HLA-DRB1 GENES AND DISEASE SEVERITY IN RHEUMATOID ARTHRITIS

Transcription

1 1802 ARTHRITIS & RHEUMATISM Vol. 39, No. 11, November 1996, pp , American College of Rheumatology HLA-DRB1 GENES AND DISEASE SEVERITY IN RHEUMATOID ARTHRITIS JOHN D. REVEILLE, GRACIELA s. ALARCON, SARAH E. FOWLER, STANLEY R. PILLEMER, ROSEMARIE NEUNER, DANIEL 0. CLEGG, ISIS S. MIKHAIL, DAVID E. TRENTHAM, JAMES C. C. LEISEN, GILBERT BLUHM, SHELDON M. COOPER, HOWARD DUNCAN, MARILYN TUTTLEMAN, STEPHEN P. HEYSE, JOHN T. SHARP, and BARBARA TILLEY, FOR THE MIRA TRIAL GROUP Objective. To examine the effect of alleles encoding the shared / rheumatoid epitope on rheumatoid arthritis (RA) disease severity in patients who participated In the minocycline in RA (MIRA) trial. Methods. Of 205 patients with a week-48 visit, blood was available for typing of HLA-DRB1 and HLA-DQB1 in 174 (85%) and successfully completed in 169 (82%). Baseline erosioes were used to assess disease severity and new erosions at the last visit served as a proxy for progression. Results. At baseline, there was no association between the presence of erosive disease or rheumatoid factor status and the dose of rheumatoid epitope (homozygous, heterozygous, none) or the specific alleles identified. At the final visit, a gradient was observed for the 3 allelic subgroups (and their gene doses) in the occurrence of new erosions among the Caucasian placebo- Supported in part by the NIAMS through research contracts N01-AR , N01-AR , N01-AR , N01-AR , N01-AR , N01-AR , N01-AR , center grant P-60- AR-20614, and grant R29-AR John D. Reveille, MD: University of Texas Health Science Center, Houston; Graciela S. Alardn, MD, MPH, Isis S. Mikhail, MPH: University of Alabama at Birmingham; Sarah E. Fowler, PhD, Barbara Tilley, PhD: Henry Ford Health Science Center, Detroit, Michigan; Stanley R. Pillemer, MD, Marilyn Tuttlernan, MS, Stephen P. Heyse, MD, MPH: NIAMS, NIH, Bethesda, Maryland; Rosemarie Neuner, MD, MPH: New York Health Science Center at Brooklyn; Daniel 0. Clegg, MD: University of Utah, Salt Lake City; David E. Trentharn, MD: Beth Israel Hospital, Boston, Massachusetts; James C. C. Leisen, MD, Gilbert Bluhm, MD, Howard Duncan, MD: Henry Ford Hospital, Detroit, Michigan; Sheldon M. Cooper, MD: University of Vermont College of Medicine, Burlington; John T. Sharp, MD: Tifton Medical Center, Tifton, Georgia. Address reprint requests to Graciela S. Alarch, MD, MPH, University of Alabama at Birmingham, Division of Clinical Immunology and Rheumatology, UAB Station, MEB 615, Birmingham, AL Submitted for publication November 20, 1995; accepted in revised form May 10, treated, but not the rninocycline-treated, patients. A treatment group/hla-dr4 epitope interaction was demonstrated in multivariate analyses. Approximately two-thirds of African-American patients did not have the rheumatoid epitope. Conclusion. HLA-DRB1 oligotyping may be useful in predicting the progression of disease in some Caucasian patients. Our study corroborates the infrequency of the epitope among African-American patients with RA. HLA-DRB1 alleles associated with rheumatoid arthritis (RA) share a sequence homology within the third hypervariable region of the DR p chain encompassing amino acids (QKRAA/QRRAA/ RRRAA). This sequence may be encoded by certain HLA-DR4 (DRB1*04) or non-dr4 alleles; thus, the concept of a shared epitope or rheumatoid epitope has emerged (1-4). A gene-dose effect has been suggested for RA, in which individuals homozygous for HLADRB1*0401 and those homozygous for the rheumatoid epitope exhibit more severe and destructive disease, particularly extraarticular manifestations, than those who are heterozygous or lack the rheumatoid epitope (5-7). The data supporting this effect were obtained at tertiary referral centers (5,6,8). The importance of the rheumatoid epitope has recently been confirmed in a study of twins and in studies of patients with early RA (9-11). Jawaheer et a1 (9) studied 91 monozygotic twins and found an increased concordance for the occurrence of RA in both HLA- DR4 positive (relative risk [RR] = 3.4) and rheumatoid epitope positive (RR = 3.7) twins. In one study of early polyarthritis, the rheumatoid epitope was found in high frequency in patients who later developed a destructive arthropathy, but was also found in patients with tran-

2 HLA-DRB1 AND RA SEVERITY 1803 sient polyarthritis and in the healthy control subjects (10). In another study of early RA, the distribution of the epitope and the alleles carrying the epitope was found to be different from that among patients with longstanding destructive arthropathy; patients without the rheumatoid epitope had minimal risk of joint destruction or rheumatoid factor (RF) seropositivity. At the other end of the spectrum, there are patients homozygous for the rheumatoid epitope encoded by HLA- DRBl*O401 with seropositive, destructive disease, usually accompanied by extraarticular manifestations. Between these two extremes, there are patients with various combinations of doses and alleles, who have disease of intermediate severity (12). The availability of a cohort of patients from the 48-week multicenter minocycline in RA (MIRA) trial (13), with disease of variable severity, offered a unique opportunity to further examine the role of the presence and dose of HLA-DR4 and non-dr4 alleles encoding the rheumatoid epitope in the severity of this disorder; we used erosions and the timing of their appearance as measures of disease severity. PATIENTS AND METHODS Patients. Patients with RA, according to the American College of Rheumatology (formerly, the American Rheumatism Association) criteria (14), with active disease, 29 tender and 26 swollen joints, and on a stable regimen of background medications were enrolled in MIRA, a multicenter, 48-week, placebo-controlled trial of minocycline (13). Race (Caucasian, African-American, Hispanic, or other) was defined by selfattribution. The presence or absence of erosions on coded hand/ wrist plain radiographs taken at baseline and the occurrence of new erosions on films taken at 48 weeks (obtained by averaging scores of 2 independent assessors) were used as proxies for disease severity. The reviewers were blinded as to radiograph sequence and treatment group. Any patient with a score >O was considered to have erosions (15). Newly involved joints were defined as those scored 0 at baseline by both readers and scored 21 by both readers at 48 weeks. Reference population. We used the frequencies of different genotypes, by ethnic group (North American Caucasians and African-Americans), from the Eleventh Histocompatibility Workshop. HLA-DRBl and DQBl typing. Blood was obtained at each participating center, refrigerated, and forwarded to The University of Texas Health Science Center at Houston. Samples were identified by the patient s study number. The buffy coat was separated, and typing and subtyping were performed on genomic DNA extracted using standard techniques (16). Data analyses. HLA-DRB1 alleles carrying the rheumatoid epitope include members of the DR4 group: DRB1*0401, *0404, *0405, *0408, *0409, *0410, *0413, *0416, *0419, and *0421 as well as certain non-dr4 alleles, specifi- cally, DRB1*0101, *0102, *1001, *1402, and *1406. We classified HLA-DR4 alleles into those encoding the epitope in a DR4 allele (DR4 epitope) or in a non-dr4 allele (non-dr4 epitope) and those not carrying the epitope (X). Thus the following allelic combinations were possible: (a) DR4 epitope/ DR4 epitope, (b) DR4 epitope/non-dr4 epitope, (c) non- DR4 epitope/non-dr4 epitope, (d) non-dr4 epitope/x, (e) DR4 epitopem, and (f) X/X. The role of HLA-DRB1*04 alleles was addressed by categorizing patients into 3 subgroups: 1) those with an HLA-DRB1*04 allele encoding the rheumatoid epitope (groups a, b, and e), or DR4 epitope, 2) those with a non-dr4 HLA-DRBl* allele encoding the epitope (groups c and d), or non-dr4 epitope, and 3) those without any allele encoding the epitope (group f), or non-epitope. The role of gene dosage was examined by categorizing patients as 1) those homozygous for the rheumatoid epitope (groups a, b, and c), or double, 2) those heterozygous for the rheumatoid epitope (groups d and e), or single, and 3) no epitope (group f), or non-epitope. The role of HLA-DRB1*0401 was examined by comparing those patients who were HLA-DRB1*0401 homozygous with all other patients. The role of HLA-DQB1 was examined by categorizing patients as having one of the HLA-DQB1 alleles linked to the rheumatoid epitope-bearing HLA-DRB1 alleles (as noted above), specifically, HL&DQB1*0301, *0302, and *0501, versus those not having them. Statistical analysis. Differences between means were examined by Student s t-test or Wilcoxon s test. Differences between proportions were examined by chi-square test or Fisher s exact test. We also conducted an exploratory analysis to determine whether the presence of the rheumatoid epitope was predictive of the response to treatment or of the development of new erosions over the course of the study. The analyses used P values of 0.05, based on univariate chi-square tests, as a guide for inclusion of variables in subsequent multivariate analyses. Adjustments for multiple comparisons were not performed because the analyses were exploratory. For predicting development of new erosions, we performed separate multiple logistic regression analyses for Caucasians and African-Americans, using a backward selection procedure starting with the following predictors: presence of the HLA-DR4 epitope, duration of RA, baseline RF positivity, treatment group, and the interaction between the DR4 epitope and treatment group. In another model, the dose of the epitope (i.e., homozygous, heterozygous) rather than the encoding allele was used as a variable. A residual 1 degree of freedom chi-square criterion of 0.05 was set for the retention of variables in the model. RESULTS Of the 219 MIRA patients, 205 completed a 48-week study visit. Blood was available for typing in 174 patients (85%); of those, 169 (82%) were successfully oligotyped. In the typed group, 104 were Caucasians, 52 were African-Americans, 10 were Hispanics, and 3 were

3 1804 REVEILLE ET AL Table 1. Percentages of patients in the minocycline in rheumatoid arthritis study with different haplotypes encoding the rheumatoid epitope, by ethnic group No. (%) No. (%) No. (%) No. (%) Caucasians African-Americans Hispanics all Haplotype (group) (n = 104) (n = 52) (n = 10) (n = 169)t (a) 12 (12) 0 (0) 0 (0) 12 (7) 04/0101, 0102, 1001, 1402, 1406 (b) 8 (8) 0 (0) 0 (0) 8 (5) 0101, 0102, 1001, 1402, 1406/0101, 0102, 4 (4) 0 (0) 0 (0) 5 (3) 1001, 1402, 1406 (c) 0101,0102, 1001, 1402, 1406/X (d) 22 (21) 10 (19) 3 (30) 36(21) 04/X (e) 25 (24) 8 (15) 4 (40) 38 (23) x/x (f) 33 (32) 34 ( (30) 70 (41) * All percentages are rounded to the nearest integer. f Three patients from other ethnic groups are included in the total. $04 includes 0401, 0404, 0405/08, 0413, 0416, 0419, and Q P < versus non-african Americans, by Fisher s exact test. from other ethnic groups. There was only minimal overlap (4 of 52) between the African-American patients participating in this study and those whose HLA allelic frequencies we have previously reported. Baseline demographic and clinical features in patients who were typed (56% were seropositive, 13% had subcutaneous nodules, 28% had joint deformities, 5% had had reconstructive joint surgery, and -50% had previously taken disease-modifying antirheumatic drugs [DMARDs]) were similar to those for whom typing could not be performed, with the exception of joint counts for swelling, which were higher among those patients who were typed (P < 0.01). HLA-DRB1* data. Reference population and MIRA patients. As expected, HLA-DRB1*0401, *0404, and *0101 alleles in Caucasians and HLA-DRB1*0401 in African-Americans were more frequent in the MIRA patients than in their respective reference population (MIRA Caucasians versus reference population 41.2% versus 16.6%; MIRA African-Americans versus reference population 6.8% versus 1.5%; P < 0.05 for both). Distribution of haplotypes and alleles in MIRA patients. Table 1 shows the proportion of MIRA patients with the different HLA-DRBl allele combinations by ethnic group. A higher proportion of African-Americans (65%), than either Caucasians (32%) or Hispanics (30%), did not carry the rheumatoid epitope in either allele (W). In addition, African-American patients who had the rheumatoid epitope were heterozygous for the epitope; their haplotypes were either DR4 epitope/x (group e) or non-dr4 epitope/x (group d). Although the number of Hispanic patients lacking the rheumatoid epitope was similar to that in the Caucasian patients, none of the small number of Hispanics had a double dose of the epitope. Percentages of MIRA patients with DR4 epitope, non-dr4 epitope, and non-epitope-encoding alleles, by ethnic group and for all patients, are shown in Table 2. Table 2. Percentages of patients in the minocycline in rheumatoid arthritis study with alleles encoding the rheumatoid epitope, by ethnic group* No. (%) No. (%) No. (%) No. (%) Epitope-encoding Caucasians African-Americans Hispanics all alleles (n = 104) (n = 52) (n = 10) (n = 169)t DR4 epitopet 4s (43) 8 (15) 4 (40) Non-DR4 epitopes 26 (25) 10 (19) 3 (30) Non-epi topell 33 (32) 34 (65) 3 (30) 58 (34) 41 (24) 70 (41) * All percentages are rounded to the nearest integer. We tested for differences among races in the distribution of the classification epitope-encoding alleles. t Three subjects from other ethnic groups are included in the total. $0401 or 0404 or 0405/08 or 0413 or 0416 or 0419 or 0421 or groups a, b: and e in Table 1. Q 0101 or 0102 or 1001 or 1402 or 1406 or groups c and d in Table 1. ll Group f in Table 1.

4 ~~ HLA-DRB1 AND RA SEVERITY 1805 Table 3. Caucasian patients who developed erosions over the trial duration, by treatment group, rheumatoid epitope dose, and rheumatoid epitope-encoding allele' No. (%) with new erosions Minocycline group Placebo group Yes No Yes No (n = 6) (n = 46) (n = 12) (n = 33) Epitope-encoding allele DR4 epitopet l(5) 19 (95) 8(42) 11 (58) Non-DR4 epitope$ 2(13) 14(87) 3(33) 6(67) Non-epitopes 3(19) 13(81) l(6) 16(94) Epitope dose Homozygous (doub1e)ll 2 (15) 11 (85) 4 (44) 5 (56) Heterozygous (single)# 1 (4) 22 (96) 7 (36) 12 (63) Non-epitopes 3 (19) 13 (81) l(6) 16 (94) * All percentages were rounded to the nearest integer. t 0401,0404,0405/08,0413,0416,0419, and 0421, and groups a, b, and e in Table 1. $0101,0102, 1001, 1402, and groups c and d in Table 1. 8 Group fin Table Groups a, b, and c in Table 1. # Groups d and e in Table 1. The frequencies of DR4 epitope, non-dr4 epitope, and non-epitope alleles were comparable for patients in the minocycline and placebo groups (data not shown). Relationship of haplotypes and alleles to baseline erosions and RF positivity. Frequencies of DR4 epitope, non-dr4 epitope, and non-epitope alleles were comparable for patients with erosive and nonerosive disease, as well as for RF-positive and RF-negative patients. Similar results were obtained when instead of the allele, the dose of the rheumatoid epitope was examined in relation to both the presence of baseline erosions and RF positivity (data not shown). Relationship of haplotypes and alleles to new erosions at 48 weeks, by univariate analyses. Table 3 shows the occurrence of new erosions at 48 weeks in Caucasian patients by treatment group, according to the alleles encoding for the epitope and the epitope dose. New erosions over the 48-week trial occurred in 12 placebotreated, but in only 6 minocycline-treated patients. A higher frequency of new erosions was observed in the DR4 epitope or double dose of the epitope (42% and 44%, respectively) than in the non-dr4 epitope or the single dose of the epitope (33% and 36%) and nonepitope (6%) alleles among the placebo-treated patients. However, the frequencies of new erosions were decreased (and were comparable) in the minocyclinetreated patients regardless of the presence of the epitope, its dose, or the allele encoding it. Of note, the baseline clinical characteristics of the patients with DR4- encoding alleles were comparable for the minocycline and placebo groups except for the erythrocyte sedimentation rate (ESR), which was higher among the placebotreated HLA-DRB1*04 patients than among their minocycline counterparts. In contrast to the Caucasians, there was a comparable number of patients who developed new erosions among the minocycline-treated and the placebo-treated African-American patients, and no trends were observed with regard to the dose of the rheumatoid epitope or the alleles encoding for the epitope. The small number of Hispanic patients and patients from other ethnic groups precluded similar analyses of those populations. Finally, when patients homozygous for HLA-DRB1*0401 (7 patients, all Caucasians) were compared with all other Caucasian patients with the rheumatoid epitope and with those without the epitope, they were more likely to be seropositive (86% versus 63% versus 44%, respectively) and to have erosive disease (71% versus 55% versus 51%, respectively). Relationship of haplotypes and alleles to new erosions at 48 weeks, by multivariate analyses. The multivariate analyses were limited to the Caucasian patients. The predictors chosen were treatment group, HLA-DR4 epitope present/absent, RF status, and disease duration. The model which best fits the data is one which uses the presence of the DR4 epitope and its interaction with treatment group. The data in Table 4 indicate that placebo-treated Caucasian patients who had an HLA- DR4 epitope-encoded allele were more likely than minocycline-treated Caucasian patients to develop new erosions over 48 weeks (odds ratio = 14, 95% confidence interval ). For the model using the rheumatoid epitope dose, there was a trend consistent with the gene-dose effect for the epitope and the treatment group, as well as the interaction of both (data not shown). HLA-DQB1* data. There was a higher proportion of HLA-DQBl*X in the African-Americans than in the Caucasians, but the differences were not statistically significant. When the distribution of the alleles of interest (those linked to HLA-DR4, DR1, and DR10) versus all others was examined with regard to erosive disease, seropositivity, or response to treatment, no differences were observed (data not shown). Therefore, HLA- DQBl alleles were not entered into the regression analyses. There was likewise no association of any HLA-DQB1 allele with RA in either the Caucasians or the African-Americans (data not shown).

5 1806 REVEILLE ET AL Table 4. Percentages of minocycline in rheumatoid arthritis study patients who developed new erosions, according to treatment group and the presence of HLA-DR for the rheumatoid epitope-encoding alleles Treatment group. rheumatoid epitopeencoding allele No. (%) with new erosions at trial completion 95% Odds confidence Yes No ratio* interval Minocycline DR4 epitope 1(5) 19 (95) Non-DR4 epitope and non-epitope 5 (16) 27 (84) Placebo DR4 epitope 8 (42) 11 (56) Non-DR4 epitope and non-epitopet 4 (15) 22 (65) - - * As compared with the reference group. t Reference group. DISCUSSION Predicting the occurrence of destructive arthropathy in patients with early RA is very appealing to the clinical rheumatologist (17). Thus, the concept of the shared or rheumatoid epitope and its gene-dose effect as described in different patient populations has been favorably received. However, in order to take these concepts to clinical practice, all major ethnic groups and patients with disease of various degrees of severity and various clinical features should be studied. In fact, the frequency of the rheumatoid epitope in other Caucasian populations has been found to be lower than in the selected RA population studied at tertiary centers (18). The availability of RA patients participating in a large multicenter clinical trial (the MIRA trial) (13) provided us with the unique opportunity to examine whether there is an association between the presence and dose of the rheumatoid epitope, as well as the alleles encoding it, and disease severity. As expected, the MIRA patients had an increased frequency of HLA- DRB1*0401 compared with healthy North American controls. However, the MIRA patients, in contrast to those described by Weyand et a1 (9, had overall shorter disease duration and less severe RA. In fact, over 40% did not have erosions at baseline because of the inclusion criteria, which called for patients who had active disease, but who had failed only 1 DMARD because of lack of efficacy. In this patient group, the surrogate marker of disease severity, radiographic evidence of erosions at baseline, tended to be associated with the presence of the rheumatoid epitope, but the association did not reach statistical significance. A baseline erosion rate was not calculated, since radiographs preceding the MIRA trial were not available in all patients, and thus, we may have failed to see the effect of the epitope since we do not know at what point in time they actually occurred prior to patient enrollment in the study. However, because of the longitudinal nature of the study, we were able to determine whether another marker of disease severity, the occurrence of new erosions over the course of the trial, correlated with HLA-DRBl alleles. For this analysis, patients receiving placebo can be considered a control group because in these patients, the occurrence of new erosions would be influenced primarily by the patient s genetic makeup and not by treatment. The data for the Caucasian patients are particularly interesting. In the placebo-treated patients, more DR4 epitope-positive patients (and epitope homozygous) developed erosions than those with either non- DR4 epitope (and epitope heterozygous)-encoding alleles or those without any rheumatoid epitopecontaining alleles. This gradient was not observed in the minocycline-treated patients; quite the contrary, new erosions occurred less frequently in those patients with the epitope than in those without it. Despite the presence of the rheumatoid epitope (or its dose), minocycline-treated patients developed fewer new erosions over the duration of the trial. It should be noted, however, that the placebo-treated patients with DR4 epitope-encoding alleles had a higher ESR at baseline than the minocycline-treated patients, possibly reflecting an added risk factor for the occurrence of disease progression. In contrast, no differential epitope effect from minocycline was observed among minocyclinetreated and placebo-treated African-American patients, in whom the majority, as previously reported in a different patient population, lacked the rheumatoid epitope (19). Taken together, our data suggest that certain treatments may be more effective for some RA

6 HLA-DRB1 AND RA SEVERITY 1807 populations than for others, and that these differences may have a genetic basis. In summary, our study further supports the notion that the presence of alleles encoding for the rheumatoid epitope (and dose) may be useful predictors of disease progression (as determined by erosions) in Caucasians. The response to treatment may also be influenced by the subject s genetic makeup. The low frequency of the rheumatoid epitope among African- American RA patients is also confirmed. REFERENCES 1. Willkens RF, Nepom GT, Marks CR, Nettles JW, Nepom BS: Association of HLA-Dwl6 with rheumatoid arthritis in Yakima Indians: further evidence for the shared epitope hypothesis. Arthritis Rheum 34:43-47, De Vries N, Ronningen KS, Tilanus MG, Bouwens-Rombous A, Segal R, Egelund T, Thorsby E, van de Putte LBA, Brautban C HLA-DR1 and rheumatoid arthritis in Israeli Jews: sequencing reveals that DRB is the predominant HLA-DR1 subtype. Tissue Antigens 41:26-30, De Juan MD, Belmonte I, Barado J, Martinez-Laso J, Figueroa M, Arnaiz-Villena A, Cuadrado E: Differential associations of HLA-DR antigens with rheumatoid arthritis (RA) in Basques: high frequency of DR1 and DRlO and lack of association with HLA-DR4 of any of its subtypes. Tissue Antigens 43: , Gregersen PK, Silver J, Winchester RJ: The shared epitope hypothesis: an approach to understanding the molecular genetics of susceptibility to rheumatoid arthritis. Arthritis Rheum 30: , Weyand CM, Hicok KC, Conn DL, Goronzy JJ: The influence of HLA-DRB1 genes on disease severity in rheumatoid arthritis. Ann Intern Med , Nepom GT, Nepom BS: Prediction of susceptibility to rheumatoid arthritis by human leukocyte antigen genotyping. Rheum Dis Clin North Am 18: , Goodman SN, Berlin JA The use of predicted confidence intervals when planning experiments and the misuse of power when interpreting results. Ann Intern Med 121: , Toda Y, Minamikawa Y, Magi S, Sugano H, Mori Y, Nishimura H, Arita S, Sugino Y, Ogawa R: Rheumatoid-susceptible alleles of HLA-DRB1 are genetically recessive to non-susceptible alleles in the progression of bone destruction in the wrists and fingers of patients with RA. Ann Rheum Dis , Jawaheer D, Thomson W, MacGregor AJ, Carthy D, Davidson J, Dyer PA, Silman AJ, Ollier WER Homozygosity for the HLA-DR shared epitope contributes the hjghest risk for rheumatoid arthritis concordance in identical twins. Arthritis Rheum 37: , Gough A, Faint J, Salmon M, Hassell A, Wordsworth P, Pilling D, Birley A, Emery P: Genetic typing of patients with inflammatory arthritis at presentation can be used to predict outcome. Arthritis Rheum 37: , Weyand CM, Goronzy JJ: HLA-DRB1 alleles as severity markers in rheumatoid arthritis. Bull Rheum Dis 43:5-8, Weyand CM, McCarthy TG, Goronzy JJ: Correlation between disease phenotype and genetic heterogeneity in rheumatoid arthritis. J Clin Invest 95: , Tilley BC, Alardn GS, Heyse SP, Trentham DE, Neuner R, Kaplan DA, Clegg DO, Leisen JCC, Buckley L, Cooper SM, Duncan H, Pillemer SR, Tuttleman M, Fowler SE, for the MIRA Trial Group: Minocycline in rheumatoid arthritis: a 48-week, double-blind, placebo-controlled trial. Ann Intern Med , Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, Healey LA. Kaplan SR, Liang MH, Luthra HS, Medsger TA Jr, Mitchell DM, Neustadt DH, Pinals RS, Schaller JG, Sharp JT, Wilder RL, Hunder GG: The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 31: , Bluhm GB, Sharp JT, Tilley B, Rusinova M, Alarcon GS, Heyse SP, Trentham DE, Neuner R, Kaplan DA, Clegg DO, Leisen JCC, Buckley L, Cooper SM, Duncan H, Pillemer SR, Tuttleman M, Fowler SE Radiographic results from the minocycline in rheumatoid arthritis trial (MIRA) (abstract). Arthritis Rheum 37 (Suppl 9):S338, Reveille JD, Durban E, MacLeod-St. Clair MJ, Goldstein R, Moreda R, Altman RD, Arnett FC: Association of amino acid sequences in the HLA-DQB1 first domain with the antitopoisomerase I: autoantibody response in scleroderma (progressive systemic sclerosis). J Clin Invest 90: , Harris EDJ: Rheumatoid arthritis: pathophysiology and implications for therapy. N Engl J Med 322: , Boki KA, Panayi GS, Vaughan RW, Drosos AA, Moutsopoulos HM, Lanchbury JS: HLA class I1 sequence polymorphisms and susceptibility to rheumatoid arthritis in Greeks the HLA-DRp shared-epitope hypothesis accounts for the disease in only a minority of Greek patients. Arthritis Rheum 35: , McDaniel DO, Alardn GS, Pratt PW, Reveille JD: Most African- Americas patients with rheumatoid arthritis do not have the rheumatoid antigenic determinant (epitope). Ann Intern Med 123: , 1995