HLA-A, B, C, DR antigens, Bf, C4 and glyoxalase I (GLO) polymorphisms in French Basques with insulin-dependent diabetes mellitus (IDDM)

The Basques were previously shown to present a high frequency of HLA—B18 and Bf Fl. which are known to be associated with insulin dependent diabetes mellitus (IDDM). During the VIII International Histocompatibility Workshop, we studied HLA-A, B, C, DR; Bf, C4 and GLO.I polymorphisms in 51 unrelated French Basque IDDM patients and in 50 controls. Haplotypes were established by family studies in all controls and some patients. Two haplotypes were frequently found in the controls: HLA- A1, Bw57, Bf S, C4 FIS, DR7 and HLA- Aw30, Cw5, B18, Bf Fl, C4 Fs°, DR3. The first one was not found in the patients. All the components of the second haplotype had increased frequencies possibly as a consequence of linkage disequilibrium with HLA— DR3 : a highly significant association between IDDM and HLA-DR3 was observed (90.2% vs 24.0%, relative risk (RR) = 29.1. P < 10⁻¹¹). The HLA-DR4 frequency was slightly increased (37.3% vs 16.0%). and HLA—DR2 was not found. The silent allele C4 s ° was particularly associated with early diagnosed IDDM (86.7% in patients with age at onset under 20 years vs 57.1% in other patients, P < 0.02). The high relative risk for HLA-DR3/DR4 heterozygous vs that of individuals, possibly HLA-DR3 homozygous, supported the hypothesis that two HLA-DR linked genetic factors could be involved in the inheritance of IDDM susceptibility.     

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.     

DNA polymorphism of major histocompatibility complex class II and class III genes in systemic lupus erythematosus

We investigated the Taq I digested DNA restriction fragment length polymorphism (RFLP) of the Major Histocompatibility Complex (MHC) class II genes: HLA-DRB, -DQA, and the class III genes: C4 and 21-hydroxylase(CYP21) in 56 caucasoid patients with systemic lupus erythematosus (SLE) and 62 control subjects in order to define the molecular variation of these genes and their association with SLE. The results showed that the gene frequencies of both HLA-DR2 and -DR3 were significantly increased in the SLE population compared to normal subjects (DR2: 21.4% vs 10.7% chi 2 = 4.5. DR3: 29.6% vs 13.3%; chi 2 = 8.3). A high frequency of C4A and CYP21A gene deletions was also found in SLE patients (SLE 52%, normals 24%). All of 22 SLE patients, and 12 of 15 normal subjects who had C4A and CYP21A gene deletions had a 10.0kb Taq 1 DRB RFLP attributable to the presence of HLA-DR3. Family studies showed linkage of C4A/CYP21A deletions with HLA-B8 and -DR3, and confirmed the previously demonstrated association of the HLA-B8, DR3, C4A*Q0, C4*B1, Bf*S, C2*C haplotype with SLE. Deletions affecting the C4A and CYP21A genes were the commonest cause of C4A null alleles in SLE. No strong association between C4 null phenotype or C4 gene deletion, as determined by RFLP, was observed in patients who possessed DR2.     

Rheumatoid arthritis and its association with HLA-DR antigens. II. Antibodies to native connective tissue antigens detected by enzyme linked immunosorbent assay

The distribution of frequencies of HLA-DR alloantigens in HLA-DR4 negative subjects was determined in patients with Rheumatoid arthritis (RA) and normal individuals. An increased incidence of HLA-DR1 alloantigen in DR4 negative RA patients (45.9%) compared with DR4 negative healthy controls (23.6%) was found. The difference became significant when the incidence of DR1 was compared between patients with severe disease stages (III-IV) (75%) in contrast to 32% of incidence in patients of the milder stages (I-II) (p less than 0.05). Using Enzyme Linked Immunosorbent Assay we have determined the incidence of serum antibodies to native bovine type I and type II collagens and proteoglycans in patients with RA. Presence of serum antibodies to native type I collagen was detected in 59% of patients with RA, 60% of sera exhibited reactivity to type II collagen and 12% had antibodies to proteoglycans. There was no correlation between the presence of antibodies to type I and II collagens and disease stages, however, the incidence of serum antibodies to proteoglycans was increased in severe disease stages. On the other hand, the presence of high levels of antibodies to type I collagen was associated to HLA-DR1 antigen, (p less than 0.05).     

MHC classes I, II, III antigens study in 70 insulin-dependent diabetics with associated auto-immune diseases

Seventy IDDM patients (insulin-dependent diabetics), 48 females and 22 males, most of them adults at the onset of diabetes, and suffering from at least one other associated autoimmune manifestation (AAM) were studied for HLA A,B,C, DR markers and Bf, C4 complement components. Comparisons were made with 108 normal controls and a series of 287 IDDM patients with juvenile onset (under 25 years) and no patent other autoimmune disease. The increase in frequency of HLA-B8 among IDDM patients with AAM was confirmed (36% versus 20% in controls) (p less than 0.04). The frequency of DR4 among diabetics with AAM (33%) was not significantly different from the normal frequency (27%), and the allelic combination DR3/4 was found in only 13% of IDDM with AAM. Corresponding frequencies in patients with IDDM alone were 66% for DR4 and 34% for DR3/4 (p less than 10(-6) and 10(-3) respectively). These results confirm the heterogeneity of IDDM and support, by genetic arguments, the concept of overlapping entities. The hypothesis of a common background of autoimmunity associated with B8 DR3 can be postulated, while the organ specific target process should be associated with various DR alleles.     

Major-histocompatibility-complex extended haplotypes in membranoproliferative glomerulonephritis

Membranoproliferative glomerulonephritis is often associated with evidence of immune derangement, especially hypocomplementemia. We studied genetic markers for membranoproliferative glomerulonephritis within the major histocompatibility complex in 34 patients and their families and in 29 normal families. We examined the frequencies of extended haplotypes (combinations of alleles that tend to occur together) in patients and controls. The extended haplotype HLA-B8,DR3,SC01,GLO2(glyoxalase I 2) was observed in 9 of 68 disease-associated haplotypes (13 percent), but in only 3 of 205 controls (1 percent) (relative risk, 14.79; P less than 0.001). An extended haplotype similar except for a different glyoxalase allotype (B8,DR3,SC01,GLO1) did not occur with increased frequency, nor did any other extended haplotypes. Patients with the extended haplotype B8,DR3,SC01,GLO2 had a higher incidence of renal insufficiency than those without it (P less than 0.01). The data support the hypothesis that a specific extended haplotype of the major histocompatibility complex is associated with susceptibility to membranoproliferative glomerulonephritis, and that patients with glomerulonephritis who have this extended haplotype have a poorer prognosis for kidney survival than those without the haplotype.     

Analysis of HLA-DP in HLA-DR/GLO recombinant families and in the population of south-western France

The existing estimates of the recombination fraction between DR and DP are quite variable and often based on anecdotal observations. We have estimated the DR/DP crossover frequency on the basis of families typed for HLA markers and GLO. The frequency of DR/GLO crossing over was 8.7% (23/264 informative meioses), maternal recombinations being about twice as frequent as paternal ones. Of 17 DR/GLO recombinant families typed for DPw1-6, DP was informative in 11 (13 recombinations) but only one of these gave rise to a DR/DP crossover. According to these data the DR/DP recombination fraction is below 1%, in contrast to some earlier published materials. HLA-DR/DP haplotypic associations on 127 informative Caucasoid haplotypes have been evaluated. In agreement with previous studies, DR3 was positively associated with DPw1 and, in addition, DR7 was found to be positively associated with DP-blank (not DPw1-6). The rare DPw6 allele is possibly associated with the DR4, Dw14 allele. The DR-DP haplotype profiles suggest other associations which might become significant if larger materials are tested. The frequency of DP alleles in a random material (N = 201) was found to be in accordance with most of the previously published frequences on European Caucasoids with DPw4 as the predominating frequency (gene frequency 40%) and a blank frequency of 27%.     

HLA class II antigens associated with lupus nephritis in Italian SLE patients

Human leukocyte antigen DR2 (HLA-DR2), namely the allelic variant HLA-DR15, have been associated with lupus nephritis (LN) in Caucasians. The study investigated the relationships between HLA class II alleles and lupus nephritis in Italian patients. Two hundred forty-four patients fulfilling the American Rheumatism Association criteria for systemic lupus erythematosus (SLE) were typed for HLA-DRB1*, -DQA1*, -DQB1*, and -DPB1* alleles by polymerase chain reaction-sequence-specific oligonucleotide and polymears chain reaction-single-strand polymorphism; 71 patients had renal damage assessed by renal biopsy. Glomerulonephritis was classified using WHO criteria. Significance was tested by X(2) on 2x2 tables. HLA-DQA1*0101 was strongly associated with LN (OR = 2.72 [1.43-5.19]; p = 0.002), whereas HLA-DRB1*1501 was only marginally associated (OR = 1.94 [0.88-4.26]; p = not significant). HLA-DQA1*0102 demonstrated a significant protective effect (OR = 0.31 [0.14-0.86]; p = 0.002). On analyzing the distribution of HLA-DRB1*1501 bearing haplotypes in our SLE patients we found that the HLA-DRB1*1501 greatly enhanced the risk of developing LN conferred by the DQA1*0101 allele (OR = 65.96 [9.35-1326.25]), whereas DQA1*0102 suppressed the nephritogenic effect of DRB1*1501. At renal biopsy, 80% of DRB1*15 positive patients were classified as having class IV LN with the remaining 20% having class III LN. The figures were 19% and 21%, respectively, among the HLA-DR15 negative patients. In the Italian population HLA-DQA and HLA-DR alleles interact in conferring susceptibility to or protection against lupus nephritis, the diffuse proliferative glomerulonephritis (i.e., the most severe form of nephritis) is associated with the HLA-DR15 bearing haplotypes.     

Immunogenetics of rheumatoid arthritis in Israel

In an attempt to study the variation of associations between HLA and rheumatoid disease a population of 44 Ashkenazi and 29 non-Ashkenazi patients with Rheumatoid Arthritis were tested for HLA-A, B, C and DR antigens and compared with the relevant control groups. In contrast to the results obtained in Middle European or North American Caucasians, Rheumatoid Arthritis in Israel is not associated with B15 and Cw3, indicating that it is very unlikely that B- and C-locus antigens are involved in coding for disease susceptibility for RA. The allele DR4 which is found associated with RA in almost all populations tested so far was in the total patient group (47.9%) slightly but not significantly more frequent than in the control group (38.3%). This difference was entirely due to a nonsignificant increase in the frequency of DR4 in the Ashkenazi patients (54.5%) compared to controls (40%), while the frequency of DR4 in non-Ashkenazi patients and controls was virtually identical (38.0% vs 36.7%). Another surprising finding was that the frequency of HLA-DR1, which has been reported to be increased in different populations of patients with RA was found to be completely normal in the present study on Israeli patients. The alleles of the Bf and the GLO system did not show any significant difference between patients and controls.(ABSTRACT TRUNCATED AT 250 WORDS)     

胃癌组织HLA－I类和DR分子免疫组化研究

本文应用免疫组化法对６４例胃癌、癌旁组织和６例胃溃疡大致正常胃粘膜冰冻和石蜡切片进行了染色．结果表明，正常胃粘膜和癌旁胃粘膜上皮细胞ＨＬＡ－Ｉ类分子表达阳性，其着色较均一，ＨＬＡ－ＤＲ染色均阴性．胃癌细胞Ｉ类分子表达缺失（２７／６４例），与癌旁上皮比较差异显著（Ｐ＜０．０１）。粘液细胞癌和低分化癌Ｉ类分子缺失率显著高于高分化癌（Ｐ＜０．０２５）．此外，发生肿瘤转移的病例Ｉ类分子缺失率（１２／１５例）显著高于无转移组（１／５例，Ｐ＜０．０２５）．ＤＲ分子在癌组织表达阳性，其阳性率高达５３．１％（３４／６４例）．低分化癌ＤＲ分子阳性率亦显著高于高分化癌和中分化癌，未分化癌ＤＲ分子阳性率亦显著高于高分化癌（Ｐ＜０．０１～０．０５）．提示（１）ＨＬＡ－Ｉ类分子表达缺失可能与癌细胞逃避宿主免疫监视发生润浸生长和转移有关；（２）分化程度不同的癌组织ＨＬＡ－Ｉ类分子表达差异显著，提示癌细胞分化可能影响Ｉ、Ⅱ类分子表达和肿癌抗原呈递；（３）ＨＬＡ－Ｉ类和ＤＲ分子表达异常可能是上皮恶性转变的标志之一．     

HLA Antigens Associated with Properdin Factor B Allotype BfF1

Linkage disequilibrium has been described between HLA-B antigens and allotypes of properdin factor B (Bf). The association between HLA-B and DR antigens and the rare allotype BfF1 was investigated. In 27 HLA-A, B and DR typed individuals only one person was found who was DR3 negative but five who were B18 negative. Family studies showed that in nine cases B18, BfF1 and DR3 segregated as a haplotype; in another family which was B18 negative, BfF1 and DR3 segregated together. It is infered that the Bf locus is more strongly associated with HLA-DR than with HLA-B.     

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.     

Reproducibility of HLA-A, B, and DR typing using peripheral blood samples: results of retyping in the collaborative corneal transplantation studies. Collaborative Corneal Transplantation Studies Group (corrected)

Recipients enrolled in the Collaborative Corneal Transplantation Studies were retyped as part of a quality assurance program. Because the observed percentage of HLA-DR homozygosity on original typing was more than twice as high as expected from CCTS allele frequencies, the sample selected for retyping was heavily weighted with patients whose original typing identified fewer than two DR antigens. Retyping was performed in a different laboratory from the laboratory performing the original typing. For the 129 patients who were retyped, agreement between the original and retyping laboratories was 88% for HLA-A, 79% for HLA-B, and 55% for HLA-DR. When criteria was relaxed to consider only discrepancies involving readily identifiable antigens, the agreement improved to 95% for HLA-A, 91% for HLA-B, and 59% for HLA-DR. Identification of a second HLA-DR antigen on retyping when only one DR antigen had been identified on original typing was by far the most common form of disagreement. There were no significant differences in the amount of disagreement among the laboratories. Of special interest is that 50% of the discrepancies involved DR3, DR5, and/or DR6, which have structural similarities. Based on the results of the project, we recommend: (1) replicate testing for all DR typing; and (2) retyping using a second source of antiserum for all subjects having DR blanks.     

Familial studies of systemic lupus erythematosus. HLA markers and complotypes

Particular susceptibility to systemic lupus erythematosis (SLE) could be due to a certain alleles of class I, II or III of the major histocompatibility complex (MHC). The existence of total hereditary deficiencies of factor 2 or 4 of the complement in this syndrome suggests the presence of silent alleles which could conceivably play a determining role in the appearance of SLE. In this study, the HLA haplotypes and complotypes (C2, C4, Bf) were determined in 20 individuals suffering from SLE, and compared with 108 healthy, genotyped individuals. The results obtained showed a significant increase in the frequency of C4 BQ0 in patients compared with that found in controls (chi 2 = 12.27, p less than 0.001, Relative Risk = 3.78), and confirm the HLA association, DR3/SLE (chi 2 = 5.45, p less than 0.02, RR = 2.53).     

HLA and Bf in idiopathic nephrotic syndrome in children: differences between corticosensitive and corticoresistant forms

An association between HLA-DR7 and the steroid sensitive idiopathic nephrotic syndrome in the children has already been reported. Immunogenetic data in the less frequent steroid resistant form of this disease have never been published. In this study, we analyse HLA-A, B and DR typing in 99 cases of nephrotic children divided in 72 with the steroid sensitive (SS) form and 27 with the steroid resistant (SR) syndrome, in comparison with those of 207 healthy controls; Bf allotypes were determined in 53 of the patients. The results show the increased frequency of DR7 in the SS syndrome (75% vs 30%, RR = 6.9, pc less than 10(-6), while the SR one is more associated to DR3 (52% vs 27%, RR = 3, p less than 0.004). In the SS patients, atopy is associated to DR7 (p less than 0.001), which is not the case in the SR group. Furthermore, a high relative risk is associated to the phenotype DR3/DR7 (30% vs 4%; RR = 9.3; pc less than 0.0004), for the SR disease; besides, this phenotype is associated to an early onset of the disease and to lesions of focal sclerosis. Thus a heterozygous effect in the SR form of idiopathic nephrotic syndrome of children has been demonstrated; the steroid sensitive and the steroid resistant forms of the disease seem to have different immunogenetic components.     

Immunogenetics of systemic lupus erythematosus in Spanish patients: differential HLA markers

HLA-DR3 antigen included in the compound phenotype B18BfF1 (but not the one linked to the B8BfS compound phenotype) was found to be significantly increased in our SLE patients. It is remarkable that in our Southern-Mediterranean population, B18BfF1DR3 individuals (but not B8BfSDR3) are prone to SLE with renal disease, in contrast with other Northern European and Caucasoid populations. Also, patients with autoantibodies to Ro/La have a significant increase of the B8DR3 compound phenotype. Production of autoantibodies against Ro alone was associated to DR2 and production of anti-Sm/nRNP to DR3 (either B18BfF1 or B8BfS associated) only in the subgroup without renal disease. The distinctive HLA and autoimmune associations to SLE with and without renal disease suggests that both clinical forms may not share a common identical pathogenesis.     

HLA-DR antigens and the antibody response against Epstein-Barr virus

Antibodies against Epstein-Barr virus (EBV) antigens, i.e. the viral capsid antigen (VCA) and the Epstein-Barr nuclear antigen (EBNA), were determined in two independent populations with known HLA-DR phenotypes. The first population consisted of 151 patients with rheumatoid arthritis; the second one of 88 healthy parents of leukemic children. Although the group of patients with rheumatoid arthritis differed significantly in the frequency of 4 DR antigens from the second group, both groups had the same correlation between HLA-DR antigens and the antibody response to EBV antigens. There was a significant correlation between HLA-DR1 and reduced titers of antibodies to VCA, whereas the persons with only one identifiable DR antigen had higher anti-VCA titers. The persons with HLA-DR5 had significantly higher anti-EBNA titers than those without DR5.     

Human MHC Class III (Bf, C2, C4) genes and GLO: their association with other HLA antigens and extended haplotypes in the Spanish population

C4 allotype frequencies and their combination with factor B and C2 alleles (complotypes) were studied in a sample of the Spanish population in relation to MHC class I, class II and GLO alleles. The shorter genetic distances found for C4 between Spaniards and North Africans and the high frequency of extended HLA haplotypes (GLO 2) HLA-DR3 F1C30 HLA-B18 HLA-Cw5 (HLA-A30) and HLA-DR7 S1C21 HLA-Bw50 HLA-Cw6 are consistent with a paleo-North African ethnic origin (about 20,000 years B.C.) of a part of present Spaniards (Iberians), and with the effect of racial admixture during late Moslem invasions (from the 8th to the 15th century). The complotype null alleles C4A QO and C4B QO may be under natural selection pressure when found in cis position, since they are never in the same haplotype in families. The underestimation of these C4 null alleles' frequencies in unrelated individuals as compared to genotyped families is shown to be a very likely event and a serious hindrance for C4-disease association studies. We have not found any C4 duplications in the Spanish population; this may be due to sample size limitations or to the degree of admixture of our population. Strikingly, no positive linkage disequilibrium between C4A and C4B alleles is detected in unrelated individuals nor in families, although strong associations are maintained among Bf, C2, C4, HLA-A, HLA-B, HLA-C and HLA-DR markers. Assuming that all MHC polymorphisms have reached equilibrium, several explanations are proposed, including the possibility of no, different or additional natural selection mechanisms operating on some MHC class III genes (Bf, C2, C4 alleles combinations for most appropriate C3 convertases), as compared to those affecting class I and class II gene clusters (most advantageous immune response genes sets).     

Lack of association between the polymorphism at the heat-shock protein (HSP70-2) gene and systemic lupus erythematosus (SLE) in the Mexican mestizo population

Major histocompatibility complex (MHC) alleles have been recognized as genetic factors for developing systemic lupus erythematosus (SLE). In the present study we analyzed whether a heat-shock protein gene (HSP70-2) is involved in determining susceptibility to develop SLE in a Mexican Mestizo population. A HSP70-2 Pst I polymorphism was detected by a restriction fragment length polymorphism analysis of polymerase chain reaction (PCR-RFLP) in 107 SLE patients and 158 healthy controls. No statistically significant differences were observed in the HSP70-2 allele distribution between patients and healthy controls. HLA-DR analysis showed an increased frequency of HLA-DR3 allele in the patients group (P < 0.05, OR = 2.26, EF = 6.0%). On the other hand, when we analyzed HSP70-2 polymorphism in relation to HLA-DR3 allele, we could only detect an increased frequency of AB genotype in the DR3 negative patients (pC < 0.05, RR = 2.6, EF = 11.3%). Linkage disequilibrium was observed for three haplotypes: HLA-DR3-HSP70-2A (D = 0.03, D' = 0.67, P < 0.01); HLA-DR1-HSP70-2A (D = 0.03, D' = 0.86, P < 0.01) and HLA-DR8-HSP70-2B (D = 0.02, D' = 0.46, P = 0.02). Our data indicate that HSP70-2 gene polymorphism as opposed to the other ethnic groups does not appear to be relevant in SLE susceptibility in Mexican patients and that the distribution of the different alleles depend on the frequency of HLA alleles associated with them.     

HLA and GM in insulin-dependent diabetes in the Netherlands: Report on a combined multiplex family and population study

This report deals with the genetic factors involved in insulin-dependent diabetes mellitus (IDD) in The Netherlands. Twenty-two Dutch multiplex families with IDD were typed for HLA-A, -B, -C, and -DR antigens for BF, C2, C4 and GLO polymorphisms, as well as for GM allotypes of immunoglobulins. In addition, 53 unrelated IDD children and 31 unrelated patients with adult onset IDD were typed for HLA-A, -B -C, and -DR antigens. A significant heterogeneity for the frequency of HLA-DR4 related to age of onset was observed. A significant deviation of the Hardy-Weinberg equilibrium was observed for the HLA-DR locus with an excess in patient of heterozygotes HLA-DR3, -DR4, HLA-B8 amd HLA-B15 were not only secondary associated, but constituted with HLA-DR3 and -DR4, respectively, a haplotype in association with IDD. Nonrandom segregation of HLA-haplotypes was observed in multiplex families ex-emplified by an excess of HLA-identical affected sibpairs. Cross-over between HLA-DR and GLO identified the HLA-DR segment as mainly invovled in the association with IDD. Three diabetic haplotypes were confirmed to occur frequently among affected sibs: (a) A1, B8, BFS, C2.1 C4AQ0, C4B1, DR3, GLO2; (b) Aw30, Cw5, B18, BFF1, 22, C4A3, C4BQ0, DR3, GLO2; (c) A2, Cw3, B15, BFS, C2.1, C4A3, C4B3, DR4, GLO1, The segregation of GM allotypes to affected sibpairs was not significantly different from random segregation. The main conclusions from this study are that significant heterogeneity for age of onset exists and that the data are not compatible with simple genetic models including dominant, recessive, and intermediate models of inheritance. The data do require more complex models, involving two different HLA-linked (sets of) susceptibility genes.