Genomic absence of the gene encoding T cell receptor V 7.2 is linked to the presence of autoantibodies in Sjgren's syndromeкод для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 50, No. 1, January 2004, pp 187–198 DOI 10.1002/art.11429 © 2004, American College of Rheumatology Genomic Absence of the Gene Encoding T Cell Receptor V␤7.2 Is Linked to the Presence of Autoantibodies in Sjögren’s Syndrome Sanil J. Manavalan,1 Jennifer R. Valiando,1 Westley H. Reeves,2 Frank C. Arnett,3 Antje Necker,4 Ronit Simantov,1 Robert Lyons,2 Minoru Satoh,2 and David N. Posnett1 Objective. It is not yet known whether the absence of certain T cell receptor V␤ (TCRBV) genes (e.g., due to genomic deletion) has functional significance. We examined this question in relation to a known 21.6-kb insertion/deletion-related polymorphism (IDRP) in the human BV locus. Methods. New polymerase chain reaction (PCR) genotyping methods were used. Monoclonal antibodies to TCRBV gene products were used to confirm the absence of the relevant proteins. Patients with Sjögren’s syndrome (SS) or systemic lupus erythematosus (SLE) were compared with normal controls with regard to TCR genotypes and serologic profiles. Results. There are 3 known haplotypes (I, D1, D2) and 6 possible genotypes related to the 21.6-kb IDRP. Novel PCR-based methods were used to define these genotypes. In subjects with deleted/deleted (D/D) genotypes, T cells could not express V␤7.2 TCRs, as assayed with a new antibody specific for V␤7.2. This was the sole significant difference between subjects without the insertion and those with either 1 or 2 copies. Surprisingly, we found that the D/D genotype was associated with primary SS, but only when pathogenic autoantibodies were present. Conclusion. These results suggest that T cells expressing TCRs with V␤7.2 are protective against a pathogenic immune response in SS. Thus, genomic polymorphism of TCR genes (along with the correct HLA alleles) determines whether T cells can direct a pathogenic autoimmune response. In both human and mouse T cell receptor (TCR) gene loci, major genomic deletions exist. These deletions result in a contracted TCR repertoire, with certain TCR V␤ genes missing in some individuals. Whether this is of any functional significance is unknown. For instance, it is well appreciated that TCR transgenic mice do not overtly suffer from immunodeficiency, although they have a severely contracted TCR repertoire that is sometimes monoclonal, as occurs when these mice are bred onto a RAGo/o background. In contrast, mice with TCR haplotypes bearing major genomic deletions have been reported to be susceptible (1) or resistant (2,3) to collagen-induced arthritis, and resistant to autoimmune disease in the NZB ⫻ NZW murine model of systemic lupus erythematosus (SLE) (4). The mechanism underlying these associations is not fully understood, and similar associations in humans have not been described. Primary Sjögren’s syndrome (SS) is an autoimmune disease characterized by keratoconjunctivitis and sialadenitis, with lymphocytic infiltration of these exocrine glands with both T and B cells (5). Extraglandular site involvement may cause lymphocytic interstitial pneumonitis, atrophic gastritis, biliary cirrhosis, vasculitis of the skin (hypergammaglobulinemic purpura), thyroiditis, neuropathy, various cytopenias, and tubulointerstitial nephritis. The autoantibodies characteristic of SS include antinuclear antibodies termed anti-Ro/SSA and anti-La/SSB (6). These antibodies are also found in patients with SLE, although less frequently. Anti-Ro antibodies are directed to 60-kd and 52-kd proteins Supported by NIH grant AR-40391, GCRC grant M01R00082, and by funding from the State of Florida to the Center for Autoimmune Disease. 1 Sanil J. Manavalan, MD, Jennifer R. Valiando, MD, Ronit Simantov, MD, David N. Posnett, MD: Weill Medical College of Cornell University, New York, New York; 2Westley H. Reeves, MD, Robert Lyons, Minoru Satoh, MD, PhD: University of Florida, Gainesville; 3Frank C. Arnett, MD: University of Texas–Houston Health Sciences Center; 4Antje Necker, PhD: Immunotech, BeckmanCoulter, Marseilles, France. Address correspondence and reprint requests to David N. Posnett, MD, Cornell University Weill Medical College, 1300 York Avenue, Box 56, Room D601, New York, NY 10021. E-mail: firstname.lastname@example.org. Submitted for publication January 22, 2003; accepted in revised form September 11, 2003. 187 188 MANAVALAN ET AL (Ro 60 and Ro 52, respectively) that associate with the Y series of small RNA through RNA binding domains and zinc-finger motifs. Anti-La antibodies are specific for a 48-kd phosphoprotein that complexes transiently with nascent RNA polymerase III transcripts. La antigens associate with Epstein-Barr virus (EBV)– encoded RNA transcripts. In the human TCR ␤-chain (TCRB) locus, there is an insertion/deletion-related polymorphism (IDRP) in the region of the variable (V) gene segments (TCRBV), which was originally described by Zhao et al (7). It is a 21.6-kb fragment, containing 3 BV gene segments. The insertion is flanked by BV13S2 (upstream) and BV6S7 (downstream). These are both polymorphic BV genes. Kay et al have described an association of SS with the BV13S2*2 allele (8). BV13S2*1 and BV13S2*2 (the 2 known alleles of BV13S2) differ, however, by only a single amino acid in the leader region (9). Because the leader region is not part of the mature TCR ␤-chain protein, there is no difference in the expressed TCR ␤-chain protein. Therefore, the data reported by Kay et al suggest that a neighboring, closely linked gene may actually be relevant, rather than the BV13S2 alleles themselves. In this study, we examined the 3 known TCRBV haplotypes in the region of BV13S2. Six possible genotypes were determined by novel polymerase chain reaction (PCR)–based genotyping methods. Genotypes and phenotypes were compared using V␤-specific monoclonal antibodies (mAb) to detect expression of specific TCRs. The specificity of a new antibody to V␤7.2 was determined. This gene is encoded within the IDRP insertion, and individuals lacking the insertion on both chromosomes also lack V␤7.2-positive T cells. Surprisingly, we observed that SS was significantly associated with the absence of BV7S2, but only when antibodies to Ro and La were present. PATIENTS AND METHODS Human subjects. Patients with SLE or SS fulfilled accepted diagnostic criteria (5,10). All patients with SS, including those without anti-Ro/La antibodies, had positive results on lip biopsies, with Daniels’ focus scores of ⬎2. Race, age, sex, and clinical features were taken into account. Normal controls were from either Marseilles, France or Houston, Texas. The former subjects were randomly selected clinic outpatients and normal volunteers of unselected ethnic background. All of the latter subjects were Caucasian and were selected as a matched population for a group of patients with primary SS. These normal adult volunteers were medical school and hospital personnel from Houston. Ethnicity was based on both grandparents being Caucasian. Other study populations included SLE patients from New York Presbyterian Hospital and the Hospital for Special Surgery in New York, and patients with SLE or SS from the University of Florida in Gainesville. All samples were obtained in conformity with protocols accepted by the respective human institutional review board committees. Peripheral blood mononuclear cells (PBMCs) were used for flow cytometry or for DNA isolation in some cases. Otherwise, stored DNA samples were used for genotyping. Stored serum or plasma was used for autoantibody tests. Analysis of patient and control groups was limited to Caucasians, because TCR genotypes in the area of the IDRP differ considerably in different ethnic groups (11,12). DNA preparations. DNA isolation was performed by several methods. For the 7.9-kb PCR, DNA was isolated using the QIAamp DNA Blood Mini Kit (Qiagen, Valencia, CA) according to the manufacturer’s instructions. Otherwise, DNA was isolated by incubating cells in lysis buffer (Tris HCl, 10 mM; EDTA, 1 mM; sodium dodecyl sulfate [SDS], 0.0001%; Triton X-100, 0.001%; proteinase K, 600 g/ml) for 1 hour at 56°C followed by 15 minutes at 95°C, or by similar protocols (13). TCR nomenclature. We used the nomenclature described by Arden et al (9), with a single exception: BV6S7 (11), as used herein, is also called BV6S4 by Arden et al (9). For BV7S2 within the IDRP and BV7S3 upstream of the IDRP, we used the nomenclature described by Arden, which coincides with GenBank nomenclature (accession no. L36092). The same genes are named in reverse (BV7S3 and BV7S2, respectively) by Zhao et al (7). TCRBV gene products (proteins) are referred to as V␤7.2, and so forth. Flow cytometry with TCR mAb. PBMCs from a FicollHypaque gradient were washed in phosphate buffered saline (PBS) and suspended in staining buffer (RPMI 1640, 2% fetal calf serum, 0.05% sodium azide). Murine mAb to TCR V␤ epitopes were added at predetermined optimal concentrations. The cells were stained at room temperature for 40 minutes on a shaker, then washed again 3 times and stained next with goat anti-mouse Ig–fluorescein isothiocyanate, followed by antiCD3–phycoerythrin, both at optimal dilutions, for 40 minutes on a shaker, at room temperature in the dark. Cells were then washed 3 times with staining buffer and suspended in 0.5 ml of the same buffer for flow analysis on a Coulter Epics XL (Beckman Coulter, Miami, FL). The mAb have been described elsewhere (14,15), except for mAb Zizou (V␤7.2). PCR-based TCR genotyping. Genotyping for the IDRP is based on 2 PCR assays. The first assay detects the presence of the insert, with primers straddling the upstream margin of the insertion (Figure 1). The primers used are 5⬘-AGGTGAACTTGGAGATGCA-3⬘ (forward) and 5⬘CTGTGGAGGCATGTACACTATGT-3⬘ (reverse). The conditions are as follows: 10 mM Tris HCl, 50 mM KCl, 0.1% Triton X-100, 1.5 mM MgCl, 200 M of each dNTP, and 1 unit of Taq polymerase (Promega, Madison, WI) at 94°C for 1 minute, 55°C for 1 minute, 72°C for 1 minute, for 36 cycles, followed by a 10-minute extension at 72°C. The result is amplification of a 169-bp fragment. The second assay is a nested PCR with primers in the BV6S7 gene (Figure 1). The outer primers are 5⬘GTCTCCAGTCCCCCAGTAACAA-3⬘ (forward) and 5⬘GCAGGAAGGAGGCGACTGTGCCA-3⬘ (reverse). The result is a 311-bp PCR fragment. The conditions are as described above, except for an annealing temperature of 65°C. The inner ABSENCE OF TCRBV7S2 ASSOCIATED WITH PRESENCE OF AUTOANTIBODIES IN SS 189 Figure 1. Map of the 3 insertion/deletion-related polymorphism haplotypes (inserted [I], deleted 2 [D2], and deleted 1 [D1]) showing the allelic variants of the T cell receptor BV genes present on each haplotype. BV genes located outside the insert are shown as open boxes, and those located within the insert are shown as filled boxes. The official nomenclature for BV genes was used: 13S2*1a means BV13S2, allele *1, copy a (there is also a copy b within the genomic insert, and an allele *2 on the D1 haplotype). Large shaded arrows indicate genomic duplication; margins of the insert are indicated by a vertical line at the end of each arrow. Primers (arrowheads) that detect the insert are shown for haplotype I; nested primers for BV6S7 are on the far right. Primers for a large 7.9-kb polymerase chain reaction that detects deleted haplotypes are shown for D2. primers are 5⬘-GGTCACAGAGAAGGGAAAGG-3⬘ (forward) and 5⬘-CGGCCGAGTCCTCCTGCTG-3⬘ (reverse), yielding a 233-bp fragment. The conditions are as described above, except for an annealing temperature of 60°C, and use of 2.0 mM of MgCl2. The 233-bp fragment is then digested with 10 units of Bam HI (17 l of PCR product, 2 l of 10⫻ buffer, 1 l of Bam HI). If the BV6S7*1 allele is present, there is a Bam HI site (11), and the 233-bp PCR product is cut into 184-bp and 49-bp fragments. The BV6S7*2 allele does not have a Bam HI site. The PCR to distinguish between inserted/inserted (I/I) and inserted/deleted 2 (I/D2) genotypes (see below) used a forward primer (5⬘-GTCTACAACTTTAAAGAACAGACTG-3⬘) paired with the reverse primer of the above PCR 5⬘-GCAGGAAGGAGGCGACTGTGCCA-3⬘) to yield a 7.9-kb product. This PCR used the Expand Long Template PCR System Kit (Roche Laboratories, Nutley, NJ). Each 50-l reaction contained 400 ng of purified genomic DNA, 350 nM of each dNTP, 400–600 ng of each primer, 20 mM of Tris HCl, 100 of mM KCl, 1 mM of dithiothreitol, 0.1 mM of EDTA, 0.5% Tween 20 (volume/volume), 0.5% Nonidet P40 (v/v), 1.75 mM of MgCl, 2.5 units of Taq, plus 2.5 units of Pwo polymerase. The cycling parameters were 92°C for 2 minutes, 92°C for 10 seconds, 65°C for 30 seconds, and an 8-minute extension at 68°C, repeated for 10 cycles; this was followed by 20 cycles of 92°C for 10 seconds, 65°C for 30 seconds, and an 8-minute extension at 68°C (with an increase of the extension time by 20 seconds with each cycle), then a final extension of 68°C for 7 minutes. This PCR was best performed using a Perkin-Elmer 9600 thermocycler (Perkin-Elmer, Emeryville, CA). Anticardiolipin antibody enzyme-linked immunosorbent assay (ELISA). The ELISA for anticardiolipin antibodies was performed with cardiolipin antigen (C1649; Sigma, St. Louis, MO) dissolved in ethanol and coated onto roundbottomed wells by evaporation. After blocking with 2% bovine serum albumin, human serum serially diluted in PBS containing 10% adult bovine serum (C1507; Sigma) was incubated at 4°C for 2–3 hours, followed by washing and further incubations, first with alkaline phosphatase–conjugated goat antihuman IgG antibody, and second with p-nitrophenylphosphate disodium substrate. Optical density (OD) was read at 405 (wavelength). Antinuclear antibodies by immunoprecipitation. K562 cells (human erythroleukemia; American Type Culture Collection, Rockville, MD) were maintained in RPMI 1640 supplemented with 10% fetal bovine serum, L-glutamine, and penicillin/streptomycin. Autoantibodies were analyzed by immunoprecipitation of 35S-radiolabeled K562 cell extract, using 8 l of human serum per sample (16). Sera containing anti–Ro 60/SSA, anti-La/SSB, anti–nuclear RNP/Sm, and other autoantibodies were identified by comparison with human reference sera, as previously described (17). Anti-Sm antibodies were detected by immunoprecipitation of the 200- and 205-kd components of the U5 small nuclear RNP, as previously described (18). Anti–Ro 52 and other antibodies by immunoblotting. IgG from 3 l of human anti–Ro 52 reference serum was 190 MANAVALAN ET AL crosslinked to protein A–Sepharose beads using dimethylpimelimidate (19). Proteins immunoprecipitated from K562 cell extract (2 ⫻ 107 cell equivalents) were fractionated by SDS–polyacrylamide gel electrophoresis and transferred to nitrocellulose. The membrane was probed with anti–Ro 52 reference serum (positive control), normal human serum (negative control), or test sera at a 1:400 dilution. Blots were developed with 1:2,000 alkaline phosphatase–labeled goat antihuman IgG (␥- and light chain–specific; Southern Biotechnology, Birmingham, AL) and developed using the Western-Star chemiluminescent system (Tropix, Bedford, MA). For immunoblot detection of anti–Ro 60/SSA and anti-La/SSB antibodies, an analogous approach was used, substituting anti–Ro 60 and anti-La reference sera, respectively. Antinuclear antibodies, anti–single-stranded DNA (anti-ssDNA), and antichromatin ELISAs. Antinuclear antibodies were detected using HEp-2 substrate, as previously described (20). ELISAs for antichromatin and anti-ssDNA antibodies were performed using sera diluted 1:500 (21,22). IgM and IgG antichromatin antibodies were considered positive when sample absorbance was more than 3 SD above the mean OD in blank wells. Anti-ssDNA antibodies were considered positive when sample absorbance was more than 3 SD above the mean OD using sera obtained from healthy volunteers. Clinical data. Platelet counts and levels of lupus anticoagulant, factor V Leiden, prothrombin mutation, and homocysteine were obtained from the clinical laboratory measurements. Statistical analysis. Statistical analysis was done by chi-square (with Yates’ correction when necessary), by 2-tailed Fisher’s exact test, or by analysis of variance. RESULTS IDRP genotypes determined by PCR. The human TCRB locus contains an IDRP in the BV region (7) (see Figure 1). It is a 21.6-kb fragment, characterized by the presence of 3 V gene segments: BV9S2P is a pseudogene, BV7S2 is a unique V gene, and BV13S2*1b is an exact copy of the BV13S2*1a gene located outside of the IDRP. There are 2 deleted haplotypes, both lacking the 21.6-kb fragment. These 2 deleted haplotypes differ in terms of the alleles of the flanking TCR BV13S2 and BV6S7 genes. Thus, the deleted 1 haplotype (D1) has BV6S7*2 and BV13S2*2 alleles, and the deleted 2 (D2) haplotype has BV6S7*1 and BV13S2*1 alleles. The inserted haplotype (I) is similar to the D2 haplotype in that it also contains BV6S7*1 and BV13S2*1. The inserted haplotype provides the total BV locus with 63 BV gene segments, of which 52 are functional, while both deleted haplotypes have 60 BV gene segments, of which 50 are functional. There are 2 additional haplotypes (23), which are not shown in Figure 1. One haplotype is thought to arise from 2 head-to-tail insertions of the 21.6-kb fragment shown in Figure 1. The other is similar to the D haplotypes shown in Figure 1 but has an additional deletion located at least 15 kb further downstream. These other haplotypes occur with gene frequencies of 0.005 and 0.008, respectively and thus are very rare. All of these TCRBV haplotypes were originally described using pulsed-field gel electrophoresis of Sfi I DNA digests (7,23). This technique is impractical when examining clinical samples obtained from a large number of patients and when only small amounts of DNA are available. Therefore, we developed simple PCRbased assays to type individuals for the IDRP and the 3 haplotypes shown in Figure 1. Several PCRs were utilized (Figure 2), with DNA from test individuals with known haplotypes. The first PCR detects the presence of the insertion. Genotypes that completely lack the insertion on both chromosomes are negative in this PCR (Figure 2A). The second PCR is a nested PCR using primers in the region of the BV6S7 gene. The external primers yield a 311-bp product, and the internal primers yield a 233-bp product (Figure 2B). The latter is then digested with the restriction enzyme Bam HI (Figure 2C) at a site that is present in BV6S7*1 but not in BV6S7*2 (11). This allows differentiation between D2 and D1 haplotypes (Figure 2C). However, these PCRs do not allow differentiation between the I/I and I/D2 genotypes (Figure 2C). Both genotypes are positive for the insertion PCR, and both have the BV6S7*1 allele. These 2 genotypes can be distinguished by Sfi I, Bgl II, or Bam HI restriction fragment length polymorphisms (RFLPs) (7), or by a special PCR involving a 7.90-kb amplification fragment. The latter approach was successful for samples with large amounts of highly purified DNA (Figure 2D) and may prove useful under select circumstances. However, this assay was not practical for most samples obtained for the population studies performed here. Therefore, results are presented for the I/I and I/D2 genotypes combined. IDRP genotypes in normal controls. First, we typed a cohort of 55 subjects with mixed ethnic backgrounds, who lived in Marseilles. The genotypes were represented as follows: for I/I and I/D2, 33.9%; for I/D1, 41.1%; for D1/D2, 7.1%; for D2/D2, 0%; and for D1/D1, 17.9%. Thus, the D2 haplotype appears to be relatively rare. Overall, these percentages fit well with the initial findings on the distribution of BV6S7 alleles, as performed in populations of Western European descent (12). A second group of 41 normal controls were from Texas. All of these subjects were Caucasian. Overall, the distribution of genotypes in this group was similar to that in the French group (Table 1). A total of 85.4% ABSENCE OF TCRBV7S2 ASSOCIATED WITH PRESENCE OF AUTOANTIBODIES IN SS 191 Figure 2. Polymerase chain reactions (PCRs) used to distinguish the 3 haplotypes. A, The 169-bp product seen only in DNA samples from individuals with the I haplotype in at least 1 chromosome. B, PCR products of the first (lane 1) and the second (lanes 2–7) nested PCRs used to determine BV6S7 alleles. Both PCRs were positive in all haplotypes. C, Bam HI digestion of the second nested PCR product (233 bp) for all 6 possible genotypes. Only the BV6S7*1 allele was digested yielding 2 bands (184 bp and 49 bp). Thus, the I/I, I/D2, and D2/D2 genotypes showed 2 bands, the I/D1 and D1/D2 genotypes showed an additional undigested fragment because the D1 haplotype does not carry BV6S7*1, and the D1/D1 genotype showed only the undigested band. D, The 7.9-kb PCR product used to distinguish between the D2 and I haplotypes. MM ⫽ molecular markers (see Figure 1 for other definitions). of subjects had at least 1 chromosome with the insert (I⫹), and 14.6% lacked the insert on both chromosomes (D/D). Use of TCR V␤ elements dependent on IDRP genotype. In 53 of 55 French subjects, peripheral blood lymphocytes (PBLs) were collected and analyzed with 4 TCR V␤–specific mAb. The percentages of positive cells per total T cells for each genotype are summarized in Figure 3. The specificity of some of the mAb has previously been described. Thus, the V␤6.7*1 mAb reacts strongly with TCRs composed of BV6S7*1 alleles (12,24). Subjects with 2 copies of the BV6S7*1 allele generally have twice as many V␤6.7*1-positive T cells as Table 1. Association of the IDRP haplotypes with Sjögren’s syndrome (SS)* Genotype Group (n) I/I & I/D2 I/D1 D1/D2 D2/D2 D1/D1 Total I⫹ Total D/D P† Control (41) SS (55) Ro⫹ (36) La⫺ (16) La⫹ (20) Any antibody (36) No antibody (19) SLE (57) Ro⫹ (21) La⫹ (9) Lupus RNP (12) Any antibody (35) No antibody (20) 31.7 12.7 13.9 12.5 15.0 13.9 10.5 19.3 4.8 0.0 8.3 14.3 30.0 53.7 45.5 33.3 37.5 30.0 33.3 68.4 50.9 66.7 66.7 50.0 54.3 50.0 9.8 25.5 30.6 25.0 35.0 30.6 15.8 15.8 9.5 11.1 25.0 17.1 15.0 0.0 1.8 2.8 0.0 5.0 2.8 0.0 5.3 0.0 0.0 0.0 0.0 0.0 4.9 14.5 19.4 25.0 15.0 19.4 5.3 8.8 19.0 22.2 16.7 14.3 5.0 85.4 58.2 47.2 50.0 45.0 47.2 78.9 70.2 71.4 66.7 58.3 68.6 80.0 14.6 41.8 52.8 50.0 55.0 52.8 21.1 29.8 28.6 33.3 41.7 31.4 20.0 0.006 0.0005 0.009 0.0019 0.0004 NS NS NS NS NS NS NS * Except where indicated otherwise, values are the percentage distribution of insertion/deletion-related polymorphism (IDRP) genotypes. D/D represents the 3 genotypes that share the complete absence of the insert with BV7S2; I⫹ represents genotypes that contain the insert. If correlations between genotypes in the different groups are considered and a Bonferroni correction is applied (online at http://home.clara.net/sisa/ bonlp.htm#Corr), the alpha level should be ⬍0.010206218 (SS), or ⬍0.007300832 (SLE). NS ⫽ not significant. † By Fisher’s exact test, comparing the I⫹ and D/D distribution with that in the control group. 192 MANAVALAN ET AL Figure 3. T cell receptor V␤ staining as a function of insertion/deletion-related polymorphism genotypes. The differences in percentages of V␤13.2 cells among the different groups were not statistically significant, and the elevation of V␤7.1 in the D1/D1 genotype is significant only when compared with the I/I & I/D2 group (P ⬍ 0.01). The differences in V␤7.2 and V␤6.7 percentages are highly significant. Bars show the mean and SEM. n ⫽ number of subjects representing each haplotype (see Figure 1 for other definitions). those who are heterozygous, and the percentage in subjects with 2 copies of BV6S7*2 is very low (12,24). The data in Figure 3 confirm these findings. The V␤13.2 mAb is also known to stain variable percentages of cells (25,26). In the presence of 2 D haplotypes (D1/D1 or D1/D2), the percentages of V␤13.2-positive T cells are smaller than in individuals who have 1 or 2 haplotypes bearing the insert (Figure 3). This confirms results of prior studies (25,26). The V␤7.1 mAb has previously been described. All subjects would be expected to have 2 functional copies of the BV7S1 gene, and percentages of positive T cells should not vary between the genotypic groups. However, the data showed an increase of V␤7.1-positive cells in D1/D1 subjects. At present, this result remains unexplained, because allelic variants of BV7S1 have not been described (9). BV alleles that differ either in the promoter region or in the recombination signal sequences (27) can affect V␤ percentages in the PBLs, and such potential variants have not been rigorously excluded. The specificity of the V␤7.2 mAb was unknown prior to this study. A transfected cell line expressing BV7S2 was used to immunize mice for production of the hybridoma. Variation in the percentage of positive cells among randomly selected subjects was noted. Figure 3 clearly shows that subjects lacking the insertion on both chromosomes rarely express the V␤7.2 determinant recognized by the mAb. It was unclear whether subjects with the I/I genotype have higher percentages of V␤7.2 T cells than do subjects with the I/D genotype, as would be expected. This is because the I/I and I/D2 genotypes were indistinguishable, and the contribution of subjects with the I/D2 genotype within this combined group was not known. However, in the I/D1 group, in which all subjects were heterozygous for the IDRP, there were slightly lower percentages of V␤7.2-positive cells than in the I/I and I/D2 group. In sum, the main specificity of this new mAb is for a V␤7.2 epitope. This mAb can therefore be used to identify subjects whose T cells are unable to use the BV7S2 gene because it is not present in their genome. Association between SS and D/D genotypes. Kay et al have described an association between homozygosity of the BV13S2*2 allele and a subset of patients with SS (8). This subset of patients had a more aggressive disease, with polyclonal hypergammaglobulinemia, extraglandular disease involvement, anti-La antibodies, and HLA–DR3. The mature proteins encoded by BV13S2*2 and BV13S2*1 are exactly the same. Thus, ABSENCE OF TCRBV7S2 ASSOCIATED WITH PRESENCE OF AUTOANTIBODIES IN SS use of BV13S2*2 by autoimmune or regulatory T cells cannot explain the observed association. The BV13S2*2 allele is found exclusively on a deleted haplotype (Figure 1). Therefore, it is possible that the observed association was actually with the deleted haplotype and with the absence of genes contained within the deleted genomic fragment. Indeed, TCR genotypes obtained in patients with SS showed a significant association between primary SS and the complete absence of the IDRP-defined genomic segment (D/D) (Table 1). D1/D2 subjects are heterozygous for BV13S2 and BV6S7 alleles, but these individuals have the IDRP deletion on both chromosomes. This genotype was also significantly increased in SS (Table 1). Thus, SS is associated with the IDRP deletion and not specifically with the BV13S2*2 allele. SS is also known to be associated with certain HLA–DQ alleles (13,28). The presence of alleles encoding Q34 on the DQA chain and L26 on the DQB chain is associated with more aggressive disease and higher titers of autoantibodies against Ro/La (13). There was a gene-dose effect, in that greater numbers of the implicated DQ alleles gave the more striking association (12). In our study, 50 patients with SS were typed for HLA–DQ. Anti-Ro/La antibodies were significantly (P ⬍ 0.025) more frequent as the number of DQ alleles encoding Q34/L26 increased. This appeared to hold true regardless of TCR D/D, I/D, and I/I genotypes, but the numbers of patients were insufficient for statistical analysis within each of these genotypic subsets. The IDRP insertion contains only 3 known TCRBV genes (Figure 1) (7). The entire sequence of this genomic segment is known (GenBank accession no. L36092), and there are no unexpected genes within it. Of the 3 TCRBV genes, BV9S2P is a pseudogene. BV13S2*1b represents an identical copy of BV13S2*1a. Thus, lack of BV13S2*1b is unlikely to be of any major consequence. Therefore, individuals who lack the IDRP insertion on both chromosomes are distinguished primarily by their inability to produce TCRs using BV7S2. It follows that T cells expressing BV7S2 may alter the nature of immune responses in SS: lack of these T cells would perhaps result in exaggerated disease-related T cell responses, and the outcome might be worsened clinical manifestations. It is conceivable that lack of BV7S2 is related to production of specific autoantibodies. Alternatively, lack of BV7S2 could be related to severity of disease (which is independently associated with high levels of autoantibodies). To examine this issue, a multiplex family with SS was genotyped (Figure 4). A mother with SS (I-2) had 5 daughters, one of whom (II-4) also had SS. This 193 Figure 4. Analysis of a multiplex family with autoimmune disease. HLA haplotypes are indicated above each symbol. The haplotypes were as follows: for a, DRB1*1301, DQB1*0603, DQA1*0103; for b, DRB1*0101, DQB1*0501, DQA1*0101; for c, DRB1*0301, DQB1*0201, DQA1*0501; for d, DRB1*1301, DQB1*0603, DQA1*0103; e and f are unrelated haplotypes. The T cell receptor (TCR) haplotypes were as follows (shading indicates a deleted haplotype): for I-1, D1/D2; for I-2, I/D2; for II-1, I/D1; for II-2, I/D1; for II-3, D1/D2; for II-4, D1/D2; for II-5, D1/D2; for III-1, D1/D1. ANA ⫽ antinuclear antibody; HT ⫽ autoimmune hypothyroid disease; SS ⫽ Sjögren’s syndrome; Ro ⫽ anti-Ro antibody; La ⫽ anti-La antibody; Raynaud’s ⫽ Raynaud’s syndrome; CVI ⫽ common variable immunodeficiency; IgA Def ⫽ IgA deficiency; IDDM ⫽ insulin-dependent diabetes mellitus (see Figure 1 for other definitions). individual has a sister (II-3) who is both HLA identical (haplotypes a and c) and TCR genotype identical (D1/ D2), yet these 2 sisters are not congruent for SS. Because II-3 has a positive antinuclear antibody test, it is possible that the 2 sisters share specific autoantibodies, excluding anti-Ro or anti-La. One may conclude that the D/D genotype does not necessarily predict a particular disease manifestation or the presence of anti-Ro/La, even when HLA genes are identical. Other polymorphic genes not shared by the 2 sisters and/or environmental factors must play a role. To examine whether the absence of the insert is associated with production of specific autoantibodies, we tested for different autoantibodies in SS patients with known genotypes (Figure 5). In the group of SS patients with the D/D genotype, antibodies to Ro and La were more frequently detected. The P value was most significant for anti-La antibodies assessed by immunoprecipitation (P ⬍ 0.006). Autoantibodies against other target antigens (ssDNA, chromatin) were not dependent on the TCR genotypes. SLE patients may also have autoantibodies to the Ro/La complex. The frequency of such autoantibodies is high in SS (⬃60–70%) but is lower in SLE (⬃30%) (6). If the D/D genotype were strictly associated with the presence of antibodies to Ro/La, we reasoned that SLE 194 MANAVALAN ET AL Figure 5. Autoantibodies in patients with Sjögren’s syndrome. Sera from 9 D/D (deleted/deleted) patients (filled bars) and 20 I/D (inserted/deleted) or I/I (inserted/inserted) patients (shaded bars) were compared. Statistical analysis was by analysis of variance. Data for antibodies to single-stranded DNA (ssDNA) and chromatin were expressed as quantitative enzyme-linked immunoassay (ELISA) units. Antinuclear antibody (ANA) was obtained with serial dilutions of serum; numbers on the Y axis represent average inverse [inv] titers. Data from immunoprecipitation (IP) and immunoblotting (IB) experiments were scored as positive (pos) or negative for the indicated antibody. Other-IP ⫽ other unidentified bands observed by immunoprecipitation. patients with such antibodies might also be predominantly of the D/D genotype. Figure 6 presents data on 56 Caucasian SLE patients with known TCR genotypes. The presence of autoantibodies to Ro/La and other specificities was not significantly different in patients with genotypes that lacked the insert (D/D) versus those with 1 or 2 copies of the insert (I/I or I/D). Autoantibodies detected by immunoprecipitation, lupus anticoagulants, and clinical manifestations of thrombotic events including fetal loss (data not shown) occurred more frequently in patients with the D/D genotype, but this difference did not reach statistical significance. Anti-La antibodies occurred in 5 of 19 patients with the D/D genotype versus 6 of 37 patients with the I/D or I/I genotype. To confirm these data, SLE sera were also tested by ELISA for the presence of La and Ro 52 antibodies (data not shown). These assays were more sensitive than the immunoprecipitation and immunoblotting assays. However, they too failed to reveal a significant association between the D/D genotype and the presence of anti-Ro/La antibodies in SLE patients. In sum, the D/D genotype and the lack of BV7S2 appeared to be associated with a subset of SS patients with high titers of Ro/La antibodies, especially those with anti-La antibodies. DISCUSSION Here we describe an association between an IDRP of the TCRBV locus and the major subset of primary SS patients with anti-Ro/La autoantibodies. The implication of our results is that the gene encoding V␤7.2 may protect against the development of severe SS or against the generation of pathogenic autoantibodies, in particular anti-La antibodies, in SS patients. The occurrence of SS per se was not related to TCR genotypes. Only those patients with autoantibodies tended to have a TCR D/D genotype (Table 1). However, in SLE patients with high titers of anti-Ro/La antibodies, the Figure 6. Features of systemic lupus erythematosus as a function of T cell receptor genotype. Two groups of patients, one with the D/D (deleted/deleted) genotype (n ⫽ 19) and one with the I/I (inserted/inserted) or I/D (inserted/deleted) genotype (n ⫽ 37), were compared. Left, Quantitative titers of antibodies to cardiolipin (aCL). Bars show the mean and SEM. Right, Percent of patients positive for lupus anticoagulant (AC), Ro 60, La 48, Sm, RNP, ribosomal P (RibP), and other unidentified immunoprecipitation bands. ABSENCE OF TCRBV7S2 ASSOCIATED WITH PRESENCE OF AUTOANTIBODIES IN SS same association with the D/D genotype was not manifest. Several lines of evidence suggest that T cells may be important in a disease such as SS. First, both activated B and activated oligoclonal T cells are found in the inflamed exocrine glandular tissues of patients with SS. Second, in a mouse model of SS, adoptive transfer of lymphocytes from a diseased animal into a SCID recipient resulted in disease that could be prevented by anti-CD4 and anti–TCR antibody treatment (29). Third, T cells, like B cells, may recognize the antigens Ro (30,31) and La (32–34). T cells with some of these specificities are found in the salivary glands of patients with SS (30). Fourth, SS is associated with certain HLA class II alleles (13,35), and this association crosses racial barriers. The DQA1 and DQB1 alleles found in SS invariably encoded a Q34 residue (DQA1) and an L26 residue (DQB1), regardless of the genetic background of the patients (13,35). Both Q34 and L26 are situated in the floor of the class II major histocompatibility complex–binding groove, suggesting a role in presenting autoantigens to CD4 T cells in SS. This HLA–DQ association is primarily with the anti-Ro/La autoantibody response and does not apply to the minority of patients without anti-Ro antibody (35). Moreover, these DQ alleles appear to control diversification of the anti-Ro/La response (i.e., intermolecular epitope spreading) (36). The simplest hypothetical model to explain the findings presented herein is that V␤7.2-positive T cells are involved in down-modulation of disease manifestations in SS. One possibility is that such T cells may participate in the regulation of epitope spreading. Epitope spreading is a process by which autoantibody specificities to autoantigen complexes spread from an initial target epitope to other epitopes. If epitope spreading is limited to the same polypeptide, it is referred to as intramolecular spreading. When epitope spreading involves noncovalently associated polypeptides, it is called intermolecular spreading. Epitope spreading is known to occur with autoantibodies to the Ro/La antigen complex (34,37). It is currently thought that Ro 60, Ro 52, La, and other proteins are transiently associated in an immunogenic RNA/protein particle (6,38,39). This allows efficient intermolecular epitope spreading, as demonstrated in mice immunized with a conserved Ro 60 epitope, which then develop autoantibodies not only to Ro 60 but also to Ro 52 and La (38,40,41). When autoantibodies to La are observed, they generally do not occur in the absence of anti-Ro antibodies. This is thought to be attributable to regulated 195 progression of intermolecular epitope spreading, in which autoantibodies to Ro 60 are the first to occur, and autoantibodies to La 48 are the last to be observed. The TCR D/D relationship described herein was most tightly associated with the presence of anti-La antibodies (Table 1 and Figure 5). However, TCR D/D genotypes were not observed in SLE patients with these antibodies. Thus, if V␤7.2 T cells somehow inhibit epitope spreading and thus formation of anti-La antibodies, this process would have to be applicable only to patients with SS. Epitope spreading can modulate disease expression in different ways. In experimental adjuvant-induced arthritis of the Lewis rat, a transient disease induced by immunization with adjuvant containing mycobacterial antigens, there is T cell epitope spreading to the carboxy-terminal peptides of mycobacterial hsp65. The T cell clones with these specificities are protective when administered prior to initial immunization. When they arise as a consequence of epitope spreading, they are thought to down-modulate disease expression, resulting in the transient nature of this disease (42). However, epitope spreading is more commonly associated with progression of disease (e.g., in experimental autoimmune encephalomyelitis [EAE], a model for human multiple sclerosis) (43–46). An alternative possibility is that V␤7.2 T cells promote antiidiotypic responses, which can mask antiLa antibodies (47). Whether these responses differ in patients with SS and those with SLE is not known. Prior studies have examined the TCR repertoire in samples from SS patients. Several reports have described elevated percentages of V␤2 and V␤13 CD4⫹ T cells in either blood or salivary gland biopsy samples (48–53), but the results are quite variable and may depend on disease activity (54) and the tissue site analyzed (55). V␤7 cells were usually not prominent in these studies, in which PCR primers were used to specifically detect the BV7 transcripts (a method that should detect all 3 BV7 genes but does not distinguish between the 3 different BV7 genes). In one study (52) using samples from Japanese patients with primary SS or with human T lymphotropic virus type I–associated SS, BV7 transcripts were elevated and clonally restricted. A conserved third complementarity-determining region motif (QDXG) was noted among these BV7 transcripts. These BV7 transcripts were present in 12 of 13 patients but not in controls, and in both blood and salivary tissue. In summary, available data do not clearly indicate whether V␤7.2 T cells participate in the autoimmune response or whether they might recognize peptide antigens from the Ro/La complex. 196 Another hypothesis to explain the present findings is that V␤7.2 T cells have a T regulatory (Treg or T suppressor) function. For example, they might downmodulate the pathogenic helper T cells that result in T cell help for autoantibody formation. There is some evidence for the presence of such T cells in the murine EAE system. For example, deletion of V␤14 and V␤3, used by such regulatory T cells, led to increased disease severity (56). But classic Treg cells are not thought to have restricted TCRBV gene usage, nor do they inhibit a specific autoimmune response (57). Finally, another hypothesis is that V␤7.2 T cells alter the production of key cytokines. T cells may direct an autoimmune response in different ways, depending on the cytokine profile produced. In EAE, pathogenic myelin basic protein–specific Th1 CD4 effector cells could be deviated to Th2 cells by immune regulatory T cells, resulting in protection from disease (58). In a mouse model of SS, there is also a regulatory population of CD4 cells that prevents disease by producing interleukin-4 (IL-4), IL-10, and transforming growth factor ␤ (59). In SLE and SS, interferon-␥ (IFN␥) appears to be one of the key pathogenic cytokines, possibly acting by stimulating antigen-presenting cells to secrete pathogenic cytokines (IL-6, tumor necrosis factor ␣) or by stimulating autoantigen presentation (60– 63). T cells reactive with Ro/La can also produce IFN␥ (32). Perhaps such a cytokine response differs if the relevant T cells use BV7S2 as opposed to other BV genes. Polymorphism in the human TCR loci was discovered soon after the TCR genes were identified (12,64). However, disease associations with TCR alleles have not been forthcoming (with few exceptions), in contrast to the many known disease associations with major histocompatibility complex polymorphisms. A TCR B association with multiple sclerosis was demonstrated by careful sibpair analysis (65). The mechanism by which TCR genes influence multiple sclerosis susceptibility remains unclear (66). Another report described a null allele of BV6S1 associated with a subset of patients with juvenile rheumatoid arthritis (67,68). How the lack of V␤6.1-expressing T cells affects susceptibility to this disease is still a mystery. Earlier studies, including studies on SLE, used RFLPs for Bgl II and Kpn I polymorphic sites in the vicinity of the TCR BC locus encoding the D␤, J␤, and C␤ segments. There is no linkage between these markers and the IDRP studied herein, which is located ⬃600 kb upstream. An early report by Frank et al described an association between Ro⫹ SLE, but not Ro⫺ SLE, with a combination of 2 RFLPs (69). Although the authors MANAVALAN ET AL could not explain the mechanism underlying this association, it is possible that polymorphisms of the BC locus are associated with different D␤ or J␤ usage and thus affect the expressed TCR repertoire. A followup report by the same authors (70) showed that the TCR BC polymorphism predicted high-titer (⬎1:106) anti-Ro antibody when patients also had particular HLA– DQA1/DQB1 alleles (DQA1*01, DQB1*0201). These alleles encode the Q34 and L26 residues mentioned above (35). Anti-Ro antibodies directed to a particular Ro 60 epitope, a 15-mer termed A480, were most strongly associated with the TCR Bgl II 9.8-kb allele (71). Next came the report (8) linking severe cases of SS (with high titers of anti-Ro/La) with TCR BV13S2*2 as opposed to BV13S2*1. BV13S2*2 is present only on the D1 haplotype (Figure 1). Thus, our data are compatible with this earlier finding. They also illuminate a potential mechanism. The BV13S2 alleles do not differ in the expressed amino acid sequence, and the relevant difference is the presence/absence of the 26-kb genomic segment, and of BV7S2 in particular. Thus, in general, there are hints throughout the literature that a combination of HLA class II alleles and the right kind of TCR is required for T cells to direct a pathogenic autoimmune response. The current findings open up new possibilities for investigating the mechanisms by which T cells exert this control in SS. ACKNOWLEDGMENT We thank Corinne Leget at Immunotech for her technical expertise. REFERENCES 1. Ortman RA, Holderbaum D, Qu XM, Banerjee S, Haqqi TM. BUB/BnJ (H-2q) is a TCR deletion mutant mouse strain (TCR V beta a, KJ16-) that is susceptible to type II collagen-induced arthritis. J Immunol 1994;152:4175–82. 2. Nabozny GH, Bull MJ, Hanson J, Griffiths MM, Luthra HS, David CS. Collagen-induced arthritis in T cell receptor V beta congenic B10.Q mice. J Exp Med 1994;180:517–24. 3. Osman GE, Hannibal MC, Anderson JP, Lasky SR, Ladiges WC, Hood L. FVB/N (H2(q)) mouse is resistant to arthritis induction and exhibits a genomic deletion of T-cell receptor V beta gene segments. Immunogenetics 1999;49:851–9. 4. Noonan DJ, McConahey PJ, Cardenas GJ. Correlations of autoimmunity with H-2 and T cell receptor beta chain genotypes in (NZB ⫻ NZW) F2 mice. Eur J Immunol 1990;20:1105–10. 5. Fox RI. Sjögren’s syndrome. In: Kelley WN, Harris ED, Ruddy S, Sledge CG, editors. Textbook of rheumatology. Philadelphia: Saunders; 1997. p. 955. 6. Chan EK, Andrade LE. Antinuclear antibodies in Sjögren’s syndrome. Rheum Dis Clin North Am 1992;18:551—70. 7. Zhao TM, Whitaker SE, Robinson MA. A genetically determined insertion/deletion related polymorphism in human T cell receptor ABSENCE OF TCRBV7S2 ASSOCIATED WITH PRESENCE OF AUTOANTIBODIES IN SS 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. beta chain (TCRB) includes functional variable gene segments. J Exp Med 1994;180:1405–14. Kay RA, Hutchings CJ, Ollier WE. A subset of Sjögren’s syndrome associates with the TCRBV13S2 locus but not the TCRBV2S1 locus. Hum Immunol 1995;42:328–30. Arden B, Clark SP, Kabelitz D, Mak TW. Human T-cell receptor variable gene segment families. Immunogenetics 1995;42:455–500. Tan EM, Cohen AS, Fries JF, Masi AT, McShane DJ, Rothfield NF, et al. The 1982 revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 1982;25:1271–7. Li Y, Szabo P, Robinson MA, Dong B, Posnett DN. Allelic variations in the human T cell receptor V␤6.7 gene products. J Exp Med 1990;171:221–30. Posnett DN, Wang CY, Friedman S. Inherited polymorphism of the human T-cell antigen receptor detected by a monoclonal antibody. Proc Natl Acad Sci U S A 1986;83:7888–92. Reveille JD, Macleod MJ, Whittington K, Arnett FC. Specific amino acid residues in the second hypervariable region of HLADQA1 and DQB1 chain genes promote the Ro (SS-A)/La (SS-B) autoantibody responses. J Immunol 1991;146:3871–6. Travaglio-Encinoza A, Chaouni I, DerSimonian H, Jorgensen C, Simony-Lafontaine J, Romagne F, et al. T cell receptor distribution in rheumatoid synovial follicles. J Rheumatol 1995;22:394–9. Posnett DN, Romagne F, Necker A, Kotzin BL, Sekaly R-P. Summary: first human TCR monoclonal antibody workshop. Immunologist 1996;4:5–8. Satoh M, Langdon JJ, Hamilton KJ, Richards HB, Panka D, Eisenberg RA, et al. Distinctive immune response patterns of human and murine autoimmune sera to U1 small nuclear ribonucleoprotein C protein. J Clin Invest 1996;97:2619–26. Reeves WH, Satoh M, McCauliffe DP. Autoantibody testing by non-FITC methods. In: Rose NR, Hamilton RG, Detrick B, editors. Manual of clinical laboratory immunology. Washington, DC: American Society of Microbiology Press; 2002. p. 933. Wang J, Satoh M, Kabir F, Shaw M, Domingo MA, Mansoor R, et al. Increased prevalence of autoantibodies to ku antigen in African American versus white patients with systemic lupus erythematosus. Arthritis Rheum 2001;44:2367–70. Schneider C, Newman RA, Sutherland DR, Asser U, Greaves MF. A one-step purification of membrane proteins using a high efficiency immunomatrix. J Biol Chem 1982;257:10766–9. Bradwell AR. Immunofluorescent antinuclear antibody tests. In: Rose NR, Hamilton RG, Detrick B, editors. Manual of clinical laboratory immunology. Washington, DC: American Society of Microbiology Press; 2002. p. 922. Satoh M, Kumar A, Kanwar YS, Reeves WH. Anti-nuclear antibody production and immune-complex glomerulonephritis in BALB/c mice treated with pristane. Proc Natl Acad Sci U S A 1995;92:10934–8. Satoh M, Weintraub JP, Yoshida H, Shaheen VM, Richards HB, Shaw M, et al. Fas and Fas ligand mutations inhibit autoantibody production in pristane-induced lupus. J Immunol 2000;165: 1036–43. Seboun E, Houghton L, Hatem CJ Jr, Lincoln R, Hauser SL. Unusual organization of the human T-cell receptor ␤-chain gene complex is linked to recombination hotspots. Proc Natl Acad Sci U S A 1993;90:5026–9. Liao L, Marinescu A, Molano A, Ciurli C, Sekaly R-P, Fraser JD, et al. TCR binding differs for a bacterial superantigen (SEE) and a viral superantigen (Mtv-9). J Exp Med 1996;184:1471–82. Choi Y, Kotzin B, Lafferty J, White J, Pigeon M, Kubo R, et al. A method for production of antibodies to human T-cell receptor ␤-chain variable regions. Proc Natl Acad Sci U S A 1991;88: 8357–61. Akolkar PN, Gulwani-Akolkar B, Robinson MA, Silver J. The influence of non-HLA genes on the human T-cell receptor repertoire. Scand J Immunol 1995;42:248–56. Posnett DN, Vissinga CS, Pambuccian C, Wei S, Robinson MA, Kostyu D, et al. Level of human TCRBV3S1 (V␤3) expression 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 197 correlates with allelic polymorphism in the spacer region of the recombination signal sequence. J Exp Med 1994;179:1707–11. Harley JB, Reichlin M, Arnett FC, Alexander EL, Bias WB, Provost TT. Gene interaction at HLA-DQ enhances autoantibody production in primary Sjögren’s syndrome. Science 1986;232: 1145–7. Hayashi Y, Haneji N, Hamano H, Yanagi K. Transfer of Sjögren’s syndrome-like autoimmune lesions into SCID mice and prevention of lesions by anti-CD4 and anti-T cell receptor antibody treatment. Eur J Immunol 1994;24:2826–31. Namekawa T, Kuroda K, Kato T, Yamamoto K, Murata H, Sakamaki T, et al. Identification of Ro(SSA) 52 kDa reactive T cells in labial salivary glands from patients with Sjögren’s syndrome. J Rheumatol 1995;22:2092–9. Sumida T, Namekawa T, Maeda T, Nishioka K. New T-cell epitope of Ro/SS-A 52 kDa protein in labial salivary glands from patients with Sjögren’s syndrome [letter]. Lancet 1996;348:1667. Halse AK, Wahren M, Jonsson R. Peripheral blood in Sjögren’s syndrome does not contain increased levels of T lymphocytes reactive with the recombinant Ro/SS-A 52 kD and La/SS-B 48 kD autoantigens. Autoimmunity 1996;23:25–34. Reynolds P, Gordon TP, Purcell AW, Jackson DC, McCluskey J. Hierarchical self-tolerance to T cell determinants within the ubiquitous nuclear self-antigen La (SS-B) permits induction of systemic autoimmunity in normal mice. J Exp Med 1996;184: 1857–70. Farris AD, Brown L, Reynolds P, Harley JB, James JA, Scofield RH, et al. Induction of autoimmunity by multivalent immunodominant and subdominant T cell determinants of La (SS-B). J Immunol 1999;162:3079–87. Reveille JD, Arnett FC. The immunogenetics of Sjögren’s syndrome. Rheum Dis Clin North Am 2000;18:539–50. Rischmueller M, Lester S, Chen Z, Champion G, Van Den Berg R, Beer R, et al. HLA class II phenotype controls diversification of the autoantibody response in primary Sjögren’s syndrome (pSS). Clin Exp Immunol 1998;111:365–71. Deshmukh US, Lewis JE, Gaskin F, Kannapell CC, Waters ST, Lou YH, et al. Immune responses to Ro60 and its peptides in mice. I. The nature of the immunogen and endogenous autoantigen determine the specificities of the induced autoantibodies. J Exp Med 1999;189:531–40. McCluskey J, Farris AD, Keech CL, Purcell AW, Rischmueller M, Kinoshita G, et al. Determinant spreading: lessons from animal models and human disease. Immunol Rev 1998;164:209–29. Scofield RH, Farris AD, Horsfall AC, Harley JB. Fine specificity of the autoimmune response to the Ro/SSA and La/SSB ribonucleoproteins. Arthritis Rheum 1999;42:199–209. Scofield RH, Kaufman KM, Baber U, James JA, Harley JB, Kurien BT. Immunization of mice with human 60-kd Ro peptides results in epitope spreading if the peptides are highly homologous between human and mouse. Arthritis Rheum 1999;42:1017–24. Scofield RH, Henry WE, Kurien BT, James JA, Harley JB. Immunization with short peptides from the sequence of the systemic lupus erythematosus-associated 60-kDa Ro autoantigen results in anti-Ro ribonucleoprotein autoimmunity. J Immunol 1996;156:4059–66. Moudgil KD. Diversification of response to hsp65 during the course of autoimmune arthritis is regulatory rather than pathogenic. Immunol Rev 1998;164:175–84. McRae BL, Vanderlugt CL, Dal Canto MC, Miller SD. Functional evidence for epitope spreading in the relapsing pathology of experimental autoimmune encephalomyelitis. J Exp Med 1995; 182:75–85. Vanderlugt CL, Miller SD. Epitope spreading in immune-mediated diseases: implications for immunotherapy. Nat Rev Immunol 2002;2:85–95. McFarland HI, Lobito AA, Johnson MM, Nyswaner JT, Frank JA, Palardy GR, et al. Determinant spreading associated with demy- 198 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. MANAVALAN ET AL elination in a nonhuman primate model of multiple sclerosis. J Immunol 1999;162:2384–90. Yu M, Johnson JM, Tuohy VK. A predictable sequential determinant spreading cascade invariably accompanies progression of experimental autoimmune encephalomyelitis: a basis for peptidespecific therapy after onset of clinical disease. J Exp Med 1996; 183:1777–88. Routsias JG, Touloupi E, Dotsika E, Moulia A, Tsikaris V, Sakarellos C, et al. Unmasking the anti-La/SSB response in sera from patients with Sjögren’s syndrome by specific blocking of anti-idiotypic antibodies to La/SSB antigenic determinants. Mol Med 2002;8:293–305. Sumida T, Yonaha F, Maeda T, Tanabe E, Koike T, Tomioka H, et al. T cell receptor repertoire of infiltrating T cells in lips of Sjögren’s syndrome patients. J Clin Invest 1992;89:681–5. Erratum in: J Clin Invest 1992;90:293. Pivetta B, De Vita S, Ferraccioli G, De Re V, Gloghini A, Marzotto A, et al. T cell receptor repertoire in B cell lymphoproliferative lesions in primary Sjögren’s syndrome. J Rheumatol 1999;26:1101–9. Yonaha F, Sumida T, Maeda T, Tomioka H, Koike T, Yoshida S. Restricted junctional usage of T cell receptor V␤2 and V␤13 genes, which are overrepresented on infiltrating T cells in the lips of patients with Sjögren’s syndrome. Arthritis Rheum 1992;35: 1362–7. Ohyama Y, Nakamura S, Matsuzaki G, Shinohara M, Hiroki A, Fujimura T, et al. Cytokine messenger RNA expression in the labial salivary glands of patients with Sjögren’s syndrome. Arthritis Rheum 1996;39:1376–84. Karim SN, Murphy EA, Sturrock RD, Goudie RB. T-cell clonality in synovial fluid from rheumatoid joints before and after culture in interleukin-2. Br J Rheumatol 1995;34:232–5. Dwyer E, Itescu S, Winchester R. Characterization of the primary structure of T cell receptor ␤-chains infiltrating the salivary gland in the sicca syndrome of HIV-1 infection: evidence of antigendriven clonal selection suggested by restricted combinations of V␤ and J␤ gene segment usage and shared somatically encoded amino acid residues. J Clin Invest 1993;92:495–502. Legras F, Martin T, Knapp AM, Pasquali JL. Infiltrating T cells from patients with primary Sjögren’s syndrome express restricted or unrestricted T cell receptor V beta regions depending on the stage of the disease. Eur J Immunol 1994;24:181–5. Sumida T, Matsumoto I, Maeda T, Nishioka K. T-cell receptor in Sjögren’s syndrome. Br J Rheumatol 1997;36:622–9. Kumar V, Stellrecht K, Sercarz E. Inactivation of T cell receptor peptide-specific CD4 regulatory T cells induces chronic experimental autoimmune encephalomyelitis (EAE). J Exp Med 1996; 184:1609–17. 57. Shevach EM. Regulatory T cells in autoimmmunity. Annu Rev Immunol 2000;18:423–49. 58. Kumar V, Sercarz E. Induction or protection from experimental autoimmune encephalomyelitis depends on the cytokine secretion profile of TCR peptide-specific regulatory CD4 T cells. J Immunol 1998;161:6585–91. 59. Saegusa K, Ishimaru N, Yanagi K, Haneji N, Nishino M, Azuma M, et al. Autoantigen-specific CD4⫹CD28low T cell subset prevents autoimmune exocrinopathy in murine Sjögren’s syndrome. J Immunol 2000;165:2251–7. 60. Lawson BR, Prud’homme GJ, Chang Y, Gardner HA, Kuan J, Kono DH, et al. Treatment of murine lupus with cDNA encoding IFN-␥R/Fc. J Clin Invest 2000;106:207–15. 61. Reininger L, Santiago ML, Takahashi S, Fossati L, Izui S. T helper cell subsets in the pathogenesis of systemic lupus erythematosus. Ann Med Interne (Paris) 1996;147:467–71. 62. Peng SL, Moslehi J, Craft J. Roles of interferon-␥ and interleukin-4 in murine lupus. J Clin Invest 1997;99:1936–46. 63. Kirou KA, Crow MK. New pieces to the SLE cytokine puzzle. Clin Immunol 1999;91:1–5. 64. Robinson MA, Kindt TJ. Segregation of polymorphic T-cell receptor genes in human families. Proc Natl Acad Sci U S A 1985;82:3804–8. 65. Seboun E, Robinson MA, Doolittle TH, Ciulla TA, Kindt TJ, Hauser SL. A susceptibility locus for multiple sclerosis is linked to the T cell receptor beta chain complex. Cell 1989;57:1095–100. 66. Wei S, Charmley P, Birchfield RI, Concannon P. Human T-cell receptor V␤ gene polymorphism and multiple sclerosis. Am J Hum Genet 1995;56:963–9. 67. Maksymowych WP, Gabriel CA, Luyrink L, Melin-Aldana H, Elma M, Giannini EH, et al. Polymorphism in a T-cell receptor variable gene is associated with susceptibility to a juvenile rheumatoid arthritis subset. Immunogenetics 1992;35:257–62. 68. Luyrink L, Gabriel CA, Thompson SD, Grom AA, Maksymowych WP, Choi E, et al. Reduced expression of a human V␤ 6.1 T-cell receptor allele. Proc Natl Acad Sci U S A 1993;90:4369–73. 69. Frank MB, McArthur R, Harley JB, Fujisaku A. Anti-Ro (SSA) autoantibodies are associated with T cell receptor beta genes in systemic lupus erythematosus patients. J Clin Invest 1990;85:33–9. 70. Scofield RH, Frank MB, Neas BR, Horowitz RM, Hardgrave KL, Fujisaku A, et al. Cooperative association of T cell beta receptor and HLA-DQ alleles in the production of anti-Ro in systemic lupus erythematosus. Clin Immunol Immunopathol 1994;72: 335–41. 71. Scofield RH, Dickey WD, Hardgrave KL, Neas BR, Horowitz RM, McArthur RA, et al. Immunogenetics of epitopes of the carboxyl terminus of the human 60-kD Ro autoantigen. Clin Exp Immunol 1995;99:256–61. Erratum in: Clin Exp Immunol 1995; 100:377.