Selective Malformation of the Splenic White Pulp Border in L1-Deficient Mice Shih-Lien Wang, Michael Kutsche, Gino DiSciullo, Melitta Schachner and Steven A. Bogen This information is current as of September 27, 2021. J Immunol 2000; 165:2465-2473; ; doi: 10.4049/jimmunol.165.5.2465 http://www.jimmunol.org/content/165/5/2465 Downloaded from

References This article cites 45 articles, 12 of which you can access for free at: http://www.jimmunol.org/content/165/5/2465.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on September 27, 2021

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2000 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Selective Malformation of the Splenic White Pulp Border in L1-Deficient Mice1

Shih-Lien Wang,* Michael Kutsche,† Gino DiSciullo,* Melitta Schachner,† and Steven A. Bogen2*

Lymphocytes enter the splenic white pulp by crossing the poorly characterized boundary of the marginal sinus. In this study, we describe the importance of L1, an adhesion molecule of the Ig superfamily, for marginal sinus integrity. We find that germline insertional mutation of L1 is associated with a selective malformation of the splenic marginal sinus. Other splenic structures remain intact. Immunofluorescence analysis of the extracellular framework of the , using an Ab to laminin, reveals that L1 knockout mice have an irregularly shaped, discontinuous white pulp margin. Electron microscopic analysis shows that it is associated with bizarrely shaped marginal sinus lining cells at the periphery of the white pulp. These abnormalities correlate with the localization of L1 in normal mice in that L1 is normally expressed on marginal sinus lining cells at the white pulp border. These Downloaded from L1-immunopositive lining cells coexpress high levels of mucosal addressin cell adhesion molecule-1 and vimentin, indicating that they are of fibroblastic lineage and express a well-characterized addressin. Our findings are the first to implicate L1 in splenic lymphoid architectural development. Moreover, these findings help define the poorly characterized sinusoidal boundary across which mononuclear cells cross to enter the splenic white pulp. The Journal of Immunology, 2000, 165: 2465–2473.

ononuclear cells enter lymphoid organs via two dis- nicates, directly or indirectly, with the blood vascular system (4– http://www.jimmunol.org/ tinct routes. Although the blood vascular route has 6). Ultrastructurally, the marginal sinus wall is more analogous to M been a major focus of interest over the last decade, the LN sinusoidal lining cells than blood vascular endothelium. there is a paucity of information about mononuclear cell passage Namely, both types of sinusoidal lining cells associate with a sim- across sinusoidal walls. Specifically, mononuclear cells, such as ilar type of protein matrix, do not assume the tall morphology often , , and dendritic cells, travel from the site associated with postcapillary venules, and do not have a known of Ag contact to a draining lymphoid organ. In the context of a intercellular junction capable of regulating fluid and solute flow. It 3 (LN), mononuclear cells drain into the subcapsular is across this boundary that mononuclear cells must cross to enter sinus, enter the cortical sinusoids, and then migrate into the lym- the splenic white pulp. by guest on September 27, 2021 phoid parenchyma by crossing the relatively poorly defined bound- We are particularly interested in identifying the underlying ad- ary of the sinusoidal lining. This lining comprises a single layer of hesion molecules that mediate structural features of sinusoidal lin- flattened cells associated with collagen fibrils and extracellular ma- ings. To this end, we have focused our attention on the L1 adhe- trix proteins, such as laminin (1, 2). Comparatively little is known sion molecule. L1 is a 1260-aa-long Ig superfamily adhesion about the physiological properties of this lining, such as the ex- molecule (7). It has been implicated in several important neurobi- pression of adhesion molecules, mechanisms of locomotion of ological processes, including neurite outgrowth, neurite fascicula- mononuclear cells along the lining surface, or whether the sinu- tion, axon-Schwann cell interaction, myelination, neuronal cell mi- soidal wall is porous or relatively impermeable to the free flow of gration, and synaptic plasticity (8–12). L1 acts homophilically and macromolecules and cells (3). heterophilically (8, 13, 14). The importance of L1 in neuronal de- In the spleen, the most analogous structure to the LN sinusoid is velopment is reflected by the fact that mutations in the human L1 the marginal sinus. The marginal sinus lining cells are a thin layer gene lead to a group of neurological syndromes (15, 16). of flattened cells that envelop the white pulp. These flattened cells Although L1 was initially identified in the nervous system, L1 is form the inner layer of the marginal sinus, a space that commu- also expressed in nonneuronal tissues. Specifically, L1 is ex- pressed by cells of hematopoietic origin (17, 18), intestinal epi- *Department of Pathology and Laboratory Medicine, Boston University School of thelial cells (19), epithelium of the male urogenital tract (20), and Medicine, Boston, MA 02118; and †Zentrum fu¨r Molekulare Neurobiologie, Univer- sita¨t Hamburg, Hamburg, Germany other cells of epithelial origin (21). The functional role of L1 on Received for publication December 22, 1999. Accepted for publication June 13, 2000. these cells is largely unexplored. L1 has been implicated in an in vitro cell binding assay between lymphocytes and bend 3 endo- The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance thelioma cells, raising the possibility of a potential role in lym- with 18 U.S.C. Section 1734 solely to indicate this fact. phocyte-endothelial cell interactions (22). It is also involved in 1 This work was supported by National Institutes of Health Grant 1R29AI34562 to kidney morphogenesis (23). S.A.B. Previous data from our laboratory implicated the Ig superfamily 2 Address correspondence and reprint requests to Dr. Steven A. Bogen, Department of adhesion molecule L1 in maintaining normal sinusoidal structure Pathology and Laboratory Medicine, Boston University School of Medicine, Boston, MA 02118. E-mail address: [email protected] in LNs during immune hypertrophy (24). Specifically, we found 3 Abbreviations used in this paper: LN, lymph node; FDC, follicular ; that in vivo administration of an L1 mAb disrupted the normal KO, knockout; MMM, marginal metallophilic ; NRIgG, normal nonim- remodeling of the cortical sinusoidal lining cells of LNs during an mune rat IgG; RT, room temperature; DAB, 3,3Ј-diaminobenzidine; SLC, sinus lining cell; RC, reticular cell; MAdCAM-1, mucosal addressin cell adhesion molecule-1; immune response. The L1 mAb did not disrupt static, quiescent PECAM-1, platelet endothelial cell adhesion molecule-1. sinusoidal lining cells. Rather, it interfered only with sinsusoidal

Copyright © 2000 by The American Association of Immunologists 0022-1767/00/$02.00 2466 MALFORMATION OF THE WHITE PULP BORDER IN L1 KO MICE

lining cells mediating the process of matrix remodeling induced by was no opportunity for confusion to arise. The brown DAB precipitate immune stimulation. A limitation with the in vivo L1 Ab study was denoting endogenous peroxidase is easily distinguished on bright-field op- that it required a xenogenic rat anti-mouse L1 mAb. Consequently, tics from the fluorescent signal identifying the location of Ags of interest. The 324 (anti-L1) mAb or NRIgG control were applied at 0.1 ␮g/ml, a mouse (host) anti-rat IgG immune response developed 10–14 Ј followed by donkey anti-rat IgG (Fab )2-biotin at a 1/1000 dilution. All days after L1 mAb administration. This model was therefore useful incubations were performed at RT for 40 min. Subsequently, streptavidin- only in short term studies, such as acute hypertrophic responses to peroxidase and tyramide-FITC were applied and incubated as per the man- immune stimulation. ufacturer’s recommendations. Between each step, the sections were washed three times for 5 min in PBS, 0.1% Tween 20. Because of the structural similarities between the splenic mar- For two-color staining of L1 and MAdCAM-1, we modified the proce- ginal sinus and LN cortical sinusoids, we hypothesize that the dure to distinguish the two rat mAbs with two distinct fluorochromes (flu- splenic marginal sinus may also be reliant on the L1 adhesion orescein and Texas Red). L1 was first detected as described above, except molecule for structural integrity. To elucidate the role of L1 in the that the 324 mAb was applied at a 10-fold lower than normal concentration ␮ development and maintenance of the splenic marginal sinus and (0.01 g/ml). This ultralow Ab concentration can easily be detected using the highly sensitive tyramide signal amplification system (fluorescein sig- the white pulp boundary, we have studied lymphoid matrix devel- nal) but is below the level of detection for the standard immunofluores- opment in an L1-null mutant mouse. cence detection system using Texas Red. Therefore, this helps ensure the fidelity of the two-color discrimination. After completing the first color Materials and Methods (fluorescein) stain, we blocked immunoreactive sites on the donkey anti-rat ␮ Animals IgG-biotin conjugate with NRIgG (60 g/ml). This step blocks unoccupied Ig binding sites on the donkey anti-rat IgG-biotin conjugate reagent. Im- Female L1 heterozygous mice of C57BL/6 background were generated as munoreactive sites on the L1 rat IgG mAb were blocked with a 40-min RT previously described (8). They were bred with either wild-type C57BL/6 incubation of rabbit anti-rat IgG Fab fragment (50 ␮g/ml). The rabbit anti- Downloaded from (The Jackson Laboratory, Bar Harbor, ME) or 129/SvEv (Taconic Farms, rat IgG Fab was prepared by digestion with papain. Fab fragments were Germantown, NY) males. The male offspring were genotyped by PCR for subsequently purified by HPLC. For the second color, sections were detection of the L1 deletion as described (8). Male L1 knockout (KO) mice stained with anti-MAdCAM-1 (4 ␮g/ml) for 40 min at RT. The anti- and their wild-type littermate controls were used at ϳ8–15 wk of age. MAdCAM-1 mAb was detected with donkey anti-rat IgG-Texas Red and Normal 8- to 12-wk-old male BALB/cByJ mice were purchased from The 30 min incubation at RT. Jackson Laboratory. Animals were housed at the Laboratory Animal Sci- For immunocytochemical identification of T and B lymphocytes, T cells were stained with a combination of anti-CD4 and anti-CD8 mAbs, whereas

ence Center, Boston University School of Medicine (Boston, MA), and http://www.jimmunol.org/ cared for in accordance with the Institutional Animal Care and Use Com- B cells were stained with anti-B220 mAb. The anti-CD4 and anti-CD8 mittee of Boston University School of Medicine. mAbs were detected with a rabbit anti-rat IgG-FITC. The anti-B220 mAb was detected with a donkey anti-rat-IgG-Texas Red. Because all the pri- Antibodies mary mAbs are of rat origin, the same previously described procedure modifications (above) were used for ensuring accurate two-color fluores- Normal nonimmune rat IgG (NRIgG) and the clone 324 rat anti-L1 mAb cence discrimination. (IgG) were prepared and purified by HPLC as previously described (24). For the double staining of laminin and marginal metallophilic macro- Other Abs used for indirect immunofluorescence studies were diluted in phages (MOMA-1 Ab), tissue sections were initially stained with PBS, 0.1% BSA: rabbit anti-mouse laminin IgG, 1/500 (Sigma, St. Louis, MOMA-1. The MOMA-1 Ab was detected with a donkey anti-rat IgG MO); normal rabbit serum, 1/1000 (Covance Research Products, Denver, (FabЈ) -biotin conjugate and visualized with avidin-FITC. Then, a rabbit PA); goat anti-rabbit IgG-FITC, 1/1000 (Organon Teknika, West Chester, 2 anti-mouse laminin Ab was applied and detected with donkey anti-rabbit- by guest on September 27, 2021 PA); rat anti-mouse mucosal addressin cell adhesion molecule-1 (MAd- ␮ Texas Red conjugate. The concentrations and incubation time for each CAM-1), MECA367, 4 g/ml (PharMingen, San Diego, CA); MOMA-1 were as previously specified. (for identifying marginal metallophilic macrophages), 1/25 (Serotec, Ra- high low ␮ For the detection of IgM IgD B cells, the sections leigh, NC); rat anti-mouse IgM-FITC, 4 g/ml (Serotec); rat anti-mouse were stained with anti-IgD and detected with donkey anti-rat IgG-Texas B220 mAb clone TIB146 culture supernatant, rat anti-mouse CD4 mAb Red. After incubation with NRIgG (60 ␮g/ml) to block the unoccupied clone TIB207 culture supernatant, rat anti-mouse CD8 mAb clone TIB105 Ig-binding sites on the donkey anti-rat IgG-Texas Red reagent, the sections culture supernatant (all from the American Type Culture Collection, Ma- were stained with anti-mouse IgM-FITC. nassas, VA); NLDC145 (for identifying interdigitating dendritic cells), For the detection of cells, tissue sections were stained hybridoma culture supernatant (25); rat anti-mouse IgD (clone JA12.5), ␮ with biotin-conjugated peanut agglutinin 1/100 (Vector Laboratories) and 3 g/ml, a gift of Dr. Fred Finkelman (26); FDC-M1 (for identifying fol- detected with avidin-FITC 1/1000 (also from Vector Laboratories). licular dendritic cells), a gift of Dr. Marie H. Kosco-Vilbois (27); rabbit For the single-color detection of all other Ags, the specific primary Abs anti-rat IgG-FITC, 1/600 (Vector Laboratories, Burlingame, CA); donkey Ј were applied and detected with appropriate FITC-conjugated secondary Abs. anti-rat IgG (Fab )2-biotin, 1/1000 (Jackson ImmunoResearch, West Grove, PA); donkey anti-rat IgG-Texas Red, 1/75 (Jackson ImmunoRe- search); donkey anti-rabbit IgG-Texas Red, 1/500 (Jackson ImmunoRe- Plastic embedding for light and transmission electron search); avidin-FITC, 1/1000 (Vector Laboratories). microscopy Immunofluorescence microscopy Organs were immersed in Karnovsky’s half-strength fixative for 30 min and then cut with a razor blade into smaller pieces (ϳ8mm3). The tissue For immunofluorescence studies, spleen tissue was embedded in OCT fragments were further fixed in Karnovsky’s half-strength fixative for Ͼ1 (Miles, Elkhart, IN), snap frozen in liquid nitrogen, cryosectioned, and day. Tissues were rinsed with 0.1 M sodium cacodylate with 5% sucrose,

fixed for 30 s in cold acetone. For the detection of L1, we used the tyramide postfixed with 1% OsO4, stained with 2% uranyl acetate and 2% lead signal amplification system (NEN Life Science, Boston, MA). Sections citrate, and then dehydrated through an ethanol gradient. The tissue was were first incubated in 1% newborn calf serum, PBS for 30 min at room then embedded in Epon 812 and cured at 60°C overnight. Semithin sec- temperature (RT). This both hydrated the tissue sections and masked sites tions (1 ␮m) were stained with 1% toluidine blue in 1% sodium borate. The of nonspecific protein adsorption. 0.06-␮m ultrathin sections were examined and photographed in an electron We masked endogenous peroxidase activity by incubating the sections microscope (model 300, Philips Electronics, Eindhoven, The Netherlands). with 3,3Ј-diaminobenzidine (DAB) 0.6 mg/ml in 50 mM Tris buffer, pH 7.6, supplemented with 0.003% hydrogen peroxide for 20 min at RT. This Flow cytometry treatment covers sites of endogenous peroxidase with DAB precipitate, physically encasing the enzyme to prevent further activity and thus effec- Splenocyte suspensions were made by dissociating the through a tively neutralizing the activity of the peroxidase enzyme. It also largely steel mesh. The RBC were lysed in Tris-buffered ammonium chloride (0.14

occludes the excitation and emission light (in a fluorescence detection sys- MNH4Cl, 0.017 M Tris-HCl, pH 7.2) for 1 min at 37°C. Cells were tem, which we used). We found this method to be more effective than resuspended in PBS at a concentration of 20 ϫ 106/ml. Splenocytes (1 ϫ quenching with high concentrations of hydrogen peroxide, especially be- 106/tube) were stained with anti-CD4-FITC, anti-CD8-PE (both from cause tyramide signal amplification is highly sensitive in detecting even PharMingen), or anti-IgM-FITC (Serotec). Cell suspensions were incu- trace levels of residual peroxidase activity. In the system that we describe, bated for 45 min at 4°C. After staining, the cells were washed, fixed in 2% even if a small amount of peroxidase activity was left unquenched, there paraformaldehyde, PBS, and stored at 4°C until assayed. The The Journal of Immunology 2467

FIGURE 1. Expression pattern of laminin in wild-type and L1 KO spleens. Splenic sections from wild type (left)orL1KO(right) mice were stained with rabbit anti-mouse laminin Ab. The double arrows outline the margin of white pulp. Opposing arrows depict the outer and inner boundaries of the marginal sinus. Opposing arrowheads depict the outer and inner boundaries of the marginal zone. The L1 KO section has a fragmented and discontinuous pattern of laminin staining (right). MS, marginal sinus; MZ, marginal zone; RP, ; WP, white pulp. Bar, 50 ␮m. subsets were quantified by flow cytometry (Coulter Profile, Coulter, laminin staining of wild-type spleen outlines the white pulp mar- Miami, FL). gins as a smooth and continuous line. Laminin staining also out- Downloaded from Immunization and ELISA lines some fine matrix material in the marginal zone. As a result, the border of the marginal zone (MZ, arrowheads) and marginal 8 The mice were immunized with 5 ϫ 10 SRBC by i.p. injection on day 0 sinus (MS, opposing arrows) can be clearly distinguished (Fig. 1, and then rechallenged with 3 ϫ 108 SRBC i.p. on day 18. Sera were collected by tail bleed on days 7, 14, and 25. left). By contrast, the laminin staining pattern of the L1 KO spleen Soluble SRBC proteins were extracted using 0.5% Triton X-100 in 300 reveals an irregular, fragmented, and discontinuous white pulp mM NaCl, 50 mM Tris-Cl. The SRBC protein extract was coated onto border. The marginal sinus and marginal zone appear to be mal- polyvinyl chloride (PVC) microtiter plates (Costar, Cambridge, MA) at 20 formed as well (Fig. 1, right). http://www.jimmunol.org/ ␮g/ml overnight at 4°C. The plates were then blocked with 5% BSA for 1 h at RT. Between steps, the plates were washed 5–6 times with PBS, 0.2% Malformation of the splenic marginal sinus in L1 KO mice Tween 20 (PBS/T). Serum, 50 ␮l, diluted either 1:30 or 1:100 in PBS/T, was incubated for1hatRT.Mouse anti-SRBC protein Abs were detected To elucidate the nature of this matrix abnormality, spleens from with an alkaline phosphatase-conjugated rabbit anti-mouse IgM or IgG both wild-type and L1 KO mice were embedded in plastic and (Sigma) after rinsing out unbound mouse serum from the PVC microtiter analyzed by light microscopy (Fig. 2) and transmission electron wells with PBS/T. The alkaline phosphatase-conjugated rabbit anti-mouse IgM or IgG was then incubated for1hatRTatadilution of 1:1000 in microscopy (Fig. 3). In a normal spleen, a flattened layer of sinus PBS/T. Colorimetric development was performed with the alkaline phos- lining cells can be traced along the margin of the white pulp. The phatase substrate, p-nitrophenyl phosphate (Sigma) and was read at 405 nm sinus lining cells have ovoid nuclei and elongated slender pro- by guest on September 27, 2021 in a microplate reader (Bio-Tek Instruments, Winooski, VT). cesses connecting to adjacent lining cells, forming a continuous lining (Fig. 2, left, and Fig. 3, top). In Fig. 2 (left ), the double Results arrows identify the white pulp border. For the most part, it is com- Matrix abnormalities in L1 knockout (KO) spleen prised of elongated slender cytoplasmic processes. A single arrow As a first step in analyzing L1 KO mice, we examined peripheral (Fig. 2, left) identifies the cell body of a marginal sinus lining cell. lymphoid organs at necropsy. LNs and Peyer’s patches were un- In Fig. 3 (top), arrows denote the cytoplasmic processes of two remarkable (data not shown). Splenic architecture, on the other flattened reticular cells. The arrow adjacent to the letters SLC iden- hand, was abnormal. Specifically, there was a striking and selec- tifies a sinus lining cell. The other arrow, adjacent to RC, denotes tive abnormality at the red-white pulp border. a reticular cell defining the outer boundary of the marginal sinus. To generate a view of the splenic white pulp framework, we In contrast, the splenic marginal sinus lining cells in L1 KO performed immunofluorescence microscopy on frozen sections of mice were difficult to find. Marginal sinus lining cells in L1 KO mouse spleen. We stained the spleen sections with a polyclonal mice did not regularly contact other sinus lining cells to form a rabbit anti-mouse laminin Ab. Because laminin is a component of continuous white pulp border. Rather, they displayed an abnormal the reticular matrix of the spleen, it outlines the margins of the stellate, rather than a flattened, linear cell shape. Their cytoplasmic white pulp and vasculature (28). As shown in Fig. 1 (left), the processes were shorter, oriented in random directions, and often

FIGURE 2. Disruption of marginal sinus in L1 KO spleen. In the wild-type spleen (left), the double arrows outline the margin of the white pulp. A pair of opposing arrowheads indicates the inner and outer boundaries of the marginal sinus. In the L1 KO spleen (right), the white pulp margin is difficult to identify. The series of arrowheads identifies what we believe is the probable location of white pulp margin. MS, marginal sinus; MZ, marginal zone; RP, red pulp; SLC, sinus lining cell; WP, white pulp. Bar, 20 ␮m. 2468 MALFORMATION OF THE WHITE PULP BORDER IN L1 KO MICE

shapen sinus lining cells, and a series of arrowheads outline the presumed white pulp border. In L1 KO mice, this border is often difficult to identify. The marginal sinus is absent. One of the stel- late-shaped sinus lining cells is shown at higher magnification in Fig. 3 (bottom) (“SLC”). Normally, the marginal sinus is delimited by sinus lining cells of the white pulp and the opposing reticular cells adjacent to marginal zone. Few RBC are found in the marginal sinus (Fig. 2, left, and Fig. 3, top). However, in the L1 KO spleen, this anatomic distinc- tion is absent. The reticular cells of the marginal zone are rare, and the marginal sinus is no longer distinguishable as a structure dis- tinct from the marginal zone. As a result, RBC migrate right up to the white pulp border (Fig. 2, right, and Fig. 3, bottom). These data demonstrate that the splenic marginal sinus in L1 KO mice is ab- normal. Moreover, the cellular abnormality spatially correlates with the abnormal pattern of laminin immunofluorescence staining as previously shown in Fig. 1.

L1 expression on sinus lining cells Downloaded from We then examined whether L1 expression in wild-type mice spa- tially correlates with the abnormally formed marginal sinus lining cells in L1 KO mice. We expect that if the absence of L1 is the cause of the structural abnormality, then those cells at the white pulp border will likely express L1 under normal circumstances. By examining the splenic structure of normal BALB/cByJ mice, we http://www.jimmunol.org/ localized L1 expression to the edge of the white pulp by immu- nofluorescence staining using an L1 mAb. L1ϩ lining cells are present in the periphery of the white pulp, at the same approximate location as the marginal sinus (Fig. 4). We sometimes found that L1ϩ staining almost completely circumscribed the entire white pulp (Fig. 4, left). However, most of time, only part of white pulp margin was stained (Fig. 4, right, and Fig. 5, top). L1 staining is denoted by double arrows at the white pulp periphery. We also by guest on September 27, 2021 noted intense staining around the central in the center of the white pulp. Two-color staining for tyrosine hydroxylase demon- FIGURE 3. Transmission electron micrographs of the splenic marginal strated that the periarteriolar staining was due to the expression of sinus. Top, wild-type spleen. The marginal sinus, a space delineated by the L1 by sympathetic neurons innervating the spleen (data not SLC and the RC, is clearly identified. Bottom, L1 KO spleen. The marginal shown). However, such neurons did not innervate the marginal sinus cells and reticular cells that delimit the borders of the marginal sinus are not readily identified. The sinus lining cell is abnormally shaped. Also sinus. In addition, lymphocytes within the white pulp also weakly note that RBC come very close to the margin of white pulp. MS, marginal expressed L1 (Fig. 4, left). Low levels of lymphocyte L1 expres- sinus; MZ, marginal zone; RC, reticular cell; SLC, sinus lining cell; WP, sion have been described previously (17). white pulp. Bar, 5 ␮m. The area of the marginal sinus/marginal zone contains a variety of specialized cell types. To pin down the precise location of L1 expression within this area, we performed colocalization studies discontinuous from the adjacent sinus lining cells. These features using two-color immunofluorescence. L1 was stained using fluo- are readily apparent in Fig. 2 (right) and Fig. 3 (bottom). In Fig. 2, rescein while other cell type-specific markers were stained with the red and white pulp can be readily distinguished by the location Texas Red. Optical filters were used that do not allow spillover of pale blue-staining erythrocytes. Arrows (“SLC”) identify mis- into the other respective color. Because the sinus lining cells of the

FIGURE 4. L1 expression on the margin of the white pulp. Splenic sections 32 ␮m thick (left)or6␮m thick (right) from BALB/c mice were stained with the rat anti-mouse L1 mAb 324. Double arrows outline the margin of the white pulp. L1 is strongly expressed on cells at the white pulp margin and cells around the cen- tral artery (left). Also note that some lymphocytes weakly expressed L1 (left). CA, central artery; RP, red pulp; WP, white pulp. Bar, 200 ␮m(left) or 50 ␮m(right). The Journal of Immunology 2469

splenic marginal sinus are described to express MAdCAM-1 (29), we compared the localization of L1 and MAdCAM-1. Fig. 5 shows the staining pattern of L1, MAdCAM-1, and a two-color overlay of both (top, middle, and bottom, respectively). As seen in the bottom panel, L1 and MAdCAM-1 colocalize along the margin of the white pulp, indicating that the L1ϩ cells at the margin of white pulp are sinus lining cells. Moreover, these L1ϩ sinus lining cells sometimes expressed high levels of vimentin (data not shown), often associated with a mesenchymal, probably fibroblast-like or- igin (30). Marginal metallophilic macrophages (MMM) that ex- press MOMA-1, situated at the margin of white pulp, did not ex- press L1 (data not shown). A separate macrophage population in the marginal zone and red pulp areas that expresses CD11b also did not express L1 (data not shown). These results indicate that L1 is expressed by MAdCAM-1ϩ sinus lining cells. Moreover, the location of L1 expression in wild type mice correlates with the location of the structural abnormal- ities (marginal sinus) in L1 KO mice. Downloaded from

Cellular analysis of L1 KO spleens We then examined whether cell populations inside the white pulp or around the marginal sinus were affected by the malformation in the marginal sinus and sinus lining cells in L1 KO mice. The individual cell types were identified with cell type-specific Abs, http://www.jimmunol.org/ either in situ or by flow cytometry. Because the abnormalities we found were concentrated in the marginal sinus and marginal zone, we examined two different cell populations known to inhabit this microanatomic region. Namely, we determined whether MMMs and marginal zone B lymphocytes were present and, if so, whether they are located in their normal anatomic location in L1 KO mice. FIGURE 5. The expression of L1 on MAdCAM-1ϩ sinus lining cells. As previously mentioned, there is a rim of MMMs situated at the Splenic sections (6 ␮m) were stained with 324 mAb (FITC, green color) margin of the white pulp, in the marginal zone. MMMs in L1 KO and anti-MAdCAM-1 (Texas Red, red color). The pictures were taken mice were examined by double staining of splenic sections with by guest on September 27, 2021 using an FITC filter (top), a Texas Red filter (middle) or a double exposure anti-laminin and MOMA-1 (Fig. 6, A and B). The laminin coun- of both filters (bottom). The arrows indicate the margin of white pulp and terstain helps identify the white pulp border. In contrast to the the staining of both L1 and MAdCAM-1. Some nerve fibers around the wild-type mice (and as previously described), L1 KO mice display central arteriole and inside the white pulp also express L1 (top, N). L1 and MAdCAM-1 staining colocalize at the margin of white pulp (bottom). RP, an irregular, poorly defined white pulp border (Fig. 6, A and B). red pulp; WP, white pulp; N, nerve fiber. Bar, 50 ␮m. Nonetheless (and similar to wild-type mice), L1KO mice demon- strate a rim of MMMs located along the white pulp border (Fig. 6, A and B). This finding suggests that L1 is not solely responsible for

FIGURE 6. Cellular analysis of L1 KO spleens. Splenic sections from wild-type (A and C)orL1 KO mice (B and D) were stained to highlight cell subsets within the marginal zone. MMMs are stained green (fluorescein) in A and B and are coun- terstained with rabbit anti-mouse laminin (Texas Red). In A and B, the arrows indicate the white pulp border and the arrowheads denote the MMMs in the marginal zone. The poorly defined white pulp border in L1 KO mice (B). IgMhigh IgDlow marginal zone B lymphocytes are highlighted in C and D. These cells are bright green after staining with for IgD (Texas Red) and IgM (FITC) (C and D). In C and D, opposing arrows point out the IgMhighIgD- low marginal zone B cells. BF, follicle; RP, red pulp; WP, white pulp. Bar, 50 ␮m. 2470 MALFORMATION OF THE WHITE PULP BORDER IN L1 KO MICE their localization in the marginal zone. Similarly, the location and Our most important finding is that L1 serves an important role the approximate number (as estimated using immunofluorescence in the structural integrity of this microanatomic region. Specifi- microscopy) of IgMhighIgDlow marginal zone B cells in L1 KO cally, L1 is expressed on the flattened sinus lining cells at the white mice were comparable with those of wild-type mice (Fig. 6, C and pulp border. The congenital absence of L1 results in nearly a com- D). These cells are those that stain bright green (IgM) along the plete absence of the marginal sinus and disorganization of RCs in edge of the follicle in Fig. 6, C and D. B lymphocytes that coex- the marginal zone. There were obvious gaps in the white pulp press high levels of IgM and IgD are stained orange, from the boundary, as discerned by light and electron microscopy. None- combination of green and red. These results indicate that cell sub- theless, the white pulp mononuclear cells still maintained their sets associated with the marginal zone are not noticeably affected cohesion to each other and the ability to exclude other blood cel- by the splenic structural abnormality in L1 KO mice. lular elements, such as erythrocytes and granulocytes. Although Table I summarizes our findings regarding follicle development the boundary was damaged, there was still a distinction between and segregation of T/B lymphocytes in L1 KO mice. We find these the white and red pulp in L1 KO mice. These observations lead us parameters to be indistinguishable from wild-type mice. Follicular to conclude that other factors besides the structural integrity of the dendritic cells (FDC) and interdigitating dendritic cells were also boundary also regulate cellular traffic into the white pulp. present in their normal locations (B and zones, respectively) The marginal zone, and possibly the marginal sinus, are the and approximate number, as estimated visually. After immuniza- major sites of termination for the branches of central arteriole (4– tion, germinal centers (highlighted using a peanut agglutinin-biotin 6). Consequently, an important function of this region is to selec- conjugate) also developed normally in L1 KO spleens (data sum- tively channel certain cell types into the white pulp. Namely, the marized in Table I). Flow cytometric analysis also showed that the Downloaded from direction of blood flow is toward the red pulp, as indicated by the major T and B lymphocyte subsets in the L1 KO spleen were left-facing arrow (Fig. 8). Representative blood cells such as eryth- present in proportions comparable with those of wild-type litter- rocytes and granulocytes are abundantly found in the red pulp and mates (Table I). marginal zone. Selected mononuclear cells, on the other hand, mi- Immune response of L1 KO mice grate into the white pulp. As a result, the marginal sinus (the area immediately adjacent to the white pulp) is rich in mononuclear To test whether the immune response is impaired in L1 KO mice, http://www.jimmunol.org/ cells but has few other blood cells, such as erythrocytes or gran- we injected L1 KO mice and littermate controls with SRBC i.p. on ulocytes. The mechanisms underlying this selective migration are days 0 and 18. The anti-SRBC IgM and IgG immune responses were measured by ELISA using extracted soluble SRBC protein as poorly understood. One possibility is that most blood cells are Ag. As shown in Fig. 7, there is no significant difference in the passively steered along with the flow of blood toward the splenic IgM and IgG titers of both primary and secondary responses. This red pulp. To enter the splenic white pulp, mononuclear cells prob- result indicates that L1 is not required for a T cell-dependent im- ably must actively locomote, possibly under the influence of che- mune response. moattractant agents such as chemokines. If this hypothesis is cor- rect, then partial defects in the reticular lining of the splenic white Discussion pulp might have minimal effect on the coalescence of mononuclear by guest on September 27, 2021 In this report, we investigate the role of L1 in the development of cells forming a white pulp. normal splenic structure using L1-deficient mice. Using these What role does L1 serve in the development and maintenance of mice, we were able to analyze the role of L1 in lymphoid organ the normal integrity of the marginal sinus lining? We envision two development over longer time periods than otherwise possible us- likely possibilities. First, L1 may be involved in the adhesion of ing L1-specific mAbs. Therefore, these data provide a molecular/ sinus lining cells to their associated reticular fibers. Laminin is a structural correlation, implicating L1 in the development of the component of the extracellular matrix at the white pulp periphery white pulp border. Specifically, we conclude that L1 serves a crit- (28) and is a known ligand for L1 (31). Second, L1 may be in- ical and highly selective role in the development of the splenic volved in the intercellular interaction between adjacent sinus lining marginal sinus. These data help define a poorly characterized si- cells. This interaction may be mediated by a homotypic interaction nusoidal boundary that defines the border of the splenic white pulp, of L1 on adjacent sinus lining cells. and across which mononuclear cells transmigrate to enter the The biological significance of marginal sinus lining cells lies in white pulp. the fact that it is the final physical barrier that mononuclear cells

Table I. Cellular analysis of L1 KO spleen

Analysis Subset Quantity/Location Wild Type (%) L1 KO (%)

Histologicala T/B cell segregation Normal FDC Normal Germinal center development Normal Marginal zone B cells Normal MMM Normal Interdigitating dendritic cells Normal Flow cytometricb CD4ϩ cells 20.4 Ϯ 2.8 18.5 Ϯ 2.7 CD8ϩ cells 15 Ϯ 2.4 17.8 Ϯ 3 IgMϩ cells 45.5 Ϯ 3 46.9 Ϯ 4.8

a Single-color or two-color immunofluorescence microscopy were performed on splenic sections of both L1 KO mice and wild-type littermate controls. The specific cell subsets were stained with cell type-specific Abs as described in Materials and Methods. For all subsets, “normal” refers to both approximate quantity and tissue location. b For flow cytometric analysis, splenocytes from three L1 KO mice and three wild-type littermate controls (in C57BL/6 background) were stained with fluorochrome-conjugated mAbs. The numbers shown respresent the mean Ϯ SD. The Journal of Immunology 2471

zone are relatively open spaces as compared with a postcapillary venule. Consequently, it is likely that the initial rolling step, ne- gotiating initial contact between a leukocyte and endothelium, is absent in the marginal sinus. Fig. 8 is a schematic representation of the microanatomy at the red-white pulp border of wild-type mice. The white and red pulp boundaries, marginal sinus, and marginal zone are delineated by flattened, elongated reticular cells. These reticular cells comprise the marginal sinus lining cells and reticular cells of the marginal zone. They are distinct from endothelium, because they do not assume a tall, activated morphology and do not express CD31 (platelet endothelial cell adhesion molecule-1) (PECAM-1), a marker for endothelial cells. Also present at the red-white pulp border are some rather unique cell types, including MMMs, a sub- set of B lymphocytes expressing the IgMhighIgDlow immunophe- notype, and a subset of marginal zone macrophages. Of minor note is that we observed the MMMs to be located just outside the white pulp boundary. A previous description of MMMs placed them just inside the white pulp boundary (5). This location was discerned Downloaded from using two-color immunofluorescence; MOMA-1 Abs identified MMMs whereas laminin identified the extracellular matrix of the white pulp boundary. Apart from L1, the only other well-characterized adhesion mol- ecule expressed by marginal sinus lining cells is MAdCAM-1 (29). MAdCAM-1 is also an Ig superfamily adhesion molecule. It is http://www.jimmunol.org/ FIGURE 7. Immune response of L1 KO mice to SRBC. Four male L1 expressed by mucosal venules and is responsible for the homing of KO mice and their littermate controls (in a C57BL/6 ϫ 129 background) lymphocytes to Peyer’s patches and the intestinal lamina propria were immunized with 5 ϫ 108 SRBC i.p. on day 0 and rechallenged with (33, 34). Its function on marginal sinus lining cells is not known. 3 ϫ 108 SRBC i.p. on day 18. Sera were collected at days 7, 14, and 25. In this regard, L1 is the only adhesion molecule that is demon- The titers of anti-SRBC IgM (A) and anti-SRBC IgG (B) were determined strated to directly serve a function in the development and pres- by ELISA. Mean Ϯ SE is represented. For both IgM and IgG titers, there ervation of the splenic white pulp boundary. are no statistically significant differences between the L1 KO mice and In this paper, we describe that L1 expression on marginal sinus littermate controls.

lining cells is somewhat heterogeneous. Namely, marginal sinus by guest on September 27, 2021 lining cells do not express L1 all the time. We believe that L1 is must traverse before entering the white pulp (32). This “transmi- temporally regulated, in response to as yet unidentified signals. gration” across the white pulp sinusoidal border is as yet poorly One of these signals may the cytokine TNF-␣. Pancook et al. (18) characterized. Certain features suggest that novel mechanisms, dis- identified TNF-␣ as capable of up-regulating L1 expression on tinct from those in blood vascular transmigration, are involved. For human dendritic cells. In addition, our previously published find- example, the shear force and blood flow velocity in the marginal ings suggest a temporal functional role for L1. We previously de- sinus may be lower than those in postcapillary venules. This sup- scribed our findings that L1 antagonism (with an L1 mAb) dis- position is based on the fact that the marginal sinus and marginal rupted lymph node sinusoidal architecture, but only in LN

FIGURE 8. Schematic representation of mi- croanatomy of white-red pulp border. PMN, polymorphonuclear leukocyte. 2472 MALFORMATION OF THE WHITE PULP BORDER IN L1 KO MICE undergoing remodeling after immune stimulation (24). Static, qui- tionally redundant molecules. Without such redundancy, the L1- escent LN were not affected. Thus, L1 appears to be important null genotype is lethal or nearly so. With increasing levels of during certain processes, such as matrix remodeling. L1 expression redundant molecule(s), the L1-null genotype has little to no phe- may therefore be temporally regulated, in response to immunolog- notypic effect. There is ample precedent for such functional re- ical or inflammatory stimuli. We speculate that the heterogeneity dundancy, such as among the selectins in mediating initial contact in L1 expression among different marginal sinus lining cells may between leukocytes and vascular endothelium. simply be due to variations in the basal rate of splenic white pulp We previously reported that administration of an L1 mAb dur- remodeling. ing an immune response disrupted the normal remodeling of the The pattern of L1 expression on sinusoids that border lymphoid fibroblastic reticular system in LNs. In these L1 KO mice, we did parenchyma suggests an analogous role to PECAM-1 (CD31) ex- not find any obvious abnormalities in LNs. This discrepancy might pression on vascular endothelium. Both interact in a homo- and be due to functional compensation by other, as yet unidentified, heterotypic fashion, are expressed at low levels on hematopoieti- L1-like molecules in LNs of L1 KO mice. Several L1 homologues cally derived cells, and are both in the Ig superfamily. PECAM-1 have been identified in the nervous system (9, 39, 40), but their serves a role in the development and maintenance of vascular en- expression outside the nervous system is largely unexplored. dothelium and leukocyte trafficking across vascular boundaries Moreover, our previous report utilized a distinct model involving (35). Based on these similarities, it is possible that L1 may serve acute hypertrophic responses after immunization in normal mice. It a similar role for mononuclear cell trafficking into the white pulp. is possible that the sudden disruption of L1 function by L1-specific We found surprisingly little functional consequence of the struc- Abs in a normal mouse may have greater structural consequences tural abnormality in L1 KO mice. All major cell populations inside than congenital absence of L1. Downloaded from the white pulp and around the marginal sinus appeared to be Several genetically targeted mutant strains of mice, notably the present in normal proportions. Immune responses to SRBC also family of lymphotoxin-␣, lymphotoxin-␤, TNF, and their recep- indicated that L1 was not required for a T dependent immune tors, have been reported to have structural abnormalities in periph- response. Immune responses to a soluble Ag, keyhole limpet he- eral lymphoid organs (41). For example, the lymphotoxin-␣ KO mocyanin, were also not appreciably different in L1 KO mice (data mice developed structural abnormalities of the spleen, and a com-

not shown). The absence of clear functional consequences suggests plete absence of LNs and Peyer’s patches (42). The TNF KO mice http://www.jimmunol.org/ that either 1) the structural abnormalities were not sufficiently se- had a decreased number of Peyer’s patches and a defect in FDC vere so as to cause functional deficits or 2) humoral immune re- development (43). We have observed that TNF and TNF receptor sponses are not a sensitive indicator of abnormalities in the mar- I/II KO mice have a similar irregularity of the white pulp border as ginal sinus. that of L1 KO mice (data not shown). As expected, these mice also In contrast to our studies in mice, the existence of the marginal show an abnormal distribution of L1ϩ cells at the white pulp bor- sinus in the human spleen is controversial. For example, van der. A naturally mutant mouse strain, aly/aly, also demonstrated Krieken et al. (36) reported finding no evidence for the presence of severe developmental abnormalities in all peripheral lymphoid or- a human splenic marginal sinus. Schmidt et al. (37), on the other gans (44). Of particular interest, aly/aly mutant mice also had de- hand, used corrosion cast scanning electron microscopy to clearly velopmental abnormalities in the splenic marginal sinus (45). by guest on September 27, 2021 demonstrate the presence of a marginal sinus in the human spleen. However, unlike all these other mutant strains, our L1 deficient Interestingly, Steiniger et al.(38) also failed to find a human marginal mice had a selective defect on splenic marginal sinus development; sinus. They attributed the findings of Schmidt et al. to capillaries in the no other structures were affected. “perifollicular zone,” a zone that they claimed to be located between In summary, our findings indicate that L1 serves a crucial role the marginal zone and the red pulp (38). In light of these contradictory in the proper development of the architecture at the white pulp conclusions over the very existence of a human splenic marginal si- border. The most notable defect resulting from this improper for- nus, it is not surprising that the function of the marginal sinus remains mation of white pulp lining was the near complete absence of the unclear. The identification of a new cell surface adhesion marker (L1) marginal sinus. These structural abnormalities correlate with the on marginal sinus lining cells may ultimately help contribute toward location of L1 expression in normal mice. By characterizing the an understanding of the function of marginal sinus lining cells and molecular and functional features of the white pulp lining, we be- thereby clarify this mystery. lieve it may help shed light on the function of the splenic marginal Throughout our investigation, we noticed a varying degree of sinus. splenic structural abnormalities in L1 KO mice. Similar variability has been noticed in CNS abnormalities of L1 KO mice. Consistent References with the previous report (8), the genetic background of L1 KO 1. Kramer, R. H., S. D. Rosen, and K. A. McDonald. 1988. Basement-membrane mice influences the resulting phenotype. The C57BL/6 background components associated with extracellular matrix of lymph node. Cell Tissue Res. yielded few viable L1 KO offspring. However, the few L1 KO 252:367. 2. Ushiki, T., O. Ohtani, and K. Abe. 1995. Scanning electron microscopic studies mice that were born in the C57BL/6 background demonstrated the of reticular framework in the rat mesenteric lymph node. Anat. Rec. 241:113. most severe phenotypic abnormalities. By contrast, backcrossing 3. Gretz, J. E., A. O. Anderson, and S. Shaw. 1997. Cords, channels, corridors and ϫ conduits: critical architectural elements facilitating cell interactions in the lymph of L1 female heterozygotes to a C57BL/6 129 F1 background node cortex. Immunol. Rev. 156:11. yielded relatively normal proportions of L1 KO male offspring. 4. Weiss, L. 1965. The structure of the normal spleen. Semin. Hematol. 11:205. However, we found few abnormalities in the lymphoid systems of 5. Kraal, G. 1992. Cells in the marginal zone of the spleen. Int. Rev. Cytol. 132:31. ϫ 6. Schmidt, E. E., I. C. MacDonald, and A. C. Groom. 1988. Microcirculatory these (C57BL/6 129 F1 background) L1 KO mice. These find- pathways in normal human spleen, demonstrated by scanning electron micros- ings are reminiscent of the broad range of phenotypic abnormali- copy of corrosion casts. Am. J. Anat. 181:253. ties found in human genetic diseases involving L1 mutations (9, 7. Moos, M., R. Tacke, H. Scherer, D. Teplow, K. Fruh, and M. Schachner. 1988. Neural adhesion molecule L1 as a member of the immunoglobulin superfamily 16). The system is therefore probably multigeneic, with at least with binding domains similar to fibronectin. Nature 334:701. two functionally redundant proteins. We hypothesize that other 8. Dahme, M., U. Bartsch, R. Martini, B. Anliker, M. Schachner, and N. Mantei. adhesion molecules may also likely contribute to the normal in- 1997. Disruption of mouse L1 gene leads to malformations of the nervous system. Nat. Genet. 17:346. tegrity of the structural features we described. Presumably, differ- 9. Hortsch, M. 1996. The L1 family of neural cell adhesion molecules: old proteins ent strains of mice may potentially express varying levels of func- performing new tricks. Neuron 17:587. The Journal of Immunology 2473

10. Seilheimer, B., E. Persohn, and M. Schachner. 1989. Antibodies to the L1 ad- 29. Kraal, G., K. Schornagel, P. R. Streeter, B. Holzmann, and E. C. Butcher. 1995. hesion molecule inhibit Schwann cell ensheathment of neurons in vitro. J. Cell Expression of the mucosal vascular addressin, MAdCAM-1, on the sinus-lining Biol. 109:3095. cells in the spleen. Am. J. Pathol. 147:763. 11. Cohen, N. R., J. S. H. Taylor, L. B. Scott, R. W. Guillery, P. Soriano, and 30. Osborn, M., E. Debus, and K. Weber. 1984. Monoclonal antibodies specific for A. J. W. Furley. 1997. Errors in corticospinal axon guidance in mice lacking the vimentin. Eur. J. Cell Biol. 34:137. ␤ neural cell adhesion molecule L1. Curr. Biol. 8:26. 31. Hall, H., S. Carbonetto, and M. Schachner. 1997. L2/HNK-1 carbohydrate and 1 12. Schachner, M. 1997. Neural recognition molecules and synaptic plasticity. Curr. integrin dependent neural cell adhesion to laminin-1. J. Neurochem. 68:544. Opin. Cell Biol. 9:627. 32. Goldschneider, I., and D. D. McGregor. 1968. Migration of lymphocytes and 13. Grumet, M., and G. M. Edelman. 1988. Neuron-glia cell adhesion molecule in- thymocytes in the rat. ⌱. The route of migration from blood to spleen and lymph teracts with neurons and astroglia via different binding mechanisms. J. Cell Biol. nodes. J. Exp. Med. 127:155. 106:487. 33. Nakache, M., E. J. Berg, P. R. Streeter, and E. C. Butcher. 1989. The mucosal 14. Burden-Gulley, S. M., M. Pendergast, and V. Lemmon. 1997. The role of cell vascular addressin is a tissue specific endothelial cell adhesion molecule for cir- adhesion molecule L1 in axonal extension, growth cone motility and signal trans- culating lymphocytes. Nature 337:179. duction. Cell Tissue Res. 290:415. 34. Briskin, M. J., L. M. McEvoy, and E. C. Butcher. 1993. MAdCAM-1 has ho- 15. Wong, E. V., S. Kenwrick, P. Willems, and V. Lemmon. 1995. Mutations in the mology to immunoglobulin and mucin-like adhesion receptor and to IgA1. Na- cell adhesion molecule L1 cause mental retardation. Trends Neurosci. 18:168. ture 363:461. 16. Fransen, E., G. V. Camp, L. Vits, and P. J. Willems. 1997. L1-associated dis- 35. Bogen, S. A., J. Pak, M. Garifallou, X. Deng, and W. A. Muller. 1994. Mono- eases: clinical geneticists divide, molecular geneticists unite. Hum. Mol. Genet. clonal antibody to murine PECAM-1 (CD31) blocks acute inflammation in vivo. 6:1625. J. Exp. Med. 179:1059. 17. Kowitz, A., G. Kadmon, M. Eckert, V. Schirrmacher, M. Schachner, and 36. van Krieken, J. H. J. M., J. Te Velde, K. Kleiverda, L. Leenheers-Binnendijk, and P. Altevogt. 1992. Expression and function of the neural cell adhesion molecule C. J. H. van de Velde. 1985. The human spleen: a histological study in splenec- L1 in mouse leukocytes. Eur. J. Immunol. 22:1199. tomy specimens embedded in methylmethacrylate. Histopathology 9:571. 18. Pancook, J. D., R. A. Reisfeld, N. Varki, A. Vitiello, R. I. Fox, and 37. Schmidt, E. E., I. C. MacDonald, and A. C. Groom. 1993. Comparative aspects A. M. Montgomery. 1997. Expression and regulation of the neural cell adhesion of splenic microcirculatory pathway in mammals: the region bordering the white molecule L1 on human cells of myelomonocytic and lymphoid origin. J. Immu- pulp. Scanning Microsc. 7:613. nol. 158:4413. 38. Steiniger, B., P. Barth, B. Herbst, A. Hartnell, and P. R. Crocker. 1997. The Downloaded from 19. Thor, G., R. Probstmeier, and M. Schachner. 1987. Characterization of the cell species-specific structure of microanatomical compartments in the human spleen: adhesion molecules L1, N-CAM and J1 in the mouse intestine. EMBO J. 6:2581. strongly sialoadhesin-positive macrophages occur in perifollicular zone, but not 20. Kujat, R., F. Miragall, D. Krause, R. Dermietzel, and K. Wrobel. 1995. Immu- in the marginal zone. Immunology 92:307. nolocalization of the neural cell adhesion molecule L1 in non-proliferating epi- 39. Holm, J., R. Hillenbrand, V. Steuber, U. Bartsch, M. Moos, H. Lubbert, thelial cells of the male urogenital tract. Histochem. Cell Biol. 103:311. D. Montag, and M. Schachner. 1996. Structural features of a close homologue of 21. Nolte, C., M. Moos, and M. Schachner. 1999. Immunolocalization of the neural L1 (CHL1) in the mouse: a new member of the L1 family of neural recognition cell adhesion molecule L1 in epithelia of rodents. Cell Tissue Res. 298:261. molecules. Eur. J. Neurosci. 8:1613.

22. Hubbe, M., A. Kowitz, V. Schirrmacher, M. Schachner, and P. Altevogt. 1993. 40. Moscoso, L. M., and J. R. Sanes. 1995. Expression of four immunoglobulin http://www.jimmunol.org/ L1 adhesion molecule on mouse leukocytes: regulation and involvement in en- superfamily adhesion molecules (L1, Nr-CAM/Bravo, Neurofascin/ABGP, and dothelial cell binding. Eur. J. Immunol. 23:2927. N-CAM) in the developing mouse spinal cord. J. Comp. Neurol. 352:321. 23. Debiec, H., E. I. Christensen, and P. M. Ronco. 1998. The cell adhesion molecule 41. Fu, Y.-X., and D. D. Chaplin. 1999. Development and maturation of secondary L1 is developmentally regulated in the renal epithelium and is involved in kidney lymphoid tissues. Annu. Rev. Immunol. 17:399. branching morphogenesis. J. Cell Biol. 143:2067. 42. DeTogni, P., J. Goellner, N. H. Ruddle, P. R. Streeter, A. Fick, S. Mariathasan, 24. DiSciullo, G., T. Donahue, M. Schachner, and S. A. Bogen. 1998. L1 antibodies S. C. Smith, R. Carlson, L. P. Shornick, J. Strauss-Schoenberger, et al. 1994. block lymph node fibroblastic reticular matrix remodeling in vivo. J. Exp. Med. Abnormal development of peripheral lymphoid organs in mice deficient in lym- 187:1953. photoxin. Science 264:703. 25. Kraal, G., M. Breel, E. M. Janse, and G. Brown. 1986. Langerhans cells, veiled 43. Pasparakis, M., L. Alexopoulou, M. Grell, K. Pfizenmaier, H. Bluethmann, and cells, and interdigitating cells in the mouse recognized by a monoclonal antibody. G. Kollias. 1997. Peyer’s patch organogenesis is intact yet formation of B lym- J. Exp. Med. 163:981. phocyte follicles is defective in peripheral lymphoid organs of mice deficient for

26. Finkelman, F. D., A. Lees, R. Birnbaum, W. C. Gause, and S. C. Morris. 1996. tumor necrosis factor and its 55-kDa receptor. Proc. Natl. Acad. Sci. USA 94: by guest on September 27, 2021 Dendritic cells can present antigen in vivo in a tolerogenic or immunogenic 6319. fashion. J. Immunol. 157:1406. 44. Miyawaki, S., Y. Nakamura, H. Suzuka, M. Koba, R. Yasumizu, S. Ikehara, and 27. Kosco, M. H., E. Pflugfelder, and D. Gray. 1992. Follicular dendritic cell depen- Y. Shibata. 1994. A new mutation, aly, that induces a generalized lack of lymph dent adhesion and proliferation of B cell in vitro. J. Immunol. 148:2331. nodes accompanied by immunodeficiency in mice. Eur. J. Immunol. 24:429. 28. van den Berg, T. K., M. van der Ende, E. A. Dopp, G. Kraal, and C. D. Dijkstra. 45. Koike, R., T. Nishimura, R. Yasumizu, H. Tanaka, Y. Hataba, T. Watanabe, ␤ 1993. Localization of 1 integrins and their extracellular ligands in human lym- S. Miyawaki, and M. Miyasaka. 1996. The splenic marginal zone is absent in phoid tissues. Am. J. Pathol. 143:1098. alymophoplastic aly mutant mice. Eur. J. Immunol. 26:669.