Immunogenetic heterogeneity in type I (insulin-dependent) diabetes among the Tunisian population. HLA-DR antigens and organ-specific autoantibodies (family study)

The frequency of HLA-DR antigens, as well as the prevalence of islet cell insulin autoantibodies and other autoimmunity disorders, were investigated in Tunisian patients with insulin-dependent diabetes mellitus (IDDM) and were compared with family members (sibs) and healthy control subjects. Cytoplasmic islet cell autoantibodies (ICA) were found in 79 of 175 (45.1%) patients with IDDM, in 23 of 126 (18.25%) unaffected first degree relatives of type I diabetes patients and in only two of 146 (1.3%) control subjects. In 79 ICA positive patients with IDDM, 46.8% presented other evidence of autoimmunity by testing for specific autoantibodies. Insulin autoantibodies were found in 86.9% of healthy ICA-positive sibs. A good correlation between HLA-DR3/DR4 heterozygous phenotypes and the presence of ICA in patients with IDDM and their unaffected sibs was observed in the Tunisian population. In fact, this heterozygous phenotype is found in 63.3% of ICA-positive diabetic patients and in 44.4% of ICA-positive unaffected sibs, whereas, HLA-DR3/DR4 antigens were noted in only 22.9% of ICA-negative diabetic patients and in no ICA-negative unaffected sibs. In these studies, we also summarized the distribution of HLA-DR antigens in patients with IDDM who presented autoimmune disorders other than ICA.     

Islet-cell antibodies (ICA-CFICA) and HLA genotypes in 107 IDD patients and their first-degree relatives

The prevalence of ICA and CFICA in relation to HLA-DR genotypes was analyzed in 107 insulin-dependent diabetic (IDD) patients with a duration of IDD ranging from onset to 30 years, and 247 nondiabetic first-degree relatives. In 46 patients tested at onset of IDD, the prevalence of ICA and of CFICA was 61 and 50%, respectively; with the longer duration of IDD, the prevalence of CFICA decreased more rapidly than that of ICA. No significant association was observed between ICA and any HLA allele in patients tested from onset till 2 years after onset. However, a higher prevalence of ICA was found in DR3 positive patients with a duration of IDD of more than 2 years (p less than 0.02). Among the healthy relatives, the prevalence of ICA was 7% in parents and 13% in siblings; one out of 101 siblings had CFICA. No association was found between ICA and any HLA marker. The presence of ICA in sibs was independent of the number of haplotypes they shared with the IDD proband: among the 13 ICA-positive healthy sibs, one shared 2 haplotypes, nine shared 1 haplotype and three shared no haplotype with the proband, which is not different from the random distribution.     

Insulin-Dependent Diabetes Mellitus in Non-DR3/Non-DR4 Subjects

ABSTRACT: Five to 20% insulin-dependent diabetes mellitus (IDDM) patients do not bear the classical HLA class II DR3 or DR4 susceptibility haplotypes. We have studied the clinical characteristics, anti-islet cell antibodies (Ab) and HLA class II genotypes in 72 non-DR3/non-DR4 Caucasian patients, mainly adults, presenting with clinically typical IDDM. The DRB1*08-DQB1*0402-DQA1*0401 haplotype frequency was increased in the patients compared to 272 non-DR3/non-DR4 controls (OR = 5.9, Pc < 0.005). This association was even stronger in the Ab-positive patients (DRB1*08: OR = 7.2, Pc < 0.005; DQB1*0402: OR = 9.2, Pc < 0.005; DQA1*0401: OR = 9, Pc < 0.02). In those subjects the DRB1*15 allele was less frequent than in controls (OR = 0.1, Pc = 0.05). By contrast, IDDM patients with no Ab showed no particular association with HLA class II allele although they had clinical and metabolic characteristics similar to that of Ab-positive subjects. The genetic basis for IDDM predisposition in the Ab-positive subgroup remains elusive since the DRB1*08-DQB1*0402 haplotype encodes an Asp57-positive DQβ chain. However, all DR8 patients had a non-Asp57 encoding DQB1 allele on the second haplotype. Thus, trans-complementation leading to peculiar predisposing DQβ heterodimers may occur. Alternatively, a direct role of the DRB1*08 allele cannot be excluded. These results show that autoimmune type 1 diabetes occurs in non-DR3/non-DR4 subjects, mainly adults. They further support that IDDM, when defined on a clinical basis, encompass different pathogenetic entities.     

Prevalence of islet cell antibodies and its correlation with glucose and insulin response to oral glucose tolerance test in 1st degree relatives of insulin dependent diabetes mellitus probands

93 first degree relatives (1st DR) of insulin dependent diabetes mellitus (IDDM) patients were investigated for detection of islet cell antibodies (ICA) and beta cell functional status. ICA were detected in 26.9% Ist DR subjects (25/93), equally in parents, siblings and offspring. Normal (n = 16), impaired (n = 5) and diabetic (n = 4), glucose curves were seen in 1st DR. Low insulin levels were observed in parents and siblings with normal glucose tolerance test (N-GTT) at 90 min (p < 0.05), and (p < 0.0005) relatives with impaired glucose tolerance and diabetes. Insulin release to glucose (IRG-insulinogenic index) in control group was 352 +/- 42 mu U/mg. From the group of 25 ICA positive cases, 4 had mean IRG of 394 +/- 70 mu U/mg (group A) comparable to control, and had N-GTT; 12 had mean IRG of 107 +/- 15.9 mu U/mg (group B) significantly low (P < 0.005) compared to controls and group A and 9 showed IRG of 75 +/- 29.3 mu U/mg, lower than group B (NS) with abnormal response to glucose load. Loss of insulin secretory ability thus can precede hyperglycemia by years. The ICA positive relatives were grouped based on the immunological status with their probands. ICA status in probands does not give an idea about ICA status in their relatives. This indepth study leads to understand the correlation of genetic, metabolic and immunological parameters for early detection of IDDM in first degree relatives.     

Echovirus 4 and type 1 diabetes mellitus

AIMS/HYPOTHESIS: To determine the association between exposure to enteroviruses and Type 1 diabetes. METHODS: We measured neutralizing antibodies to the following enteroviruses: Coxsackievirus CA9, CB1, CB2, CB3, CB4, CB5, CB6, and Echovirus E4, E6, E9, E11 in the sera of (1) Type 1 diabetic patients at diagnosis (n = 33), (2) healthy offspring of parents with Type 1 diabetes without islet cell antibodies (ICA) (n = 43) and (3) normal controls (n = 57). All subjects were less than 20 years old. We performed the neutralization test determining the cytopathogenic effect on Vero cells. HLA DR serotyping was also performed in Group 2. RESULTS: Type 1 diabetic patients showed a higher frequency (21.2%, p < 0.01) of neutralizing antibodies to E4 in relation to controls (1.8%), although there were no differences comparing with offspring of Type 1 diabetic patients (20.9%). Healthy offspring carrying Type 1 diabetes HLA DR susceptibility genes were also exposed to E4 (15.0%). High frequencies of neutralizing antibodies to most enteroviruses were found in the control group. CONCLUSION: This study shows the association between Type 1 diabetes and the presence of neutralizing antibodies to Echovirus 4, suggesting the possible participation of this virus as an environmental trigger of this autoimmune disease. Interestingly, our population displays high frequencies of exposure to enterovirus (including CB4) although the incidence for Type 1 diabetes is low (2.9 per 100,000 inhabitants).     

Risk of progression to diabetes of low titer ICA-positive first-degree relatives of type I diabetics in southern Germany

In a prospective study to evaluate the prevalence and predictive potential of circulating cytoplasmatic islet cell antibodies (ICA) and competitive insulin autoantibodies (CIAA), we screened 406 non-diabetic first-degree relatives of patients with Type I diabetes mellitus (n = 154 for CIAA). The prevalence of ICA was 2.5% (10/406) and of CIAA 0.6% (1/154) in ICA- and 10% (1/10) in ICA+ relatives at initial screening. The titer of ICA positivity in all relatives varied between 1:1 and 1:4. Values of elevated CIAA were 256 nU/ml of the CIAA+/ICA+, and 97 nU/ml of the CIAA+/ICA- relatives (normal range less than or equal to 39 nU/ml). Sera for repeat ICA and CIAA determination was obtained, and 70% of relatives were found to be again ICA+ after 1.5 years, 40% after 3 years, and 10% after 5.7 years. Both CIAA+ relatives were found to be again CIAA+ on follow-up. Intravenous glucose tolerance tests (IVGTT) were performed in all antibody-positive relatives. No decrease in first-phase insulin secretion (1 + 3 min) below the 1st percentile was observed in any of the ICA+ relatives during follow-up. No ICA+, but one CIAA+/ICA- relative had developed Type I diabetes after 5.6 years of follow-up. In summary, these results indicate that low titer ICA (less than 40 JDF units) are often transient and relatives with low titer ICA rarely progress to Type I diabetes. Elevated CIAA appear to be constant over time and associated with increased progression to overt diabetes.     

The analysis of in vitro transforming growth factor-beta1 (TGF-beta1) production by peripheral blood in overt and pre-clinical type 1 diabetes mellitus

The alterations of TGF-beta1 production are believed to contribute to the development of insulin-dependent diabetes mellitus (IDDM) in animal models as well as in humans. There is also increasing evidence about the role of this cytokine in the pathogenesis of diabetic vascular complications. The aim of our study was to evaluate in vitro TGF-beta1 production by peripheral blood of newly diagnosed type 1 diabetes patients and subjects in the pre-clinical stage of the disease in comparison to healthy controls and relatives of IDDM patients with low genetic risk for diabetes development. The study was carried out in three groups of subjects: 22 patients with a recently diagnosed type 1 diabetes, their 24 first degree relatives with a different genetic risk of IDDM development and 18 healthy volunteers (control group). In all studied groups whole blood was taken for morphology parameters. HbA1C and for 72 h cultures with PHA stimulation for the estimation of TGF-beta1 in vitro production. TGF-beta1 concentration in supernatants were quantified by ELISA. In the first degree relatives HLA typing (for DR3, DR4 and DQB1*0602 alleles), measurements of anti-pancreatic antibodies (ICA, GADA, IA-2A, IAA) and intravenous glucose tolerance tests were performed. The levels of TGF-beta1 in the supernatants were significantly higher in diabetic patients (P < 0.0002) and in their first degree relatives (P < 0.05) in comparison to the control group. In the group of first degree relatives TGF-beta1 levels were highest in subjects with the presence of two or more pancreatic autoantibodies and/or with impaired insulin release in IVGTT, but lowest in relatives with protective DQB1*0602 alleles (P < 0.01). There was also a significant positive correlation between the TGF-beta1 levels and HbA1C in the IDDM subjects and first degree relatives (P < 0.03). Our study suggests that the alterations of TGF-beta1 levels could be associated with the activity of autoimmune process leading to pancreatic B cells destruction and may have a role in the pathogenesis of diabetic complications, but further studies in humans are needed.     

Correlation between islet cell antibodies and anti-cytomegalovirus IgM and IgG antibodies in healthy first-degree relatives of type 1 (insulin-dependent) diabetic patients

To investigate whether cytomegalovirus (CMV) infection may be related to islet cell antibodies (ICA) production and/or to insulin-dependent diabetes mellitus (IDDM) development, we have analyzed the prevalence of anti-CMV, IgM, and IgG antibodies and of ICA in 80 healthy siblings of IDDM patients (HSIDDP) and in 60 control subjects with negative familiar anamnesis of IDDM. HSIDDP and controls were also typed for HLA-A-B-C and DR antigens. IgM and IgG anti-CMV were detected by an ELISA method, whereas the ICA assay was performed by standard indirect immunofluorescence on 5-microns unfixed sections of human pancreas. HLA-A-B and C antigens were studied by standard microlymphocytotoxicity; DR antigens were also studied by a standard microlymphocytotoxicity on a B-enriched lymphocyte population. Our results indicate a significant association (P less than 0.0001) between high titers of anti-CMV IgG antibodies and ICA in HSDIDDP, whereas no correlation was found between the presence of any HLA-A-B-C and DR antigens and the prevalence of anti-CMV IgM and IgG antibodies and/or ICA. Thus, these data may support the hypothesis that a chronic CMV infection may be associated with ICA production whereas other factors seem to be needed for the complete development of type 1 diabetes.     

Autoantibodies to glutamic acid decarboxylase (GAD) detected by an immuno-trapping enzyme activity assay: relation to insulin-dependent diabetes mellitus and islet cell antibodies.

It has recently been proposed that the islet 64,000 Mr protein autoantigen (64K) of insulin-dependent diabetes mellitus (IDDM) is glutamic acid decarboxylase (GAD). We evaluated, by means of a newly developed immunotrapping enzyme activity assay (ITEAA), the prevalence of circulating GAD-autoantibodies (Ab) in a large population of IDDM patients (n = 168), blood donors (n = 87) and non-diabetic autoimmune patients (n = 40). The latter two groups were used as controls. Overall, GAD-Ab were found in 22% of IDDM patients, but in none of the two control groups (P = 0.007). These specificities were invariably associated with islet cell antibodies (ICA) (31.6% in IDDM with ICA vs 0 in IDDM without ICA, P = 0.0001), and this prevalence was higher in sera with high titer ICA (54.5% in IDDM with ICA greater than 80 JDF-units vs 22.6% of IDDM with ICA 5-80 JDF units; P = 0.002). Moreover, GAD-Ab were associated with the female sex (P = 0.002) and the concomitant presence of thyroid and/or gastric antibodies (P = 0.002). No correlation was observed between GAD-Ab and age of the patients, duration of IDDM, or associated non-organ specific antibodies. Our study indicates that GAD-Ab measured by ITEAA are: (1) detected in a proportion of IDDM patients; (2) strongly associated with ICA; (3) preferentially found in IDDM female patients with autoimmune polyendocrine serology; and (4) detected with lower frequency than that reported for 64K-Ab in IDDM.     

HLA associations in type 1 diabetes among patients not carrying high-risk DR3-DQ2 or DR4-DQ8 haplotypes

Type 1 diabetes is a complex disease where numerous genes are involved in the pathogenesis. Genes that account for approximately 50% of the familial clustering of the disease are located within or in the vicinity of the HLA complex on chromosome 6. Some DRB1, DQA1 and DQB1 genes are known to be involved, in addition to as yet unidentified HLA-linked genes. The DR4-DQ8 and DR3-DQ2 haplotypes are known to confer high risk for developing the disease, particularly when occurring together. Approximately 10% of patients, however, do not carry any of these high-risk HLA class II haplotypes. We have performed genotyping of DRB1, DQA1 and DQB1 alleles in non-DR3-DQ2/non-DR4-DQ8 patients and controls from Sweden and Norway to test if any HLA associations were observed in these patients. Our results clearly demonstrate several statistically significant differences in the frequency of HLA haplotypes between patients and controls. Case-control analysis including the relative predispositional effect test, and transmission disequilibrium test (TDT) analysis in Norwegian type 1 diabetes families revealed that the DQA1*03-DQB1*0301, DQA1*0401-DQB1*0402, DQA1*0101-DQB1*0501, DQA1*03-DQB1*0303 and DQA1*0102-DQB1*0604 haplotypes may also confer risk. Our analyses also supported independent risks of certain DRB1 alleles. The study clearly demonstrates that HLA associations in type 1 diabetes extends far beyond the well-known associations with the DR4-DQ8 and DR3-DQ2 haplotypes. Our data suggest that there is a hierarchy of HLA class II haplotypes conferring risk to develop type 1 diabetes.     

Restriction fragment length polymorphism analysis of HLA haplotypes in families with Type I diabetes mellitus

HLA Class II polymorphisms were analysed in 27 families with at least one Type I diabetic proband using Southern blotting technique according to 10th Histocompatibility Workshop Standards. The probes used were DRB, DQA1, DQB1 and DOB. We have studied 108 haplotypes and performed segregation analysis with HLA serology and restriction fragment length polymorphism (RFLP) data and compared "affected" with "non-affected" haplotypes (not inherited by IDDM patients). RFLPs correlated well with DR and DQ serology and detected additional polymorphisms. In particular, DQB polymorphism analysis showed segregation of the DQw3 splits with 88.5% of the DR4 affected haplotypes bearing the DQw3.2 split (now DQw8) and 11.5% the DQw3.1 split (now DQw7) while in the non-affected DR4 haplotypes 33.3% were DQw3.2 and 66.6% were DQw3.1. Haplotype analysis showed that DR4-DQw3.2 was in strong linkage with the U fragment (2.1 kb Taq I) of DQA2 (DX alpha) and with the L fragment (5.4 kb BamH I) of DOB. This study confirms previous observations of DQB polymorphisms in heterozygous IDDM patients, supports the protective effect of DQw3.1 (DQw7) against the development of the disease and demonstrates the importance of DQw3.2 (DQw8) for susceptibility to Type I diabetes.     

Proinsulin autoantibodies: association with type I diabetes but not with islet cell antibodies, insulin autoantibodies or HLA-DR type

Antibodies reacting with proinsulin but not with insulin determinants have been observed recently in Type I diabetes. We describe here that ELISA-determined proinsulin autoantibodies (IgG-PAA) also occur in first-degree relatives of IDDM patients (38/513, 7.4% vs 1.9% in controls, P less than 0.025). In contrast to insulin autoantibodies (IgG-IAA) and islet cell antibodies (ICA) no association with HLA type was found. Furthermore, IgG-PAA occur independently of IgG-IAA and ICA. We conclude that the humoral autoimmune response to proinsulin determinants is under separate genetic control.     

Transracial evidence for the influence of the homologous HLA DR-DQ haplotype on transmission of HLA DR4 haplotypes to diabetic children

The HLA-DQB1*0302 allele on DR4 haplotypes is a marker for type 1 diabetes susceptibility and it is an especially high-risk allele in DR3/4 because of its preferential distribution in Caucasian DR3/4 patients. In Asians, not only DQB1*0302 but also DQB1*0401 on DR4 haplotypes are associated with type 1 diabetes. We investigated whether the contribution of these DQ molecules was also genotype-dependent in Asians. Although the prevalence of the DR4-DQB1*0302 haplotype did not differ in patients vs. controls, the DR3/4-DQB1*0302 genotype had a RR of 12 (P<10(-4)). Moreover, a significant association of DQB1*0302 with the DR3/4 genotype was found (RR=3, P<10(-2)). In contrast, the distribution of DQB1*0401 alleles of DR4/X (X: other than 3, 4) is different from that of DR3/4 and DR4/4. Especially a significant association of DQB1*0401 with DR4/X (X: other than 1, 3, 4) was found (RR=3, P<10(-3)). The frequency of transmission of the DR4-DQB1*0302 haplotypes to diabetic offspring with DR3 was 80%, while to those without DR3 was 40%. In contrast, the transmission of the DR4-DQB1*0401 to those with DR3 was 60%, while to those without DR3 was 80%. High-risk DR4 subtypes were predominant in DR4/X (RR=7, P<10(-3)), whereas protective DR4 subtypes were observed mainly in the DR3/4 (RR=3, P<0.05). The association with diabetes and transmission to a diabetic offspring of DR4 haplotypes varies depending on the haplotype borne on the homologous chromosome. This might contribute not only to the synergistic effect of DR3/4, but also to the susceptibility influence of HLA DQB1*0401 alleles confined to DR4/X.     

Autoantibodies and High-Risk HLA Susceptibility Markers in First-Degree Relatives of Brazilian Patients with Type 1 Diabetes Mellitus: A Progression to Disease Based Study

Purpose

The objective of this study was to determine the frequencies of autoantibodies to heterogeneous islet-cell cytoplasmic antigens (ICA), glutamic acid decarboxylase₆₅ (GAD₆₅A), insulinoma-associated antigen-2 (IA-2A) and insulin (IAA)–and human leukocyte antigen (HLA) class II markers (HLA-DR and -DQ) in first degree relatives of heterogeneous Brazilian patients with type I diabetes(T1DM). A major focus of this study was to determine the influence of age, gender, proband characteristics and ancestry on the prevalence of autoantibodies and HLA-DR and -DQ alleles on disease progression and genetic predisposition to T1DM among the first-degree relatives.

Methods

IAA, ICA, GAD₆₅A, IA-2A and HLA- class II alleles were determined in 546 first-degree-relatives, 244 siblings, 55 offspring and 233 parents of 178 Brazilian patients with T1DM.

Results

Overall, 8.9% of the relatives were positive for one or more autoantibodies. IAA was the only antibody detected in parents. GAD₆₅ was the most prevalent antibody in offspring and siblings as compared to parents and it was the sole antibody detected in offspring. Five siblings were positive for the IA-2 antibody. A significant number (62.1%) of siblings had 1 or 2 high risk HLA haplotypes. During a 4-year follow-up study, 5 siblings (expressing HLA-DR3 or –DR4 alleles) and 1 offspring positive for GAD₆₅A progressed to diabetes.

Conclusions

The data indicated that the GAD₆₅ and IA-2 antibodies were the strongest predictors of T1DM in our study population. The high risk HLA haplotypes alone were not predictive of progression to overt diabetes.     

Effects of irradiation on diabetes in the BB/Wor rat

Lymphoid irradiation is known to prevent spontaneous autoimmune diabetes in susceptible BB rats. The present studies investigated further the effects of radiation in diabetes prone (DP) and resistant (DR) BB/Wor rats, and histocompatible Yoshida (YOS) rats. Single doses of total body gamma irradiation (125-600 rads) induced diabetes within 22-44 days in 20 of 102 (20%) 30 day old DR rats, less than 1% of which develop the disease. Radiation was also associated with (1) a reduction in the ratio of W3/25+ to OX8+ peripheral blood lymphocytes within 2 weeks, and (2) a decreased percentage of lymph node cells expressing the RT6.1 surface alloantigen 3-4 weeks after treatment. Similar doses of irradiation did not alter the frequency or age at onset of diabetes in DP rats, and did not induce diabetes in YOS rats. When a single dose of 250 or 500 rads of gamma irradiation was followed by injection of mitogen activated spleen cells from acutely diabetic rats to adoptively transfer diabetes, 16 of 19 (84%) DR and 8 of 14 (57%) YOS rats became diabetic. Long term exposure to ultraviolet irradiation (UVB) did not alter the frequency or age at onset of diabetes in either DP or DR rats. We conclude that there may exist a population of regulatory cells relatively sensitive to gamma irradiation that play a role in determining the susceptibility of rats to autoimmune diabetes mellitus.     

Humoral beta-cell autoimmunity in relation to HLA-defined disease susceptibility in preclinical and clinical type 1 diabetes

We have studied the relationship between diabetes-associated autoantibodies and various HLA genotypes and alleles in patients with newly diagnosed type 1 diabetes, in non-diabetic siblings and in children representing the general population to test the hypothesis that specific HLA genes regulate the humoral immune response to various autoantigens. Among the newly diagnosed patients we observed that those carrying the DR4-DQB1*0302 haplotype had increased levels of insulin autoantibodies (IAA) and IA-2 antibodies (IA-2A), but low levels of GAD antibodies (GADA). In contrast, those with the DR3-DQB1*02 haplotype had increased GADA titers but low IAA and IA-2 levels. DQB1*02-homozygous patients had a conspicuously low frequency and low levels of IA-2A. Among the siblings there was an apparent association between genetic risk and the prevalence of islet cell antibodies (ICA), IAA, GADA, IA-2A and multiple autoantibodies, the latter being detectable at a frequency of 24.1% in high risk siblings and at a frequency of only 0.9% in those with genotypes conferring disease protection. Among 7-16-year-old Finnish schoolchildren there was an association between GADA and the DQB1*02/*0302 genotype and the DQB1*0302 allele. Among young children identified from the general population based on high (DQB1*02/0302 heterozygosity) or moderate (DQB1*0302/x; x not equal *02, *0301, *0602, *0603) genetic risk for type 1 diabetes the high risk children seroconverted more often to positivity for ICA, GADA and IA-2A over their first 2 years of life. Among older sibs of these children we observed an obvious relationship between the degree of genetic risk and the frequency of ICA, IAA, GADA, IA-2A, and multiple antibodies. Taken together these observations suggest that HLA genes have a strong impact on the appearance of diabetes-associated autoantibodies both in first-degree relatives of affected children and in the general population. In patients with newly diagnosed disease IA-2A may be a more specific marker of beta-cell damage, whereas GADA might reflect a propensity to autoimmunity in general.     

Preferential transmission of diabetic alleles within the HLA gene complex

Several studies suggest a higher incidence of insulin-dependent diabetes mellitus (IDDM) among the offspring of men with the disease than among those of female diabetics. Differential transmission by the father of genes that predispose to diabetes may explain this phenomenon. To test this hypothesis, we examined parent-to-offspring transmission of HLA haplotypes and DR (D-related) alleles in 107 nuclear families in which a child had IDDM. We observed that fathers with a DR4 allele were significantly more likely to transmit this allele to their diabetic or nondiabetic children than were mothers with a DR4 allele (72.1 vs. 55.6 percent, P less than 0.001). No differences between parents were observed for HLA-DR3; however, DR3 was transmitted significantly more than 50 percent of the time from either parent (P less than 0.001). These data suggest that differential parental transmission of the HLA-DR4-linked diabetes-predisposing allele may explain the higher risk of diabetes among children of diabetic fathers than among those of diabetic mothers. In addition, the excess transmission of diabetogenic HLA alleles from parent to offspring may explain how these deleterious genes continue to recur at such high frequencies in the general population.     

Enhanced percentage of CD5+ B lymphocytes in newly diagnosed IDDM patients.

The percentage of CD5+ B lymphocytes, the prevalence of islet cell antibodies (ICA) and of anti-insulin autoantibodies (IAA) and HLA-A-B-C and DR antigens were studied in 32 newly diagnosed insulin-dependent diabetes mellitus (IDDM) patients, in 12 non-insulin-dependent diabetes mellitus (NIDDM) patients and in 12 healthy subjects. The percentage of CD5+ B lymphocytes ranged from 18% to 51.2% (mean 40.3 +/- 11%) in IDDM patients, whereas in NIDDM patients and in controls it ranged from 20% to 25.2%, (mean 21.3 +/- 4.1%) and from 16% to 24%, (mean 19.3 +/- 1.9%), respectively (P less than 0.01 vs. NIDDM patients and vs. controls). There was no correlation between a higher percentage of CD5+ B lymphocytes and the presence of ICA and/or IAA, and their titres, and/or of any HLA-A-B-C and DR antigens. Thus, an enhanced percentage of CD5+ B lymphocytes may be present in newly diagnosed IDDM patients; the possible role of this cell type in the pathogenesis of IDDM needs further investigation.     

Detection of islet cell antibodies by a microcytotoxicity method

A microcytotoxicity test for antibodies against islet cells (ICA) is described. Sera from patients with insulin-dependent diabetes, their healthy first degree relatives, and normal controls, genotyped for HLA-A, -B and -DR, were tested by 4 different methods. Cytoplasmic ICA and complement fixing ICA were detected by indirect immunofluorescence with human pancreas sections, and cytotoxic complement dependent ICA and surface ICA were tested against murine beta cell suspensions. Strong correlation was found between cytotoxic and surface antibodies (P less than 10(-7). The technique described is appropriate for use in the screening of large numbers of sera.     

Beta-cell, thyroid, gastric, adrenal and coeliac autoimmunity and HLA-DQ types in type 1 diabetes.

The autoimmune attack in type 1 diabetes is not only targeted to beta cells. We assessed the prevalence of thyroid peroxidase (aTPO), parietal cell (PCA), antiadrenal (AAA) and endomysial antibodies (EmA-IgA), and of overt autoimmune disease in type 1 diabetes, in relation to gender, age, duration of disease, age at onset, beta-cell antibody status (ICA, GADA, IA2A) and HLA-DQ type. Sera from 399 type 1 diabetic patients (M/F: 188/211; mean age: 26 +/- 16 years; duration: 9 +/- 8 years) were tested for ICA, PCA, AAA and EmA-IgA by indirect immunofluorescence, and for IA2A (tyrosine phosphatase antibodies), GADA (glutamic acid decarboxylase-65 antibodies) and aTPO by radiobinding assays. The prevalence rates were: GADA 70%; IA2A, 44%; ICA, 39%; aTPO, 22%; PCA, 18%; EmA-IgA, 2%; and AAA, 1%. aTPO status was determined by female gender (beta = - 1.15, P = 0.002), age (beta = 0.02, P = 0.01) and GADA + (beta = 1.06, P = 0.02), but not by HLA-DQ type or IA2A status. Dysthyroidism (P < 0.0001) was more frequent in aTPO + subjects. PCA status was determined by age (beta = 0.03, P = 0.002). We also observed an association between PCA + and GADA + (OR = 1.9, P = 0.049), aTPO + (OR = 1.9, P = 0.04) and HLA DQA1*0501-DQB1*0301 status (OR = 2.4, P = 0.045). Iron deficiency anaemia (OR = 3.0, P = 0.003) and pernicious anaemia (OR = 40, P < 0.0001) were more frequent in PCA + subjects. EmA-IgA + was linked to HLA DQA1*0501-DQB1*0201 + (OR = 7.5, P = 0.039), and coeliac disease was found in three patients. No patient had Addison's disease. In conclusion, GADA but not IA2A indicate the presence of thyrogastric autoimmunity in type 1 diabetes. aTPO have a female preponderance, PCA are weakly associated with HLA DQA1*0501-DQB1*0301 and EmA-IgA + with HLA DQA1*0501-DQB1*0201.