Molecular analysis of rheumatoid factor (RF)-negative B cell hybridomas from rheumatoid synovial tissue: evidence for an antigen-induced stimulation with selection of high mutated IgV_H and low mutated IgV_L/λ genes

Abstract

The mutational pattern of IgV_H and IgV_L genes from synovial tissue B cell hybridomas (n = 8) of patients (n = 4) with rheumatoid arthritis (RA) was analysed, which had been produced by the electrofusion technique without prior in vitro stimulation. The molecular data were correlated with immunohistopathological data and parameters of local disease activity. The IgV_H genes of the B cell hybridomas belonged to the V_H3 family (DP42; DP47, n = 2; DP53), the V_H1 family (DP75), the V_H4 family (DP71) and the V_H5 family (DP73); 7/7 IgV_H genes showed somatic mutations, the R/S ratio (CDR) was > 3 in 4/7 IgV_H genes and the mean R/S ratio of all IgV_H genes was 9.3 (CDR) and 1.0 (FR), suggesting an antigen-dependent selection. The IgV_L/λ genes belonged to the Vλ1 family (DPL2, DPL5, DPL8nf), the Vλ2 family (DPL11, n = 2) and to the Vλ6 family (IGLV6S1); 6/6 IgV_L genes showed somatic mutations, the R/S ratio (CDR) was > 3 in 3/6 IgV_L genes and the mean R/S ratio of all IgV_L was 3.0 (CDR) and 2.3 (FR), suggesting an antigen-dependent selection. The synovial tissue exhibited germinal centres in the follicles (3/4), with the unique distribution of Ki-M4⁺ follicular dendritic cells and Ki-67⁺ proliferating cells and a dominance of IgA⁺ plasma cells (3/3). All patients were positive for RF in serum and exhibited severe local symptoms (swelling 4/4; warmth 4/4; effusion 2/4), whereas the hybridomas were negative for RF. Since B cell hybridomas showed hypermutation and affinity selection for IgV_H and IgV_L/λ genes and the patients exhibited severe local symptoms with germinal centres in synovial tissue, this study indicates that an antigen-driven process is behind the B cell expansion in the synovial tissue of clinically affected joints. These mutated B hybridomas were negative for RF, thus suggesting that antigens different from RF are also involved in the local B cell expansion and in the chronic synovitis of RA.

INTRODUCTION

In rheumatoid arthritis (RA) the synovial tissue shows a complex inflammatory infiltration of macrophages, lymphocytes and antigen-presenting cells (APC). Although B lymphocytes and plasma cells may—depending on their histopathological type [1]—be amongst the dominant classes of infiltrating cells [2,3], it is still unclear whether these cells are trapped antigen-dependent or antigen-independent solely due to a complex pattern of B cell-specific cytokines [4,5].

Since somatically mutated IgV genes are suggestive of an antigen-dependent response [6], in various studies the mutational pattern of IgV genes has been analysed. Studies carried out with synovial tissue isolates [7–11] demonstrated predominantly mutated IgV_H and IgV_L genes, indicating local B cell maturation. Of course, these studies exclude the specificities of B cells and in most cases either IgV_H or IgV_L genes were analysed.

B cell hybridomas have been analysed from different compartments: peripheral blood, synovial fluid [12,13] and bone marrow [14], and from synovial tissue with prior in vitro simulation/Epstein–Barr virus stimulation [12,15,16]. B cells from the haematological compartment may not directly correspond to those under the pathological conditions in the affected joints and in vitro stimulation modifies and selects B cells due to polyclonal stimulation [17]. Moreover, a correlation of molecular data with immunohistochemical and clinical data, which could help to complete the picture of B cell activity in RA, is lacking.

In this study IgV_H and IgV_L genes were obtained for analysis from RF⁻ B cell hybridomas which where produced by the electrofusion method from synovial tissue of four patients with definite RA. Electrofusion may be carried out without prior in vitro stimulation owing to high fusion yields [18] and therefore allows an artefact-free analysis of the B cell response. The molecular data were compared with the immunohistopathological findings in the synovial tissue and with data from local disease activity. This comparison should help to clarify whether the B cell response in synovial tissue is antigen-dependent and linked to local manifestations of RA.

PATIENTS AND METHODS

Patients, disease activity and tissue samples

Synovial tissue samples (two from the knee, one from the elbow, and one from the wrist) from four patients with confirmed seropositive RA [19] were obtained at synovectomy and snap-frozen. The patients (one man and three women) were between 54 and 79 years of age. All patients were receiving anti-rheumatic medication (gold, methotrexate and sulphasalazine). In the present study, the degree of local disease activity was scored according to the method described in [20,21] on the basis of: (i) warmth, (ii) effusion, and (iii) swelling. In all cases, part of the synovial tissue (approx. 50%) was fixed in formalin and embedded in paraffin (Giemsa, haematoxylin–eosin (H–E) staining) to be used for diagnosis and scoring of the degree of inflammation. The only material used for immunohistochemical analyses was tissue which exhibited macroscopic signs of inflammation, taken from at least three different regions of the resected synovial membrane.

Immunohistochemistry

For immunohistochemical staining 7-μm cryosections, 5-μm deparaffinated tissue sections (mounted on poly l-lysine-coated slides) and air-dried cytospin preparations were used which had been produced after tissue dissociation. Immediately before staining, the cryosections and cytospin preparations were treated with acetone for 10 min, air-dried at room temperature (10–20 min), and the following immunohistochemical procedures were performed as described [21]. In brief:

(i) The indirect immunoperoxidase technique involved incubation of the slides with the following primary MoAbs: CD3, CD22 (To15), CD68 = Ki-M8, Ki-M4, κ, λ, IgM, IgA, IgG and Ki-67 (Dianova, Hamburg, Germany) used in dilutions of 1:300, 1:200, 1:5000, 1:500, 1:10, 1:10, 1:200, 1:200, 1:300 and 1:500, respectively. Negative controls were always obtained on parallel slides by replacing the primary MoAb with PBS or neutral MoAb (pancytokeratin).

(ii) The alkaline phosphatase anti-alkaline phosphatase (APAAP) technique was performed in the same manner using a rabbit anti-mouse bridging serum. Fast blue salt was used, giving a bright, intense blue staining (CD22, Ki-M4). No counterstaining was performed.

(iii) The combined immunostaining was performed by combining (i) the indirect immunoperoxidase technique (Ki-67) with (ii) the APAAP technique (Ki-M4). No counterstaining was performed. In all cases, control staining was performed and single stainings were compared with double stainings in order to ascertain that the pattern of immunohistochemical reaction remained unaltered.

Semiquantitative analysis of κ- and λ-expressing cells

The quantity of κ- and λ-expressing B cells was analysed in immunoperoxidase-stained, serially produced paraffin sections (κ, λ) by counting the number of positive cells in 10 different high power fields. The numbers of cells from all fields were added, and the data from the patients were compared statistically by Student's t-test. The evaluation was performed by V.K.

Immunophenotyping of isolated synovial tissue cells in cytocentrifuge preparations

In order to determine the composition of an isolated synovial tissue cell population immunohistochemical staining was performed in cytospin preparations as previously described [21]. The following MoAbs were used: CD22 (To15), IgG, IgM, IgA, CD3 and CD68 (Ki-M8). Other cell types than leucocytes, such as fibroblasts, smooth muscle cells and endothelial cells, were excluded in the evaluation.

Histopathological score of inflammatory infiltration

In all cases, a portion of synovial tissue (approx. 50%) was fixed in formalin and paraffin-embedded (Giemsa, H–E staining) for use in diagnosis and scoring of the degree of the inflammatory infiltration, which in this study was performed according to [21] on a semiquantitative 1–5 scale: 1, very low inflammatory infiltration: the synovial intima is slightly enlarged (two to three cell layers thick); the degree of lymphocytic infiltration is very low, showing a diffuse pattern; and the subsynovial region exhibits chronic oedema with slight fibrosis; 2, low inflammatory infiltration: the synovial intima is slightly enlarged (two to three cell layers thick) and the degree of inflammatory infiltration is low with a diffuse perivascular lymphocytic and plasma cell infiltration; the subsynovial region shows chronic tissue granulation with moderate fibrosis; 3, moderate inflammatory infiltration: the synovial intima is moderately enlarged (three to five cell layers thick) and the degree of lymphocytic infiltration is moderate, with small follicle-like aggregates near small blood vessels; there is moderate cellularity of the subsynovial region, which exhibits slight fibrosis; 4, strong inflammatory infiltration: the synovial intima is extensively enlarged (five to 10 cell layers) and multinucleated giant cells are frequently detected. Lymphocytes exhibit a dense follicle-like pattern; the ‘interfollicular’ area exhibits high cellularity with fibrosis; granulocytes may be detected in the perivascular area as well as near the synovial lining cells; 5, very strong inflammatory infiltration: the synovial intima is extensively enlarged and the distribution of lymphocytes exhibits a dense follicle-like pattern with formation of germinal centres; granulomas and hemigranulomas can be seen in the subsynovial region. The interfollicular area exhibits a very high cellularity with activated fibroblasts, and granulocytes show a focal and diffuse distribution. In each histopathological analysis, 10 fields were examined, and the most prominent finding in a given field determined the score.

Electrofusion protocol

The protocol was carried out as previously described [18]. Briefly, the cell suspensions (lymphocytes, HAB-1 and Ag8) were harvested and mixed in a ratio of 1:1. The cell mixture was washed with fusion medium and after the total suspension density had been adjusted to 2 × 10⁶ cells/ml, 200 μl were poured into the helical fusion chamber and aligned electrically using the Biojet CF equipment (Biomed, Theses). Fusion was induced 30 s after alignment by injection of a field pulse of 30 V amplitude and 15 μs duration followed by post-alignment with a constant alternating field of the same amplitude and frequency for a further 30 s. Ten minutes later the pulsed cell suspension was removed carefully from the fusion chamber, diluted with complete growth medium and transferred to a 96-well cloning plate. Twenty-four hours later the medium was replaced by HAT-selection medium. After 3–4 weeks the supernatants of the growing hybridomas were screened for antibody production by an ELISA. The positive B cell hybridomas were tested again in the same way and recloned by limiting dilution as described [22]. Sterile supernatant which was used for further analysis was concentrated using a membrane filter system (Amicon; Centrion, USA; 100 kD). The concentrated antibodies were adjusted to approx. 1–3 μg/ml in PBS (pH 7.4) and stored at −70°C pending further analysis.

Screening for rheumatoid factor reactivity

The concentrated supernatants of the B cell hybridomas positive for IgM/λ, IgA/λ and IgG/λ and the sera were screened for RF reactivity by the latex agglutination ELISA test. The latex assay was performed by application of 4 μl of human IgG-coated latex particles (Rapi Tex-RF; Behring, Marburg, Germany) to 50 μl of unknown supernatants. The agglutination was determined microscopically after 20 min. For positive control the RF⁺ B cell hybridoma E1 was used in identical concentration. This B cell hybridoma was obtained from prestimulated peripheral blood B cells from a RA patient and had germ-line configuration IgV_H and IgV_L genes which belonged to germ-lines DP-75 and DPL11, respectively.

Molecular analysis of B hybridoma IgV genes

Amplification and cloning of DNA

RNA was prepared from approx. 1 × 10⁶ hybridoma cells and cDNA was synthesized using the following primers: Cg-(5′) and CL-(5′). DNA was amplified with the following primers [14,23] specific to the rearranged H- and L-chain variable regions (all given in the 5′-3′ direction): V_H1 CCTCAGTGAAGT(CT)TCCTGCAAGGC;

V_H2 GTCCTGCGCTGGTGAAA(GC)CCACACA;

V_H3 GGGGTCCCTGAGACTCTCCTGTGCAG;

V_H4 GACCCTGTCCCTCACCTGC(AG)CTGTC;

V_H5 AAAAAGCCCGGGGAGTCTCTGA(AG)GA;

V_H6 ACCTGTGCCATCTCCGGGGACAGTG;

J_H1–5 GGTGACCAGGGT(TGC)CC(TC)TGGCCCCAG;

J_H6 GGTGACCGTGGTCCCTTGCCCCCAG;

V_L2–3 ACCTAGGACGGTCAGCTTGGTCCC;

V_L4–5 ACCTAAAACGGTGAGCTGGGTCCC;

L1 CAGTCTGTGTTGACGCAGCCGCC;

J_L2 CAGTCTGCCCTGACTCAGCCTGC;

J_L3a TCCTATGTGCTGACTCAGCCACC;

J_L3B TCTTCTGAGCTGACTCAGGACCC;

J_L6 AATTTTATGCTGACTCAGCCCCA. After 30 polymerase chain reaction (PCR) cycles, amplified DNA was separated on an agarose gel and a band of approx. 350 bp was cut out and cloned onto the pCR II vector using the TA cloning system (Invitrogen, San Diego, CA).

DNA sequence determination

Variable regions were sequenced by a modified Taq sequencing protocol (Unites States Biochemical, USA). In short, inserts were sequenced from both sides by using a −40 or +40 digoxigenin (DIG) end-labelled primer (TIB; Mol Biol, Berlin, Germany). Sequencing reaction products were applied to a direct blotting machine (MWG Biotech, Ebersberg, Germany) and run onto Nylonbind A membrane (Serva, Heideberg, Germany). DNA was stained with a DIG detection kit (Boehringer, Mannheim, Germany)

Sequence analysis

To identify germ-line genes, searches of the EMBL/GenBank database (February 1998, Heidelberg, Germany) were performed. In addition, sequences were compared with the database V BASE (J. M. Tomlinson, Medical Research Council, Centre for Protein Engineering, Cambridge, UK) using the Macintosh program DNA PLOT 1.4 (W. Müller, Institut für Genetik, Cologne, Germany).

RESULTS

Disease activity in patients

All patients (three female, one male) exhibited severe local symptoms with the combinations of swelling, warmth and effusion (2/4) and swelling and warmth (2/4). All patients were positive for RF (4/4) and had moderate (2/4) and high (2/4) erythrocyte sedimentation rate (ESR) values and had long duration of disease (8–16 years).

Histopathology and immunohistochemistry of rheumatoid synovial tissue

Histopathological/immunohistochemical findings and inflammatory score are shown in Table 1. One patient (71) showed low inflammatory infiltration (score 2) with absence of follicular dendritic cells (FDC). Inflammatory cells were diffusely distributed, and lymphocytic aggregates with perivascular localization but no follicles could be recognized. One patient (56) had a strong inflammatory infiltration (score 4) with several B cell follicles (Fig. 1a,b) having germinal centres exhibiting an FDC-rich compartment (Fig. 1c) with a low content of proliferating Ki-67⁺ cells (Fig. 1e) and a compartment with few FDC but a high content of Ki-67⁺ proliferating cells. Two patients (54,58) exhibited a very strong inflammatory infiltration (score 5) with several germinal centres contained within B follicles. The synovial intima was extensively enlarged and contained several multinucleated giant cells.

Table 1.

Histopathology of synovial tissue

Fig. 1.

Histopathology and immunohistochemistry (d,f, immunoperoxidase; b,c, alkaline phosphatase anti-alkaline phosphatase; e, a combination of both) from rheumatoid synovial tissue of patient 56, exhibiting a follicular inflammatory infiltrate with enlarged synovial intima (a, Giemsa staining, arrow points at synovial intima) and large aggregates of CD22⁺ B lymphocytes (b) which exhibit a network of follicular dendritic reticulum cells (c) with a compartment of Ki-67⁺ proliferating cells (e, arrow points at Ki-67⁺ cells) and a compartment of follicular dendritic reticulum cells (e, intense blue area). Expression of κ (d) and λ (f) in plasma cells of 5-μm serial sections exhibiting no significant dominance of κ⁺ cells (original mag.: a–f × 250).

No statistically significant differences between the quantity of κ- or λ-expressing B cells could be found in any case (Fig. 1d,f).

Cytospin preparation

Immunophenotyping of isolated cells from synovial tissue used for electrofusion showed that the inflammatory infiltrate consisted of 9–12% CD22⁺ B lymphocytes, 1–6% IgM⁺ plasma cells, 7–11% IgA⁺ plasma cells, 1–3% IgG⁺ plasma cells. Percentage of B cells (sum of plasma cells and B lymphocytes) was 21–28%, CD3⁺ T lymphocytes 26–33% and CD68⁺ macrophages 31–40%.

Fusion yields

A total of 16 immunoglobulin-producing B cell hybridomas was obtained from series 54, five immunoglobulin-producing B cell hybridomas from series 56, 25 immunoglobulin-producing B cell hybridomas from series 58 and 20 immunoglobulin-producing B cell hybridomas from series 71. During recloning, a part of the B cell hybridomas ceased immunoglobulin production and molecular IgV and IgH analyses demonstrated further that B cell hybridomas which were regarded as different clones exhibited the identical mutational pattern of IgV_H and IgV_L genes. Finally, two IgG/λ B cell hybridomas could be obtained from series 54, four IgG/λ B cell hybridomas from series 56, one IgM/λ B cell hybridoma from series 58 and one IgA/λ B cell hybridoma from series 71.

ELISA for RF

None of the concentrated supernatants from the B cell hybridomas IgG/λ (n = 6), IgA/λ and IgM/λ showed positive reactivity (0/8) in the latex agglutination ELISA for RF. The supernatant of the positive control B cell hybridoma E1 and the sera of all patients (4/4) showed positive reactivity in the latex agglutination ELISA for RF.

Molecular analysis of IgV_H and IgV_L

The ratio of replacement to silent mutation (R/S) within the V_H genes can be regarded as an indication of antigen selection of high-affinity mutants [6]. Values of R/S ratios > 3 in the complementarity-determining regions (CDRs) may be used as an indicator of selectivity of B cell affinity [24,25]. R/S rations of framework regions (FR) are usually lower than in the CDRs, assuming random mutation in order to preserve the antibody structure [6,8].

As displayed in Table 2, in 4/7 IgV_H genes and in 3/6 IgV_H/λ genes the R/S ratios on the CDRs were > 3. In all these cases the R/S values in the CDRs were higher than the corresponding values for the FRs. Figures 2 and 3 show two examples that illustrate the higher accumulation of replacement mutations in the CDRs than in the FRs, characteristic of affinity-selected antibodies [8].

Table 2.

Molecular analysis of IgV_H and IgV_L genes* from synovial B cell hybridomas

Fig. 2.

The V_H sequence (IgG/λ) of B cell hybridoma ELB13/3/66 (V_H4, DP71) exhibiting in complementarity-determining regions (CDRs) 10 replacement and three silent mutations and in the framework region (FR) five replacement and five silent mutations, and the V_L sequence (Vλ1, DPL2) exhibiting in CDRs four replacement and no silent mutations and in the FR two replacement and one silent mutation.

Fig. 3.

The V_H sequence (IgA) of B cell hybridoma S71 (V_H5, DP73) exhibiting five replacement and one silent mutation in complementarity-determining regions (CDRs) and six replacement and five silent mutations in the framework region (FR).

Homology to known germ-line sequences was between 85% and 95% for the IgV_H genes and between 93% and 97% for the IgVλ. The assignment of V_H sequences to their closest germ-line counterparts indicated the biased usage of DP-47 by hybridomas 202 and ELE 36/6, and a random distribution of other germ-line genes among the remaining five functional IgV_H genes. For the Vλ sequences there was a biased usage of DPL11 by hybridomas 202 and ELE 36/6, and a random distribution among the other IgVλ genes.

All sequences were deposited in the EMBL database: AJ010334-AJ010337.

DISCUSSION

Germinal centre-like reaction in rheumatoid synovial tissue

Histopathological analysis of the patients' synovial tissue demonstrated an intense inflammatory infiltration with a high proportion of B cells, T lymphocytes and macrophages. Immunohistochemically, a germinal centre-like pattern of Ki-67⁺ proliferating cells and FDC could be demonstrated in follicles of the synovial tissue, so that these follicles represent an equivalent of lymphatic secondary follicles. Lymphatic secondary follicles are morphologically characterized by two separate compartments [26,27]. This pattern could be seen in three out of four cases, suggesting that a germinal centre-like reaction had taken place in the synovial tissue of these patients. Similar patterns of B cell follicles in synovial tissue have been described by other authors [3,8,28], but the definite morphological proof of a germinal centre can only be the demonstration of the two specific compartments, which can be achieved unequivocally by means of this double-staining technique [21].

Somatic mutation and IgV_H and IgV_L usage

Since germinal centres in lymphatic organs are regarded as the site of B cell affinity maturation [29], and as Schröder et al. [8] demonstrated in the germinal centre-like structures of RA synovium both B cell affinity maturation through hypermutation and antigen selection and memory cell diversification, one would expect to find somatically mutated and affinity-selected IgV_H and IgV_L genes in B cell hybridomas obtained from synovial tissue with germinal centres. This was actually the case: all of the IgV_H and IgV_L/λ showed somatic mutations and in 4/7 IgV_H and 3/6 IgV_L the R/S values of CDRs were > 3, which may be interpreted as a selection for affinity [29]. Although none of the B cell hybridomas was clonally related to others from the same patient, we may speculate that such affinity selection may have occurred in germinal centres of synovial tissue. The demonstration of somatically mutated and affinity-selected IgV_H and IgV_L/λ genes corroborates the data obtained from synovial tissue B cell hybridomas generated by polyethylene glycol (PEG) fusion with prior in vitro stimulation [15,16] and the data obtained from synovial tissue cell isolates [7,8,10,11].

Three out of seven IgV_H genes from B cell hybridomas showed no selection for affinity although germinal centres were demonstrable in the synovial tissue. The presence of germinal centres in the synovial tissue with no selection for affinity could mean that the germinal centre reaction had started only recently in the synovial tissue and the hypermutation mechanism was just beginning, or that part of the B cells did not enter the germinal centre reaction and remained in an unmutated state or in germ-line configuration. Our finding of a very low mutated IgM B cell hybridoma is therefore in line with the findings of other authors demonstrating low mutated IgM B hybridomas with RF activity in RA patients [15,16].

The fact that one affinity-selected, class-switched IgA B cell hybridoma originated in synovial tissue without germinal centres may be explained by the fact that germinal centres are transitional structures, and plasma cells, which have gone through the germinal centre reaction, are long-lived [6,30], or that already affinity-selected memory B cells had migrated to the synovia from elsewhere.

Since the mean R/S ratio of all IgV_H CDRs is 9 and the R/S ratio of all IgV_L is 3, this analysis suggests that, in synovial tissue of RA patients, an affinity selection with predominance in IgV_H genes has occurred. Samples with high numbers of somatic mutations in IgV_H showed relatively high numbers of somatic mutations in IgV_L. The same association was true for samples with low numbers of somatic mutations in IgV_H and IgV_L, indicating that somatic mutations in synovial tissue occur in a synchronized manner which is apparently necessary for efficient binding of the antigens involved in the local B cell expansion.

In our small collection there was a predominance of the V_H3 family (5/7) but not of V_H4 as described by other authors [9]. The DP47 gene was involved in 2/7 cases, here in association with the IgV_H gene DPL11. Interestingly, the DP47 gene was found to be the most frequently used DP gene in single-cell isolates from the synovial tissue of RA patients [11]. The DP49 and DP63 genes, which have been shown to be expanded in the synovial tissue of RA patients [10], could not be detected in this study.

Finally, the question arises why the B cell hybridomas of all patients used λ chains to the exclusion of κ chains. This predominance of λ obviously does not reflect the conditions seen in tissue sections, since κ and λ showed a balanced relation in plasma cells. The predominance of λ could possibly have contributed to the results of the fusion method which was shown preferentially to immortalize B lymphocytes, but not plasma cells [18]. The immunohistochemical expression of λ and κ chains in B lymphocytes is much lower than light chain expression in plasma cells and consequently a predominance of λ on the B lymphocyte level in vivo cannot be excluded by this method.

Antigens involved in local B cell expansion

The molecular analysis of these B cell hybridomas suggests that in FDC-containing germinal centres of synovial tissue there occurs somatic mutation of IgV_H and IgV_L genes, as well as selection for affinity. This process occurs in joints with severe clinical symptoms, so that it appears that the B cell response in synovial tissue is directly related to the specific manifestation of RA and may therefore mirror the antigenic scenario of synovitis. This interpretation is supported by experimental data in a mouse model for RA [31].

Although the B cell hybridomas have been analysed by various methods, their specificities are not clear at present: one hybridoma showed a specificity for a mitochondrial antigen [32]. The complete sequences of the IgV_H and IgV_L genes of the B cell hybridomas have been determined, hence it will be possible to express these genes experimentally, characterizing in this way the specificities of the synovial B cell response. Since we have shown that these B cell hybridomas are somatically mutated with the features of affinity selection originating from joints with germinal centres in synovial tissue and severe local symptoms, the further molecular analysis may help to identify antigens which are involved in the local B cell expansion and pernicious inflammatory process of RA.