By Follicular Dendritic Cells FDC-SP, a Novel Secreted Protein Expressed

The Journal of Immunology

FDC-SP, a Novel Secreted Protein Expressed by Follicular Dendritic Cells1

Aaron J. Marshall,2*† Quijiang Du,* Kevin E. Draves,† Yasufumi Shikishima,* Kent T. HayGlass,* and Edward A. Clark†

To define better the molecular basis for follicular dendritic cell (FDC) function, we used PCR-based cDNA subtraction to identify genes specifically expressed in primary FDC isolated from human tonsils. In this work we report the discovery of a novel gene encoding a small secreted protein, which we term FDC-SP (FDC secreted protein). The FDC-SP gene lies on chromosome 4q13 adjacent to clusters of proline-rich salivary peptides and C-X-C chemokines. Human and mouse FDC-SP proteins are structurally unique and contain a conserved N-terminal charged region adjacent to the leader peptide. FDC-SP has a very restricted tissue distribution and is expressed by activated FDCs from tonsils and TNF-␣-activated FDC-like cell lines, but not by B cell lines, primary germinal center B cells, or anti-CD40 plus IL-4-activated B cells. Strikingly, FDC-SP is highly expressed in germinal center light zone, a pattern consistent with expression by FDC. In addition, FDC-SP is expressed in leukocyte-infiltrated tonsil crypts and by LPS- or Staphylococcus aureus Cowan strain 1-activated leukocytes, suggesting that FDC-SP can also be produced in response to innate immunity signals. We provide evidence that FDC-SP is posttranslationally modified and secreted and can bind to the surface of B lymphoma cells, but not T lymphoma cells, consistent with a function as a secreted mediator acting upon B cells. Furthermore, we find that binding of FDC-SP to primary human B cells is markedly enhanced upon activation with the T-dependent activation signals such as anti-CD40 plus IL-4. Together our data identify FDC-SP as a unique secreted peptide with a distinctive expression pattern within the immune system and the ability to specifically bind to activated B cells. The Journal of Immunology, 2002, 169: 2381–2389.

lymphocyte development, survival, and activation are structure that supports B cell survival, affinity maturation of Ab re- critically dependent on interactions with specialized non- sponses, and the generation of long-lived B cell memory (7, 14–19). B hematopoietic cells of mesenchymal origin. In primary Primary human or mouse FDCs (20–23) and human FDC cell lymphoid organs such as bone marrow and fetal liver, developing lines (24–27) can directly stimulate B lymphocyte survival and pre-B lymphocytes undergo their differentiation program in close proliferation, Ig secretion, and expression of costimulatory mole- proximity with cells referred to as stromal cells (1–4). Stromal cules in vitro. However, the molecular basis for FDC-B cell inter- cells support survival, proliferation, and differentiation of B cell actions remains poorly characterized. Several mAbs have been precursors using both contact-dependent (membrane bound) and generated that differentially stain FDCs in tissue sections (28–31). soluble growth factors, several of which have been identified and Most of the markers recognized by these mAbs are unknown, but well characterized. In secondary lymphoid organs, mature B lym- they have been useful for FDC identification and study of FDC phocytes interact with cells known as follicular stromal cells or development. Expression cloning studies led to identification of follicular dendritic cells (FDCs),3 which are present in primary and markers recognized by two such mAbs as a novel splicing isoform secondary follicles (5–9). One defining characteristic of FDCs is of the complement receptor CD21 (32) and a novel transmembrane their ability to trap immune complexes and act as Ag depots during protein that appears to be involved in FDC-B lymphocyte interac- the germinal center (GC) response (10–13). While immune complex tions (33, 34). FDC also produce the chemokine B lymphocyte deposition on FDCs may not be essential for initiation of GC re- chemoattractant (BLC; also known as CXCL13) (35), which is the sponses (14), FDCs play a central role in organization of the follicular most efficacious chemoattractant for B lymphocytes known and is indispensable for organizing the follicular structure in lymphoid organs (36, 37). Induction of mature, BLC-producing FDC net- † *Department of Immunology, University of Manitoba, Winnipeg, Canada; and De- works requires interactions between FDC precursors and B cells partment of Microbiology, University of Washington, Seattle, WA 98195 expressing the TNFR family ligands lymphotoxin ␣ ␤ and TNF Received for publication November 1, 2001. Accepted for publication June 24, 2002. 1 2 (38–40) and, reciprocally, BLC can induce expression of mem- 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 brane lymphotoxin 1 2 on B cells (36). Thus, bidirectional sig- with 18 U.S.C. Section 1734 solely to indicate this fact. naling mechanisms between FDCs and B cells are critical for es- 1 This work is supported by National Institutes of Health Grant AI44257 (to E.A.C.) tablishment of a functional follicular structure. Furthermore, and a Manitoba Health Research Council Operating Grant (to A.J.M.). A.J.M. is FDC-B cell interactions may play a critical role in organizing the supported by a Canadian Institutes of Health Research New Investigator award. K.T.H. holds a Canada Research Chair in Immune Regulation. follicle-like structures observed in multiple nonlymphoid tissues 2 Address correspondence and reprint requests to Dr. Aaron J. Marshall, Department during chronic immune reactions (41–47). of Immunology, University of Manitoba, 611 Basic Medical Sciences Building, Win- To define better the molecular basis for FDC-B cell interactions, we nipeg, MB R3E 0W3, Canada. E-mail address: [email protected] devised a strategy to isolate genes specifically expressed in primary 3 Abbreviations used in this paper: FDC, follicular dendritic cell; FDC-SP, FDC se- human FDCs. In this work we report the identification of a set of creted protein; GC, germinal center; EST, expressed sequence tag; SAC, Staphylo- coccus aureus Cowan; SSH, suppression subtractive hybridization; BLC, B lympho- FDC-associated genes cloned by this approach and the initial charac- cyte chemoattractant. terization of FDC-SP, a gene encoding a novel secreted protein.

Materials and Methods were randomly picked and screened for differential expression using tester Isolation of FDC-associated genes and driver cDNA probes. Approximately 70% of the clones appeared to be differentially expressed as assessed by hybridization with tester and driver Highly enriched FDC populations were obtained from human tonsils ac- cDNA probes (data not shown). Twenty-nine differentially expressed cording to the method of Liu et al. (32). Briefly, tonsils were minced in clones were sequenced by dye-terminator sequencing. Sequences are sum- RPMI 1640 medium containing FCS and collagenase IV (Sigma-Aldrich, marized in Table I. St. Louis, MO) and the released cells were harvested. The low buoyant density fraction was obtained using a Percoll gradient (40/45% interface), Northern blot, in situ hybridization, and RT-PCR then cells were stained for CD21 and CD14 expression and the FDC pop- A 32P-labeled FDC-SP cDNA probe was hybridized to human multiple ulation was isolated by FACS using the indicated gate (Fig. 1). Recoveries ϩ ϩ tissue Northern blots (Clontech Laboratories) and to a Northern blot con- of CD14 ϩ CD21 cells from the low-density population varied between 0.5 taining 2 ␮g of poly(A) RNA isolated from human tonsils, according to and 2%. Total RNA was isolated and pooled RNA from three sorted prep- the manufacturer’s protocol. For RT-PCR analysis of FDC-SP expression, arations (1 ϫ 106 cell equivalents) was used to generate double-stranded 0.5–1 ϫ 106 cell equivalents of total RNA from the indicated cells was cDNA by the SMART (switching mechanism at RNA termini) PCR cDNA reverse transcribed using avian myeloblastosis virus reverse transcriptase synthesis method (Clontech Laboratories, Palo Alto, CA). A double- and random hexamer primers (Promega, Madison, WI). One-tenth of the stranded driver cDNA was concomitantly produced by the same method resulting cDNAs were subjected to PCR amplification for 30 cycles using using an equal mixture of RNA from the human fibroblast cell line HFF FDC-SP-specific primers CAGCGTCAGAGAGAAAGAACTGACTG and the epithelial line HeLa (Fig. 1). A suppression subtractive hybridiza- and TACTTTTCGCTAGGAAGGGGAGTTG or G3PDH control primers tion (SSH) PCR subtraction procedure was conducted according to the TGAAGGTCGGAGTCAACGGATTTGGT and CATGTGGGCCATGA manufacturer’s protocol (PCR Select; Clontech Laboratories), using the GGTCCACCAC. The resulting PCR products were run on agarose gels FDC cDNA as tester and the fibroblast/epithelial cell cDNA as driver. and either stained with ethidium bromide or Southern blotted and hybrid- The pool of differentially expressed gene fragments generated were then ized with appropriate oligonucleotide probes. cloned into the pCRII vector (Invitrogen, Carlsbad, CA) and 60 clones In situ hybridization was conducted essentially as described (48). Briefly, tonsil sections were deparaffinized, treated with proteinase K (Sig- ma-Aldrich), washed, prehybridized, and then hybridized overnight at 50°C with hybridization solution containing 106 cpm of 35S-labeled probe. Sense and antisense probes were generated by in vitro transcription of plasmid templates containing the FDC-SP cDNA in forward and reverse orientations in the presence of [35S]UTP. After hybridization, slides were rinsed, digested with RNase A (type XI; Sigma-Aldrich), and washed twice with 2ϫ SSC at room temperature and twice with 0.1ϫ SSC at 50°C for 40 min each. After washing, slides were dehydrated with successive eth- anol rinses, air dried, dipped in emulsion (Kodak NTB2 diluted 1/1 in

Table I. Identity of differentially expressed clonesa

No. of Gene Sequencesb Description

HLA DR, DQ 5 MHC class II Ig ␭ L chain 4 Ig chain CD53 3 Tetraspanin family receptor on hematopoeitic cells CD52/CAMPATH 2 PI-linked surface molecule, lymphocyte specific Jaw1 2 Endoplasmic reticulum membrane protein, lymphocyte specific, involved in Ag processing CD20 2 B cell differentiation Ag CD69 1 Activation Ag of lymphocytes, monocytes, NK cells Ly-GDI 1 Inhibitor of Rho family GTPases, lymphoid specific CD83/HB15 1 Ig superfamily receptor, on dendritic cells and activated lymphocytes GM2 activator 1 Activator of ganglioside degradation in lysosomes Ig ␬ L chain 1 Ig chain RGS1/BL34 1 Regulator of heterotrimeric G proteins, expressed in activated B cells L-plastin 1 Actin-binding protein in hematopoietic cells, involved in cell adhesion HEAB 1 Putative ATP-binding protein, unknown function FDC-SP 1 Novel secreted protein Bam32 1 Novel B lymphocyte-specific FIGURE 1. Strategy for isolation of FDC-associated genes. A, Low buoy- signal transduction protein ant density tonsil cells were isolated and stained for CD21 and CD14 expres- DCAL-1 1 Novel C-type lectin sion. The gate used for isolation of FDC by cell sorting is shown. Single-color a controls confirmed that the CD14 staining is not due to bleedthrough between Differentially expressed clones were isolated as described in Fig. 1 and 29 clones were sequenced. channels, but represents specific staining of a population with very high for- b The number of independent clone sequences that corresponded to the indicated ward scatter. B, Flow chart showing the gene cloning strategy. genes. The Journal of Immunology 2383 water), exposed for 3–10 days at 4°C, and then developed. Every second preparation, immunoprecipitation, and Western blotting were conducted as slide in the series was counterstained with H&E to confirm GC structure. described (52).

Cell cultures FDC-SP binding assay For analysis of FDS-SP expression and binding in primary human leuko- L cells were transfected with wild-type or mutant FDC-SP-myc vector cytes, mononuclear cells were prepared from human blood as described using Lipofectamine (Invitrogen) and stable clones were generated by (49) and cultured for 1–3 days at 1 ϫ 106 per milliliter in RPMI/10% FCS G418 selection. The indicated cell lines or primary cells were added to medium. The following concentrations of stimuli were used: 100 ng/ml semiconfluent L cell cultures, cocultured overnight, and harvested. Cells LPS (Sigma-Aldrich), a 1/25,000 dilution of Staphylococcus aureus were then stained for detection of myc-tagged protein on the cell surface Cowan strain 1 (SAC; Calbiochem, La Jolla, CA), 10 ng/ml TNF-␣ (Pepro- using FITC-labeled anti-myc Ab (Invitrogen) or FITC-labeled Mopc21 iso- Tech, Rocky Hill, NJ), or 1 ␮g/ml anti-CD40 (BD PharMingen, San Diego, type control Ab (BD PharMingen) and analyzed on a FACSCalibur instru- CA) plus 2 ng/ml IL-4 (PeproTech). Cells were harvested at the indicated ment (BD Biosciences, Mountain View, CA). In some experiments, cells times and counted, and RNA was prepared using TRIzol (Life Technolo- were also stained with biotinylated anti-CD19 (BD PharMingen) followed gies, Rockville, MD). For some experiments, after 3 days of stimulation by streptavidin-PE (Jackson ImmunoResearch Laboratories, West Grove, PA). PBMC were then cocultured overnight with L cell transfectants as described below. FDC-1 or HK cells were cultured as described (25, 26) with or without the addition of LPS or TNF-␣. Results Cloning and sequence analysis of FDC-SP Expression constructs, transfection, and Western blotting To identify genes specifically expressed by human FDC, we first FDC-SP cDNA was cloned into pcDNA3.1 myc/his (Invitrogen) in frame generated FDC-enriched cell preparations from human tonsils us- 3 with the myc epitope tag. The RR GG mutation was introduced by prim- ing the cell sorting method described by Liu et al. (21). This er-directed mutagenesis. All constructs were confirmed by sequencing before use. COS cell culture and transfection by the DEAE-dextran method method identifies FDC by coexpression of the markers CD21 were performed as described (50). The day after transfection, medium was (complement receptor 2) and CD14 (Fig. 1). Total RNA isolated replaced with fresh medium containing 1% FCS. Cells and supernatants from enriched FDCs were pooled from three separations and used were then collected at day 4 after transfection for immunoprecipitation to prepare double-stranded cDNA. PCR analysis of the resulting analysis. BJAB cells were cultured and transfected by electroporation as described (51). The day after transfection cells were centrifuged and re- cDNA using primers that distinguish FDC and B cell isoforms of plated at high density (106 cells/ml) in medium containing 1% FCS. Cells CD21 (32) indicated that both isoforms were present in approxi- and supernatants were then collected at day 4 after transfection. Cell lysate mately equal proportions (data not shown), suggesting that our cell

FIGURE 2. FDC-SP sequence and chromosomal location. A, Sequence alignment of human and mouse FDC-SP proteins. The putative secretion signal is underlined. Locations of conserved intron/exon borders are indicated by arrows. Note the proline-rich sequence of the central peptide and the conserved N-terminal charged sequence. FDC-SP sequences can be accessed in GenBank (human, accession no. AF435080; mouse, accession no. BF300338). B, Chromosomal location of human FDC-SP. The relative location of FDC-SP exons on chromosome 4 are indicated, based on basic local alignment search tool (BLAST) comparison of the cDNA sequence with the public draft human genome sequence (http://www.ncbi.nlm.nih.gov/genome/seq/). Approximate scale, as determined from the public genome sequence, is indicated on the left. Note that a similar five-exon structure was noted for mouse FDC-SP (data not shown). 2384 NOVEL FDC-ASSOCIATED SECRETED PROTEIN preparations contained adhering GC B cells in addition to FDCs. Restricted expression of FDC-SP mRNA For our cDNA subtraction strategy we chose to perform two sub- The expression of FDC-SP was examined by hybridization of the tractions: one using sort-purified B cells isolated from the same FDC-SP cDNA with Northern blots containing mRNA from a donors as the source of the driver cDNA (because B cells were large variety of human tissues (Fig. 3A). FDC-SP mRNA was de- likely the major contaminant of our FDC preparations) and the tected as a single 0.5-kb band and was abundantly expressed in other using a mix of cDNA from human fibroblast and epithelial tonsils (a mucosa-associated lymphoid tissue), lymph nodes, and cell lines (because FDC are generally thought to be nonhemato- trachea. Interestingly, FDC-SP expression in other lymphoid tis- poietic cells of mesenchymal origin). Surprisingly, the former sub- sues such as spleen, peripheral blood leukocytes, and bone marrow traction failed to give a robust SSH PCR and the clones that were is much lower or absent. FDC-SP is also significantly expressed in isolated were largely widely expressed housekeeping genes (data prostate, and at much lower levels in colon, stomach, and thyroid, not shown), while the latter subtraction resulted in a robust SSH but not in Ͼ10 other tissue types, including heart, brain, lung, PCR and a high frequency of differentially expressed genes (see liver, or skeletal muscle. Thus, FDC-SP appears to have a re- Table I). We speculate that the B cell-specific genes present in the stricted tissue distribution. FDC preparations may have provide a “carrier” effect during the To determine which cell types within the secondary lymphoid subtraction procedure by increasing the number of genes differen- tissues may be expressing FDC-SP, we examined FDC-SP expres- tially expressed between the tester and driver. The majority of sion in a panel of cell lines comprising B lymphocytes, T lym- sequences corresponded to “named” genes present in the GenBank phocytes, myeloid cells, and epithelial cells, as well as FDCs iso- database and these included several B lymphocyte genes known to lated from tonsils as a positive control (Fig. 3B). Our results be up-regulated upon activation, such as Ig L chains, MHC class II, showed strong expression of FDC-SP in tonsillar FDCs but no CD69, and RGS1/BL34 (Table I). Several novel genes were expression in any of the cell lines examined. In addition, FDC-SP present among this initial set of sequences, one of which encodes expression was not detected in myeloid-lineage dendritic cells Bam32, a novel CD40-inducible B lymphocyte signal transduction (data not shown), which are fundamentally distinct from FDCs in molecule that we have described elsewhere (51), and another their lineage derivation, phenotype, and function (53–55). These which encodes a novel C-type lectin, DCAL-1 (64). data support the hypothesis that FDC-SP expression in lymphoid One of the novel sequences contained a short open reading tissues is primarily due to expression in FDCs. frame that encodes a small protein containing the hallmarks of a signal peptide at its N terminus (Fig. 2A). Based on the evidence FDC-SP is expressed by FDC, but not B cells, within tonsillar GCs described below, we have named this gene product FDC-SP (FDC To further characterize the expression of FDC-SP within lymphoid secreted protein). Inspection of the predicted amino acid sequence tissues, we analyzed FDC-SP expression in human tonsils by in of FDC-SP shows a highly charged N-terminal sequence adjacent situ hybridization (Fig. 4, A and B). An antisense RNA probe gave to the putative leader peptide and a moderately proline-rich central region (Fig. 2A). FDC-SP appears to be structurally unique and has no significant similarity to known genes in GenBank. More than 30 sequences matching our FDC-SP sequence are present in the human expressed sequence tag (EST) database, and these derive largely from fetal tissue or tumor cell libraries. Notably, none of the EST clones were derived from B cell cDNA libraries, in contrast to Bam32, for which nearly one-third of the EST clones are derived from normal or GC B cell libraries. Searches of the public human genome sequence revealed that FDC-SP is a single-copy gene that maps to chromosome 4q13 (Fig. 2B). Interestingly, FDC-SP is closely linked with genes encoding three proline-rich salivary/tear gland proteins as well with as the cluster of C-X-C chemokines, which includes the B lymphocyte chemoattractant BLC. The FDC-SP gene is spread over ϳ10 kb and contains five exons which encode the 5Ј untranslated region (exon 1), the leader peptide (exon 2), most of the N-terminal charged region (exon 3), the remainder of the coding sequence (exon 4), and the 3Ј untranslated sequence (exon 5). Three murine ESTs with significant homology to FDC-SP were recently deposited into the database. We have confirmed the sequence of two of these clones and they encode a putative murine homolog of FDC-SP (Fig. 2A). Murine FDC-SP shows 45% amino acid identity and 54% amino acid similarity with human FDC-SP, with the highest degree of conservation (Ͼ80% identity) evident in the FIGURE 3. Expression of FDC-SP in human tissues and cell lines. A, ϩ charged N-terminal sequence. The proline-rich region is conserved Northern blots containing 2 ␮g of poly(A) RNA from the indicated tissue 32 in terms of proline content (21.4% proline overall vs 18.8% for source were hybridized with a P-labeled FDC-SP cDNA probe, then ␤ human); however, the primary amino acid sequence of this region stripped and reprobed with a -actin cDNA probe. B, RNA from the indicated cell lines were examined for FDC-SP expression using RT-PCR is only modestly conserved. The sequence of an unmapped murine followed by Southern blot analysis. cDNA from the FDC-enriched fraction genomic clone containing murine FDC-SP was recently deposited of human tonsil was used as a positive control for FDC-SP expression. in GenBank and sequence alignments show that the intron/exon Note that all of the cell lines examined, including B and T lymphocytes, structure is virtually identical to the human gene, further support- fibroblasts, and epithelial cells, are negative for FDC-SP expression, but all ing a close evolutionary relationship (data not shown). express the housekeeping gene G3PDH. The Journal of Immunology 2385 a very robust signal on tonsil sections, while a sense RNA probe FDC-SP is also expressed in leukocyte-infiltrated tonsillar gave no significant signal (data not shown). Strong signals were crypts and in activated, but not resting, peripheral blood observed in all GCs, while no significant expression was observed leukocytes in the intervening T cell areas. The intensity of expression is In addition to GCs, FDC-SP was also highly expressed in inflamed clearly higher in the central “light zone,” which is known to con- tonsillar crypts (Fig. 5, A and B). These are deep invaginations of tain the highest density of FDCs, than in the “dark zone,” which the epithelia, which open into the oral cavity and are often the site contains closely packed B lymphocyte blasts, suggesting that of inflammatory responses in tonsillitis. Crypt epithelia also con- FDC-SP is expressed in FDC but not GC B cells. Consistent with tain secretory ducts and may be important sites of Ab secretion this interpretation, sort-purified GC B cells are virtually devoid of into the oral cavity. FDC-SP expression was not observed in the FDC-SP expression, while unsorted tonsillar cells from the same patient express high levels of FDC-SP (Fig. 4C). The follicular organized epithelial layer of the exterior capsule (data not shown), mantle, which contains primarily naive B lymphocytes not partic- suggesting that expression is restricted to the leukocyte-infiltrated ipating in the response and a few FDC cytoplasmic processes, epithelial areas. The signal is clearly present through several cell showed a much lower intensity of FDC-SP expression. Indeed, layers, suggesting that it is not derived from activated epithelial sort-purified naive B cells we also found to express little or no cells but may derive from either locally activated stromal cells FDC-SP (Fig. 4C). To confirm expression by FDC, we also as- and/or infiltrating leukocytes. sayed FDC-SP expression in two independently derived tonsillar To test whether activated human leukocytes can express FDC- FDC-like cell lines, namely FDC-1 (25) and HK (26). It was found SP, we isolated PBMC from human subjects, cultured them for that these lines expressed detectable levels of FDC-SP but can be 24–72 h alone or with LPS or SAC, and assessed FDC-SP expres- induced to high-level expression after brief exposure to TNF-␣ sion by RT-PCR. These stimuli were used at concentrations shown (Fig. 4D). This TNF-activated expression of FDC-SP appears to be to potently induce expression of type I cytokines and chemokines specific to FDC, because TNF-treated B cells (Fig. 4D)orfibro- such as IL-12, IL-18, IFN-␥-inducible protein-10, and monokine blasts (data not shown) did not express detectable levels of FDC- induced by IFN-␥ (data not shown). We found that the PBMC cells SP. Collectively these data strongly support the conclusion that do not express significant levels of FDC-SP directly ex vivo, or FDC-SP is expressed by FDCs, rather than B cells, within when cultured without addition of stimulating agents, consistent active GCs. with our Northern blot analysis (Fig. 3A, lane 8). However,

FIGURE 4. FDC-SP is highly expressed in tonsillar GCs but not in GC B cells. Expression of FDC-SP in human tonsil sections was examined by in situ hybridization with a 35S-labeled antisense RNA probe. No significant signal was observed with a sense RNA probe (data not shown). A, Low- magnification view of section with no counterstaining, showing expression of FDC-SP in GCs. Note the preponderance of FDC-SP signal in the central area of the GCs corresponding to the light zone and reduced expression in areas corresponding to the dark zone (confirmed by counterstaining of serial sections). B, High-magnification view of counterstained GC, showing rapid decrease in signal at the border of the light zone. C, GC B cells do not express FDC-SP. Tonsil cell suspensions were depleted of T cells by rosetting (ErϪ population), stained for CD38 and IgD expression, and sorted into naive (IgDϩCD38Ϫ), GC (IgDϪCD38ϩ), and memory cell (IgDϪCD38Ϫ) populations, and FDC-SP expression was determined by RT-PCR. Note that all of the sorted B cell populations are greatly depleted of FDC-SP expression compared with the unsorted population, which contains FDCs. D, Two FDC-like cell lines express high levels of FDC-SP after stimulation with TNF-␣. Cells were cultured with medium (M) or for the indicated time with LPS or TNF-␣, then expression of FDC-SP was determined by RT-PCR. Note that BJAB cells fail to express FDC-SP after TNF-␣ treatment. 2386 NOVEL FDC-ASSOCIATED SECRETED PROTEIN

FDC-SP is a secreted protein To study FDC-SP protein biosynthesis and function, we prepared expression constructs encoding FDC-SP fused to a myc epitope tag and transfected these constructs into COS cells. Cell lysates and supernatants of the transfected cells were harvested and the myc- tagged FDC-SP proteins were immunoprecipitated and detected by Western blot analysis (Fig. 6). As a control, cells were transfected with expression constructs encoding Bam32-myc (51) or untagged FDC-SP. In the cell lysates, FDC-SP-myc was observed as a band running at 14–15 kDa on SDS-PAGE. This is slightly larger than the predicted molecular mass of FDC-SP-myc (13.1 kDa including the putative secretion signal). In the supernatants, FDC-SP-myc was observed as a broad band running at ϳ19–22 kDa. Bam32- myc transfectants show the expected 35-kDa band in the lysates, but no band in the supernatants, confirming that there is minimal nonspecific release of cytoplasmic proteins into the supernatants. Similar results were observed when FDC-SP-myc was expressed in the human B lymphocyte line BJAB (Fig. 6). These results demonstrate that FDC-SP is a secreted protein and suggest that FDC-SP undergoes heterogenous posttranslational modification during biosynthesis and trafficking through the secretory pathway. We hypothesized that the pairs of basic residues flanking the proline-rich sequence might be important for the function of FDC- SP, because such dibasic motifs can be the target sites for cleavage by proteases, such as has been well characterized for peptide hor- mone precursors (56). Therefore, we generated constructs encoding mutant FDC-SP-myc in which either the N-terminal lysine- arginine (K28R29) or the C-terminal arginine-arginine (R65R66) were converted to glycines. Interestingly, we reproducibly failed to detect protein expression of the K28R293G28G29 mutant, de- spite correct sequence insertion and mRNA expression (data not shown), suggesting that this mutation within the N-terminal

FIGURE 5. FDC-SP is expressed at sites of leukocyte infiltration within the tonsillar epithelium and expression can be induced in human leukocytes by innate immunity signals but not T-dependent B cell activation signals. A, FDC-SP is also expressed in leukocyte-infiltrated tonsillar crypts. Low- magnification view showing a crypt with intense FDC-SP expression adjacent to a GC. B, High-magnification view of a counterstained tonsillar crypt, showing irregular areas of FDC-SP expression within the infiltrated areas. C, Human PBMC can be induced to express FDC-SP in the presence of LPS or SAC, but not CD40 plus IL-4. Freshly prepared PBMC were cultured for the indicated time in medium alone or in the presence of LPS, SAC, or anti-CD40 plus IL-4. Expression was determined by extracting RNA, performing RT-PCR, and separating the resulting PCR products on ethidium bromide-stained gels. Data are representative of six RT-PCR ex- FIGURE 6. FDC-SP is a secreted protein. Left panel, COS-1 cells were periments and three independent cell isolations. transfected by the DEAE-dextran method using pcDNA3 vectors encoding native or myc-tagged human FDC-SP, or myc-tagged Bam32 as an intracellular protein control. Cells and culture supernatants were harvested 6 days after transfection. Expression of tagged proteins was analyzed by immunoprecipitation and Western blotting with 9E10 anti-myc Ab. Intra- FDC-SP expression is clearly induced within 48 h after culture cellular FDC-SP-myc protein runs as a 14- to 15-kDa protein, while se- with LPS or SAC (Fig. 5C). In contrast, stimulation of PBMC with creted FDC-SP-myc runs as a 19- to 22-kDa protein (indicated by arrows). a T-dependent B cell activation signal, anti-CD40 plus IL-4, did The predominant band running at 28 kDa represents the L chain of the not induce expression of FDC-SP (Fig. 5C). This result indicates immunoprecipitating Ab. The 17-kDa background band present in the that FDC-SP expression in leukocytes can be up-regulated by in- BJAB blots does not appear to be related to FDC-SP because it is present in the empty vector transfectants. Results are representative of four exper- nate immunity signals but not T-dependent activation signals. In- iments. Right panel, BJAB B lymphoma cells were transfected by electro- terestingly, the PBMC cells also did not express FDC-SP in re- poration using native or myc-tagged human FDC-SP, harvested 3 days sponse to TNF, indicating that the TNF-regulated expression of after transfection, and analyzed as in the left panel. Results are represen- FDC-SP may be a unique property of FDCs. tative of three experiments. The Journal of Immunology 2387 charged region renders the protein unstable or insoluble. However, the R65R663G65G66 FDC-SP-myc was expressed well in BJAB cells and resulted in both intracellular and secreted protein (Fig. 6). The molecular mass of the protein was substantially altered, suggesting that this mutation may affect the posttranslational processing of FDC-SP.

FDC-SP can bind to the surface of B cells We hypothesized that FDC-SP may function by binding to and influencing the behavior of immune cells. Because FDC-SP mRNA is present within GCs where B lymphocytes are undergoing proliferation, differentiation, and affinity selection (8, 57, 58), we assessed whether B cells might represent one such cellular target of FDC-SP. We established a FACS-based assay for FDC-SP binding using stably transfected fibroblast cells as a source of epitope-tagged FDC-SP and assayed binding of FDC-SP to B cells (BJAB) or T cells (Jurkat). We found that FDC-SP showed significant binding to B cells, but not T cells, and this binding was blocked by the G65G66 mutation (Fig. 7). We then used this same assay to examine whether FDC-SP can bind to FIGURE 8. FDC-SP binding to peripheral blood B cells is enhanced by primary human leukocytes. We found that fresh PBMC cells show T-dependent activation signals. PBMC were cultured for 3 days in medium very little FDC-SP binding (Fig. 8). However, we reasoned that if or with anti-CD40 plus IL-4 and then cocultured with FDC-SP-myc L cell GC B cells represent a specific target of FDC-SP it is possible that transfectants for 18 h before analysis. A, Cells were stained with FITC- resting peripheral B cells may only acquire the ability to bind anti-myc or isotype control and analyzed by flow cytometry. Results are FDC-SP after undergoing T-dependent activation. Therefore, we representative of three experiments. B, After 3 days of culture with anti- examined the ability of cells stimulated with anti-CD40 Ab and CD40 plus IL-4, PBMC cells were cocultured with FDC-SP-myc transfec- IL-4 to bind FDC-SP (Fig. 8A). It was found that CD40 plus IL- tants for 18 h and then stained with FITC-anti-myc and PE-anti-CD19. 4-activated PBMC cultures reproducibly developed a large FDC-SP binding population. To determine whether this population represents B cells, we performed two-color analysis and found that Discussion virtually all of the FDC-SP-binding cells coexpress the B cell Together, our findings on FDC-SP expression, biosynthesis, and marker CD19 (Fig. 8B), suggesting that activated B cells represent cell binding activity lead us to hypothesize that this molecule is a a specific target for FDC-SP. novel secreted mediator that regulates immune responses. While the immunoregulatory function(s) of FDC-SP remain to be determined, its exquisite pattern of expression, its regulation by TNF-␣, and its specific binding to activated B cells suggest that one important function may be to regulate B cell responses in GCs. Be- cause FDC-SP expression in spleen is Ͻ5% that in tonsil or lymph nodes and spleen tissue contains many primary follicles, we favor the hypothesis that FDC-SP expression is low or absent in primary follicles but is up-regulated in the follicles during GC formation. It is known that FDC networks expand and display phenotypic changes during the GC response (59–61). We propose that FDC-SP is one of the genes that is turned on during FDC activation. FDC-SP is unique in this regard, because only a handful of genes, and even fewer encoding secreted proteins, are known to be selectively expressed by activated FDCs. Together with the unique structure of the FDC-SP protein, it is tempting to speculate that FDC-SP has a novel, nonredundant function in regulating B cell responses. Our results indicate that FDC-SP is specifically and strongly expressed within GCs, in a pattern that is consistent with expression by FDCs. Several other lines of evidence support this interpretation, including data showing FDC-SP is not expressed by a FIGURE 7. FDC-SP binds to the surface of B lymphoma cells but not variety of B cell lines examined or by primary human B cells, T lymphoma cells. B or T lymphoma cells were cocultured with mouse CD40 plus IL-4-activated B cells, or sort-purified GC B cells. In fibroblasts (L cells) stably transfected with myc-tagged human FDC-SP contrast, we find that FDC-SP is expressed by sort-enriched FDCs proteins. Nonadherent cells were harvested after 18 h of coculture and and two independently derived FDC-like cell lines. It will be im- analyzed for myc-tagged protein at the cell surface by staining with FITC- portant to determine the molecular signals controlling FDC-SP ex- labeled anti-myc mAb. Gray open traces indicate anti-myc staining and filled traces represent staining with the isotype control mAb. Residual L pression within GCs. In this regard it is interesting that high-level cells were excluded from the analysis by side scatter gating. A, BJAB cells. expression of FDC-SP by FDC-like cells requires a brief exposure B, Jurkat cells. Note that BJAB cells, but not Jurkat cells, show binding to to TNF-␣. This result is consistent with previous work showing FDC-SP and the G64G65 mutant FDC-SP showed no significant binding to that TNF-␣ stimulation of FDC-like cells can up-regulate expres- either cells. sion of chemokines, including some of the chemokines genetically 2388 NOVEL FDC-ASSOCIATED SECRETED PROTEIN linked to FDC-SP on chromosome 4 (62) and adhesion molecules Alpers for help with tissue sectioning and advice on in situ hybridization, such as ICAM-1 (24). Because TNF expression on B cells has been Andrew Farr for assistance with imaging of histological specimens, and shown to provides critical signals for the generation and mainte- Dr. Y. Choi for providing the HK cells. nance of mature, functional FDC networks (36, 38, 40), it is tempting to speculate that FDC-SP may represent one of the TNF target References genes involved in this bidirectional signaling process. Up-regula- 1. Cumano, A., K. Dorshkind, S. Gillis, and C. J. Paige. 1990. The influence of S17 tion of FDC-SP expression by cytokines such as TNF may also stromal cells and interleukin 7 on B cell development. Eur. J. Immunol. 20:2183. 2. Cumano, A., A. Narendran, and C. J. Paige. 1991. Factors produced by stromal explain why FDC-SP is expressed at much higher levels in lymph cells involved in B-cell development. Adv. Exp. Med. Biol. 303:191. nodes and tonsils than in spleen and peripheral blood, in that the 3. Kincade, P. W. 1991. Molecular interactions between stromal cells and B lym- former tissues might be expected to be generally more “active” in phocyte precursors. Semin. Immunol. 3:379. 4. Jacobsen, K., and D. G. Osmond. 1990. Microenvironmental organization and immunological terms than the latter. stromal cell associations of B lymphocyte precursor cells in mouse bone marrow. Interestingly, FDC-SP expression is also up-regulated during the Eur. J. Immunol. 20:2395. inflammatory process occurring in tonsil crypts. Because FDC may 5. Liu, Y. J., G. Grouard, O. de Bouteiller, and J. Banchereau. 1996. Follicular dendritic cells and germinal centers. Int. Rev. Cytol. 166:139. derive from cells of mesenchymal origin that are induced to take 6. Tew, J. G., J. Wu, D. Qin, S. Helm, G. F. Burton, and A. K. Szakal. 1997. on the FDC phenotype by activated B lymphocytes (7, 38, 41), one Follicular dendritic cells and presentation of antigen and costimulatory signals to possibility is that stromal cells present under the crypt epithelium B cells. Immunol. Rev. 156:39. 7. Cyster, J. G., K. M. Ansel, K. Reif, E. H. Ekland, P. L. Hyman, H. L. Tang, are induced to express FDC-SP during the inflammatory response, S. A. Luther, and V. N. Ngo. 2000. Follicular stromal cells and lymphocyte perhaps in response to local production of TNF or other cytokines. homing to follicles. Immunol. Rev. 176:181. Our finding that circulating leukocytes can be induced to express 8. MacLennan, I. C. 1994. Germinal centers. Annu. Rev. Immunol. 12:117. 9. Szakal, A. K., M. H. Kosco, and J. G. Tew. 1989. Microanatomy of lymphoid FDC-SP in the presence of microbial products suggests that the tissue during humoral immune responses: structure function relationships. Annu. expression in tonsillar crypts may also be at least partially derived Rev. Immunol. 7:91. from infiltrating leukocytes. We speculate that FDC-SP is pro- 10. Szakal, A. K., and M. G. Hanna. 1968. The ultrastructure of antigen localization and virus like particles in mouse spleen germinal centers. Exp. Mol. Pathol. 8:75. duced in response to the local infection within the crypts and may 11. Tew, J. G., T. E. Mandel, and A. W. Burgess. 1979. Retention of intact HSA for play a role in controlling the infection, perhaps by regulating local prolonged periods in the popliteal lymph nodes of specifically immunized mice. mucosal B cell responses. In this context it is interesting to note Cell. Immunol. 45:207. 12. Nossal, G. J. V., A. P. Abbot, J. Mitchell, and Z. Lummus. 1968. Antigens in that memory B cells may specifically home to submucosal areas of immunity. XV. Ultrastructural features of antigen capture in primary and sec- the tonsil (63). ondary lymphoid follicles. J. Exp. Med. 127:227. Our working hypothesis is that FDC-SP functions in inter- 13. Klaus, G. G. 1978. The generation of memory cells. II. Generation of B memory cells with preformed antigen-antibody complexes. Immunology 34:643. cellular communication by binding target cells through a specific 14. Hannum, L. G., A. M. Haberman, S. M. Anderson, and M. J. Shlomchik. 2000. signaling receptor, analogous to cytokines/chemokines. While Germinal center initiation, variable gene region hypermutation, and mutant B cell selection without detectable immune complexes on follicular dendritic cells. FDC-SP does not share significant primary sequence homology J. Exp. Med. 192:931. with cytokines or chemokines, the overall amino acid composition, 15. Koni, P. A., and R. A. Flavell. 1999. Lymph node germinal centers form in the molecular mass, and charge (pI) of FDC-SP is similar to known absence of follicular dendritic cell networks. J. Exp. Med. 189:855. 16. Fu, Y. X., G. Huang, Y. Wang, and D. D. Chaplin. 2000. Lymphotoxin-␣-de- inflammatory mediators such as IL-8. Furthermore, FDC-SP is pendent spleen microenvironment supports the generation of memory B cells and linked to the cluster of ␣-chemokines on chromosome 4q13 that is required for their subsequent antigen-induced activation. J. Immunol. 164: includes IL-8 and other proinflammatory chemokines. Our data 2508. 17. Karrer, U., C. Lopez-Macias, A. Oxenius, B. Odermatt, M. F. Bachmann, indicate that FDC-SP can bind B cells, but not T cells or any other U. Kalinke, H. Bluethmann, H. Hengartner, and R. M. Zinkernagel. 2000. An- cell types present in the circulation, suggesting that B cells may be tiviral B cell memory in the absence of mature follicular dendritic cell networks Ϫ Ϫ the primary target of FDC-SP. The finding that FDC-SP binding and classical germinal centers in TNFR1 / mice. J. Immunol. 164:768. 18. Matsumoto, M., S. F. Lo, C. J. Carruthers, J. Min, S. Mariathasan, G. Huang, activity is up-regulated upon B cell activation with anti-CD40 plus D. R. Plas, S. M. Martin, R. S. Geha, M. H. Nahm, and D. D. Chaplin. 1996. IL-4 suggests that B cells activated during a T-dependent response Affinity maturation without germinal centres in lymphotoxin-␣-deficient mice. may represent the more specific target of FDC-SP. This result fits Nature 382:462. 19. Koni, P. A., R. Sacca, P. Lawton, J. L. Browning, N. H. Ruddle, and R. A. Flavell. with a model where FDC-SP regulates GC initiation and mainte- 1997. Distinct roles in lymphoid organogenesis for lymphotoxins ␣ and ␤ revealed in nance by acting upon newly activated B cells transiting from the T lymphotoxin ␤-deficient mice. Immunity 6:491. zones. To our knowledge, FDC-SP is the first FDC-restricted gene 20. Kosco-Vilbois, M. H., D. Gray, D. Scheidegger, and M. Julius. 1993. Follicular dendritic cells help resting B cells to become effective antigen-presenting cells: product that binds to activated but not resting B cells. Identification induction of B7/BB1 and upregulation of major histocompatibility complex class of the specific FDC-SP receptor will clearly be essential for full II molecules. J. Exp. Med. 178:2055. understanding of the distribution and function of FDC-SP binding. 21. Grouard, G., O. de Bouteiller, J. Banchereau, and Y. J. Liu. 1995. Human follicular dendritic cells enhance cytokine-dependent growth and differentiation of Production of recombinant, bioactive PDC-SP protein and Ab re- CD40-activated B cells. J. Immunol. 155:3345. agents for detecting and blocking the activity of secreted FDC-SP 22. Lindhout, E., M. L. Mevissen, J. Kwekkeboom, J. M. Tager, and C. de Groot. 1993. Direct evidence that human follicular dendritic cells (FDC) rescue germinal will be key for dissecting the potential functions of FDC-SP in centre B cells from death by apoptosis. Clin. Exp. Immunol. 91:330. regulating immunity. In addition, our identification of the murine 23. Burton, G. F., D. H. Conrad, A. K. Szakal, and J. G. Tew. 1993. Follicular homolog of FDC-SP will facilitate definitive in vivo analyses of dendritic cells and B cell costimulation. J. Immunol. 150:31. 24. Clark, E. A., K. H. Grabstein, and G. L. Shu. 1992. Cultured human follicular FDC-SP function in immune responses. dendritic cells: growth characteristics and interactions with B lymphocytes. J. Im- In summary, we report the identification of a novel secreted munol. 148:3327. mediator that displays a restricted expression pattern, is specifi- 25. Clark, E. A., K. H. Grabstein, A. M. Gown, M. Skelly, T. Kaisho, T. Hirano, and G. L. Shu. 1995. Activation of B lymphocyte maturation by a human follicular cally expressed at sites of ongoing immune responses by FDC and dendritic cell line, FDC-1. J. Immunol. 155:545. activated leukocytes, and can bind to the surface of activated B 26. Kim, H. S., X. Zhang, E. Klyushnenkova, and Y. S. Choi. 1995. Stimulation of cells. germinal center B lymphocyte proliferation by an FDC-like cell line, HK. J. Im- munol. 155:1101. 27. Lindhout, E., A. Lakeman, M. L. Mevissen, and C. de Groot. 1994. Functionally Acknowledgments active Epstein-Barr virus-transformed follicular dendritic cell-like cell lines. J. Exp. Med. 179:1173. We thank Geraldine Shu and Dario Magaletti for technical assistance and 28. Johnson, G. D., D. L. Hardie, N. R. Ling, and I. C. Maclennan. 1986. Human Ted Yun for advice on preparing site-directing mutants. We also thank follicular dendritic cells (FDC): a study with monoclonal antibodies (MoAb). Cathy Allen for help with flow cytometry, Kelly Hudkins and Charlie Clin. Exp. Immunol. 64:205. The Journal of Immunology 2389

29. Butch, A. W., B. A. Hug, and M. H. Nahm. 1994. Properties of human follicular 46. Kratz, A., A. Campos-Neto, M. S. Hanson, and N. H. Ruddle. 1996. Chronic dendritic cells purified with HJ2, a new monoclonal antibody. Cell. Immunol. inflammation caused by lymphotoxin is lymphoid neogenesis. J. Exp. Med. 183: 155:27. 1461. 30. Naiem, M., J. Gerdes, Z. Abdulaziz, H. Stein, and D. Y. Mason. 1983. Production 47. Luther, S. A., T. Lopez, W. Bai, D. Hanahan, and J. G. Cyster. 2000. BLC of a monoclonal antibody reactive with human dendritic reticulum cells and its expression in pancreatic islets causes B cell recruitment and lymphotoxin-depen- use in the immunohistological analysis of lymphoid tissue. J. Clin. Pathol. 36: dent lymphoid neogenesis. Immunity 12:471. 167. 48. Segerer, S., Y. Cui, F. Eitner, T. Goodpaster, K. L. Hudkins, M. Mack, 31. McNagny, K. M., R. P. Bucy, and M. D. Cooper. 1991. Reticular cells in pe- J. P. Cartron, Y. Colin, D. Schlondorff, and C. E. Alpers. 2001. Expression of ripheral lymphoid tissues express the phosphatidylinositol-linked BP-3 antigen. chemokines and chemokine receptors during human renal transplant rejection. Eur. J. Immunol. 21:509. Am. J. Kidney Dis. 37:518. 32. Liu, Y. J., J. Xu, O. de Bouteiller, C. Parham, G. Grourd, O. Djossou, 49. Gangur, V., F. E. Simons, and K. T. Hayglass. 1998. Human IP-10 selectively promotes dominance of polyclonally activated and environmental antigen-driven B. de Saint-Vis, S. Lebecque, J. Banchereau, and K. Moore. 1997. Follicular ␥ dendritic cells specifically express the long CR2/CD21 isoform. J. Exp. Med. IFN- over IL-4 responses. FASEB J. 12:705. 185:165. 50. Law, C. L., K. A. Chandran, S. P. Sidorenko, and E. A. Clark. 1996. Phospho- lipase C-␥1 interacts with conserved phosphotyrosyl residues in the linker region 33. Li, L., X. Zhang, S. Kovacic, A. J. Long, K. Bourque, C. R. Wood, and of Syk and is a substrate for Syk. Mol. Cell. Biol. 16:1305. Y. S. Choi. 2000. Identification of a human follicular dendritic cell molecule that 51. Marshall, A. J., H. Niiro, C. G. Lerner, T. Y. Yun, S. Thomas, C. M. Disteche, stimulates germinal center B cell growth. J. Exp. Med. 191:1077. and E. A. Clark. 2000. A novel B lymphocyte-associated adaptor protein, Bam32, 34. Zhang, X., L. Li, J. Jung, S. Xiang, C. Hollmann, and Y. S. Choi. 2001. The regulates antigen receptor signaling downstream of phosphatidylinositol 3-ki- distinct roles of T cell-derived cytokines and a novel follicular dendritic cell- nase. J. Exp. Med. 191:1319. signaling molecule 8D6 in germinal center-B cell differentiation. J. Immunol. 52. Leprince, C., K. E. Draves, J. A. Ledbetter, R. M. Torres, and E. A. Clark. 1992. 167:49. Characterization of molecular components associated with surface immunoglob- 35. Gunn, M. D., V. N. Ngo, K. M. Ansel, E. H. Ekland, J. G. Cyster, and ulin M in human B lymphocytes: presence of tyrosine and serine/threonine pro- L. T. Williams. 1998. A B-cell-homing chemokine made in lymphoid follicles tein kinases. Eur. J. Immunol. 22:2093. activates Burkitt’s lymphoma receptor-1. Nature 391:799. 53. Banchereau, J., F. Briere, C. Caux, J. Davoust, S. Lebecque, Y. J. Liu, 36. Ansel, K. M., V. N. Ngo, P. L. Hyman, S. A. Luther, R. Forster, J. D. Sedgwick, B. Pulendran, and K. Palucka. 2000. Immunobiology of dendritic cells. Annu. J. L. Browning, M. Lipp, and J. G. Cyster. 2000. A chemokine-driven positive Rev. Immunol. 18:767. feedback loop organizes lymphoid follicles. Nature 406:309. 54. Steinman, R. M., and K. Inaba. 1999. Myeloid dendritic cells. J. Leukocyte Biol. 37. Cyster, J. G., V. N. Ngo, E. H. Ekland, M. D. Gunn, J. D. Sedgwick, and 66:205. K. M. Ansel. 1999. Chemokines and B-cell homing to follicles. Curr. Top. Mi- 55. Mellman, I., S. J. Turley, and R. M. Steinman. 1998. Antigen processing for crobiol. Immunol. 246:87. amateurs and professionals. Trends Cell Biol. 8:231. 38. Endres, R., M. B. Alimzhanov, T. Plitz, A. Futterer, M. H. Kosco-Vilbois, 56. Hook, V. Y., A. V. Azaryan, S. R. Hwang, and N. Tezapsidis. 1994. Proteases S. A. Nedospasov, K. Rajewsky, and K. Pfeffer. 1999. Mature follicular dendritic and the emerging role of protease inhibitors in prohormone processing. FASEB J. cell networks depend on expression of lymphotoxin ␤ receptor by radioresistant 8:1269. stromal cells and of lymphotoxin ␤ and tumor necrosis factor by B cells. J. Exp. 57. Liu, Y. J., C. Arpin, O. de Bouteiller, C. Guret, J. Banchereau, H. Martinez-Valdez, Med. 189:159. and S. Lebecque. 1996. Sequential triggering of apoptosis, somatic mutation and 39. Fu, Y. X., G. Huang, Y. Wang, and D. D. Chaplin. 1998. B lymphocytes induce isotype switch during germinal center development. Semin. Immunol. 8:169. the formation of follicular dendritic cell clusters in a lymphotoxin ␣-dependent 58. Kelsoe, G. 1996. The germinal center: a crucible for lymphocyte selection. Se- fashion. J. Exp. Med. 187:1009. min. Immunol. 8:179. 40. Ngo, V. N., H. Korner, M. D. Gunn, K. N. Schmidt, D. S. Riminton, 59. Yoshida, K., T. K. van den Berg, and C. D. Dijkstra. 1993. Two functionally different follicular dendritic cells in secondary lymphoid follicles of mouse M. D. Cooper, J. L. Browning, J. D. Sedgwick, and J. G. Cyster. 1999. Lym- ␥ photoxin ␣/␤ and tumor necrosis factor are required for stromal cell expression spleen, as revealed by CR1/2 and FcR II-mediated immune-complex trapping. of homing chemokines in B and T cell areas of the spleen. J. Exp. Med. 189:403. Immunology 80:34. 60. Imazeki, N., A. Takeuchi, A. Senoo, and Y. Fuse. 1994. New monoclonal anti- 41. Lindhout, E., M. van Eijk, M. van Pel, J. Lindeman, H. J. Dinant, and bodies directed against mouse follicular dendritic cells. J. Histochem. Cytochem. C. de Groot. 1999. Fibroblast-like synoviocytes from rheumatoid arthritis patients 42:329. have intrinsic properties of follicular dendritic cells. J. Immunol. 162:5949. 61. Maeda, K., G. F. Burton, D. A. Padgett, D. H. Conrad, T. F. Huff, A. Masuda, 42. Chvatchko, Y., M. H. Kosco-Vilbois, S. Herren, J. Lefort, and J. Y. Bonnefoy. A. K. Szakal, and J. G. Tew. 1992. Murine follicular dendritic cells and low 1996. Germinal center formation and local immunoglobulin E (IgE) production in affinity Fc receptors for IgE (Fc⑀RII). J. Immunol. 148:2340. the lung after an airway antigenic challenge. J. Exp. Med. 184:2353. 62. Husson, H., S. M. Lugli, P. Ghia, A. Cardoso, A. Roth, K. Brohmi, E. G. Carideo, 43. Watanabe, S., K. Watanabe, T. Ohishi, and K. Kageyama. 1979. The develop- Y. S. Choi, J. Browning, and A. S. Freedman. 2000. Functional effects of TNF ment of extranodal lymphoid follicles in experimental bronchopneumonia. Acta ␣ ␤ and lymphotoxin 1 2 on FDC-like cells. Cell. Immunol. 203:134. Pathol. Jpn. 29:533. 63. Liu, Y. J., C. Barthelemy, O. de Bouteiller, C. Arpin, I. Durand, and 44. Takemura, S., A. Braun, C. Crowson, P. J. Kurtin, R. H. Cofield, W. M. O’Fallon, J. Banchereau. 1995. Memory B cells from human tonsils colonize mucosal ep- J. J. Goronzy, and C. M. Weyand. 2001. Lymphoid neogenesis in rheumatoid ithelium and directly present antigen to T cells by rapid up-regulation of B7-1 and synovitis. J. Immunol. 167:1072. B7-2. Immunity 2:239. 45. Schroder, A. E., A. Greiner, C. Seyfert, and C. Berek. 1996. Differentiation of B 64. Ryan, E. J., A. J. Marshall, D. M. Magaletti, H. E. Floyd, K. E. Draves, N. E. cells in the nonlymphoid tissue of the synovial membrane of patients with rheu- Olson, and E. A. Clark. D cell-asssociated molecule 1, a novel C-type lectin-like matoid arthritis. Proc. Natl. Acad. Sci. USA 93:221. molecule enhances T cell secretion of IL-4. J. Immunol. In press.