Stromal cells in the lymphoid organs provide a microenvironment where lymphocytes undergo various biological processes such as development, homing, clonal expansion, and differentiation. Follicular dendritic cells (FDCs) in the primary and secondary follicles of the peripheral lymphoid tissues interact with lymphocytes by contacting directly or producing diffusible molecules. To understand the biological role of human FDC at the molecular level, we developed a mAb, 3C8, that recognizes FDC but not bone marrow-derived cells. Through expression cloning and proteome analysis, we identified the protein that is recognized by 3C8 mAb, which revealed that FDC expresses prostacyclin synthase. The 3C8 protein purified from FDC-like cells indeed displayed the enzymatic activity of prostacyclin synthase and converted PGH2 into prostacyclin. In addition, prostacyclin significantly inhibited proliferation of T cells but delayed their spontaneous apoptosis. These findings may help explain why T cells constitute only a minor population compared with B cells in the germinal center.

T cell-dependent humoral immune responses culminate in the germinal centers (GCs) 3 of secondary lymphoid follicles. B cells are the major constituent of the GC and found closely associated with fine dendritic structures of specialized stromal cells called follicular dendritic cells (FDCs) (1). FDC is thought to provide B cells with essential survival signals because B cells detached from FDC undergo rapid apoptosis (2, 3). The essential requirement of FDCs for normal immune responses has been demonstrated in vivo with lymphotoxin-α knockout animals. Mice deficient in lymphotoxin-α lack the FDC network and the ability to mount isotype-switched memory Ig response (4, 5, 6). Accumulated evidence suggests that FDCs play central roles in B cell survival, proliferation, and differentiation into memory cells (1, 7, 8, 9). FDCs also appear to interact with another cellular component of the GC, T cells, which are found much less frequently than B cells (10). T cells stimulate proliferation of FDC-like cells (11). In contrast, FDCs inhibit T cell proliferation (12). However, little is known of the molecular basis of these cellular interactions.

FDCs constitute only a minor cellular component of the GC, which makes extensive functional studies problematic due to the difficulty of isolating cells to a significant purity and maintaining them in culture. Several research groups including ours have developed mAbs against FDC for the purpose of using them in immunohistological studies and in purification of FDCs (13, 14, 15, 16, 17, 18, 19, 20, 21). Most FDC-reactive mAbs developed to date cross-react with bone marrow-derived cells. For example, DRC-1, Ki-M4, 7D6, and HJ2 stain B cells (13, 16, 18), while 12B1 cross-reacts with monocytes (15), AS02 with blood stem cells (22), and CNA4.2 with T cells (17). In addition, FDC-SP, a recently discovered novel FDC molecule, is induced on cultured PBMC (23). Because mAb 3C8 that was developed in our laboratory did not display cross-reactivity with bone marrow-derived cells but specifically recognized FDCs in the GC (24, 25), we reasoned that identification of the Ag would reveal novel molecular mechanisms to explain biological functions of FDC. In this study, we have attempted molecular identification of the 3C8 Ag by expression cloning and proteome analyses, which led to the conclusion that human FDC expresses PGI2 synthase (PGIS, EC 5.3.99.4). PGIS belongs to the cytochrome P450 superfamily. This enzyme is localized in the endoplasmic reticulum and converts PGH2 into PGI2 or prostacyclin, the most potent endogenous inhibitor of platelet aggregation and a strong vasodilator (26, 27). With regard to the biological function of PGIS in the GC, we found that prostacyclin inhibits proliferation and spontaneous apoptosis of T cells.

The mAb 3C8 was developed as previously described (21), purified, and conjugated with biotin using a biotinylation kit (Pierce) according to the manufacturer’s instructions. Abs and reagents used in this work were anti-β-actin, biotin-conjugated mouse IgG1, streptavidin-FITC (DAKO), mouse IgG1, PGI2 (Sigma-Aldrich), alkaline phosphatase (AP)-conjugated goat anti-mouse IgG1, beraprost, PGH2, rabbit anti-human PGIS (Cayman Chemical), unconjugated anti-human CD3 (64.1, Dr. Y. S. Choi, Ochsner Medical Foundation, New Orleans, LA), and HRP-conjugated anti-rabbit Ig (Pierce).

HK cells were prepared and maintained as described by Kim et al. (11). Synovial fibroblasts were a kind gift from Dr. C.-S. Cho (Division of Rheumatology, Center for Rheumatic Diseases in St. Mary’s Hospital, Seoul, Korea). HK cells and fibroblasts were grown in RPMI 1640 (Invitrogen Life Technologies) supplemented with 10% FBS, 2 mM l-glutamine (Invitrogen Life Technologies), and 80 μg/ml gentamicin (Sigma-Aldrich). T cells were isolated from tonsillar mononuclear cells by rosetting with SRBC; the resulting cells contained >98% CD3+ cells as analyzed by FACScan (BD Biosciences). T cells were cultured under the various conditions as described in the figure legends in the presence or absence of PGI2 and beraprost. For the proliferation assay, 96-well flat-bottom microtiter plates were coated with 10 μg/ml 64.1 Ab, followed by incubation with T cells for the indicated periods of time. The degree of cellular proliferation was measured by using cell-counting kit-8 reagent (Dojindo Laboratories) according to the manufacturer’s instructions. The cellular proliferation was also measured by pulsing with 0.5 μCi [3H]thymidine (DuPont NEN) during the last 16 h culture period. The cultures were harvested onto glass-fiber filters, and [3H]thymidine incorporation was measured by a liquid scintillation counter (Packard). In the survival assay, T cell survival was measured with the cell-counting kit-8 reagent or by directly counting viable cells after trypan blue staining.

A Uni-ZAP XR expression cDNA library of human rheumatoid arthritis-positive synoviocyte was purchased from Stratagene. The cDNA library was immunoscreened according to the manufacturer’s protocols. Briefly, the library at the concentration of 1 × 104 plaques in 1 μl of SM buffer (100 mM NaCl, 10 mM MgSO4, 50 mM, Tris-HCl, pH 7.5) was preincubated with 600 μl of XL1-Blue Escherichia coli (OD600 = 0.5) resuspended in 10 mM MgSO4 for 15 min at room temperature. Then 7 ml of top agar was added, and the mixture was plated immediately into a 150-mm agar plate. Plates were incubated at 37°C until small size plaques appeared. Nitrocellulose membranes (Schleicher & Schuell) impregnated in 10 mM isopropyl β-d-thiogalactoside (USB) were overlaid on the agar surface for at least 5 h at 37°C. The membranes were removed, washed in TTNT (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.05% Tween 20), and blocked with 3% skim milk in TTNT for 1 h at room temperature. The membranes were then incubated overnight at 4°C with 3C8 ascites diluted 1/1000 in blocking buffer. After thorough washing three times for 10 min in TTNT, AP-conjugated goat anti-mouse IgG1 (0.3 mg/ml diluted 1/5000 in blocking buffer) was applied for 2 h at room temperature. The membranes were washed five times and preincubated in AP buffer (0.1M Tris-HCl, pH 9.5, 0.1M NaCl, 5 mM MgCl2 · 6H2O) for 5 min. Color development was conducted at the dark within 20 min by immersing in NBT/5-bromo-4-chloro-3-indolyl phosphate substrate solution (0.3 and 0.15 mg/ml, respectively, in AP buffer). A positive plaque was cored from the agar plate and transferred to 500 μl of SM buffer containing 20 μl of chloroform. To release the plaque particles into the buffer, the plaque was vortexed vigorously and incubated overnight at 4°C. After the third round of screening, a clone was obtained. And then, 250 μl of the positive phage stock containing over 1 × 105 phage particles was combined with 200 μl of XL1-Blue bacteria (A600 = 1.0) and 1 μl of ExAssist helper phage containing over 1 × 106 PFU. The mixture was incubated at 37°C for 15 min, followed by the addition of 3 ml of Luria-Bertani broth and further incubation for 3 h at 37°C with shaking. After heating at 70°C for 20 min, the mixture was centrifuged at 1000 × g for 15 min. To plate the excised phagemid, 100 μl of the supernatant was added to 200 μl of freshly grown SOLR cells (A600 = 1.0) for incubation at 37°C for 15 min. Two hundred microliters of the mixture was then spread on the Luria-Bertani/ampicillin (50 μg/ml) plates for incubation overnight at 37°C. Colonies appearing on the plate contain the pBluescript phagemid with the cloned cDNA insert. Phagemid DNA was extracted from the 3C8 clone using the Plasmid Mini kit (Qiagen) according to the manufacturer’s protocols. The insert size was determined on an agarose gel after digesting phagemid DNA with EcoRI and XhoI.

Sequencing of the insert cDNA was performed using a BioDye Terminator version 3.1 Cycle Sequencing kit (Applied Biosystems) and an automatic sequencer ABI 3730xl DNA Analyzer (Applied Biosystems). The sequencing primers used were the standard pBluescript T3 (5′-AATTAACCCTCACTAAAGGG-3′) and T7 (5′-AATACGACTCACTATAG-3′) universal primers and the internal custom-synthesized primer (5′-GATGTCTTCCACACCTTTCG-3′). For identification, sequences were manually edited to remove vector sequences and compared with the National Center for Biotechnology Information (NCBI) nonredundant GenBank nucleotide database. Sequence alignments were conducted using the ClustalW alignment procedure (28) with the BioEdit software.

Western blotting was conducted as previously described (24). 3C8 Ags were immunoprecipitated and visualized on 10% SDS-PAGE gel by Coomassie blue-staining. The 3C8 Ag band was excised from the gel with a sterile scalpel, subjected to in situ digestion with trypsin, and analyzed by MALDI-TOF-MS. The peptide mass fingerprints obtained were allowed to search against a mammalian subset in the NCBI nonredundant protein database. Sequence alignments were conducted using the ClustalW program.

To measure PGIS enzymatic activity of 3C8 Ags, 3C8-immunoprecipitate and its supernatant were incubated with PGH2 for 30 min. The produced amount of 6-keto-PGF, chemically stable metabolite of prostacyclin, was measured using an enzyme immunoassay kit (Cayman) according to the manufacturer’s instructions.

In the survival assay, T cells were cultured with HK cells that were transfected with PGIS-specific siRNA. The construction of siRNA was conducted according to the manufacturer’s instructions (Ambion). For the construction of PGIS siRNA, sense-(5′-AAGGTTCTAGACAGCACACCTCCTGTCTC-3′) and antisense-(5′-AAAGGTGTGCTGTCTAGAACCCCTGTCTC-3′) oligonucleotides were synthesized. The sequences of control siRNA with random PGIS nucleotides are as follows; sense- (5′-AAACTGACTCTTCTCCACAGACCTGTCTC-3′) and antisense-(5′-AATCTGTGGAGAAGAGTCAGTCCTGTCTC-3′). These siRNAs were transfected into HK cells by using FuGENE6 transfection reagent (Boehringer Mannheim) according to the manufacturer’s directions. The degree of gene silencing was assayed by Western blotting.

Because the 3C8 Ag was expressed in synovial fibroblasts as well as in FDC (24, 25), we attempted to isolate 3C8 cDNA from the synoviocyte cDNA library by expression cloning. A 3.2-kb cDNA insert was obtained from a positive plaque (Fig. 1). The sequence of the 3C8 cDNA insert matched exactly that of PGIS (data not shown).

FIGURE 1.

Expression cloning with cDNA library of synovial fibroblasts reveals the molecular identity of the 3C8 Ag. With a cDNA library of human rheumatoid arthritis positive synoviocytes, expression cloning was conducted by immunoscreening with 3C8 mAb. A positive plaque (A) was subjected to the second (B) or third (C) rounds of immunoblotting. A replica of B was immunoblotted with isotype-matched control Ab (D).

FIGURE 1.

Expression cloning with cDNA library of synovial fibroblasts reveals the molecular identity of the 3C8 Ag. With a cDNA library of human rheumatoid arthritis positive synoviocytes, expression cloning was conducted by immunoscreening with 3C8 mAb. A positive plaque (A) was subjected to the second (B) or third (C) rounds of immunoblotting. A replica of B was immunoblotted with isotype-matched control Ab (D).

Close modal

To rule out the possibility that 3C8 Ab bound the isolated cDNA clone by cross-reactivity, we purified 3C8 protein from HK cells by immunoprecipitation for biochemical analysis of the 3C8 Ag. SDS-PAGE followed by immunoblotting with 3C8 Ab revealed two specific bands (Fig. 2,A), a major and a minor band with molecular masses of around 55 kDa. The nominal molecular mass of PGIS is 57 kDa. These two specific bands were identified both as PGIS by MALDI-TOF-MS analysis (Fig. 2, B and C), confirming the expression cloning result. The appearance of 3C8 as a doublet may have been resulted from protein fragmentation during immunoprecipitation procedure since a single band was obtained with Western blot analysis that was performed with 3C8 Ab and commercially available polyclonal Ab against PGIS (Fig. 3,A). Immunoprecipitation with 3C8 Ab followed by immunoblotting with PGIS polyclonal Ab further confirmed that 3C8 mAb and anti-PGIS polyclonal Ab recognize the same molecule (Fig. 3,B). Finally, because 3C8 Ag was successfully isolated in pure forms by immunoprecipitation, we investigated whether the isolated Ag would exhibit the enzymatic activity of PGIS, that is, conversion of PGH2 into PGI2. After incubation of 3C8 Ab with HK cell lysates followed by precipitation with protein A-Sepharose beads, we measured the PGIS activity of precipitates and supernatants. Compared with the control precipitate that had been incubated with isotype-matched control Ab, the 3C8 precipitate converted 15 times more PGH2 into PGI2 (Fig. 4,A). At the same time, the PGIS activity of the control supernatant that had been incubated with control Ab was comparable to that of 3C8 precipitate while 3C8 supernatant lost most of the PGIS activity after incubation with 3C8 Ab (Fig. 4,A). The production of PGI2 by 3C8 precipitate was dependent on the concentration of PGH2. However, increased addition of PGH2 to control precipitate did not result in a significant increase of PGI2 production (Fig. 4 B). These results suggest that the protein that precipitated with 3C8 Ab is indeed the enzyme PGIS.

FIGURE 2.

MALDI-TOF-MS analysis of immunoprecipitated 3C8 protein confirms the cloning result. A, 3C8 molecules were purified from HK cells by immunoprecipitation, and then the presence of 3C8 molecules in the supernatants (lanes 1, 2, 5, 6) or precipitates (lanes 3, 4, 7, 8) was examined by Coomassie blue staining (lanes 14) and Western blotting (lanes 58) after SDS-PAGE under the nonreducing condition. The supernatants or precipitates were obtained after incubation with control Ab (lanes 1, 3, 5, 7) or 3C8 mAb (lanes 2, 4, 6, 8). Arrows indicate specific bands obtained with 3C8 mAb. B, MALDI-TOF mass spectrum of a digest of the upper band of the gel (band 1) with trypsin was obtained as described in Materials and Methods. C, The amino acid sequences of eight selected peptides of the two 3C8-specific bands were aligned in boxes with the sequence of human PGIS protein.

FIGURE 2.

MALDI-TOF-MS analysis of immunoprecipitated 3C8 protein confirms the cloning result. A, 3C8 molecules were purified from HK cells by immunoprecipitation, and then the presence of 3C8 molecules in the supernatants (lanes 1, 2, 5, 6) or precipitates (lanes 3, 4, 7, 8) was examined by Coomassie blue staining (lanes 14) and Western blotting (lanes 58) after SDS-PAGE under the nonreducing condition. The supernatants or precipitates were obtained after incubation with control Ab (lanes 1, 3, 5, 7) or 3C8 mAb (lanes 2, 4, 6, 8). Arrows indicate specific bands obtained with 3C8 mAb. B, MALDI-TOF mass spectrum of a digest of the upper band of the gel (band 1) with trypsin was obtained as described in Materials and Methods. C, The amino acid sequences of eight selected peptides of the two 3C8-specific bands were aligned in boxes with the sequence of human PGIS protein.

Close modal
FIGURE 3.

mAb 3C8 and anti-PGIS polyclonal Ab recognize the same molecule. A, Immunoblotting of HK cell lysate was conducted with 3C8 mAb (lane 1) or anti-PGIS polyclonal Ab (lane 2) after SDS-PAGE under the reducing condition. B, Immunoblotting with anti-PGIS polyclonal Ab after immunoprecipitation of HK cell lysate with control (lane 1) or 3C8 mAb (lane 2). SDS-PAGE was conducted under the nonreducing condition

FIGURE 3.

mAb 3C8 and anti-PGIS polyclonal Ab recognize the same molecule. A, Immunoblotting of HK cell lysate was conducted with 3C8 mAb (lane 1) or anti-PGIS polyclonal Ab (lane 2) after SDS-PAGE under the reducing condition. B, Immunoblotting with anti-PGIS polyclonal Ab after immunoprecipitation of HK cell lysate with control (lane 1) or 3C8 mAb (lane 2). SDS-PAGE was conducted under the nonreducing condition

Close modal
FIGURE 4.

Immunoprecipitated 3C8 Ag exhibits PGIS activity. A, After incubation of HK cell lysates with control or 3C8 mAb, the ability of supernatants and precipitates to convert PGH2 to 6-keto PGF was measured by ELISA. B, Dose-response result of control or 3C8 precipitates. The results are expressed as the ability of each fraction relative to control precipitate and in terms of the mean ± SEM of three independent experiments. Asterisks indicate significant difference compared with controls (∗∗, p < 0.01; ∗∗∗, p < 0.001).

FIGURE 4.

Immunoprecipitated 3C8 Ag exhibits PGIS activity. A, After incubation of HK cell lysates with control or 3C8 mAb, the ability of supernatants and precipitates to convert PGH2 to 6-keto PGF was measured by ELISA. B, Dose-response result of control or 3C8 precipitates. The results are expressed as the ability of each fraction relative to control precipitate and in terms of the mean ± SEM of three independent experiments. Asterisks indicate significant difference compared with controls (∗∗, p < 0.01; ∗∗∗, p < 0.001).

Close modal

Widespread distribution of PGIS expression in the GC suggests its critical role in the GC reaction. The report that indomethicin reversed FDC-induced inhibition of T cell proliferation (12) prompted us to investigate the effect of prostacyclin on T cell proliferation. Because prostacyclin is a very unstable molecule (29), we also examined the effect of beraprost, a more stable prostacyclin analog. T cells were stimulated to proliferate on the plate coated with anti-CD3 Ab and continuously proliferated up to 80 h poststimulation (Fig. 5,A). Addition of prostacyclin resulted in a slight decrease of T cell proliferation, whereas beraprost significantly reduced the proliferation from 48 h (Fig. 5,C). The inhibition of T cell proliferation was dose-dependent, which was more evident with beraprost than prostacyclin at the equivalent concentrations (Fig. 5, B and D). This inhibitory effect of prostacyclin and beraprost was confirmed because addition of beraprost gave rise to ∼50% inhibition of [3H]thymidine uptake by proliferating T cells (Fig. 5 E).

FIGURE 5.

Prostacyclin and beraprost suppress T cell proliferation. Kinetics and dose response of the effect of PGI2 (A and B) or beraprost (C and D) on tonsillar T cell (1 × 105 cells/well) proliferation stimulated by immobilized anti-CD3 Ab. The degree of cellular proliferation was measured colorimetrically by cell counting kit-8 at the indicated time points or at the end of 72 h culture (B and D). E, The effect of prostacyclin and beraprost on the [3H]thymidine uptake by proliferating T cells. Tonsillar T cells (1 × 105 cells/well) were cultured in the presence or absence of prostacyclin or beraprost for 72 h, and [3H]thymidine incorporation during the last 16 h was measured. Results are presented as the mean ± SEM of three independent experiments. Asterisks indicate significantly different effect compared with controls (∗, p < 0.05; ∗∗, p < 0.01; ∗∗∗, p < 0.001).

FIGURE 5.

Prostacyclin and beraprost suppress T cell proliferation. Kinetics and dose response of the effect of PGI2 (A and B) or beraprost (C and D) on tonsillar T cell (1 × 105 cells/well) proliferation stimulated by immobilized anti-CD3 Ab. The degree of cellular proliferation was measured colorimetrically by cell counting kit-8 at the indicated time points or at the end of 72 h culture (B and D). E, The effect of prostacyclin and beraprost on the [3H]thymidine uptake by proliferating T cells. Tonsillar T cells (1 × 105 cells/well) were cultured in the presence or absence of prostacyclin or beraprost for 72 h, and [3H]thymidine incorporation during the last 16 h was measured. Results are presented as the mean ± SEM of three independent experiments. Asterisks indicate significantly different effect compared with controls (∗, p < 0.05; ∗∗, p < 0.01; ∗∗∗, p < 0.001).

Close modal

We next examined the effect of beraprost on T cell survival. T cells underwent spontaneous apoptosis, and only small numbers of viable cells were observed after 48 h of the culture. In contrast to its inhibitory effect on T cell proliferation, beraprost enhanced T cell survival during the 3-day culture period. The presence of beraprost resulted in significantly higher number of viable cells at 24 and 48 h (Fig. 6,A). Beraprost enhanced the T cell survival in a dose-dependent manner, and significant difference was obtained at the concentration of 1 μM (Fig. 6,B). Because HK cells express PGIS and beraprost enhances T cell survival, we investigated whether knockdown of PGIS from HK cells would modulate the effect of HK cells on T cell survival. The PGIS gene was silenced by transfection with siRNA against PGIS. The HK cells that were transfected with 100 nM PGIS siRNA expressed significantly reduced PGIS proteins, whereas endogenous expression of PGIS protein was slightly modulated by transfection with control siRNA (Fig. 6,C). Freshly isolated T cells were cultured with equal numbers of normal or transfected HK cells. T cells cultured in the absence of HK cells exhibited 48% of viable cell recovery at the end of 48 h culture. Both normal HK cells and HK cells that were transfected with control siRNA increased the recovery to 57%. However, PGIS siRNA-transfected HK cells did not support the T cell survival and yielded 49% of viable cell recovery (Fig. 6 D). This result suggests that HK cells enhance T cell survival by producing prostacyclin. The effect of transfected HK cells on the T cell proliferation was not examined because the coating of culture plate with anti-CD3 Ab was technically problematic due to the presence of adherent HK cells.

FIGURE 6.

Beraprost enhances T cell survival, which is impaired by silencing of PGIS. Resting T cells (1 × 105 cells/well) were cultured in the presence or absence of beraprost, and the kinetics (A) and dose response (B at 24 h) of the effect of beraprost on the survival were measured by cell-counting kit 8 as described in Materials and Methods. HK cells were transfected with control or PGIS siRNA for 72 h, and immunoblotting was conducted to examine the effect of transfection on the endogenous expression of PGIS (C). The effect of PGIS knockdown on the T cell survival was assayed by culturing resting T cells (5 × 105 cells/well) for 48 h with or without HK cells that were nontreated or transfected with control or PGIS siRNA. The number of viable cells was counted after staining with trypan blue (D). Results are presented as the mean ± SEM of three independent experiments. An asterisk indicates significantly different effect compared with controls (∗, p < 0.05).

FIGURE 6.

Beraprost enhances T cell survival, which is impaired by silencing of PGIS. Resting T cells (1 × 105 cells/well) were cultured in the presence or absence of beraprost, and the kinetics (A) and dose response (B at 24 h) of the effect of beraprost on the survival were measured by cell-counting kit 8 as described in Materials and Methods. HK cells were transfected with control or PGIS siRNA for 72 h, and immunoblotting was conducted to examine the effect of transfection on the endogenous expression of PGIS (C). The effect of PGIS knockdown on the T cell survival was assayed by culturing resting T cells (5 × 105 cells/well) for 48 h with or without HK cells that were nontreated or transfected with control or PGIS siRNA. The number of viable cells was counted after staining with trypan blue (D). Results are presented as the mean ± SEM of three independent experiments. An asterisk indicates significantly different effect compared with controls (∗, p < 0.05).

Close modal

Based on the following results presented previously (25, 26) and in this paper, we conclude that human FDCs express prostacyclin synthase: 1) FDCs express 3C8 Ag. 2) Molecular cloning and proteome analysis reveal that 3C8 Ag is prostacyclin synthase. 3) 3C8 Ag is recognized by anti-PGIS polyclonal Ab. 4) Purified 3C8 protein displays PGIS activity.

Although PGs, particularly PGE2, have been suggested as potent modulators of immunity (30), their cellular source and biological functions at the culminating site of the immune response, the GC, were not studied. Strong expression of PGIS throughout the GC suggests that prostacyclin is produced by FDC and is the major PG species in this microenvironment. This is, to our best knowledge, the first report that FDC expresses PGIS. Therefore, 3C8 is the first mAb raised against human PGIS. Furthermore, the result that prostacyclin and beraprost inhibited T cell proliferation is extending the previous report by Butch et al. (12) that PGs are involved in FDC-mediated inhibition of T cell proliferation and specifies the molecular identity of the responsible PG. Our results may explain how T cell numbers are controlled in the GC. GC T cells are mostly CD4+ and specific to the inducing Ag (31, 32), suggesting cognate interactions between T and B cells. Considering that B cells are efficient APCs to T cells, it has been puzzling that GC T cells do not normally proliferate and are defective for IL-2 production in response to Ag stimulation (12, 33). Our current data and recent results by Tenner-Racz and colleagues (31) that showed 70–90% GC T cells express CTLA-4 suggest that the number of T cells in the GC appears to be tightly regulated by FDC and B cells. B cells may induce negative signals in T cells during the cognate interactions through CTLA-4. At the same time, FDCs in close contact with T cells inhibit T cell proliferation by producing prostacyclin.

FDC expression of PGIS is more evidence supporting the view that FDC originates from local fibroblast as reviewed by Heinen and Bosseloir (34). Ultrastructural (35, 36), enzymological (35, 37), and functional studies (38, 39) support this view. We demonstrated recently that fibroblasts as well as FDCs express 3C8 Ag (25). Because 3C8 Ag has turned out to be PGIS in the current study, fibroblast and FDC exhibit another common feature of expressing PGIS. Fibroblasts are the major producer of prostacyclin (40). When the requirement for FDC increases in the GC during immune response, fibroblasts in connective tissues may migrate to lymphoid follicles in response to certain chemokines and differentiate to FDCs as a result of interaction with GC B cells. FDC may produce prostacyclin, which controls the number of T cells by inhibiting their proliferation but promoting their survival. Lindhout et al. (41) have demonstrated that synovial fibroblasts could be induced with IL-1β and TNF-α to express the FDC phenotype. Because fibroblasts are ubiquitous cells, it remains to be elucidated whether all kinds of fibroblasts are capable of differentiating to FDCs or whether this property is limited to certain types of fibroblasts. The distinct staining pattern of 3C8 Ab from those of other fibroblast-specific Abs (25) implies that only a subset of fibroblasts may differentiate to FDCs.

The authors have no financial conflict of interest.

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 with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported by the Vascular System Research Center grant and a grant (R05-2003-000-11631-0) from the Korea Science and Engineering Foundation.

3

Abbreviations used in this paper: GC, germinal center; FDC, follicular dendritic cell; PGIS, PGI2 synthase; AP, alkaline phosphatase; MALDI-TOF-MS, MALDI-TOF mass spectrometry; siRNA, small interfering RNA.

1
MacLennan, I. C. M..
1994
. Germinal centers.
Annu. Rev. Immunol.
12
:
117
-139.
2
Kosco, M. H., E. Pflugfelder, D. Gray.
1992
. Follicular dendritic cell-dependent adhesion and proliferation of B cells in vitro.
J. Immunol.
148
:
2331
-2339.
3
Koopman, G., R. M. J. Keehnen, E. Lindhout, W. Newman, Y. Shimizu, G. A. van Seventer, C. de Groot, S. T. Pals.
1994
. Adhesion through the LFA-1 (CD11a/CD18)-ICAM-1 (CD54) and the VLA-4 (CD49d)-VCAM-1 (CD106) pathways prevents apoptosis of germinal center B cells.
J. Immunol.
152
:
3760
-3767.
4
Hir, M. L., H. Bluethmann, M. H. Kosco-Vilbois, M. Muller, F. di Padova, M. Moore, B. Ryffel, H.-P. Eugster.
1996
. Differentiation of follicular dendritic cells and full antibody responses require tumor necrosis factor receptor-1 signaling.
J. Exp. Med.
183
:
2367
-2372.
5
Fu, Y.-X., H. Molina, M. Matsumoto, G. Huang, J. Min, D. D. Chaplin.
1997
. Lymphotoxin-α (LTα) supports development of splenic follicular structure that is required for IgG responses.
J. Exp. Med.
185
:
2111
-2120.
6
Fu, Y.-X., G. Huang, Y. Wang, D. D. Chaplin.
2000
. Lymphotoxin-α-dependent spleen microenvironment supports the generation of memory B cells and is required for their subsequent antigen-induced activation.
J. Immunol.
164
:
2508
-2514.
7
Choe, J., H.-S. Kim, R. J. Armitage, Y. S. Choi.
1997
. The functional role of BCR stimulation and IL-4 in the generation of human memory B cells from germinal center B cells.
J. Immunol.
159
:
3757
-3766.
8
Choe, J., L. Li, X. Zhang, C. D. Gregory, Y. S. Choi.
2000
. Distinct role of follicular dendritic cells and T cells in the proliferation, differentiation, and apoptosis of a centroblast cell line, L3055.
J. Immunol.
164
:
56
-63.
9
Tew, J. G., J. Wu, M. Fakher, A. K. Szakal, D. Qin.
2001
. Follicular dendritic cells: beyond the necessity of T-cell help.
Trends Immunol.
22
:
361
-367.
10
Jacobson, E. B., L. H. Caporale, G. J. Thorbecke.
1974
. Effect of thymus cell injections on germinal center formation in lymphoid tissues of nude (thymusless) mice.
Cell Immunol.
13
:
416
-430.
11
Kim, H.-S., X. Zhang, Y. S. Choi.
1994
. Activation and proliferation of follicular dendritic cell-like cells by activated T lymphocytes.
J. Immunol.
153
:
2951
-2961.
12
Butch, A. W., K. A. Kelly, M. S. Willbanks, X. Yu.
1999
. Human follicular dendritic cells inhibit superantigen-induced T-cell proliferation by distinct mechanisms.
Blood
94
:
216
-224.
13
Naiem, M., J. Gerdes, Z. Abdulaziz, H. Stein, D. Y. Mason.
1983
. Production of a monoclonal antibody reactive with human dendritic reticulum cells and its use in the immunohistological analysis of lymphoid tissue.
J. Clin. Pathol.
36
:
167
-175.
14
Parwaresch, M. R., H. J. Radzun, M.-L. Hansmann, K.-P. Peters.
1983
. Monoclonal antibody Ki-M4 specifically recognizes human dendritic reticulum cells (follicular dendritic cells) and their possible precursor in blood.
Blood
62
:
585
-590.
15
Farace, F., N. Kieffer, S. Caillou, W. Vainchenker, T. Tursz, M. C. Dokhelar.
1986
. A 93-kDa glycoprotein expressed on human cultured monocytes and dendritic reticulum cells defined by an anti-K562 monoclonal antibody.
Eur. J. Immunol.
16
:
1521
-1526.
16
Butch, A. W., B. A. Hug, M. H. Nahm.
1994
. Properties of human follicular dendritic cells purified with HJ2, a new monoclonal antibody.
Cell Immunol.
155
:
27
-41.
17
Raymond, I., T. A. Saati, J. Tkaczuk, S. Chittal, G. Delsol.
1997
. CNA. 42, a new monoclonal antibody directed against a fixative-resistant antigen of follicular dendritic reticulum cells.
Am. J. Pathol.
151
:
1577
-1585.
18
Liu, Y.-J., J. Xu, O. de Bouteiller, C. L. Parham, G. Grouard, O. Djossou, B. de Saint-Vis, S. Lebecque, J. Banchereau, K. W. Moore.
1997
. Follicular dendritic cells specifically express the long CR2/CD21 isoform.
J. Exp. Med.
185
:
165
-170.
19
Li, L., X. Zhang, S. Kovacic, A. J. Long, K. Bourque, C. R. Wood, Y. S. Choi.
2000
. Identification of a human follicular dendritic cell molecule that stimulates germinal center B cell growth.
J. Exp. Med.
191
:
1077
-1084.
20
Bofill, M., A. Akbar, P. Amlot.
2000
. Follicular dendritic cells share a membrane-bound protein with fibroblasts.
J. Pathol.
191
:
217
-226.
21
Lee, I. Y., J. Lee, W. S. Park, E.-C. Nam, Y. Shin, J. Choe.
2001
. 3C8, a new monoclonal antibody directed against a follicular dendritic cell line, HK.
Immune Network
1
:
26
-31.
22
Saalbach, A., R. Kraft, K. Herrmann, U.-F. Haustein, U. Anderegg.
1998
. The monoclonal antibody AS02 recognizes a protein on human fibroblasts being highly homologous to Thy-1.
Arch. Dermatol.
290
:
360
-366.
23
Marshall, A. J., Q. Du, K. E. Draves, Y. Shikishima, K. T. HayGlass, E. A. Clark.
2002
. FDC-SP, a novel secreted protein expressed by follicular dendritic cells.
J. Immunol.
169
:
2381
-2389.
24
Lee, I. Y., K. S. Ha, J. Choe.
2003
. 3C8 antigen is a novel protein expressed by human follicular dendritic cells.
Biochem. Biophys. Res. Commun.
303
:
624
-630.
25
Lee, I. Y., J. Choe.
2003
. Human follicular dendritic cells and fibroblasts share the 3C8 antigen.
Biochem. Biophys. Res. Commun.
304
:
701
-707.
26
Yokoyama, C., T. Yabuki, H. Inoue, Y. Tone, S. Hara, T. Hatae, M. Nagata, E.-I. Takahashi, T. Tanabe.
1996
. Human gene encoding prostacyclin synthase (PTGIS): genomic organization, chromosomal localization, and promoter activity.
Genomics
36
:
296
-304.
27
Chevalier, D., C. Cauffiez, C. Bernard, J.-M. Lo-Guidice, D. Allorge, F. Fazio, N. Ferrari, C. Libersa, M. Lhermitte, J.-C. D’Halluin, F. Broly.
2001
. Characterization of new mutations in the coding sequence and 5′-untranslated region of the human prostacyclin synthase gene (CYP8A1).
Hum. Genet.
108
:
148
-155.
28
Thompson, J. D., D. G. Higgins, T. J. Gibson.
1994
. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice.
Nucleic Acids Res.
22
:
4673
-4680.
29
Pola, R., E. Gaetani, A. Flex, T. R. Aprahamian, M. Bosch-Marce, D. W. Losordo, R. C. Smith, P. Pola.
2004
. Comparative analysis of the in vivo angiogenic properties of stable prostacyclin analogs: a possible role for peroxisome proliferator-activated receptors.
J. Mol. Cell Cardiol.
36
:
363
-370.
30
Harris, S. G., J. Padilla, L. Koumas, D. Ray, R. P. Phipps.
2002
. Prostaglandins as modulators of immunity.
Trends Immunol.
23
:
144
-150.
31
Castan, J., K. Tenner-Racz, P. Racz, B. Fleischer, B. M. Broker.
1997
. Accumulation of CTLA-4 expressing T lymphocytes in the germinal centres of human lymphoid tissues.
Immunology
90
:
265
-271.
32
Fuller, K. A., O. Kanagawa, M. H. Nahm.
1993
. T cells within germinal centers are specific for the immunizing antigen.
J. Immunol.
151
:
4505
-4512.
33
Velardi, A., M. C. Mingari, L. Moretta, C. E. Grossi.
1986
. Functional analysis of cloned germinal center CD4+ cells with natural killer cell-related features: divergence from typical T helper cells.
J. Immunol.
137
:
2808
-2813.
34
Heinen, E., A. Bosseloir.
1994
. Follicular dendritic cells: whose children?.
Immunol. Today
15
:
201
-204.
35
Heusermann, U., K. H. Zurborn, L. Schroeder, H. J. Stutte.
1980
. The origin of the dendritic reticulum cell: an experimental enzyme-histochemical and electron microscopic study on the rabbit spleen.
Cell Tissue Res.
209
:
279
-294.
36
Maeda, K., M. Matsuda, Y. Imai.
1995
. Follicular dendritic cells: structure as related to function.
Curr. Top. Microbiol. Immunol.
201
:
119
-139.
37
Rademakers, L. H. P. M., R. A. de Weger, P. J. M. Roholl.
1989
. Identification of alkaline phosphatase positive cells in human germinal centers as follicular dendritic cells.
Adv. Exp. Med. Biol.
237
:
165
-169.
38
Humphrey, J. H., P. Grennan, A. Sundaram.
1984
. The origin of follicular dendritic cells in the mouse and the mechanism of trapping of immune complexes on them.
Eur. J. Immunol.
14
:
859
-864.
39
Heinen, E., M. Braun, P. G. Coulie, J. van Snick, M. Moeremans, N. Cormann, C. Kinet-Denoel, L. J. Simar.
1986
. Transfer of immune complexes from lymphocytes to follicular dendritic cells.
Eur. J. Immunol.
16
:
167
-172.
40
Harvey, W., F. Guat-Chen, D. Gordon, S. Meghji, A. Evans, M. Harris.
1984
. Evidence for fibroblasts as the major source of prostacyclin and prostaglandin synthesis in dental cyst in man.
Arch. Oral Biol.
29
:
223
-229.
41
Lindhout, E., M. van Eijk, M. van Pel, J. Lindeman, H. Dinant, C. de Groot.
1999
. Fibroblast-like synoviocytes from rheumatoid arthritis patients have intrinsic properties of follicular dendritic cells.
J. Immunol.
162
:
5949
-5956.