-mem-
brane-specific macromolecules and potentially could con-
tribute to the synthesis and deposition of these components.
Several cell types have been observed in direct contact with
reticular fibers, including sinus lymphatic endothelial cells
(Luk et al. 1973; Bairati et al. 1964; Forkert et al. 1977),
macrophages (Forkert et al. 1977; Fossum 1980), lympho-
cytes (Bairati et al. 1964), and plasma cells (Bairati et al.
1964; Rodin 1974). Conceivably all of these cell types could
use the laminin- and type IV collagen-rich surface of reticu-
lar fibers as a scaffolding or a pathway for their migration
374
t h r o u g h o u t the lymph node. Recirculating lymphocytes ex-
travasate in large n u m b e r s through high endothelial venules
(HEV), migrate through the node and eventually leave
through efferent lymphatics. After invading the HEV and
associated basement membrane, these lymphocytes enter
the nodal c o m p a r t m e n t and may both bind to and migrate
along the extensive network of reticular fibers that are
k n o w n to radiate from the basal lamina of the HEV (Clark
1962; F r e e m o u n t et al. 1986). In this regard, Stoolman et al.
(1985) recently reported that lymphocytes can adhere to
l a m i n i n and display a chemotactic response to this ligand.
The natural killer (NK) cells, one subclass of lymphocytes,
have recently been shown to express a laminin-like sub-
stance on their surface (Hiserodt et al. 1985). Macrophages,
another cell type frequently observed in contact with reticu-
lar fibers, have been shown to adhere to laminin-coated
surfaces, and a specific cell-surface receptor for this ligand
has been identified on these cells (Huard et al. 1986).
The role of reticular fibers in various pathological condi-
tions should also be considered. F o r example, early cancer
metastasis in m a n proceeds initially through the lymphatic
system, with infiltration of t u m o r cells through the lymphat-
ic vessels followed by eventual arrest of t u m o r cells in the
subcapsular sinus. These arrested cells rapidly proliferate
in this highly vascularized and permissive environment and
eventually invade and replace nodal parenchyma. Most hu-
m a n cancers are carcinomas derived from various epithelial
cell layers, which normally adhere to basement-membrane
matrices (Liotta et al. 1986). The high density of basement-
membrane-specific components on nodal reticular fibers
may provide a mechanism for promoting t u m o r cell arrest
and invasion of this tissue (Kramer et al. 1987b).
Acknowledgments. We wish to thank E. Leash for her expert edito-
rial assistance during the preparation of this manuscript. The work
was supported by grants from the American Heart Association
and the National Cancer Institute, National Institutes of Health
(CA 33834).
References
Abrahamson DR (1986) Recent studies on the structure and pa-
thology of basement membranes. J Pathol 149:257-278
Amenta PS, Gay S, Vaheri A, Martinez-Hernandez A (1986) The
extracellular matrix is an integrated unit: Ultrastructural local-
ization of collagen type I, III, IV, V, and VI, fibronectin, and
laminin in human term placenta. Coll Rel Res 6:125-152
Bachinger HP, Flessler LI, Flessler JH (1982) Mouse procollagen
IV: characterization and supramolecular association. J Biol
Chem 257 : 9796-9803
Bairati A, Amantie L, de Petris S, Pernis B (1964) Studies on
the ultrastructure of the lymph nodes. 1. The reticular network.
Z Zellforsch 63 : 644-672
Barsky SH, Baker A, Siegal GP, Togo S, Liotta LA (1983) Use
of anti-basement-membrane antibodies to distinguish blood
vessel capillaries from lymphatic capillaries. Am J Surg Pathol
7 : 667-677
Becker U, Nowack H, Gay S, Timpl R (1976) Production and
specificity of antibodies against the aminoterminal region in
type III collagen. Immunology 31:57 65
Beckstead JH (1983) The evaluation of human lymph nodes, using
plastic sections and enzyme histochemistry. Am J Clin Pathol
80:131-139
Beckstead JH, Wood GS, Fletcher V (1985) Evidence for the origin
of Kaposi's sarcoma from lymphatic endothelium. Am J Pathol
119:294-300
Burgeson RE (1982) Genetic heterogeneity of collagens. J Invest
Dermatol 79:25s-30s
Chen L-T, Weiss L (1972) Electron microscopy of the red pulp
of the human spleen. Am J Anat 134:425-458
Clark SL, Jr (1962) The reticulum of lymph nodes in mice studied
with the electron microscope. Am J Anat 110:217-258
Drenckhahn D, Wagner J (1986) Stress fibers in the splenic sinus
endothelium in situ: Molecular structure, relationship to the
extracellular matrix, and contractility. J Cell Biol
102:1738-1747
D'Ardenne AJ, Burns J, Sykes BC, Kirkpatrick P (1983) Compara-
tive distribution of fibronectin and type II1 collagen in normal
human tissues. J Pathol 141:155-169
Ewenstein BM, Warhol HL, Handin R1, Pober JS (1987) Composi-
tion of the yon Willebrand factor storage organelle (Weibel-
Palade body) isolated from cultured human umbilical vein en-
dothelial cells. J Cell Biol 104:1423-1433
Farquhar MG (1978) Structure and function in glomerular capillar-
ies: role of the basement membrane in glomerular filtration.
In: Kefalides, NA (ed.) Biology and Chemistry of Basement
Membranes. Academic Press, New York, NY
Forkert P-G, Thliveris JA, Bertalanffy FD (1977) Structure of sin-
uses in the human lymph node. Cell Tissue Res 183:115-130
Fossum S (1980) The architecture of rat lymph nodes. II. Lymph
node compartments. Scand J Immunol J 2:411-420
Freemont AJ, Stoddart RW, Steven F, Jones CJ, Matthews S
(1986) The structure of the basement-membrane of human
lymph node high endothelial venules: an ultrastructural, histo-
chemical and immunocytochemical study. Histochem J
18 :421-428
Fresen F, Wellensiek RB (1959) Zur elektronoptischen Struktur
des Lymphknotens. Verb Dtsch Ges Pathol 42:353 363
Gridley MF (1951) A modification of the silver impregnation meth-
od of staining reticular fibers. Am J Clin Pathol 21:89%899
Hassel JR, Robey RB, Barrach H J, Wilczek J, Rennard SI, Martin
GR (1980) Isolation of heparan sulfate proteoglycan-containing
proteoglycan from basement-membrane. Proc Natl Acad Sci
USA 77 :4494-4498
Heathcote JG, Grant ME (1981) The molecular organization of
basement-membranes, l nt Rev Connect Tissue Res 9:191-264
Hiserodt JC, Laybourn KA, Varani J (1985) Expression of a la-
minin-like substance on the surface of murine natural killer
(NK) lymphocytes and its role in NK recognition of tumor
target cells. J Immunol 135 : 1484--1487
Hoyer LW (1981) The factor VIII complex. Structure and function.
Blood 58:1-13
Hynes RO (I 987) Integrins: a family of cell surface receptors. Cell
48:549 554
Huard TK, Malinoff HL, Wicha MS (1986) Macrophages express
a plasma membrane receptor for basement membrane laminin.
Am J Pathol 123:365 370
Jaffe EA, Nachman RL, Becker CG, Minick GR (1973) Culture
of human endothelial cells derived from umbilical cord. J Clin
Invest 52:2745-2756
Junqueira LCU, Cossermelli W, Brentani R (1978) Differential
staining of collagens type I, II, and III by Sirus Red and polar-
ization microscopy. Arch Histol Jpn 41 : 267 274
Kalimo H, Lehto M, Nanto-Salonen K, Jalkanen M, Risteli L,
Risteli J, Narva EV (1985) Characterization of the perivascular
reticulin network in a case of primary brain lymphoma. Immu-
nohistochemical demonstration of collagen types I, IIl, IV, and
V; laminin; and fibronectin. Acta Neuropathol 66:299-305
Kleinman HK, Klebe RJ, Martin GR (1981) Role of collagenous
matrices in the adhesion and growth of cells. J Cell Biol
88:473-485
Kleinman HK, McGarvey ML, Liotta LA, Gehron-Robey P,
Tryggvason K, Martin GR (1982) Isolation and characteriza-
tion of type IV collagen, laminin, and heparan sulfate proteo-
glycan from the EHS sarcoma. Biochemistry 21:4969-4974
Kramer RH, Bensch KG, Karasek KA (1987a) Isolation and char-
acterization of human vascular endothelial cells with applica-

tion to studies of the subendothelial matrix. In: Pretlow TG,
II, Pretlow TP (ed) Cell Separation : Methods and Selected Ap-
plications, Academic Press, Orlando, Vol. 5, pp 1-27
Kramer RH, McDonald KA, Berston E, Damsky CH, Ramos
D (1987 b) Model for lymphatic metastasis: Tumor cell interac-
tion with basement-membrane-containing reticular fibers of the
lymph node. J Cell Biol 91a
Laurie GW, Leblond CP (1983) What is known of the production
of basement membrane components. J Histochem Cytochem
31 : 159-163
Laurie GW, Leblond CP, Martin G R (1982) Localization of type
IV collagen, laminin, heparan sulfate proteoglycan, and fibro-
nectin to the basal lamina of basement-membranes. J Cell Biol
95 : 340-344
Lennert K (1978) Malignant lymphomas other than Hodgkin's
disease. Springer, Berlin Heidelberg New York
Linder E, Stenman S, Lehto V-P, Vaheri A (1978) Distribution
of fibronectin in human tissues and relationship to other con-
nective tissue components. Ann N Y Acad Sci 312:151-159
Liotta LA, Rao CN, Barsky SH (1986) Biochemical interactions
of tumor cells with the basement membrane. Annu Rev Bio-
chem 55:1037-1057
Luk SC, Nopajaroonsri C, Simon G T (1973) The architecture of
the normal lymph node and hemolymph node. A scanning and
transmission electron microscopic study. Lab Invest
29 : 258-265
Macarek EJ, Howard PS, Lally ET (1986) Production and charac-
terization of a monoclonal antibody to human type III collagen.
J Histochem Cytochem 34:1003-1011
Martinez-Hernandez A (1980) The hepatic extracellular matrix.
I. Electron immunohistochemical studies in normal rat liver.
Lab Invest 57: 57-74
Montes GS, Krisztan RM, Shigihara KM, Komoro R, Mourano
PAS, Junqueira LCU (1980) Histochemical and morphological
characterization of reticular fibers. Histochemistry 65 : 131-141
Nopajaroonsri C, Luk SC, Simon GT (1971) Ultrastructure of
the normal lymph node. Am J Pathol 65 : 1-24
375
Rodin J A G (1974) Histology, A Text and Atlas. Oxford University
Press, New York, pp 395
Ruoslahti E, Pierschbacher MD (1986) A r g - - G l y - - A s p : A versa-
tile cell recognition signal. Cell 44:517-518
Sakashita S, Engval E, Ruoslahti E (1980) Basement-membrane
glycoprotein laminin binds to heparin. FEBS Lett 116: 243-246
Sorenson G D (1960) An electron microscopic study of popliteal
lymph nodes from rabbits. Am J Anat 107:73-96
Stenman S, Vaheri A (1978) Distribution of a new major connective
tissue protein, fibronectin, in normal human tissues. J Exp Med
147:1054-1064
Stern R (1979) Experimental aspects of hepatic fibrosis. Prog Liver
Dis 7:173-175
Stoolman LM, Frazein A, Burton M (1985) Laminin selectively
enhances the migratory activity of rodent lymphoid cells. J Cell
Biol 101:259a
Timpl R, Fujiwara S, Dziadek M, Aumailley M, Weber S, Engel
J (1984) Laminin, proteoglycan, nidogen, and collagen IV:
structural models and molecular interactions. In: Bernfield, M
(ed) Basement Membranes and Cell Movement. Pitman, pp
25-36
Todd RB, Bowman W (1857) The physiological anatomy and phys-
iology of man. Blanchard and Lea, Philadelphia
Unsworth D J, Scott DL, Walton KW, Walker-Smith JA, Holbrow
EJ (1984) Failure of R1 type anti-reticulin antibody to react
with fibronectin, collagen type III or the non-collagenous com-
ponent (NCRC). Clin Exp Immunol 57:609-618
Van Vliet E, Melis M, Foidart JM, Van Ewijk W (1986) Reticular
fibroblasts in peripheral lymphoid organs identified by a mono-
clonal antibody. J Histochem Cytochem 34:883-890
v o n d e r Mark K (1981) Localization of collagen types in tissues.
Int Rev Connect Tissue Res 9:265-324
Vracko R (1975) Basal lamina scaffold - anatomy and significance
for maintenance of orderly tissue structure. Am J Pathol
77: 313-346
Accepted October 26, 1987