HLA-DP in rheumatoid arthritis

Frequencies of HLA-DR, Dw and DPw specificities were compared between rheumatoid arthritis (RA) patients, Felty's syndrome (FS) patients and normal controls. It was confirmed that the frequency of DR4 was increased in RA patients (54% (n = 111) vs 23% (n = 272), relative risk (RR) = 3.98, P less than 0.001). Cellular typing showed a highly significant increase in HLA-Dw14 in the entire RA population (17% (n = 32) vs 2% (n = 242), RR = 11.90, P less than 0.001), and a tendency towards an increase of HLA-Dw14 in DR4+ RA patients compared to DR4+ controls (28% (n = 32) vs 11% (n = 47), RR = 3.29, P less than 0.05). Regarding DPw specificities, the only significance was for a negative association with DPw3 (13% vs 22% (n = 254), RR = 0.51, P less than 0.05), with an additional tendential decrease of DPw1 (11% vs 19%, RR = 0.53, not significant (NS]. The decrease of DPw3 was more marked in DR4- RA patients (RR = 0.33, P less than 0.05) than in DR4+ RA patients (RR = 0.69, NS). In FS patients, 96% of whom were DR4+, decreased DPw1 was very marked, whereas the frequency of DPw3 was unaltered compared to DR4+ normals. These alterations in frequencies were not caused by linkage disequilibria between HLA-DR and -DP alleles. Thus, taken together, these data suggest that, in the presence of the major DR4-associated "susceptibility" gene(s) for RA, DPw1 may have "protective" effects, whereas in the absence of DR4, the presence of DPw3 has significant "protective" activity.     

HLA-DQ beta 3.1 allele is a determinant of susceptibility to DR4-associated rheumatoid arthritis

Rheumatoid arthritis is associated with HLA-DR4 in several ethnic groups. Since DR4 haplotypes encode a diverse array of class II molecules, it is of interest to characterize the nature of the primary association. We have examined molecular polymorphisms of HLA class II gene products expressed by normals and rheumatoid arthritis patients using monoclonal antibodies and two-dimensional electrophoresis. Most homozygous DR4 rheumatoid arthritis patients express DR beta 1 molecules associated with Dw4 or Dw14 mixed lymphocyte culture determinants. In Caucasoids, two DR4-linked DQw3-associated beta-chain alleles are defined by two-dimensional electrophoresis. These variants, designated DQ beta 3.1 and 3.2, are associated with the serologic determinants DQw7 and DQw8, respectively. A panel of 40 DR4-positive normals was also examined for nucleotide sequence polymorphisms associated with DQB3.1 and 3.2 genes using the polymerase chain reaction and specific oligonucleotide probes. At the DQ beta level the rheumatoid arthritis panel was distinguished by enrichment for the DQ beta 3.1 allele with 100% of patients positive for DQw7. Results presented here suggest that specific DQ beta alleles may modify the effect of HLA-DR4 beta 1 alleles in conferring susceptibility to rheumatoid arthritis in a phenotype-specific fashion.     

The molecular basis for HLA class II associations with rheumatoid arthritis

The association of HLA-DR4 with rheumatoid arthritis strongly implicates genes of the major histocompatibility complex (MHC) as contributing to disease susceptibility. Molecular analysis of MHC genes expressed on haplotypes in association with HLA-DR4 reveals that at least six different alleles of the DR₁ locus and at least three different alleles of the DQ locus occur on different DR4+ haplotypes. Some of these allelic differences are quite substantial, and others are rather subtle, involving as few as two amino acids. The analysis of individual DR and DQ alleles in rheumatoid arthritis identifies some DR4+ genes strongly associated with disease susceptibility and some DR4+ genes which are not. The Dw4(DR4) and Dw14(DR4) DR₁ genes appear to represent specific alleles which confer disease risk in RA; other DR₁ genes, such as Dw10(DR4), may represent DR genes altered during evolution which have lost their contribution to RA susceptibility. DQ 3.1(DQw3) and DQ 3.2(DQw3) DQ genes, which are present on DR4+ haplotypes, represent discrete variable alleles not directly implicated in RA, but which account for HLA-DR4 associations with other diseases, such as the association of DQ 3.2(DQw3) with Type I diabetes.     

Analysis of DR and DQ gene products of the DR4 haplotype in patients with IDDM: Possible involvement of more than one locus

Fifteen DR4-bearing haplotypes from twelve patients with insulin-dependent diabetes mellitus (IDDM) were analyzed serologically, cellularly, and biochemically. The HLA-Dw composition of these DR4-positive haplotypes was Dw4 (46%), Dw14 (22%), and Dw10 (33%). The biochemical analysis by two-dimensional electrophoresis (2D-PAGE) of the DR beta chains showed that each Dw specificity is characterized by a specific DR4 beta chain that appears to be identical in normal and diabetic individuals. Analysis of DQ beta chains in the DR4-bearing haplotypes revealed that certain Dw specificities such as Dw4 are characterized by the presence of either the DQw7 (formerly DQw3.1) or DQw8 (formerly DQw3.2) alleles, which generate the Dw4.1 or Dw4.2 subtypes, respectively. Others such as Dw14 and Dw10 are characterized by the presence of the DQw8 allele. In our sample of 12 patients the Dw4.2 (Dw4, DR4βI-4 DQw8) and Dw10 (Dw10, DR4βI-1, DQw8) subtypes were predominant. It is concluded that individual DR β and DQ β gene products from the DR4-bearing haplotype of IDDM patients are identical to those of normal control subjects and that Dw14 as well as Dw10 are involved in disease susceptibility. We suggest that disease susceptibility may be influenced by more than one locus within the HLA-D region.     

HLA DQ alpha and DQ beta restriction fragment length polymorphisms associated with Felty's syndrome and DR4-positive rheumatoid arthritis

HLA DQ alpha and DQ beta cDNA probes were used to study TaqI generated restriction fragment length polymorphisms (RFLPs) in DR4-positive patients with Felty's syndrome (FS), seropositive rheumatoid arthritis (RA), and in HLA-DR4 positive controls. The results of this analysis revealed two DQ beta RFLP patterns (DQ beta 3a and DQ beta 3b) associated with DR4, of which DQ beta 3b was found at significantly higher frequency in patients with FS (73%) or with RA (52%) than in DR4 controls (29%). Hind III generated RFLPs provide evidence that DQ beta 3b is in strong linkage disequilibrium with the gene encoding the serologically recognized epitope TA10. Results obtained using a DQ alpha chain probe revealed polymorphic differences between DQ alpha chain genes associated with different DR types, thereby providing a possible explanation for the lack of association between RA and other DR haplotypes in linkage disequilibrium with TA10. We conclude that both DQ alpha and DQ beta genes may be important in determining HLA-linked susceptibility to severe forms of RA.     

HLA antigens in Tlingit Indians with rheumatoid arthritis

HLA-DR4 has been described in association with rheumatoid arthritis (RA) in multiple populations. We have studied HLA antigens in Alaskan Tlingit Indians. HLA-DR4 was decreased in the RA group (n = 32) compared with controls (n = 62) (6% vs 21% p = 0.07). The predominant DR4 allele observed was DRB1*0403 (Dw13.1). The most striking observation in these studies was a marked predominance of the DRB1*1402 allele encoding Dw16 (DRw14). This allele was present in 91% of RA cases, but was also highly prevalent in controls (80%, OR = 2.4 p = 0.20). DRB1*1402 only was observed in 47% of cases and 31% of controls. The DRB3*0101 (DRw52), and the DQA*0501 and DQB*0301 alleles encoding a subset of DQw3 were associated with DRB1*1402 in cases and in controls. HLA-Bw62 was increased in RA cases (28%) compared with controls (8%) (OR = 4.5, p = 0.01, corrected p = ns).     

Full-length DQβ cDNA sequences of HLA-DR2/DQw1 subtypes: Genetic interactions between two DQβ loci generate human class II HLA diversity

Two-dimensional gel electrophoresis of DQ molecules from three different Dw subtypes (Dw2, Dw12, and Dw21/FJO) of the HLA-DR2/DQw1 haplotype reveals that one β heterodimer of DQ molecule is expressed by each subtype and the DQB chain is electrophoretically variable among the three DR2/DQw1 subtypes. We have constructed cDNA libraries from the same homozygous typing cells used in the two-dimensional polyacrylamide gel electrophoresis analyses (HTC VYT for Dw2, HTC DHO for Dw12, and HTC FJO for Dw21/FJO) and isolated DQβ cDNA clones with full-length coding sequences for each subtype. The deduced amino acid sequences show that the DQβ chains of these three DR2/DQw1 subtypes are highly polymorphic and confirm their electrophoretic heterogeneity: for a mature protein of 229 amino acids, they differ with each other by 10–17 amino acids in the first domain and by 3–7 residues in the C-terminal sequence. Comparison among the available DQβ sequences representing the four major DQ specificities (DQw1, DQw2, DQw3, and DQw4) in the DQ subregion as defined by serologic method suggests that (1) DR2, Dw2, DQw1 and DR3, DQw2 haplotypes probably interact with each other to generate the DQw3 and DQw4 β alleles and (2) an evolutionary scheme may be proposed to relate the various β alleles of the four major DQ specificities.     

HLA-DR4 associated Dw types in rheumatoid arthritis

Frequencies of HLA-DR4 and its related Dw types were compared between randomly selected normal controls and the index cases of multiplex rheumatoid arthritis (RA) families. A DR4 frequency of 68.3% was observed in index cases (n = 57) compared to 31.2% in normal controls (n = 96). Cellular typing with homozygous typing cells (HTCs) revealed significant increases of Dw4 (49.1% vs 22.9% RR = 3.2 p less than 0.001) and Dw14 (22.8% vs 2.1% RR = 13.9 p less than 0.001) in the index cases. A non-significant increase was seen for Dw13 (8.8% vs 4.1%). When DR4 positive patients and controls were compared, a significant increase was seen only for Dw14 (34.2% vs 6.6% RR = 7.3 p less than 0.01). Data from HLA genotyped RA and normal families allowed an examination of haplotype combinations of HLA-B antigens and DR4/Dw types to be made. HLA-Dw4 was predominantly found with B44 and Bw62 with nearly all DR4/Bw62 haplotypes being Dw4 positive. HLA-Dw13 was associated with B44 and Dw14 with Bw60, B44 and B27. Based on HTC and normal family data. Dw10 was found to be strongly associated with B38 containing haplotypes. Analysis of 69 C4A, C4B complement typed DR4 haplotypes failed to show any statistically significant association between Dw type and "complotype". However, there was a suggestion of C4A3. BQO being associated with Dw4 (34.2% vs 16.1% X2 = 2.9 p = ns) and C4A3, B1 with Dw14 (45.5% vs 27.6% X2 = 2.1 p = ns).(ABSTRACT TRUNCATED AT 250 WORDS)     

Rheumatoid arthritis may be primarily associated with HLA-DR4 molecules sharing a particular sequence at residues 67–74

Genomic typing of in vitro amplified DNA with sequence-specific oligonucleotide (SSO) probes was performed for DRB1, DQA1, DQB1, DPA1 and DPB1 alleles in 54 random Norwegian rheumatoid arthritis (RA) patients and 181 healthy controls. DRB1 alleles encoding the serological specificity DR4 were found in 80% of the patients, compared to 34% of the controls (relative risk = 7.9, p less than 0.0001). All DR4-positive RA patients carried either DRB1*0401 (Dw4), 0404 (Dw14), or 0405 (Dw15), while no patients were found to carry DRB1*0402 (Dw10) or 0403 (Dw13). The frequency of the DRB1*0101 allele encoding DR1 was not increased, even among DR4-negative RA patients, and we were unable to detect any sharing of other class II alleles among DR4-negative patients. No contribution of any DQA1, DQB1, DPA1 or DPB1 alleles to RA susceptibility could be detected. The results suggest that in the Norwegian population RA is primarily associated with a shared sequence at residues 67-74 of the DR beta 1 chain, but only when this sequence is expressed on DR4 molecules.     

Functional analysis of MHC class II-restricted T cells derived from a Caucasian with a DR4, Dw15, DQw8 haplotype

Rabies virus-specific CD4+ T lymphocyte clones were isolated from a Caucasian male vaccine recipient (DR4/7, DQw2/w8; DPw4) and studied for their major histocompatibility complex restricting elements. None of the rabies-specific T-cell clones could be induced to proliferate to antigen by either lymphoblastoid cells or DR-transfected L cells expressing DR4 molecules of the Dw subtypes commonly found on Caucasian individuals (Dw4, Dw10, Dw13, Dw14). The HLA-Dw subtype of the rabies vaccine recipient was determined by conventional mixed lymphocyte culture, and the results revealed that this individual had a DR4 (Dw15), DR7 (Dw7) phenotype. The presence of the DR4, Dw15 antigen was confirmed by nucleotide sequencing of the DR4B1 gene corresponding to the DRB1*0405 allele. Significant antigen-induced T-cell proliferative responses were obtained with two DR4, Dw15, DQw4 homozygous lymphoblastoid cell lines of Japanese origin (HAS-15 and KT-3) and with a L-cell transfectant expressing the DR4, Dw15 molecule. The existence of the DR4, Dw15 antigen in the Japanese has been reported to be associated with the DQw4 specificity. However, the presence of DQw8 (previously designated DQw3.2) and the absence of DQw4 in the lymphoblastoid cells of the Caucasian rabies vaccine was confirmed with monoclonal antibodies IVD12 (anti-DQw7 + DQw8 + DQw9) and HU46 (anti-DQw4) and by the reactivity of a DQw8-restricted antigen-specific T-cell clone. These studies indicate, contrary to previous findings, that the DR4, Dw15 molecule may be present in Caucasian (non-Japanese) individuals in association with DQw8.     

T-cell-defined DR4 subtypes as markers for type 1 diabetes

In most populations studied, HLA-DR4, a DRB1 (formerly DR beta I) allele, is increased in frequency among patients with insulin-dependent diabetes mellitus (IDDM). Using T-cells, one can distinguish five subtypes of DR4 (Dw4, Dw10, Dw13, Dw14, and Dw15). Two of these (Dw4 and Dw10) are IDDM-associated in the populations studied here. Therefore, Dw4 and Dw10 could be causative or merely markers for a linked diabetes allele. If they are causative, one might expect them to share some unique structural element or at least to associate consistently with IDDM in different populations. Published sequence data show no structural element unique to Dw4 and Dw10; moreover, the associations of these DR4-Dw subtypes with diabetes vary considerably in different populations. Thus the DRB1 locus probably cannot account for the DR4 association in IDDM. The strong linkage disequilibrium between IDDM and Dw4 and Dw10 suggests that the diabetes susceptibility locus should be in the vicinity of the DR region or the DQ region of the HLA complex. Alternative hypotheses are discussed, relating DR- and DQ-region alleles to IDDM. We further postulate that the evolutionary event that produced the Dw10 allele occurred on a Dw4 haplotype that happened to carry a diabetes susceptibility allele at another locus.     

Heterogeneity of HLA-DR4 in Greeks including a unique DR4-DQw2 association.

It has been shown that Greek RA patients do not show an increased frequency of HLA-DR4 compared with Greek controls. As only the Dw4 and Dw14 subtypes of DR4 are implicated in RA, we have characterised DR4-positive Greek RA patients and controls using serology, MLC typing and RFLP analysis to see whether the distribution of DR4 subtypes or other HLA antigens associated with DR4 could account for these findings. In the 10 patients and 12 controls studied, Dw10 was common, being present in almost half the controls. Dw4 was not detected in the controls, but was present in three of the RA patients, indicating that there may be some relationship between Dw4 and RA in this population; Dw13 and Dw14 were also found, Dw15 was not detected. Eight of the subjects did not type as any of the HLA-D antigens mentioned. Three controls had unusual DR4-DQ associations, two having DR4-DQw2 and one possessing DR4-DQw4. Analysis of the TaqI DRB RFLP of the subjects showed that one control did not have the 6.1 kb band characteristic of DR4s. All these results indicate that DR4 is a heterogeneous antigen in this population and that several as yet undescribed variants of DR4 may be present.     

HLA class II genes poIymorphism in DR4 giant cell arteritis patients

We have previously reported a significant increase of HLA-DR4 antigen frequency in giant cell arteritis (GCA). This finding suggested an important role of immunogenetic factors in this syndrome. Recent data suggest that inherited susceptibility to several autoimmune diseases was associated with specific DR4 associated DQ beta alleles. DNAs from 27 DR4 positive patients with GCA were digested with Taq I and Bam HI, analysed on 0.7% agarose gel and hybridized with DR beta, DQ alpha and DQ beta probes. DR beta hybridization produced no variant detectable within DR4. DQ beta probe confirmed two clusters among DR4 associated DQW3 alleles: DQW 3.1 (Bam HI 360 Kb) and DQw 3.2 (Taq I 1.9 Kb and Bam HI 11 Kb). Among our 27 DR4 positive patients, 34% were DQW 3.1 and 66% were DQW 3.2. These frequencies are the same as those observed in healthy controls.     

Diversity in HLA-DR4-related DR,DQ haplotypes in Australia, Oceania, and China

The relative distributions of 12 HLA-DR4-related DRB1 alleles in indigenous populations of Australia, Melanesia, Micronesia, Polynesia, and northern and southern China have been determined by analysis of oligonucleotide hybridization patterns of 406 examples of HLA-DR4. DRB1^*0405 and DRB1^*0410 were common DR4 alleles in Australian aborigines and in Melanesians, while DRB1^*0403 was the predominant DR4 allele in coastal Melanesians, Micronesians, and Polynesians; DRB1^*0406 was confined to Chinese. A novel DR4 allele, found in 30% of DR4-positive Australian aborigines but exclusive to one aboriginal population, was a combination of DRB1^*04 and 0803 nucleotide sequences and was carried on a haplotype with DR4-like DQ linkage arrangements. DQA1 and DQB1 typing generated 12 DR4-related haplotypes; the population distributions of these reflected the ancestral affinities of aborigines and Melanesians, the overlaping of coastal Melanesia with pre-Polynesian DR4 alleles and the colonization of Micronesia by an independent, non-Polynesian group. DR4-related autoimmune disorders such as rheumatoid arthritis (RA) and insulin-dependent diabetes mellitus (IDDM) are virtually unknown in indigenous populations of Australian and Oceania and this study confirmed that high-risk RA determinants, Dw4 and Dw14, occurred rarely. However, the DQw8 allele, thought particularly to predispose to IDDM, was present in the majority of DR4-positive Polynesians and Micronesians.     

Cells homozygous for two HLA-DR7-associated HLA-D specificities express highly similar DR molecules but different DQ molecules

The Ia molecules expressed by cells homozygous for two distinct HLA-DR7-associated HLA-D specificities, Dw7S and Dw11L, were compared. The complete Ia phenotypes of these cells are DR7, DRw53, DQw2, Dw7S, DPw4 and DR7, DRw53, DQw3, Dw11L, DPw4, respectively. Immunofluorescence analysis revealed that three DQ-specific monoclonal antibodies (Leu-10, 33.1, and HK-19), which detect polymorphic DQ determinants that do not correspond to known serologic specificities, are nonreactive with DR7, Dw7S cells but are reactive with DR7, Dw11L cells. The DR molecules isolated from Dw7S and Dw11L cells are very similar and comigrate when analyzed together by two-dimensional gel electrophoresis. In contrast, the DQ molecules isolated from these cells are structurally distinct: the DQ beta chains of DQw2-bearing molecules from Dw7S cells are very basic, while those of DQw3-bearing DQ molecules from Dw11L cells are more acidic. The finding that two DR7, D-different cells express indistinguishable DR molecules and structurally distinct DQ molecules documents a unique pattern of Ia molecular organization which is different from those previously described for the DR2-, DR4-, or DRw8-associated HLA-D specificities.     

Are HLA class II genes controlling susceptibility and resistance to Brazilian pemphigus foliaceus (fogo selvagem)?

In order to investigate the possible association between Brazilian pemphigus foliaceus and HLA, we studied 48 patients and 74 matched controls, all Brazilian Caucasoids, for HLA-A,B,C; DR1 to DRw8 and DQw1 to DQw3. The frequencies of DR1, DR4 and B16 were significantly increased, while DR7 was significantly decreased among the patients. Furthermore DQw2, likewise the DR specificities associated with it - DR3 and DR7 - never occurred among the patients in the absence of the susceptibility markers DR1, DQw1 or DR4, DQw3. Acting on these findings, we suggest that at least two MHC-class II genes are involved in the pathogenesis of Brazilian pemphigus foliaceus: at least one gene, associated to DR1,DQw1 and to DR4,DQw3, confers susceptibility and at least one gene, associated to DR7,DQw2 and DR3,DQw2, confers resistance. The susceptibility gene(s) seem(s) to be epistatic to or dominant over (if allelic) the resistance gene(s). Both are dominant over other alleles at their locus (or loci).     

The HLA-DR4-associated DQw8 allele is cod1.ned to HLA-DR3/DR4 heterozygous type I (insulin-dependent) diabetics

The HLA-DR4 specificity revealed a relative risk of 8.5 (chi 2 = 99.6; p less than 0.0001) when 193 type I diabetics were compared to 305 controls. Prevalence of the HLA-DR4-associated DQ types, i.e. DQw7 and DQw8, were determined, using a restriction fragment length polymorphism (RFLP) typing that combines the probe/enzyme combinations DQB/Taq I and DQB/Bam HI. The HLA-DQw8 specificity was confined to HLA-DR3/DR4 heterozygous patients when compared to controls (chi 2 = 4.9; p less than 0.025) or to all other DR4-heterozygous patients (chi 2 = 6.7; p less than 0.01). No association with HLA-DQw8 was seen in HLA-DR1/DR4 or HLA-DR"X"/DR4 (X not equal to 1,3,4) heterozygous patients. Due to the excess of HLA-DR3/DR4 patients the DQw8 allele is a risk factor in type I diabetics, but in HLA-DR1/DR4 and DRX/DR4 heterozygotes one might suggest that DQB1 and DRB combinations confer HLA-associated susceptibility.     

Association of tumor necrosis factor (TNF) and class II major histocompatibility complex alleles with the secretion of TNF-a and TNF-0 by human mononuclear cells: a possible link to insulin-dependent diabetes mellitus

We have investigated the correlation between different tumor necrosis factor (TNF) and class II major histocompatibility complex alleles in the lipopolysaccharide- or phytohemagglutinin-induced secretion of TNF-α and TNF-β by human monocytes and peripheral blood mononuclear cells in 87 unrelated Danish male individuals. Significant differences in TNF-α secretory capacity between TNF Ncol restriction fragment length polymorphisms, TNFa and TNFc micro-satellite alleles and DR alleles were identified. No correlation with TNF-β secretory capacity was found for any of the markers studied. TNF genotyping allowed us to define four extended HLA haplotypes which correlate with TNF-α secretory capacity. Two of these are DR4 positive: DQw8, DR4, TNFB*1, TNFa6, B44, A2 and DQw8, DR4, TNFB*2, TNFa2, B15, A2. Individuals carrying the TNFB*2, TNFa2 haplotype had a higher TNF-α secretory capacity than those carrying the TNFB*1, TNFa6 haplotype. In a group of DR3/DR4 heterozygous patients with insulin-dependent diabetes mellitus (IDDM), the frequency of the TNFa2 allele was higher than in HLA-DR matched controls, whereas theTNFa6 allele was more frequent in control individuals. In the DR3/DR4 heterozygous diabetic group 12/26 had the alleles combination DQw8, DR4 (Dw4), C4A3, TNFB*2, TNFa2, B15, whereas only 1/18 controls had this haplotype. This diabetogenic haplotype is identical to the DR4 haplotype which correlates with a higher TNF-α response. These observations suggest a direct role for the TNF locus in the pathogenesis of IDDM.     

B27-bearing HLA haplotypes in rheumatoid arthritis: characterization in Finnish patients

The frequency of human leukocyte antigen (HLA)-B27 has been found to be increased in rheumatoid arthritis (RA) in Finland and marginally also in some other populations. In the present study HLA-B27-bearing haplotypes in RA patients were found to carry DR1 and DR4 genes more often than do B27 haplotypes in control population. B27;DR4,DW14;DQw7 was a typical B27-DR4 haplotype whereas DR4 in the majority of other haplotypes occurs with Dw4 and DQw8 genes. The result indicates that the B27 association with RA is not independent of DR1 and DR4, but whether the B27;DR4;DQw7 haplotype subjects a person to a higher disease risk than do other DR4 haplotypes, or is associated with a more severe course of the disease, remains to be investigated.     

HLA DQα and DQβ restriction fragment length polymorphisms associated with Felty's syndrome and DR4-positive rheumatoid arthritis

HLA DQα and DQβ cDNA probes were used to study TaqI generated restriction fragment length polymorphisms (RFLPs) in DR4-positive patients with Fetly's syndrome (FS), seropositive rheumatoid arthritis (RA), and in HLA-DR4 positive controls. The results of this analysis revealed two DQβ RFLP patterns (DQβ3a and DQβ3b) associated with DR4, of which DQβ3b was found at significantly higher frequency in patients with FS (73%) or with RA (52%) than in DR4 controls (29%). Hind III generated RFLPs provide evidence that DQβ3b is in strong linkage disequilibrium with the gene encoding the serologically recognized epitope TA10. Results obtained using a DQα chain probe revealed polymorphic difference between DQα chain genes associated with different DR types, thereby providing a possible explanation for the lack of association between RA and other DR haplotypes in linkage disequilibrium with TA10. We conclude that both DQα and DQβ genes may be important in determining HLA-linked susceptibility to serve forms of RA.