HLA-Dw2 as a marker of resistance against juvenile diabetes mellitus

Genetic susceptibility to type 1 diabetes in childhood - estimation of HLA class II associated disease risk and class II effect in various phases of islet autoimmunity

OBJECTIVE

The HLA-DR/DQ region remains the major determinant of susceptibility to type 1 diabetes (T1D) despite the more than 50 risk affecting loci outside human leukocyte antigen (HLA) region that have been identified. We aimed at developing a simple risk estimation based on HLA class II genotyping, which was also tested by analyzing HLA class II effect on the autoantibody seroconversion and further progression to diabetes.

SUBJECTS AND METHODS

A total of 2991 trio families with a diabetic child from the Finnish Pediatric Diabetes Register were genotyped and the risk contributed by each DR-DQ haplotype calculated through transmission analysis. The genotype risk was estimated based on the summary effect of haplotypes. Genotype grouping was further tested in a subcohort of the Finnish Type 1 Diabetes Prediction and Prevention (DIPP) Study

RESULTS

The summary effect of haplotypes was generally seen in genotypes, while the expected synergistic effect of DR3-DQ2 and DR4-DQ8 (DRB1*04:03 excluded) combination was also clear in the T1D risk association analysis. This highest risk DR/DQ genotype was found in 21.6% of patients and 2.0% of controls, odds ratio (OR) = 13.2 (10.1-17.2), whereas the lowest risk genotype contained only 0.8% of patients and 28.0% of controls, OR = 0.02 (0.01-0.03). In the subcohort from the DIPP study the risk grades correlated clearly with seroconversion for islet autoantibodies and T1D development. In contrast, DR/DQ risk groups did not associate with the progression rate from advanced autoimmunity to clinical diabetes.

CONCLUSIONS

Class II HLA genotype groups improve the estimation of T1D risk. Class II effect is limited to the early phase of the disease process characterized by seroconversion for islet autoantibodies.

HLA-DR2-associated DRB1 and DRB5 alleles and haplotypes in Koreans

There are considerable racial differences in the distribution of HLA-DR2-associated DRB1 and DRB5 alleles and the characteristics of linkage disequilibrium between these alleles. In this study, the frequencies of DR2-associated DRB1 and DRB5 alleles and related haplotypes were analyzed in 186 DR2-positive individuals out of 800 normal Koreans registered for unrelated bone marrow donors. HLA class I antigen typing was performed by the serological method and DRB1 and DRB5 genotyping by the PCR-single strand conformational polymorphism method. Only 3 alleles were detected for DR2-associated DRB1 and DRB5 genes, respectively: DRB1(*)1501 (gene frequency 8.0%), (*)1502 (3.2%), (*)1602 (0.9%); DRB5(*)0101 (8.0%), (*)0102 (3.2%), and (*)0202 (0.9%). DRB1-DRB5 haplotype analysis showed an exclusive association between these alleles: DRB1*1501-DRB5*0101 (haplotype frequency 8.0%), DRB1(*)1502-DRB5(*)0102 (3.2%), and DRB1(*)1602-DRB5(*)0202 (0.9%). The 5 most common DR2-associated A-B-DRB1 haplotypes occurring at frequencies of > or = 0.5% were A24-B52-DRB1(*)1502 (1.8%), A2-B62-DRB1(*)1501, A2-B54-DRB1(*)1501, A26-B61-DRB1(*)1501, and A24-B51-DRB1(*)1501. The remarkable homogeneity in the haplotypic associations between DR2-associated DRB1 and DRB5 alleles in Koreans would be advantageous for organ transplantation compared with other ethnic groups showing considerable heterogeneity in the distribution of DRB1-DRB5 haplotypes.

Polymorphism and distribution of HLA-DR2 alleles and haplotypes in a Greek population

HLA-DR2 serological subtyping has indicated that the DR16 serotype appears at a higher frequency relative to the DR15 serotype in the Greek population, differing from the distribution observed in most other Caucasian groups. In this study, we have analyzed by the PCR-SSP technique a DR2-positive group of unrelated Greek individuals selected from our normal control panel for the different DRB1, DRB5, DQB1 and DQA1 DR2-associated alleles present. Six of the 50 individuals analyzed were homozygous for DR2, contributing a total of 56 haplotypes for DR2. The observed frequencies of the DR2-related DRB1 alleles were as follows: 58.9% for the DRB1*1601, 7.1% for the DRB1*1602, 25.0% for the DRB1*1501 and 7.1 % for the DRB1*1502 allele. The rare allele DRB1*1605 was detected in one heterozygous sample and its presence was definitively established by DNA sequencing. The alleles *1503, *1504, *1505, *1603 and *1604 were not detected. Three DRB5 alleles were identified: DRB5*0202 (67.8%), DRB5*0101 (25.0%) and DRB5*0102 (7.1%). Ten different DRB1/DQB1/ DQA1 DR2-associated haplotypes were defined. The most frequently observed haplotype was DRB1*1601-DQB1*0502-DQA1*0102 (relative frequency=57%) followed by DRB1*1501-DQB1*0602-DQA1*0102 (relative frequency=14.3%). In conclusion, the refined analysis of the DR2-associated DRB1 alleles in the Greek population revealed the prevalence of the DRB1*1601 allele. The rare allele DRB1*1605 was demonstrated once. A considerable variety of different DR2-related DR/DQ haplotypes was detected and the overall haplotypic frequencies in the Greek population are distributed differently compared to those reported for most other Caucasian populations.

The Dai minority population of southwest China: heterogeneity of DR2-associated HLA-DRB1,DRB5,DQA1, and DQB1 haplotypes

HLA-DR2 is the most common DR specificity (60.3%) identified in the Dai minority population of Xishuangbanna, Yunna Province, China. We characterized the DRB1, DRB5, DQA1, and DQB1 alleles of 44 unrelated DR2-positive individuals, 11 of whom (15%) were DR2 homozygous. Four DRB1 and four DRB5 alleles encoding DR2 were identified in this population. The most frequent DR2-associated DRB1 alleles were *1602 (gf = 0.164) and *1502 (gf = 0.151). DRB1*1501 (gf = 0.048) and a new allele designated DRB1*1504 (gf = 0.014) were also detected, but *1601 and *1503 were absent. The most frequent DR2-associated DRB5 alleles were *0101 (gf = 0.233) and *0102 (gf = 0.110). Nine different DR2-associated DR/DQ haplotypes were identified. The two most common DR2 haplotypes were DRB1*1602, DRB5*0101, DQA1*0102, DQB1*0502 (hf = 0.142) and DRB1*1502, DRB5*0102, DQA1*0101, DQB1*0501 (hf = 0.075). The new DRB1*1504 allele was found on a single haplotype: DRB1*1504, DRB5*0101, DQA1*0102, DQB1*0502 (hf = 0.017). The Dw2, Dw12, Dw21, and Dw22 haplotypes, present in many other Asian and Mongoloid populations, were not identified in this unique group. However, the Dai minority population is characterized by a relatively large number of diverse DR2 haplotypes and a new DRB1 allele encoding DR2.

Self and non-self antigen in diabetic autoimmunity: molecules and mechanisms

In this article, we have summarized current facts, models and views of the autoimmunity that leads to destruction of insulin-producing beta-cells and consequent Type 1 (insulin-dependent) diabetes mellitus. The presence of strong susceptibility and resistance gene loci distinguishes this condition from other autoimmune disorders, but environmental disease factors must conspire to produce disease. The mapping of most of the genetic risk (or disease resistance) to specific alleles in the major histocompatibility locus (MHC class II) has direct functional implications for our understanding of autoimmunity in diabetes and directly implies that presentation of a likely narrow set of peptides is critical to the development of diabetic autoimmunity. While many core scientific questions remain to be answered, current insight into the disease process is beginning to have direct clinical impact with concerted efforts towards disease prevention or intervention by immunological means. In this process, identification of the critical antigenic epitopes recognized by diabetes-associated T cells has achieved highest priority.

The role of genetic predisposition to type I (insulin-dependent) diabetes mellitus

The aetiology of insulin-dependent diabetes (IDDM) involves genetic predisposition, a major component of which has been mapped in the HLA complex, near to or identical with genes encoding class II molecules. In Caucasian populations IDDM is strongly associated with the serologically defined HLA-DR3 and DR4 antigens, which are widely recognised as markers of susceptibility. The particularly high risk of DR3/DR4 heterozygotes suggests that susceptibility is determined by two genes acting synergistically. The development of recombinant DNA technology has allowed a finer description of the class II region and provided evidence that DQ rather than DR determinants may primarily influence IDDM susceptibility. The search for specific structural changes of the DQA and DQB genes has shown that susceptibility correlates with the absence of aspartic acid at position 57 on the DQ beta chain (DQ beta 57 Asp--) and/or the presence of arginine at position 52 on the DQ alpha chain (DQ alpha 52 Arg+). In Caucasians the formation of a putative DQ susceptibility molecule (DQ alpha 52 Arg+, DQ beta 57 Asp-) accounts best for the disease associations when transcomplementation molecules consisting of DQ alpha and beta chains encoded by different haplotypes are postulated to explain the excess risk of heterozygotes. The HLA-IDDM associations in the Japanese, however, are not explained by this model. These and other unresolved questions indicate that other residues of the DQ alpha and beta chains or other class II molecules (DR beta chains), as well as non-MHC genes, may also contribute to the susceptibility.

DNA typing for class II HLA antigens with allele-specific or group-specific amplification. IV. Typing for alleles of the HLA-DR2 group

In this study we present an approach for the definition of the alleles belonging to the HLA-DR2 group by DNA typing with oligonucleotide probes. Following methodology similar to that we used previously for the definition of other HLA-DR subsets, we have now developed primers for DR2-DRB1 and DR2-DRB5 amplification, and probes for the identification of sequences that distinguish the subtypes of this group of genes. The method used defines all the previously described alleles at both DR2-associated DRB loci. In addition, we have identified a variant of DRB1-DR2-Dw2. This new allele has been called DRB1*15.3. It is different from DRB1*1501 in codon 30, where it carries histidine instead of tyrosine. Eight different haplotype combinations of DRB5, DRB1, and DQB1 were identified within the DR2 group and their occurrence in four normal panels of different ethnic origin has been described. Haplotypes containing DRB1*15.3 occurred most frequently in black panel members in whom it was associated with either DQB1*0602 or DQB1*0501. Two unusual haplotypes were observed: one containing elements of DR2-Dw21 (DQB1*0502) and of DR2-Dw22 (DRB1*1602) and one containing elements of Dw21 (DRB1*1601, DQB1*0502) and Dw2 (DRB5*0101). The methods described permit simple and rapid determination of the alleles of the HLA-DR2 group and should be useful for population studies and for investigation of DR2-associated diseases.

HLA antigens and complotypes in insulin-dependent diabetes mellitus

One hundred and thirty-six Finnish patients with insulin-dependent (type I) diabetes mellitus were investigated for the HLA-A, B, D and DR antigens as well as the Bf and C4 allotypes. The statistically significant increase in the frequencies of HLA-A9, B8, B15, Dw3, Dw4, DR3, DR4, C4A0 and C4B3 was observed when compared with the healthy controls. About 79% of the patients had HLA-DR4, and 53% had HLA-DR3 antigens. A rare C4 allele C4B3 was found in 21% of the patients, whereas only in 2% among the controls (relative risk 16.35). The etiological fraction (EF) values indicated that HLA D/DR alleles were the best markers for IDDM, the observed EF for HLA-DR4 in diabetes was as high as 0.70. Examination of HLA, Bf and C4 phenotypes suggested that at least two supratypes "B15 BfS C4A3B3 D(R)4" and "B8 BfS C4A0B1 D(R)3" were markers for the susceptibility to type I diabetes, one third of our patients had either of these supratypes. The protective role of DR2 and Dw2 antigens was also confirmed: no HLA-Dw2 positive patients and only one with HLA-DR2 was found.

An analysis of epidemiological data in HLA-typed diabetic children

To clarify the heterogeneity of Type 1 (insulin-dependent) diabetes mellitus, differences between patients with different HLA risk antigens were investigated with regard to sex, age at diagnosis, season of year and calendar year at diagnosis of the disease. The study consisted of 293 HLA-typed patients from the Department of Paediatrics, University of Oulu, Oulu, Finland. HLA-Dw2 was extremely rare among diabetic patients, whereas Dw3 and Dw4 were associated with increased risk in this as in other series. Male patients more often had the HLA-A1 antigen than females. On comparison of the Dw3 positive patients, boys more frequently had the combination A1,B8 than girls. A1,B8-positive patients were more often diagnosed during the warm months, in the late summer and autumn. Patients with both Dw3 and Dw4 were younger at diagnosis when compared with the rest of the patients. The results support the concept of heterogeneity in the pathogenesis of Type 1 diabetes associated with HLA-linked genetic determinants.

Segregation of HLA-DR2 among affected and non-affected offspring of 66 families with type 1 (insulin-dependent) diabetes

Segregation of HLA-DR2 among affected and unaffected offspring was studied in 66 HLA-genotypes families with Type 1 diabetes in whom at least one parent carried DR2. The frequency of DR2-positive parents (21%) was not different from that of control families (29%). Among the diabetic probands, the gene frequency of DR2 was significantly decreased compared with control subjects (0.05 versus 0.17, p less than 0.001) as were DR5 (0.07 versus 0.17, p less than 0.01) and DR7 (0.06 versus 0.13, p less than 0.003). Twenty probands carried DR2, in 11 or whom (55%) it was found in combination with either DR3 or DR4. The nine cases who carried another DR allele included one who was DR2 homozygous. Transmission of DR2 was reduced in affected offspring, and random in unaffected siblings, compared with the expected ratio. However, when the DR2 transmission was analysed separately for parents bearing DR2 with DR3, DR4 or another DR allele, it appeared that DR2 transmission to affected offspring was random when the parents carried neither DR3 or DR4, the transmission deficit being due to over-transmission of DR3 and DR4. The haplotype analysis showed that the haplotype A3, Cw7, B7, GfS, DR2, found in 19% of "non-diabetic" DR2 haplotypes was practically absent among "diabetic" DR2-haplotypes (4%). In conclusion, population and segregation analysis could not demonstrate a specific protective effect of DR2.

HLA DR antigens and disease expression in rheumatoid arthritis

Ninety-four patients with rheumatoid arthritis who possessed one or more of the HLA DR alloantigens 2, 3, or 4 were studied to investigate the genetic influence on disease severity and prognosis. In those with a disease duration of less than 10 years radiological damage was less in patients with DR2 than in those without this antigen. When current joint scores were compared, patients with this antigen had less evidence of disease than patients with DR3 or 4, DR3 patients having the highest scores. The presence of nodules and Sjögren's syndrome were less common in the DR2 patients. Variability in response to disease modifying drugs according to the patient's HLA DR antigen status may explain these differences. It is concluded, however, that possession of HLA DR2 may be an indicator of good prognosis in patients with rheumatoid arthritis.

HLA-DR2 in two sibships with insulin-dependent diabetes mellitus

We report two families selected from 124 genotyped Caucasian insulin-dependent diabetes mellitus (IDDM) families because of unusual features. In both families, all offspring are affected and four out of six bear the allele HLA-DR2 which is an uncommon phenotype among diabetic patients. Onset before the age of 1 year in all the patients of one family, association with optic atrophy in the other, and the existence of pairs of affected sibs of different HLA types in both, are infrequent findings and support the evidence of heterogeneity in IDDM.

HLA-DR specificities among Japanese with several autoimmune diseases

HLA-A, -B, -C, and -DR antigens were determined in order to study the association of HLA in Japanese patients with several autoimmune diseases, hypertrophic cardiomyopathy, and Hodgkin's disease. The frequency of HLA-DR4 was significantly increased in the patients with rheumatoid arthritis, juvenile-onset insulin-dependent diabetes mellitus (IDDM) and hypertrophic obstructive cardiomyopathy. In this study, no significant associations with A, B, or C specificities were observed except BW22 in IDDM. In contrast, the negative association with HLA-DR2 was observed in Hashimoto's thyroiditis, pemphigus vulgaris and hypertrophic non-obstructive cardiomyopathy.

Genetic heterogeneity of diabetes and HLA

Histocompatibility (HLA) antigens and genotypes B, D and DR were studied in large sample of Caucasian insulin dependent diabetic (IDD) probands. The associations between IDD and B8, B15, Dw3, Dw4, and DR3, and DR4 were measured by relative risks (RR) and delta values. Both the homozygotes (B8/8: RR 10, B 15/15: RR 7, DR3/3: RR32, DR4/4: RR34) and the heterozygotes (B8/15: RR 11, DR3/4: RR22) for the high-risk antigens showed highly significant elevation of the relative risks, yet there were no statistically significant differences between the homo- and the heterozygotes. The delta measurements supported the RR results. RR and delta were found significantly decreased for B7, Dw2, and DR2. There were no relationships observed between age at diagnosis or family history and HLA. Although we were unable to demonstrate a statistically significant difference between the RR for the high-risk antigens heterozygote vs. the high-risk antigen homozygotes, our study like many others shows that the RR is higher for the heterozygotes. Thus our data are compatible with genetic heterogeneity of IDD.

HLA types in American black juvenile diabetics strong associations with Dw3 and Dw4

The distributions of HLA-A, -B, -C and -D antigens in 38 black American insulin-dependent juvenile diabetics were studied. Antigens A1, A2, B8 and Cw3 were slightly increased, but the corrected probability values were not statistically significant. As determined by mixed lymphocyte culture, the frequency of Dw3 was 89% in the juvenile diabetics and that of Dw4 was 42% in comparison with 14 and 8%, respectively, in the controls. The relative risks for juvenile diabetes were 52 for Dw3 (p = 10(-8) and 9 for Dw4 (p = 10(-6). Dw2 was significantly decreased in the diabetics (p equals 0.008). All of these deviations in A, B, C and D locus specificities have been previously reported by others in white juvenile diabetics. Because there are white genes in the American black gene pool and juvenile diabetes is rare in blacks in western Africa, many cases of juvenile diabetes in American blacks could be the result of genes ultimately derived from the white genes. This hypothesis is supported by the similar HLA associations in juvenile diabetes in the black and white ethnic groups.

HLA antigens and susceptibility to juvenile diabetes: do additive relative risks imply genetic heterogeneity?

The relative risk of B8/B15 heterozygotes for juvenile-onset diabetes is higher than the risk for people having B8 or B15 alone. This has been cited as evidence for genetic heterogeneity in juvenile diabetes. However, the observed relative risks are compatible with a single susceptibility allele. If disease susceptibility is recessive, for example, then an individual with two disease associated antigens is more likely to be susceptible than an individual with only one associated antigen. The relative risk for an HLA heterozygote should be intermediate between that of the respective homozygotes, so that an interaction effect of two alleles can only be supported if the heterozygote risk is significantly greater than both homozygote risks. The estimated relative risks for B8 and B15 homozygotes, based on data from four different populations, is approximately equal to the risk for B8/B15 heterozygotes. Moreover, disease manifestations which are differentially associated with B8 and B15, such as antibody production to exogenous insulin, may be due to linkage disequilibrium between HLA and other loci which are not directly related to susceptibility of juvenile diabetes. Therefore, while the susceptibility to juvenile diabetes may have several genetic forms, there is no support for distinct B8-associated and B15-associated forms of susceptibility.

On the significance of C2, C4, and factor B polymorphisms in disease

In this review article, recent evidence is presented that some diseases like insulin-dependent diabetes mellitus, multiple sclerosis, and idiopathic membranous nephropathy, which are primarily associated with HLA-D,DR, are also related to the rare C2, C4, and Factor B alleles. Circumstantial evidence is available that at least some of these rare variants may be functionally deficient. Based on the concept of functionally interacting gene clusters, mutant complement genes may lead to impaired effector mechanisms in virus neutralization or lysis of virus-infected cells. Other mechanisms such as alteration of vascular permeability may be involved in the development of proliferative retinopathy and familial hypertension. In lepromatous lepra, an impaired cell-mediated lysis of M. leprae may be related to the hemolytically inactive C4F1 allelic product.

HLA-D antigen frequencies in Sjögren's syndrome. Differences between the primary and secondary form

HLA-A, B, C and D typing was performed in 19 patients with primary Sjögren's syndrome (primary SS) and in 15 patients with rheumatoid arthritis (RA) and secondary Sjögren's syndrome (RA-SS). In the primary SS group, the frequency of HLA-Dw2 was increased (p less than 0.01; "corrected" p greater than 0.05) while the frequency of Dw3 was non-significantly increased. In the RA-SS patients, the frequency of HLA-Dw4 was increased to 84.6% (relative risk = 22.8; p less than 0.001) and the frequency of Dw2 was non-significantly decreased. An increased frequency of Dw2 in primary SS has not been reported before. To investigate if this observation was due to chance, a new series of 16 patients with primary SS were HLA typed. In this new group, the frequencies of both Dw2 and Dw3 were significantly increased, while the frequency of Dw4 was significantly decreased. In the whole group of primary SS patients, the frequencies of Dw2 were 56.3% (relative risk = 3.7; p less than 0.001). Dw3:50.0% (relative risk = 2.8; p less than 0.01) and Dw4: 6.3% (relative risk = 0.28; p less than 0.05). We conclude that genetic factors associated with the HLA-system are involved in the development of Sjögren's syndrome and that these genetic factors are different in primary and secondary Sjögren's syndrome. In primary Sjögren's syndrome, the association with both Dw2 and Dw3 might suggest a further heterogeneity of the syndrome.

Association of B cell alloantigen with juvenile onset diabetes mellitus in the Japanese

Sixty-four Japanese insulin dependent juvenile onset diabetes mellitus (JOD) were studied in relation to HLA-A, B, and DR. Significant deviations were observed. HLA-Bw54 was increased (PF = 49.2%, RR = 6.4) and HLA-B5 was decreased (PF = 7.9%, RR = 0.19). Using radioimmunoassay, two HLA-DR antigens were investigated. Hon 7 antigen, so-called MT3 (WIA4x7), which has linkage disequilibrium between HLA-BW54, is highly associated (PF = 96.9%, RR = 27.8) with JOD found in the Japanese.

HLA genetics of insulin-dependent juvenile-onset diabetes mellitus in northern Finland

HLA--A, B and C antigens were determined from 63 cases of juvenile-onset, insulin-dependent diabetes mellitus (IDDM) in northern Finland. There was a very strong association between IDDM and HLA--B8 and B15 antigens in this area where the incidence of IDDM is also high. Relative risks for B8 and B15 were 4.8 and 3.8, respectively. In a small group of subjects typed for HLA--D antigens Dw3 and Dw4 appeared to be stronger risk factors than associated B antigens, whereas Dw2 was almost totally absent in the patients. The effect of combinations of B8 and B15 antigens in unrelated patients and in diabetic families was analyzed. B8, B15 was found in 11 out of 63 unrelated diabetics; however, this did not differ significantly from the expected value. In diabetic families there was no increased rate of intra-HLA recombinations.

HLA patterns in Australian patients with insulin dependent diabetes mellitus (IDDM)

HLA antigens were determined in 169 Australian patients with insulin dependent diabetes mellitus (IDDM). HLA-B8 (47.9% v. 23.8%) and B15 (18.9% v. 8.2%) were significantly more frequent in the IDDM patients than in 1460 controls. The relative risks for developing IDDM in people carrying these antigens were 2.94 and 2.59 respectively. Carriers of the B8/B15 genotype had a relative risk of 5.48. These HLA associations in Australian IDDM patients are similar to those reported in other predominantly Caucasian populations.

HLA-Dw antigens in unrelated juvenile, insulin-dependent diabetics

Forty-one unrelated juvenile, insulin-dependent diabetics have been HLA tissue typed for A, B and Dw anitgens and compared with a normal control population. We have found statistically significant increases in the frequencies of B8, B18, and Dw3, and significant decrements in the frequencies of B7, B12 and Dw2. The log-linear modeling technique was used to study the association of JIDD with Dw3 and B8 antigens. We confirmed that the B8 excess seen in diabetics is secondary to the excess of Dw3. The decrements of B7, B12 and Dw2 could reflect an association of these antigens with a protective factor for the disease, or could be due to an artifact. The latter possibility was excluded for B7 and Dw2 by adjusting for the excess antigen frequencies. These findings suggest that the associations between the HLA and diabetes are compatible with the existence of genes which are concerned with the pathogenesis of the disease and are closely associated with the D locus of the major histocompatibility system.

HLA-DR specificities among black Americans with juvenile-onset diabetes

To study the association of histocompatibility (HLA) genes in black persons with juvenile-onset diabetes, we determined HLA-A, HLA-B, HLA-C and HLA-DR specificities in 40 black Americans with this disease and in 67 unaffected black Americans. Marked increases in the frequencies of HLA-DRw3 and HLA-DRw4 were found in the patients as compared with the unaffected persons: DRw3 was found in 72.5 per cent of patients versus 29.9 per cent of unaffected persons and DRw4 in 72.5 per cent versus 25.4 per cent (corrected P values each less than 0.0007). DRw2 was not found in any of the patients but was present in 26.9 per cent of unaffected persons (P corrected less than 0.035). There is thus a negative correlation between this specificity and juvenile-onset diabetes. By contrast, no meaningful differences were found in the frequencies of A, B, or C locus antigens. Studies in white persons with juvenile-onset diabetes have suggested that the reported HLA-B associations are due to HLA-D region specificities, and our results also support the premise that D region specificities are the primary associations with juvenile-onset diabetes.

HLA--Dw4 and rheumatoid arthritis

Forty-seven patients with a "definite" or "classical" rheumatoid arthritis according to the ARA criteria were typed for the serologically detectable HLA--A, --B, and --C antigens and 36 of these patients were typed for the HLA--D antigens, Dw1, 2, 3, 4, 6, 7, and 8 by the MLC technique. The frequency of Dw4 was increased to 44.4% in the patients compared to 17.2% in normal controls (P = 8 X 10(-4)). The frequency of Dw1 and Dw7 was also increased although this was only of borderline significance. The frequency of Dw2 was remarkably low, especially in females, which is of interest, as the same antigen has a low frequency in some other autoimmune diseases. No significant deviations of the frequencies of HLA--A, --B, and --C antigens were found in rheumatoid arthritis patients.