| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Published online before print June 5, 2007
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Compton Laboratory, Institute for Animal Health, Compton, United Kingdom;
AgResearch Ltd., Wallaceville Animal Research Centre, Upper Hutt, New Zealand;
Pirbright Laboratory, Institute for Animal Health, Pirbright, United Kingdom; and
Service d’Immuno-Virologie, Commissariat à l’Energie Atomique, DSV/iMETI, Fontenay aux Roses, France
1 Correspondence: Institute for Animal Health, Compton Laboratory, High Street, Compton, near Newbury, Berkshire RG20 7NN, UK. E-mail: eric.lefevre{at}bbsrc.ac.uk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: FDC subpopulations • B cells • secondary follicle • lymphoid organs
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
FDC constitute an important cell population within the GC. They possess long and delicate cytoplasmic extensions, forming a reticular network, which makes close contact with adjacent lymphocytes. Using ex vivo-isolated FDC or cells from FDC-derived cell lines, numerous in vitro studies have shown that FDC potently increase the proliferation, survival, and differentiation of B cells [6 7 8 9 10 ]. Moreover, knock-out mice in which the FDC network is absent show an impairment of affinity maturation upon immunization with a T-dependent antigen [11 , 12 ].
FDC do not internalize and process antigen but trap them on their surface in the form of antigen-antibody complexes (immune complexes) for long periods of time [13 ]. The ability to trap immune complexes is linked directly to their expression of complement receptors (CD21, CD35) and FcRs (CD23, CD32, CD16) [14 , 15 ] and is believed to play a major role in the FDC/B cell interactions [1 , 16 ]. However, other noncognate interactions are involved in the cross-talk between FDC and B cells [17 ]. In particular, FDC-bound, complement components interact with CD21 on B cells and provide a costimulatory signal for their activation and proliferation [18 , 19 ]. Adhesion molecules also play a major role in the interaction between FDC and B cells, mainly via the LFA-1/ICAM-1 and VLA-4/VCAM-1 pathways [20 ]. Moreover, Park et al. [21 ] reported recently a membrane-bound form of IL-15 on FDC, which plays an important role in the support of GC B cell proliferation. However, most of these molecular mechanisms are involved in processes taking place within the LZ of GC. In stark contrast, our understanding of the specific molecules expressed by FDC involved in FDC-driven B cell responses within the DZ of secondary follicles is still poor.
Two major subtypes of FDC can be identified within the GC on the basis of their localization, morphology, and phenotype. FDC in the LZ display abundant cytoplasmic extensions with a high level of membrane-bound immune complexes, whereas DZ FDC display fewer cytoplasmic extensions and exhibit a low capacity to trap immune complexes [22 ]. Electron microscopy and phenotypic analysis have shown clearly that the reticular cells within the DZ were FDC. Moreover, these studies demonstrated a great heterogeneity of FDC, as up to seven subtypes were identified [23 24 25 ]. These FDC subtypes appear to derive from a common precursor under the influence of signals delivered within the different GC compartments [24 ].
FDC-specific mAb have been raised in different species and are available to study the FDC network in normal and pathological conditions. These mAb stain FDC throughout the GC or specifically in the LZ of the GC, but so far, none has been described, which stain DZ FDC specifically. Amongst these FDC-specific mAb, DRC-1, KiM4, and 7D6 mAb have been shown to recognize the long isoform of CD21 and FDC-M2 mAb to recognize the complement component C4 [26 , 27 ]. However, thus far, only one molecule expressed specifically by FDC has been described to be involved in FDC-driven B cell functions: the novel protein recognized by the FDC-specific 8D6 mAb, which stimulates B cell proliferation and differentiation [28 ]. We generated a new mAb, which recognized FDC specifically in the DZ of GC. Here, using this mAb, we describe the presence of fibrinogen on the surface of DZ FDC, and we show that fibrinogen is able to stimulate in vitro the proliferation of cells of a centroblastic cell line, which originates from the GC.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
+ T cells and was carried out at the Basel Institute for Immunology (Switzerland). Briefly, BALB/c mice were immunized with a + T cell line derived from efferent lymph of a sheep vaccinated with bacillus Calmette-Guerin (BCG). The cell line was maintained in culture by stimulating with purified protein derivative (PPD) and recombinant human (rh)IL-2. Immunization of mice and the production and screening of hybridomas followed standard procedures. During screening, one mAb (initially called Du-6-184) was observed to have unusual staining properties in the GC of ovine lymph nodes (LN). This hybridoma was transferred to the Institute for Animal Health (Compton, UK), where it was recloned, renamed as mAb D46, and characterized further as described in this paper.
Cells and tissues
Bovine and ovine tissues obtained from healthy animals raised at the Institute for Animal Health and palatine tonsils removed from children with chronic tonsilitis were embedded in Tissue-Tek® O.C.T. compound (Sakura, Zoeterwoude, The Netherlands), snap-frozen, and stored at –70°C. All the experiments were approved by the Institute’s ethical review process and were in accord with national guidelines about animal use.
Bovine PBMC were obtained by density gradient centrifugation (1200 g for 30 min over Histopaque®, 1.083 g/ml, purchased from Sigma-Aldrich, Poole, UK). Bovine B cells were subsequently isolated from these PBMC by magnetic cell sorting using a mouse anti-bovine CD21 mAb (Clone CC21) [29
] coupled to rat anti-mouse IgG1 beads (Miltenyi Biotec, Surrey, UK) and MACS® LS separation columns (Miltenyi Biotec) according to the manufacturer’s instructions. The resulting cell population consistently contained 
96% CD21+ cells.
The human hepatocyte carcinoma (Hep G2) and the human FDC-derived (FDC-1) cell lines were obtained from ECACC (Salisbury, UK) and Ed A. Clark (University of Washington, Seattle, WA, USA), respectively. The L3055 cell line, originated from a patient with Burkitt lymphoma, was obtained from Chris Gregory (The University of Edinburgh, Centre for Inflammation Research, Edinburgh, UK). These cell lines were cultured routinely in RPMI-1640 medium with glutamax-I and 25 mM Hepes containing penicillin (100 units/ml), streptomycin (100 µg/ml), 1x nonessential amino acids, 1 mM sodium pyruvate, and 10% heat-inactivated FCS, subsequently termed complete medium.
For the coculture assay, L3055 cells (5x103/well) were cultured alone or with 5 x 103 FDC-1 cells/well in 96-well plates in 200 µl complete medium in the presence of various stimuli: 1 µg/ml anti-human CD40 mAb (Clone G28.5), 50 IU/ml rhIL-2, 20 ng/ml rhIL-4, 50 ng/ml rhIL-10, 10 µg/ml polyclonal anti-IgM antibody coupled to beads (Irvine Scientific, Santa Ana, CA, USA), 1–100 µg/ml Type I fibrinogen from human plasma (Sigma-Aldrich), 1 unit/ml hirudin (American Diagnostica, Stamford, CT, USA), 10 or 20 µg/ml anti-fibrinogen mAb (Clones FG-21 and 85D4, Sigma-Aldrich), or a combination of these. Cultures were supplied with a pulse of 1 µCi/well [3H]-thymidine for the last 12 h on the 3rd day of incubation, and proliferation was measured by counting [3H]-thymidine incorporation into the cellular DNA. Results are expressed in cpm (mean cpm of triplicates±SD).
Immunofluorescence and confocal microscopy
Frozen tissue sections were incubated for 30 min with D46 mAb (IgG2a, 0.5 µg/ml), CNA42 mAb (IgM, 1/10 of hybridoma supernatant, kindly given by Georges Delsol, Toulouse, CHU Purpan, Laboratoire d’anatomie et cytologie pathologiques, France), anti-bovine CD21 mAb (CC21, IgG1, 1/500 of hybridoma supernatant or CC51, IgG2b, 1/10 of hybridoma supernatant) [29
], anti-bovine CD62L mAb (CC32, IgG1, 1/10 of hybridoma supernatant) [30
], anti-Ki67 mAb (Clone MIB-5, IgG1, 1/50, Dako A/S, Glostrup, Denmark), anti-bovine CD3 mAb (Clone MM1A, IgG1, 1 µg/ml, VMRD Inc., Pullman, WA, USA), anti-bovine CD41/61 mAb (Clone CAPP2a, IgG1, 10 µg/ml, VMRD Inc.), 12B1 mAb (IgG2a, 1/20, Beckman Coulter, Fullerton, CA, USA), anti-human CD23 mAb (Clone 9P25, IgG1, 1/50, Beckman Coulter), anti-human and/or -bovine fibrinogen mAb (Clone FG-21, IgG2a, 1/4000, or Clone 85D4, IgG1, 1/4000, Sigma-Aldrich), FITC-conjugated polyclonal rabbit anti-human fibrinogen antibody (1/200, Dako A/S), isotype-matched control TRT6 mAb (IgG2a) [31
], or a combination of these. Sections were subsequently stained for 30 min with isotype-specific, secondary antibodies: goat anti-mouse IgG1, IgG2a, IgG2b, and IgM antibodies coupled to FITC (1/200), tetramethylrhodamine isothiocyanate (TRITC; 1/200), or Alexa Fluor® 633 (1/1000). In some sections, nuclei were also stained for 5 min with 4',6-diamidino-2-phenylindole (DAPI; 1/3000). Slides were mounted in Dako® fluorescent mounting medium (Dako A/S) and analyzed using a Leica confocal microscope and software (Leica Microsystems, Knowlhill, UK).
Flow cytometry
L3055 cells (106/well) were cultured for 2 days, alone or with 5 x 104 FDC-1 cells/well in six-well plates in the presence of 1 unit/ml hirudin and various stimuli used at the same concentrations as described above. Subsequently, cell surface antigen expression was analyzed by indirect immunofluorescence using the following reagents: biotin-conjugated mouse anti-human CD95, anti-human CD80, or anti-human CD86 mAb (all from BD Biosciences, San Jose, CA, USA), followed by incubation with PE-conjugated streptavidin. Biotin-conjugated mouse control IgG1 was also purchased from BD Biosciences. TACSTM Annexin V-FITC apoptosis detection kit (R&D Systems, Minneapolis, MN, USA) was used according to the manufacturer’s instructions. Five thousand viable cells/sample (surface antigen expression) and 10,000 total cells/sample (apoptosis) were analyzed using a FACScan flow cytometer (BD Biosciences).
Preparation and biotinylation of bovine FDC-enriched LN cells and platelets
Bovine mesenteric LN were macerated in PBS and subsequently enriched for the presence of FDC by centrifugation over a 35% Percoll gradient for 30 min at 800 g. Cell surface proteins were biotinylated using the EZ-linkTM Sulfo-NHS-LC-biotinylation kit (Pierce, Rockford, IL, USA), according to the manufacturer’s instruction. Bovine platelets were obtained from blood samples of healthy calves using a method described previously [32
].
Immunoprecipitation
Bovine B cells and FDC-enriched LN cells (108/ml) or platelets (109/ml) were resuspended in lysis buffer (50 mM Tris-HCl, pH 6.8, 150 mM NaCl, 1% Nonidet P-40, 0.1% SDS, 1/100 protease inhibitor cocktail, Sigma-Aldrich) for 30 min at 4°C and then centrifuged for 30 min at 13,000 g. The total protein concentration of the bovine cell and bovine platelet lysates was similar (1.5 mg/ml, data not shown). These lysates were precleared overnight at 4°C by the addition of 10 µl Protein A-agarose beads (Santa Cruz Biotechnology, Santa Cruz, CA, USA) and 10 µg TRT3 mAb [31
] per ml lysate (equivalent to 108 cells or 109 platelets). The samples were then centrifuged at 13,000 g, and 1 ml PBS or supernatant was incubated at 4°C for 2 h with 10 µl Protein A-agarose beads, previously incubated with 10 µg TRT3 mAb, IL-A88 mAb [33
], or D46 mAb. The beads were washed three times, boiled in loading buffer, and resolved using SDS-PAGE.
Western blotting
Following immunoprecipitation and SDS-PAGE resolution, proteins were transferred electrophoretically onto a polyvinylidene fluoride (PVDF) membrane in transfer buffer (25 mM Tris, 192 mM glycine, 20% methanol) and probed with HRP-conjugated streptavidin or with D46, FG21, or 85D4 mAb followed by an incubation with HRP-conjugated goat anti-mouse Ig (Dako A/S). The membrane was exposed to an ECL detection system (Amersham Biosciences, UK), and ECL detection was visualized on a Kodak scientific imaging film (Eastman Kodak Co., Rochester, NY, USA).
Amino acid sequence analysis of D46 mAb-immunoprecipitated protein
SDS-PAGE resolved proteins were transferred electrophoretically onto a PVDF membrane in 10 mM 3-(cyclohexylamino)-1-propanesulfonic acid buffer (pH 11.0), visualized by amido black staining, excised from the membrane, and N-terminally sequenced on an Applied Biosystems 492 automated protein sequencer, with on-line chromatography of phenylthiohydantoin (PTH) amino acids by using the manufacturer’s 140C microgradient system and 220 x 2.1 mm reverse-phase columns.
RT-PCR
Total cellular RNA was isolated from Hep G2 and FDC-1 cell lines or from calf liver and mesenteric LN of healthy animals using the Purescript® RNA isolation kit (Gentra Systems, Minneapolis, MN, USA). Potentially contaminating, genomic DNA was removed by treatment with RNase-free DNase I (Promega, Madison, WI, USA). RT of 2 µg total RNA was performed in a final volume of 40 µl using the AMV RT system (Promega) in an automated thermal cycler (iCyclerTM, Bio-Rad, Hercules, CA, USA). One-twentieth of the resulting cDNAs or 2 µg human follicular lymphoma or human spleen cDNA libraries (Invitrogen, Carlsbad, CA, USA) was subjected to 30 PCR cycles for human GAPDH, bovine ß-actin, and bovine fibrinogen chain, and 50 PCR cycles for human fibrinogen chain and PCR products were resolved in a 1% agarose gel.
Primer sequences were as follows: human GAPDH sense, 5'-TGT TGC CAT CAA TGA CCC CTT-3', and antisense, 5'-CTC CAC GAC GTA CTC AGC G-3'; bovine ß-actin sense, 5'-ACT GGG ACG ACA TGG AGA AG-3', and antisense, 5'-AGG AAG GAA GGC TGG AAG AG-3'; bovine fibrinogen chain sense, 5'-GAA TTT TGG CTG GGA AAT GA-3', and antisense, 5'-ATC ATC GCC AAA ATC GTA GC-3'; human fibrinogen chain sense, 5'-GAA TTT TGG CTG GGA AAT GA-3', and antisense, 5'-ATC ATC GCC AAA ATC AAA GC-3'.
Statistical analysis
Differences between groups were assessed using an ANOVA followed by a Dunnett simultaneous test, and P values <0.01 were considered significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
D46 mAb stained cells in specific areas of bovine mesenteric LN with a reticular pattern reminiscent of FDC. To further characterize the cells recognized by this mAb, tissue sections were triple-stained with D46, anti-bovine CD21, and CNA42 mAb. Our results showed that D46 mAb stained CD21lowCNA42– reticular cells specifically in the DZ of GC within secondary follicles (Fig. 1A ). No staining was observed when using TRT6 mAb, the isotype-matched control for D46 mAb (Fig. 1B) . The characteristic reticular pattern of staining obtained with D46 mAb as well as the observation that D46+ cells expressed low levels of CD21 (Fig. 1C) strongly suggested that the cells recognized by D46 mAb in the DZ were indeed FDC. The localization of these cells within the DZ was confirmed further by their presence in the Ki67+-rich zone of the bovine GC (Fig. 1D) .
|
D46 mAb also stained reticular cells in the DZ of secondary follicles within bovine discrete Peyer’s patches (Fig. 1G) and palatine tonsils (data not shown) but was not observed in the thymus or in nonlymphoid tissues such as the liver, skin, kidney, or cerebellum (data not shown). In ovine mesenteric LN sections, D46 mAb also stained reticular cells in the DZ of secondary follicles, which was identified as the Ki67+ area within the GC (Fig. 1H) . D46 mAb did not stain human tonsil or mouse spleen and LN sections (data not shown).
Characterization of D46 antigen as a multimeric protein by immunoprecipitation
To characterize D46 antigen, we biotinylated the cell surface proteins of bovine LN cells enriched for FDC prior to their lysis. Proteins in the lysate were immunoprecipitated subsequently with D46 or control mAb and detected by Western blotting using HRP-conjugated streptavidin. An immunoprecipitation with the MHC Class I-specific IL-A88 mAb was used as a positive control and retrieved two bands under reducing conditions, one of 
40 kDa corresponding to the 
chain of MHC Class I antigen and another of less than 25 kDa corresponding to ß2-microglobulin. Under reducing conditions, the D46 mAb immunoprecipitation led to the detection of three bands of 60, 57, and 50 kDa, which were not observed in the lane corresponding to the immunoprecipitation performed with the negative control TRT3 mAb (Fig. 2A
). However under nonreducing conditions, the D46 mAb immunoprecipitation led to the detection of a single band of more than 250 kDa (Fig. 2B)
. Therefore, our results suggest that these bands represent different subunits, which associate to form a multimeric antigen. However, insufficient quantities of D46 antigen were present in LN lysates to obtain protein sequences of the different subunits.
|
|
53 kDa, corresponding to the Ig heavy chain of D46 mAb (Fig. 4A
, Lanes 1 and 3, respectively). When the D46 mAb immunoprecipitation was performed on platelet lysate, three other bands were detected at 62, 57, and 50 kDa (Fig. 4A
, Lane 2). This pattern of bands was thus similar to that obtained when the D46 mAb immunoprecipitation was performed on bovine mesenteric LN lysate. These three bands were excised to determine the N-terminal sequences of the corresponding proteins, termed S1, S2, and S3, using an automated protein sequencer. Although no sequence could be ascribed to the S2 protein, we determined the N-terminal sequences of the two other subunits of D46 antigen (Fig. 4B)
. Amino acids in parentheses within the S1 subunit sequence represent positions where more than one possible amino acid was identified. However, a FASTA search of Swissprot using the sequences DGSDPPSGDFLT (for S1 protein) and YVATRDN (for S3 protein) identified a 100% match to the bovine fibrinogen and chains, respectively (Fig. 4C)
. These results indicate that D46 mAb immunoprecipitates fibrinogen from a bovine platelet lysate, consistent with the finding that D46 antigen is a multimeric antigen of more than 250 kDa. Indeed, fibrinogen is a 340-kDa hexamer, which contains two sets of nonidentical (, ß, and ) disulfide-linked chains. It is likely that the S2 subunit (57 kDa), from which no sequence could be ascribed, corresponds to the ß chain of bovine fibrinogen (expected m.w., 56 kDa).
|
Fibrinogen is present in the DZ of bovine mesenteric LN and human tonsils
To determine whether D46 mAb recognized bovine fibrinogen in platelet lysates but a different ligand in secondary lymphoid organ sections, we double-stained bovine mesenteric LN sections with D46 mAb and a commercially available, anti-human fibrinogen mAb, which also cross-reacts with bovine fibrinogen (85D4 mAb). As shown in Figure 5A
, there is a colocalization of the reticular staining obtained with both mAb in the DZ within the GC, thus confirming that fibrinogen is present on DZ FDC in situ and that D46 antigen does indeed correspond to bovine fibrinogen.
|
Fibrinogen is produced mainly by hepatocytes in the liver, although mRNA transcripts have also been detected in other tissues such as the lung, ovary, and bone marrow under normal and pathological conditions [35
36
37
]. Therefore, we next determined whether fibrinogen was expressed in situ within the secondary lymphoid organs. For this purpose, we designed primers specific for the -chain transcripts of bovine or human fibrinogen and optimized the RT-PCR conditions for their detection. cDNA prepared from Hep G2, a human hepatocyte carcinoma cell line, and calf liver tissue was used as positive controls for these PCR reactions, and -chain mRNA transcripts were detected in both. However, we did not detect such transcripts in cDNA prepared from calf LN tissue, from the human FDC-derived cell line FDC-1, or in spleen and follicular lymphoma cDNA libraries (Fig. 6
).
|
Addition of 50 µg/ml fibrinogen significantly increased the L3055 cell proliferation obtained in the presence of anti-IgM antibody and IL-4 by 1010 ± 590% (range, 360–1760%; n=6; P<0.01; Fig. 7A , upper, left-most panel). Upon proteolysis by thrombin, fibrinogen polymerizes and generates fibrin, an insoluble extracellular matrix. To determine whether fibrinogen or fibrin was responsible for modulating the L3055 cell proliferation, we supplemented the complete medium with 1 unit/ml hirudin, a pharmacologic thrombin antagonist. Addition of hirudin to the culture did not significantly modulate the anti-IgM antibody and IL-4-induced proliferation of L3055 cells in the absence (data not shown) or presence of fibrinogen (Fig. 7A , upper left-most panel). Therefore, these results strongly suggest that the effect of fibrinogen on L3055 cell proliferation did not require its conversion to fibrin. We subsequently showed that boiling fibrinogen for 5 min reduced its functional activity significantly (Fig. 7A , upper, second from left). As potential contaminants such as endotoxins are heat-resistant [39 ], these results excluded a potential effect of endotoxins in the enhancement of L3055 cell proliferation in our culture conditions.
|
In contrast, the L3055 cell proliferation induced by anti-CD40 mAb and IL-4 (Fig. 7A , lower panels) or by anti-CD40 mAb and IL-2 + IL-10 (data not shown) was not modulated significantly by the addition of any of the fibrinogen doses tested. These results, showing that fibrinogen does not synergize with T cell-mediated signals (such as CD40-triggering) to increase B cell proliferation, are consistent with the localization of fibrinogen in the DZ of GC, where T cells are mostly absent.
We next assessed whether fibrinogen could modulate the expression of the costimulatory molecules CD80 and CD86 or of the proapoptotic molecule CD95 on L3055 cells. Stimulation in the presence of anti-CD40 mAb and IL-4 increased the level of CD80, CD86, and CD95 expression on L3055 cells. However, the increased CD95 expression was most striking. Addition of fibrinogen to the culture did not further modulate the expression of these molecules (Fig. 7B) . Stimulation in the presence of anti-IgM antibody and IL-4 differentially modulated the expression of the costimulatory and proapoptotic molecules on L3055 cells, decreasing CD86 expression but not affecting CD80 or CD95 expression significantly. As shown in Figure 7B , the addition of fibrinogen counteracted the anti-IgM antibody and IL-4-induced decrease of CD86 expression (mean fluorescence intensities of 76, 42, and 68 when cells were cultured with media alone, with anti-IgM antibody+IL-4, and with anti-IgM antibody+IL-4+fibrinogen, respectively). This result suggests fibrinogen could be involved in the maintenance of CD86 costimulatory molecule expression within the GC. The addition of FDC-1 cells to L3055 cells did not further modulate the effects of the various stimuli and fibrinogen on CD80, CD86, or CD95 expression (data not shown).
It has been shown previously that L3055 cells were prone to apoptosis when cultured in the presence of anti-IgM antibodies [38 , 40 ]. As fibrinogen induced proliferation and maintained CD86 expression of BCR-stimulated L3055 cells, we determined whether these activities could correlate with the ability of fibrinogen to protect L3055 cells from BCR-dependent apoptosis. In the absence of any stimuli, L3055 cells only displayed a low level of apoptosis, and 66% of the cells were viable by Day 2, as shown by their Annexin Vneg propidium iodideneg phenotype (Fig. 7C) . When stimulated with anti-IgM antibody and IL-4 for 2 days, the percentage of viable L3055 cells was decreased significantly to 6% of the total population. The addition of fibrinogen partly counteracted the proapoptotic effect of BCR stimulation on these cells, as the percentage of viable cells increased with the addition of increasing amounts of fibrinogen in the culture media. Indeed, addition of 50 µg/ml and 100 µg/ml fibrinogen increased the percentage of viable L3055 cells to 20% and 25%, respectively (Fig. 7C) . Although the percentages of viable cells were slightly higher when FDC-1 cells were added to the culture, fibrinogen had a similar antiapoptotic effect on BCR-stimulated L3055 cells in the absence and presence of FDC-1 cells (data not shown).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
We generated a mAb, D46, and showed that it recognizes reticular cells, specifically within ovine and bovine GC. Our observations strongly suggest that these cells are FDC: The cells recognized by D46 mAb are localized within the B cell area and express CD21; the reticular staining pattern obtained with D46 mAb is characteristic of an FDC-specific mAb; and the finding that D46 mAb stains reticular cells in the DZ of secondary follicles but not within the LZ of secondary follicles, the primary follicles, the extrafollicular area, or the nonlymphoid tissues strongly suggests that these cells are not epithelial cells or fibroblasts. In summary, we describe here for the first time a mAb, which stains FDC specifically in the DZ of the GC within LN, tonsils, and Peyer’s patches.
Different subsets of FDC have been described previously within the secondary follicles on the basis of their localization, morphology, and phenotype [23 24 25 ]. These studies show major differences between FDC in the DZ and LZ, but there is little data about differences between FDC within the primary follicles and those within the DZ of secondary follicles. Bofill et al. [43 ] reported that FDC within primary follicles and DZ formed loose reticular networks (in contrast to the more extensive networks observed in the LZ). However, they found that FDC in primary follicles stained with a fibroblastic marker, AS02, whereas DZ FDC had lost most of their reactivity to this marker. Our data confirm that FDC in the DZ and primary follicles are phenotypically distinct, as D46 antigen is localized only on FDC within the DZ of secondary follicles.
We identified the D46 antigen as bovine fibrinogen, a soluble, 340-kDa hexamer formed by two sets of three distinct, disulfide-linked chains (, ß, and ), which is known to play an essential role in blood coagulation, as it can be cleaved by thrombin to form an insoluble fibrin clot [44
]. Indeed, D46 mAb immunoprecipitated a multimeric antigen of greater than 250 kDa formed by the association of three distinct subunits from bovine platelets and bovine LN lysates. We obtained the N-terminal sequences of two of these three subunits, which matched perfectly with the sequences of bovine fibrinogen and chains. We also confirmed that D46 mAb recognizes purified fibrinogen from bovine plasma by Western blotting. Finally, D46 mAb staining colocalized with anti-bovine fibrinogen mAb staining in bovine mesenteric LN sections. This unique pattern of fibrinogen localization was not restricted to ruminant tissues, as we showed human fibrinogen is also present on DZ FDC within the palatine tonsil.
We could not detect any mRNA transcripts specific for the fibrinogen chain by RT-PCR in samples prepared from calf LN tissue, from the human FDC-derived cell line, FDC-1, or in human spleen and follicular lymphoma cDNA libraries. Fibrinogen is not only present in the blood but also in the lymph, which percolates through the secondary lymphoid organs [45
, 46
], and it can be hypothesized that fibrinogen from the lymph could be trapped by the DZ FDC. Therefore, it can be proposed that FDC within the DZ of the GC express a receptor for fibrinogen, which is not expressed by FDC present in primary follicles or in the LZ of secondary follicles. Several adhesion molecules, such as Mß2 (CD11b/CD18, membrane-activated complex-1, complement receptor 3), Xß2 (CD11c/CD18, complement receptor 4), CD41/61, and ICAM-1 (CD54), have been shown to interact with fibrinogen, and most of these are expressed by FDC [41
, 42
, 47
48
49
]. We stained human tonsillar sections with mAb specific for CD11b, CD11c, CD18, and CD54. CD11c expression was restricted to DC and/or macrophages within the GC, whereas CD11b, CD18, and CD54 expression was low on DZ FDC and high on LZ FDC (data not shown). These results are consistent with previously published data reporting the distribution of adhesion molecules on FDC subsets [50
]. Therefore, the mechanism of fibrinogen trapping by DZ FDC cannot be explained by differential expression of these adhesion molecules by DZ and LZ FDC. Alternatively, it is possible that fibrinogen is produced in situ within the secondary lymphoid organs. However, if a low frequency of mRNA transcripts specific for the fibrinogen chain were present in our samples, these could not have been detected by our PCR assay. Further work will be necessary to elucidate the mechanisms involved in fibrinogen localization on DZ FDC.
It is known that there is cross-talk between inflammatory responses and coagulation, which modulates the response to infections [51 ]. In particular, fibrinogen is able to enhance the cytokine/chemokine responses of PBMC and macrophages [39 , 52 ], and its accessory role in inflammation has been documented by numerous in vivo studies [48 ]. Moreover, fibrinogen is an acute-phase protein whose plasma concentration can increase by up to 200% 2 weeks after an inflammatory stimulus [53 ]. It is interesting that changes in the blood fibrinogen concentration following an inflammatory stimulus (increasing within the first 14 days poststimulation and subsequently decreasing to base level by Day 21) mirror the time-frame of formation of the mature, secondary follicle within the lymphoid organs during a primary immune response to a T-dependent antigen. Indeed, mature secondary follicles consisting of a DZ and a LZ within the GC appear by Day 4 following an antigenic stimulation, remain until Day 12, and gradually decrease in size thereafter in the absence of restimulation [2 ]. The detection of fibrinogen within the DZ of GC suggests that these two events could be interconnected. Further time-course studies in mice following an immunization with a T-dependent antigen will be needed to confirm this hypothesis.
As mentioned previously, fibrinogen has been shown to modulate the function of PBMC and macrophages [39 , 52 ]. As fibrinogen is localized on FDC in the DZ of GC, it was tempting to speculate that fibrinogen could also modulate the function of GC B cells and, in particular, that of the B cell blasts present in the DZ of GC. We show in this report that fibrinogen strongly enhances the proliferation and survival of BCR-stimulated L3055 cells, which express CD10, CD20, CD38, and CD77 and bind peanut agglutinin, features that are typical for GC centroblasts. As L3055 cells also express surface IgM [38 ], they likely derive from the GC founder cells, which are the first immigrant cells into the DZ of GC [54 ]. We show that fibrinogen only enhances the proliferation of L3055 cells when stimulated in the presence of anti-IgM antibody and IL-4. This effect is not modulated significantly in the presence of hirudin, which prevents the formation of a translucent, insoluble gel through the gradual conversion of fibrinogen to fibrin. This suggests that fibrinogen but not fibrin is active on anti-IgM antibody and IL-4-stimulated L3055 cells. It has been shown previously that stimulation with an anti-Ig antibody induces the apoptosis of L3055 cells [38 , 40 ]. Accordingly, we show that anti-IgM antibody and IL-4 stimulation induce a strong apoptosis of L3055 cells and decrease CD86 expression. Both of these effects are counteracted partly by the addition of fibrinogen in the culture. We show that even 5 µg/ml fibrinogen increases L3055 cell proliferation significantly, but a higher concentration (50 µg/ml) is needed to protect these cells from apoptosis. The localization of fibrinogen on FDC in the DZ could therefore be important in situ to create a microenvironment, where the concentration of fibrinogen is sufficiently high to maintain the proliferation and survival of the DZ B cell blasts. It is important that such concentrations of fibrinogen could be expected locally within the secondary lymphoid organs, as the fibrinogen concentration in rat and human plasma averages 3 mg/ml [36 , 44 , 55 ], and up to 0.5 mg/ml fibrinogen has been detected in rabbit and rat lymph [45 , 46 ]. Finally, it was not possible to show that the effects of fibrinogen on L3055 cells were modified in the presence of the FDC-1 cell line. However, it is not clear whether the FDC-1 cell line is representative of DZ FDC, and therefore, this cell line may not be able to interact appropriately with fibrinogen. To our knowledge, no DZ FDC-derived cell line has been described so far. In contrast to studies performed with other molecules expressed on FDC throughout the GC such as 8D6 [28 ], where an effect could be demonstrated on L3055 cells and total GC B cells, fibrinogen is localized only on a minor population of DZ FDC, and we could not detect any effect of fibrinogen on total GC B cell populations (data not shown). Further studies will be required to identify the discrete B cell population present in GC, which responds to fibrinogen.
Recent findings have identified some discrepancies in the role of deposition and persistence of antigen in the form of immune complexes on the surface of FDC in GC development, in affinity maturation, and in formation and maintenance of B cell memory [56 57 58 59 ]. However, it is clear that FDC also provide essential, noncognate signals to the GC B cells. Therefore, the elucidation of new FDC-specific molecules is an important step toward the resolution of the biological role of these cells. Here, we describe for the first time that fibrinogen is localized in vivo on DZ FDC in the GC and that the addition of exogenous fibrinogen in vitro increases the survival and proliferation of cells of a centroblastic cell line, which originates from the GC. The factors that sustain the survival and extensive proliferation of B cell blasts in the DZ of GC in vivo are still largely unknown. Our results suggest that fibrinogen could represent one of these factors.
|
ACKNOWLEDGEMENTS |
|---|
Received January 22, 2007; revised April 20, 2007; accepted May 4, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-deficient mice Nature 382,462-466[CrossRef][Medline] II-mediated immune-complex trapping Immunology 80,34-39[Medline]
B transcription factors in mononuclear phagocytes J. Immunol. 161,1462-1470(E)C domain of human fibrinogen-420 is a novel ligand for leukocyte integrins (M)ß(2) and (X)ß(2) Blood 98,2448-2455
| ||||||||||
