Recently human cartilage gp-39 (HC gp-39) was identified as a candidate autoantigen in rheumatoid arthritis (RA). To further investigate the relevance of this Ag in RA, we have generated a set of five mAbs to a combination epitope of complexes of HC gp-39263–275 and the RA-associated DRαβ1*0401 HLA class II molecules. FACS studies revealed that these mAb recognize specific complexes on homozygous DRαβ1*0401-positive B lymphoblastoid cells pulsed with HC gp-39263–275. The best mAb, 12A, was further characterized using a set of irrelevant DRαβ1*0401-binding peptides and truncated/elongated versions of HC gp-39263–275 itself. The minimal epitope recognized in combination with DRαβ1*0401 was HC gp-39263–273. Peptides not encompassing HC gp-39263–273 were not recognized. Three of five mAb were able to inhibit (up to 90%) the response of HC gp-39263–275-specific DRαβ1*0401-restricted T cell hybridomas to peptide-pulsed APC or purified complexes. Using mAb 12A, we have been able to identify and localize dendritic cells that present DRαβ1*0401/HC gp-39263–275 complexes in synovial tissue of DRαβ1*0401-positive RA patients, indicating local presentation of the HC gp-39263–275 epitope in the inflamed target tissue by professional APC. These data support a role of HC gp-39 in the local autoimmune response that leads to chronic inflammation and joint destruction.
T cells recognize specific antigenic peptide combined with MHC molecules by virtue of their specific TCR. The signal generated by the MHC/peptide/TCR complex leads to T cell activation. This trimolecular complex is a key element in the adaptive immune response and in autoimmunity. It is currently believed that the pathogenesis of some autoimmune diseases like rheumatoid arthritis (RA)2 includes at some point the presentation of MHC-bound processed autoantigens to the TCR of CD4+ T cells. To investigate this, two different strategies can be applied. The classical strategy is to analyze the presence of Ag-specific T cells at the primary site of inflammation, i.e., the synovial membrane in RA. However, the number of Ag-specific T cells in synovial tissue may be small and therefore difficult to detect using classical technologies. An alternative strategy would be to study the specific presentation of known autoantigens in a well-defined MHC context. This strategy should include four steps. First, identify a candidate autoantigen. Second, demonstrate that presentation of this Ag in the specific MHC context indeed is functional and can activate peripheral blood T cells in diseased individuals. Third, develop a tool to detect specifically these Ag-MHC complexes and validate this tool, e.g., by inhibition of T cell clone activation. Finally, use this tool to analyze specific Ag presentation at the primary site of inflammation in humans, namely in the synovial membrane of diseased joints.
There are several indications that human cartilage gp-39 (HC gp-39) is a candidate autoantigen in RA. Serum and synovial fluid HC gp-39 levels are elevated in inflammatory diseases (1) and correlate with disease activity in RA (2). HC gp-39 mRNA has been found in the cartilage and inflamed synovium of RA patients (3). Recently, it has been found that the presence of HC gp-39-positive monocytes in RA synovium is correlated with the degree of joint destruction (4). Furthermore, immunization of BALB/c mice with HC gp-39 resulted in the development of a chronic, relapsing arthritis (5). Interestingly, intranasal administration of the protein resulted in: 1) suppression (delayed onset and reduced inflammation) of HC gp-39-induced arthritis in BALB/c mice and 2) reduced incidence and severity of collagen type II-induced arthritis in DBA/1 mice using a semitherapeutic regime (6).
The association of RA with certain genes of the HLA class II complex, DR4, DR1 and DR10, has well been documented (7, 8). In northern Europe, DRαβ1*0401 is the most strongly associated haplotype occurring in 60–70% of RA patients. Using a DRαβ1*0401 peptide binding motif, six peptides of HC gp-39 were identified as T cell epitopes in the context of DRαβ1*0401. At least four of these peptides (103–116, 259–271, 263–275, and 326–338) are recognized by T cells of RA patients. Interestingly, HC gp-39263–275 was more prominently recognized in RA patients than in healthy controls (5), suggesting a role for this T cell epitope in initiation or maintenance of RA. This peptide was also defined as an immunodominant T cell epitope in HLA- DRα*0101/DRβ1*0401+/+/human CD4+/+/I-Aβ−/− transgenic mice immunized with HC gp-39 (9).
This report describes the generation of Abs to DRαβ1*0401/HC gp-39263–275 complexes. The specificity of these Abs is extensively analyzed, including functional blocking experiments using specific T cell clones. Finally, experiments to detect DRαβ1*0401/HC gp-39263–275 complexes in the synovium of RA patients are reported.
Materials and Methods
The medium used in all culture experiments was DMEM/HAM’s F12 modified (cat. no. 041-91825; Life Technologies, Paisley, U.K.). For hybridomas, this culture medium was supplemented with 10−4 M hypoxanthine (Sigma-Aldrich, St. Louis, MO) and 1.6 × 10−5 M thymidine (Sigma-Aldrich) and 1% of IL-6 containing supernatant of a human bladder carcinoma cell line T24 (HTB-4; American Type Culture Collection, Manassas, VA). Peptides were synthesized by solid phase peptide synthesis using an automated MilliGen 9050 synthesizer (Burlington, MA) and were purified by reverse phase HPLC. The DRαβ1*0401/HC gp-39263–275 complex (further referred to as Org38948: purified protein from BSM cells loaded for 80–90% with HC gp-39263–275) and DRαβ1*0401 protein (purified protein from BSM cells) were obtained from Anergen (Redwood City, CA). Anti-HLA/DR, L243 was purified from the hybridoma (HB 55; American Type Culture Collection) supernatant by affinity chromatography using protein G-Sepharose.
Generation of Org38948-specific mAbs
Six-week-old female BALB/c mice were immunized with either 10 or 100 μg Org38948 in the presence or absence of Freund’s adjuvant. Each mouse received four injections of the Ag at 3-wk intervals. In the absence of adjuvant, the injections were given i.p. In the presence of adjuvant, the first injection with CFA was given s.c. Then, two injections were given s.c. with IFA and the final injection was given i.p. without adjuvant. Seven days after the third injection, a blood sample was taken by orbita punction. Five days after the final injection, mice were sacrificed and erythrocyte-depleted spleen cell populations were prepared as described previously (10, 11).
For the generation of mAb, 2 × 107 erythrocyte-depleted spleen cells from all mice were pooled and incubated in DMEM/HAM’s F12, 10% calf serum (HyClone Laboratories, Logan, UT) for 1 h at 37°C on a plastic culture flask to remove the majority of monocytes. Subsequently, the nonadherent cells were submitted to three subsequent cycles of panning on DRαβ1*0401-coated culture dishes to remove B cells to irrelevant HLA-DRαβ1*0401 complexes. Subsequently, B cells directed to DRαβ1*0401/HC gp-39263–275 complexes were selected by incubating the resulting cell suspension on Org38948-coated culture dishes for 90 min at 37°C. Unbound cells were removed by careful washing. Finally, bound cells were harvested by trypsin treatment (11).
mAb producing hybridomas were generated from these selected B cells by clonal expansion and minielectrofusion as described previously (11). Briefly, selected B cells were mixed with T cell supernatant and 50,000 irradiated (2500 rad) EL-4 B5 cells at a final volume of 200 μl of DMEM/HAM’s F12, 10% calf serum in 96-well flat-bottom tissue culture plates. At day 8, supernatants of these cultures were tested for reactivity against the specific complex in an ELISA using either Org38948- or DRαβ1*0401-coated plates and goat anti-mouse Ig-HRP for detection. B cell cultures producing mAb reactive with Org38948 but not with DRαβ1*0401 were submitted to a minielectrofusion procedure. The specific B cells from these cultures were mixed with 106 NS-1 myeloma cells and serum was removed by washing with DMEM/HAM’s F12. Next, the cells were treated with pronase solution for 3 min and subsequently washed with fusion medium. Electrofusion was performed in a 50-μl fusion chamber by an alternating electric field of 30 s, 2 MHz, 400 V/cm followed by a square, high field pulse of 10 μs, 3 kV/cm and again an alternating electric field of 30 s, 2 MHz, 400 V/cm. Finally, the contents of the fusion chamber were transferred to selection medium and plated into a 96-well microtiter plate under limiting dilution conditions. At day 13 after fusion, the cultures were examined for hybridoma growth and screened again in an ELISA using either Org38948- or DRαβ1*0401-coated plates.
Well-characterized, homozygous, EBV-transformed B lymphoblastoid cell lines (BLCL) were used for Ag presentation studies (HLA-DRB1* initalics). Priess: DRα, DRβ1*0401; BSM: DRα*0101, DRβ1*0401, DRβ4*0101, DQα1*0301, DQβ1*0302, DPα1*01, DPα1*01012; YAR: DRα*0101, DRβ1*0402, DRβ4*0101, DQα1*0301, DQβ1*0302, DPα1*01, DPβ1*0401; SA9001: DRα, DRβ1*0101, DQ1, DP4; BM92: DRα*0101, DRβ1*0404, DRβ4*0101, DQα1*0301, DQβ1*0302, DPα1*01, DPβ1*0402; MGAR: DRα*0102, DRβ1*1501, DRβ5*0101, DQα1*0102, DQβ1*0602, DPα1*01, DPβ1*0401; JHAF: DRα, DRβ1*0407, DRβ4*0101, DQα1*0301, DQβ1*0301, DPα1*01, DPβ1*0301; AMALA: DRα*0102, DRβ1*1402, DRβ3*0101, DQα1*0501, DQβ1*0301, DPα1*01, DPβ1*0402; EK: DRα*0102, DRβ1*1401, DRβ3*0202, DQα1*0101, DQβ1*0503, DPα1*01, DPβ1*0402.
Using FACS analysis, binding of the mAb to different MHC-expressing EBV-transformed BLCL pulsed with various peptides was established.
Briefly, 106 BLCL were incubated with 40 μg of peptide in 500 μl of DMEM/HAM’s F12 or blank medium for 4 h at 37°C. After this incubation, the cells were washed with PBS, 2% FCS, 0.02% sodium azide. Approximately 2 × 105 cells were incubated for 1 h, 4°C with 130 μl of mAb-containing hybridoma supernatant plus 20 μl of PBS, 20% FCS, 0.02% sodium azide. After washing the cells twice with PBS, 2% FCS, 0.02% sodium azide, they were incubated for 1 h at 4°C with 50 μl of PBS, 20% FCS, 0.02% sodium azide plus 10 μl of goat anti-mouse Ig/FITC (BD Biosciences, Mountain View, CA). Subsequently, the cells were washed three times with PBS, 2% FCS, 0.02% sodium azide and finally resuspended in 400 μl of PBS, 2% p-formaldehyde.
As a control for peptide binding, cells were incubated with biotinylated HC gp-39263–275 and stained with streptavidin/PE (BD Biosciences).
Stained cells were analyzed with the FACScan (BD Biosciences). In all cases, forward and side scatter analysis was applied to eliminate dead cells and debris from further analysis.
Functional inhibition of T cell hybridomas by Org38948-specific mAb
Mouse T cell hybridomas recognizing HC gp-39263–275 were generated following immunization of HLA-DRα*0101/DRβ1*0401+/+, human CD4+/+, I-Aβ−/− transgenic mice with HC gp-39 (9, 12). In this study, three different HC gp-39263–275-specific, DRαβ1*0401-restricted T cell hybridomas were used: HCDA.5G11.4G8, HCDA.14G11.1H7, and HCDA.8B12.1D8. These were abbreviated as 5G11, 14G11, and 8B12, respectively. The minimal epitopes for stimulation of these hybridomas were established by using homozygous DRαβ1*0401-positive BLCL loaded with N-terminal- and C-terminal-truncated peptides (see below). For 5G11 the minimal epitope was found to be 265–275, and for 8B12 and 14G11 it appeared to be 264–274. Single alanine substitutions in the 263–275 region revealed further differences in specificity.
Inhibition of Ag-induced activation of T cell hybridomas by anti-Org38948 mAb was measured in two different assays. In the first assay, the T cell hybridomas were stimulated with MHC/peptide complexes in the absence of APC. Notably, the hybridomas were found to respond to the specific complex without costimulation. In the second assay, BLCL loaded with HC gp-39263–275 were used as APC.
For stimulation with MHC/peptide complexes, flat-bottom microwells were coated overnight at 4°C with 100 μl of Org38948 at a concentration of 200 ng/ml in PBS. Excess complex was removed by washing twice with PBS. Then, the wells were incubated for 1 h at 37°C with various concentrations of mAb in 100 μl of DMEM/HAM’s F12, 10% FCS. After preincubation with mAb, 100 μl of T cell hybridoma suspension was added (5G11 and 14G11 at 2 × 104 c/well; 8B12 at 104 cells/well in DMEM/HAM’s F12, 10% FCS). Cultures were incubated for two days at 37°C and finally supernatant was harvested for measurement of mouse IL-2 as readout for T cell activation.
In the second assay, homozygous DRαβ1*0401-positive BSM cells were loaded with HC gp-39263–275 by incubation of 1.2 × 106 cells with 10 μg peptide in 1 ml DMEM/HAM’s F12 for 5 h at 37°C. Then, excess peptide was washed out and the cells were irradiated with a dose of 15,000 rad. Subsequently, 2 × 104 peptide-loaded BSM cells were preincubated for 1 h at 37°C in round bottom microwells with various concentrations of mAb in a final volume of 100 μl DMEM/HAM’s F12. Then, 2 × 104 T cell hybridomas were added in 100 μl DMEM/HAM’s F12, 20% FCS. After two days incubation at 37°C, supernatant was harvested and tested for mouse IL-2.
Mouse IL-2 was determined in a double sandwich ELISA using anti-mouse IL-2 (BD PharMingen 18161D; San Diego, CA) for capture and anti-mouse IL-2/biotin (BD PharMingen 18172D) as a second Ab. Streptavidin conjugated to Europium (Wallac 1244-360; Gaithersburg, MD) was used for detection of IL-2 binding in a time-resolved fluorometer.
Synovial tissue samples
Synovial membrane biopsy tissues were obtained from 16 patients with RA, according to the American College of Rheumatology criteria (13). Their mean age was 52 years (range 28–78 years). Mean disease duration in this group was 4.5 years (range 0.1–22 years). Four spondyloarthropathy (SpA) patients with a mean age of 34 years (range 22–41 years) and a mean disease duration of 3.7 years (range 0.25–12 years), as well as four osteoarthritis (OA) patients with a mean age of 69 years (range 60–86 years) and a mean disease duration of 6.5 years (range 1–15 years), were included as patient controls. The biopsies were obtained from clinically involved knee joints by needle arthroscopy as described by Baeten et al. (14), or during joint replacement surgery. Synovial biopsies were snap-frozen and mounted in Jung tissue freezing medium (Leica Instruments, Nussloch, Germany). All patients gave their informed consent as approved by the Ethics Committee of Ghent University Hospital (Ghent, Belgium).
Immunohistochemistry on synovial sections was performed as described by Baeten et al. (4). Briefly, synovial biopsies were snap-frozen in liquid nitrogen and 5-μm cryostat sections were made. After fixation in acetone for 10 min and blocking of endogenous peroxidase with 1% hydrogen peroxide, the sections were incubated for 30 min with a pool of three different mAb to HC gp-39 (06A, 08B, and 10B), or mAb 12A. Parallel sections were incubated with irrelevant isotype-matched mAb as a negative control. The sections were subsequently incubated with biotinylated anti-mouse secondary Ab, followed by a streptavidin-peroxidase complex (DAKO, Glostrup, Denmark). The color reaction was developed using 3-amino-9-ethylcarbazole chromogen substrate. Finally, the sections were counterstained with hematoxylin. The stained synovial sections were blinded and scored independently by two observers.
To identify the cells that stain with mAb 12A, double stainings were performed. Frozen sections of synovial tissue were incubated for 30 min with mAb 12A, followed by 30-min incubation with a swine anti-mouse Ig-tetramethylrhodamine isothiocyanate (DAKO). Subsequently the sections were incubated for 30 min with one of the following FITC-labeled mouse mAb: anti-CD1a (follicular dendritic cells, clone NA1/34; DAKO), anti-CD3 (T cells, clone UCHT1; DAKO), anti-CD14 (monocytes, clone TÜK4; DAKO), anti-CD20 (B cells, clone B-Ly1; DAKO), anti-CD38 (plasma cells, clone At13/5; DAKO), or anti-CD68 (monocytes and macrophages, clone KP1; DAKO). Finally, the nuclei were counterstained with diamidinophenylindole. All incubations were conducted at room temperature in the dark and the sections were washed with PBS between all steps. Parallel sections were incubated with irrelevant isotype and concentration-matched Ab as negative control.
Generation of Org38948-specific mAbs
For the generation of mAb to Org38948, four mice were immunized with either a high or a low dose of Org38948 in the presence or absence of Freund’s adjuvant. At day 48, high titers of Abs to Org38948 and DRαβ1*0401 (ranging from 22,000 to 46,000) were found in an ELISA. No significant differences were found between the immunization schedules used.
Spleen cells of immunized mice were pooled and submitted to three cycles of panning on DRαβ1*0401-coated dishes to remove HLA-DRαβ1*0401-specific B cells. Then Org38948-specific B cells were selected on Org38948-coated dishes. After B cell culture and minielectrofusion, 5 mAb (01A, 04B, 08A, 11B, and 12A) were found that showed reactivity with Org38948, but not with DRαβ1*0401 in an ELISA (Fig. 1⇓). All Abs were of the IgG1κ isotype except mAb 04B which was IgAκ. As these mAb also do not react with HC gp-39263–275 peptide coated on polystyrene plates, and because reactivity to Org38948 could not be inhibited by free HC gp-39263–275, these mAb are likely to be directed to a combination epitope of HC gp-39263–275 bound to DRαβ1*0401. Further, absence of binding to peptide HC gp-39263–275 alone was confirmed in a BIAcore experiment (data not shown).
For all mAb, the reactivity to Org38948 was confirmed by immunoprecipitations. In these experiments, the mAb immunoprecipitated a molecule of 60 kDa which dissociates into two molecules of 33 and 28 kDa when run on a SDS/PAGE under reducing conditions (data not shown). This finding is consistent with the molecular mass of the α-chain and β-chain of HLA-DR, respectively.
Anti-Org38948 mAb recognize DRαβ1*0401-positive BLCL loaded with HC gp-39263–275
As the Abs are intended for the detection of MHC/peptide complexes on APC, we investigated whether the Abs also recognize DRαβ1*0401/HC gp-39263–275 on BLCL. For this purpose, two DRαβ1*0401-positive BLCL (BSM and Priess) were loaded with HC gp-39263–275, and reactivity of the Abs to these cells was compared with cells that were not loaded with this peptide. Fig. 2⇓, a–f, shows that all Org38948-specific Abs, except mAb 11B, discriminate between peptide-loaded and nonloaded BSM. The best and most specific staining was obtained with mAb 12A, making this Ab a powerful tool for monitoring MHC/HC gp-39 complexes by immunohistochemistry on synovial tissue and investigating their role in autoimmune responses. Similar results were obtained with Priess cells (data not shown).
In a next experiment, it was investigated whether the anti-Org38948 mAb cross-react with DRαβ1*0401 loaded with a set of different peptides. Priess cells were pulsed with peptides that bind well to DRαβ1*0401 (5) and were subsequently stained with mAb. The FACS plots in Fig. 2⇑, g and h, and the data in Table I⇓ show that no cross-reaction was observed with DRαβ1*0401 loaded with HC gp-39-derived peptides, other than HC gp-39263–275, that accommodate a DRαβ1*0401-binding motif. Also no cross-reaction was found with DRαβ1*0401 loaded with unrelated peptides from IHA and Mycobacterium leprae. Poor reactivity of mAb 11B and background staining with mAb 08A is in line with the previous experiment (Fig. 2⇑). mAb 12A also recognizes biotinylated HC gp-39263–275 in the context of DRαβ1*0401 whereas mAb 01A, mAb 04B, and mAb 08A do not. As the biotin is coupled to the N terminus of the peptide, this suggests that the latter mAb recognizes an epitope in the proximity of the N terminus of the peptide in the complex.
Fine specificity of mAb 12A with respect to the peptide
In the previous experiment, it has been shown that mAb 12A recognizes both biotinylated-HC gp-39263–275 and HC gp-39263–275 in the context of DRαβ1*0401. By studying binding to various truncated peptides of HC gp-39263–275 in the context of DRαβ1*0401, we further mapped the epitope recognized by mAb 12A. At the C terminus two amino acids can be removed without loss of binding, whereas at the N terminus no truncations are allowed (Table II⇓). Thus, the minimal epitope recognized by mAb 12A is HC gp-39263–273. Therefore, recognition of DRαβ1*0401/peptide complexes by mAb 12A is not restricted to DRαβ1*0401/HCgp-39263–275, but also extends to DRαβ1*0401/HC gp-39263–274 and DRαβ1*0401/HC gp-39263–273. For clarity we will only use the term DRαβ1*0401/HC gp-39263–275. Further, mAb 12A does not react with a peptide that is elongated by two amino acids at the N terminus (DRαβ1*0401/HC gp-39261–275). The epitope recognized by mAb 12A is different from the epitope recognized by hybridoma 8B12. Besides truncation of two amino acids at the C terminus, hybridoma 8B12 reactivity also allows removal of two amino acids at the N terminus.
Fine specificity of mAb 12A with respect to the MHC molecule
To further establish the specificity of mAb 12A, we investigated whether binding of this Ab is restricted to HC gp-39263–275 in the context of DRαβ1*0401. For this purpose, well-characterized homozygous BLCL carrying different HLA-DR molecules (Table III⇓) were loaded with HC gp-39263–275 and subsequently stained with mAb 12A. In a series of experiments, mAb 12A stained HC gp-39263–275 bound to DRαβ1*0401 or DRαβ1*0407 (Table III⇓). At normal concentrations of Ab (3 μg/ml), no staining was observed of the peptide bound to DRαβ1*0101, DRαβ1*0404, DRαβ1*1402 (RA-susceptible haplotypes), DRαβ1*0402, DRαβ1*1301, DRαβ1*1401 (closely related, not RA-susceptible haplotypes), or DRαβ1*1501 (more distantly related, not RA-susceptible haplotype). At extremely high concentrations of Ab (>200 μg/ml), a weak reactivity was found with HC gp-39263–275 bound to DRαβ1*0404 or DRαβ1*1401 suggesting a low affinity of mAb 12A for these complexes. In the controls of these experiments, it was established that 1) mAb 12A does not bind to nonloaded BLCL, 2) HC gp-39263–275 binds to the BLCL and 3) all BLCL show a high level of DR-expression (data not shown).
Anti-Org38948 mAb inhibit DRαβ1*0401/HC gp-39263–275-induced activation of T cell hybridomas
In the next experiments, we investigated whether the Abs (mAb 01A, mAb 08A, and mAb 12A) are able to inhibit activation of HC gp-39263–275-specific, DRαβ1*0401-restricted T cell hybridomas.
In the first experimental set-up, T cell hybridomas were stimulated with Org38948 complexes coated on polystyrene microplates. Fig. 3⇓a shows that all mAb inhibited activation of hybridoma 5G11 in a dose-related fashion. Using mAb 01A, a partial inhibition was obtained as compared with a control IgG mAb. In contrast, incubation with mAb 08A and mAb 12A resulted in complete inhibition at a concentration of 25 μg/ml. The complex-specific Abs are less potent inhibitors of T cell hybridoma activation than anti-HLA/DR mAb, L243, which may be due to differences in affinity of the Abs. Similar results were obtained using hybridoma 14G11 (Fig. 3⇓b). Hybridoma 8B12 was inhibited less well (Fig. 3⇓c) which is in agreement with the observation that this hybridoma requires less complexes to become fully stimulated (data not shown).
In another experimental set-up, T cell hybridomas were stimulated by BSM BLCL pulsed with HC gp-39263–275. As can be deduced from Fig. 4⇓, all Abs were found to inhibit peptide-induced activation of hybridomas 5G11 and 14G11 at a concentration of 10 μg/ml. Again, stronger inhibition was obtained with anti-HLA/DR mAb, L243.
DRαβ1*0401/HC gp-39263–275 complexes are presented on APC in the synovia of RA patients
With the availability of a well-characterized Org38948-specific Ab (mAb 12A), we have a tool to investigate the in situ presence and localization of specific MHC/peptide complexes at the level of single cells in synovial tissue. Synovial tissue sections of 16 RA patients, 4 OA patients, and 4 SpA patients were tested by immunohistochemistry for HC gp-39 expression and for presentation of HC gp-39263–275 in the context of DRαβ1*0401, using a set of three anti-HC gp-39 mAb and mAb 12A, respectively.
HC gp-39 expression was found in seven of eight DRαβ1*0401-positive RA patients, in five of eight shared epitope-negative RA patients, and in six of eight non-RA controls (Table IV⇓). Both in RA and controls, multiple HC gp-39-positive mononuclear cells were detected throughout the synovial lining and sublining layer (Fig. 5⇓, d, f, and h). Both the presence and the pattern of positive staining were concordant with the previous description of HC gp-39 expression in synovium (4).
In contrast, staining with mAb 12A showed clearly different results (Table IV⇑). DRαβ1*0401/HC gp-39263–275 complexes were detected in five of eight DRαβ1*0401-positive RA patients. mAb 12A stained single cells with a dendritic cell-like morphology (Fig. 5⇑, a and b). For all stainings, the isotype controls were completely negative (Fig. 5⇑c). The number of mAb 12A-positive cells was relatively small, mostly three to five cells per biopsy sample. These cells were located in the sublining layer of the synovium, mostly in the vicinity of lymphoid aggregates, whereas no positive cells were seen in the lining layer. Thus, the staining pattern was clearly different from the anti-HC gp-39 staining of the same samples (Fig. 5⇑d). Because both the morphology and the localization within the tissues suggested the possibility that mAb 12A recognizes specific dendritic cells in the synovial membrane, double immunofluorescence stainings were performed to identify the phenotype of the mAb 12A-positive cells. There was no colocalization between mAb 12A staining (Fig. 5⇑i) and phenotypic markers of T cells (CD3), B cells (CD20), plasma cells (CD38), monocytes (CD14), and macrophages (CD68). However, mAb 12A-positive cells stained clearly for the dendritic cell marker CD1a (Fig. 5⇑j).
Interestingly, no staining with mAb 12A was observed in eight DRαβ1*0401-negative RA patients which is in agreement with the specificity of mAb 12A. Moreover, no staining was observed in 8 SpA and OA controls (Fig. 5⇑, e and g).
In the context of the association of RA with the Ag-presenting HLA molecule DRαβ1*0401 and the possible role of HC gp-39 as autoantigen in this disease, the present report describes the generation of specific mAb to DRαβ1*0401/HC gp-39263–275 and their application in the study of autoimmune inflammation.
Five mAb were generated to a combination epitope of HC gp-39263–275 and DRαβ1*0401. The mAb also recognize specific complexes on intact DRαβ1*0401-positive cells loaded with HC gp-39263–275, and do not react with other HC gp-39 peptides or unrelated peptides in the context of the same HLA class II molecule on these cells. However, three of five mAb (mAb 01A, 08A, and 11B) have shown some background reactivity with at least one of the BLCL that was not loaded with the specific peptide. Such background staining of cells not exposed to the cognate foreign Ag has also been observed by other investigators with most Abs directed against other peptide-MHC class II and peptide MHC class I (15, 16, 17, 18). This background staining might be due to cross-reactivity with DRαβ1*0401 occupied with naturally processed peptides derived either from serum proteins or endogenous peptides from the BLCL. Two Abs (mAb 04B and 12A) have shown high specificity for BLCL loaded with HC gp-39263–275. The most intense staining was obtained with mAb 12A which probably makes this Ab the best candidate for further studies on autoantigen presentation. However, mAb 04B might be suitable as well because the lower intense staining can be explained by the isotype (IgA) of the Ab which probably does not react equally well with the anti-Ig-FITC conjugate used for detection.
Because mAb 12A seemed the most interesting candidate, the epitope recognized by this mAb was further mapped and the restriction was characterized. The minimal epitope recognized in the context of DRαβ1*0401 is RSFTLASSETG (HC gp-39263–273). Recognition by mAb 12A seems restricted to the epitope in the context of DRαβ1*0401 and DRαβ1*0407. Using a normal Ab concentration for staining, no recognition was observed of the same epitope in the context of other RA-susceptible alleles (DRαβ1*0101, DRαβ1*0404, and DRαβ1*1402) and a number of nonsusceptible alleles (DRαβ1*0402, DRαβ1*1301, DRαβ1*1401, and DRαβ1*1501). At 100-fold higher concentrations of mAb 12A, HC gp-39263–275 is also recognized in the context of DRαβ1*0404 or DRαβ1*1401 which might be explained by a low affinity of mAb 12A for these complexes. Cross-reactivity of mAb 12A for HC gp-39263–275 in the context of DRαβ1*0401 and DRαβ1*0407 may be explained by the high level of homology in these MHC molecules. Both molecules differ only in two amino acids, namely lysine and arginine at position 71, and alanine and glutamate at position 74. Moreover, one of these changes is conservative in nature. The fact that mAb 12A recognizes the HC gp-39 epitope in the context of DRαβ1*0401 and DRαβ1*0407, and not with DRαβ1*0404, DRαβ1*1401, DRαβ1*0402, DRαβ1*1301, and DRαβ1*1501 may be explained by the glycine at position 86 of the β-chain which is located in the region where the N terminus of the peptide binds to the MHC molecule. The other MHC molecules tested, except DRαβ1*0101 and DRαβ1*1402, have a valine at this position. DRαβ1*0101 and DRαβ1*1402, however, have more differences in the rest of the molecule which may influence the conformation of the MHC/peptide complex.
Considering the high specificity of these mAb and the importance of the MHC/peptide/TCR reactions in autoimmune inflammation in general and in RA in particular, several applications for such mAb can be envisaged: quantification of DRαβ1*0401/HC gp-39263–275 complexes on APC, purification by affinity chromatography of MHC/peptide complexes, and studying T cell recognition and MHC function. The present study investigated two particular applications, namely the inhibition of T cell responses and the in situ presentation in tissue sections of peptides encompassing HC gp-39263–275 in the context of DRαβ1*0401.
In the first set of functional experiments, three of five Abs (mAb 01A, 08A, and 12A) inhibited the response of specific T cell hybridomas to purified Org38948 complexes or BLCL loaded with HC gp-39263–275. As it is assumed that the epitope encompassing 263–275 of HC gp-39 is involved in the pathogenesis of RA, such Abs may have potential for therapeutic studies. Therapy using Abs against defined MHC/peptide complexes would definitely be more specific than treatment with anti-DR Abs (19, 20, 21) or immunization with hypervariable domains of DRαβ1*0401 (22, 23) which would probably result in unacceptable immunosuppresion. Previous studies in a mouse experimental autoimmune encephalomyelitis model have already demonstrated that specific Abs to myelin basic protein/Ias complexes can specifically block the response to an encephalitogenic determinant and inhibit the in vivo manifestations of experimental autoimmune encephalomyelitis in H-2s mice (24). The high Ab concentration required for complete inhibition of the T cell hybridomas does not necessarily reflect a potential therapeutic situation in vivo because it is expected that T cell hybridomas need less MHC-peptide complexes to become activated than normal T cells.
The second functional application was a proof-of-concept study of specific autoantigen presentation at the primary site of autoimmune inflammation in RA, namely the synovial membrane. In situ localization of APCs bearing particular TCR ligands would be valuable in characterizing the cell-cell interactions involved in initiation, propagation, and maintenance of T cell immune responses. The conventional way to detect specific MHC-peptide complexes relies on the activation of T cells bearing relevant TCR. However, such functional assays cannot be used to identify TCR-ligand-bearing APC in tissue sections. Using mAb 12A, we have been able to identify and localize cells within the synovial tissue of RA patients that process and present HC gp-39263–275 in the context of DRαβ1*0401. Both the morphology of these cells and staining with the dendritic cell marker CD1a suggest a dendritic cell phenotype. Other MHC class II-expressing cell types, e.g., B cells and macrophages, do not seem to be involved in presentation of this epitope. The location of MHC/HC gp-39263–275 expressing cells is clearly distinct from the location of HC gp-39-positive cells which were previously identified as CD14dimCD16+ monocytes (4). This suggests that MHC/HC gp-39263–275-expressing cells have taken up the protein from the environment (rather than produced it themselves), and subsequently loaded MHC molecules with endogenously processed peptides. The formal demonstration of HC gp-39263–275 presentation by synovial cells in RA, strongly supports the role of HC gp-39 as candidate autoantigen in RA as previously proposed by Verheijden et al. (5).
Moreover, it indicates that autoantigen presentation in RA can occur at the primary site of inflammation and not exclusively in lymph nodes at distant sites, although the present study did not assess whether this Ag presentation resulted in a functional interaction with T cells. If such an interaction takes place, it may be an interesting therapeutic strategy to induce regulatory T cells recognizing the same MHC/peptide complex, e.g., by mucosal Ag administration. Indeed, these regulatory T cells would than be expected to be restimulated at the specific site of inflammation. Independently of the putative T cell contact, the local presentation of HC gp-39263–275 by specific HLA-DR molecules may be a specific event for RA synovitis compared with other types of inflammatory synovitis. Interestingly, HC gp-39263–275 bound to RA-associated DRαβ1*0401 molecules have been found in the synovium of five of eight RA patients. They are present in both early (2 mo) and late (22 years) disease stages. Further studies will be performed to analyze staining with mAb 12A in a large number of RA patients and patients with other inflammatory joint diseases to analyze both the specificity of the autoantigen presentation in the synovium and possible applications in the histological diagnosis of early arthritis.
We thank J. Vermeersch for immunohistochemical studies, Dr. A. Cope and H. Hubers for generation of T cell hybridomas, Dr. G. Verheijden and B. Westrek for peptide binding studies, and Dr. C. van Staveren for peptide synthesis.
↵1 Address correspondence and reprint requests to Dr. Peter Steenbakkers, Department of Pharmacology, RH 1123, N. V. Organon, P. O. Box 20, 5340 BH Oss, The Netherlands. E-mail address:
↵2 Abbreviations used in this paper: RA, rheumatoid arthritis; HC gp-39, human cartilage gp-39; BLCL, B lymphoblastoid cell line; SpA, spondyloarthropathy; OA, osteoarthritis; IHA, influenza hemagglutinin.
- Received December 4, 2002.
- Accepted March 21, 2003.
- Copyright © 2003 by The American Association of Immunologists