Search

Aims/hypothesis: We examined the prevalence of islet autoantibodies and their relationship to glycaemic control over 10 years in patients diagnosed clinically with new-onset type 2 diabetes., Methods: Patient clinical characteristics and autoantibody status were determined at entry to the UK Prospective Diabetes Study (UKPDS) before randomisation to different glucose control policies. Patients were followed for 10 years., Results: Data available on 4,545 of the 5,102 UKPDS patients showed that 11.6% had antibodies to at least one of three antigens: islet cell cytoplasm, glutamic acid decarboxylase and islet autoantibody 2A (IA-2A). Autoantibody-positive patients were younger, more often Caucasian and leaner, with lower beta cell function and higher insulin sensitivity than autoantibody-negative patients. They also had higher HbA1c, and HDL-cholesterol levels, and lower blood pressure, total cholesterol and plasma triglyceride levels. Despite relative hyperglycaemia, autoantibody-positive patients were less likely to have the metabolic syndrome (as defined by the National Cholesterol Education Program Adult Treatment Program III), reflecting a more beneficial overall risk factor profile. Of 3,867 patients with post-dietary run-in fasting plasma glucose (FPG) values between 6.0 and 14.9 mmol/l and no hyperglycaemic symptoms, 9.4% were autoantibody-positive, compared with 25.1% of 678 patients with FPG values of 15.0 mmol/l or higher, or hyperglycaemic symptoms. In both groups, no differences were seen between those with and without autoantibodies in changes to HbA1c over time, but autoantibody-positive patients required insulin treatment earlier, irrespective of the allocated therapy (p<0.0001)., Conclusions/interpretation: Autoantibody-positive patients can be treated initially with sulphonylurea, but are likely to require insulin earlier than autoantibody-negative patients.

Aims/hypothesis: Established autoimmune markers of type 1 diabetes, including islet cell autoantibodies (ICA) and autoantibodies to glutamic acid decarboxylase (GADA) have been used to screen people presenting with type 2 diabetes for latent autoimmune diabetes in adults. We have examined the prevalence of autoantibodies to protein tyrosine phosphatase isoforms IA-2 (IA-2A) and IA-2beta/phogrin (IA-2betaA) in a cohort of adult UKPDS patients thought to have type 2 diabetes, and investigated the possible role of these autoantibodies in predicting requirement for insulin therapy., Methods: IA-2A and IA-2betaA were measured by a validated radioimmunoassay with human recombinant autoantigens in 4,169 white Caucasian patients aged 25-65 years and newly diagnosed with type 2 diabetes. The clinical requirement for insulin therapy within 6 years was examined in 2,556 patients not randomised to insulin., Results: IA-2A and IA-2betaA were present in 2.2 and 1.4%, respectively, of these patients. IA-2A were more prevalent in younger patients (p for trend <0.00001), more often associated with the HLA-DR4 allele (26.3 vs 8.0%, p<0.0001), and their presence increased the likelihood of insulin therapy requirement within 6 years from diagnosis [relative risk (95%CI) 12.2 (9.8-15.3)]. The presence of IA-2A together with GADA increased the relative risk of requiring insulin therapy from 5.4 (4.1-7.1) for GADA alone to 8.3 (3.7-18.8) and the corresponding positive predictive value from 33 to 50%., Conclusions/interpretation: In type 2 diabetes, the presence of IA-2A is infrequent, associated with the HLA-DR4 haplotype, and highly predictive of future need for insulin therapy. The measurement of IA-2betaA does not provide additional information.

Aims/hypothesis: Our aim was to characterise the genetic and immunological features associated with Type I (insulin-dependent) diabetes mellitus in a cohort of Indo-Aryan children resident in the United Kingdom., Methods: Children with Type I diabetes (n = 53), unaffected first-degree relatives (n = 146) and unrelated healthy control children (n = 54) were typed for alleles of the HLA-DRB1, HLA-DQA1 and HLA-DQB1 genes. Islet cell antibodies and antibodies to glutamic acid decarboxylase, protein tyrosine phosphatase-2 (IA-2ic) and insulin were measured in the diabetic and control children., Results: The DRB1*03.DQA1*05.DQB1*02 haplotype was positively associated with the disease, occurring in 78% of diabetic children compared with 22.6 % of healthy children (p(c) < 2.4 x 10(-5)). In simplex families, this haplotype was transmitted more frequently to the diabetic children than to their unaffected siblings (p < 1 x 10(-4)). The DRB1*04.DQA1* 03.DQB1*0302 haplotype was also transmitted preferentially to the diabetic probands (p < 0.025) but was not associated with disease in the case control study. Islet-related autoantibodies were detected in 89.6 % of diabetic patients compared with 11.8 % of control children (p < 1 x 10(-6)). Although protein tyrosine phosphatase-2 autoantibodies were detected more frequently among DRB1*04-positive diabetic patients compared with patients lacking this allele, the overall frequency of these autoantibodies was lower than observed in Europid diabetic subjects. This could reflect the absence of a disease association with DRB1*04 in the Indo-Aryan cohort., Conclusion/interpretation: Type I diabetes in our Indo-Aryan cohort is similar to the disease observed in Anglo-Europeans but has important immunogenetic differences. The low frequency of protein tyrosine phosphatase-2 autoantibodies among the Indo-Aryan diabetic children could have important implications for the design of future strategies for disease prediction in this population.

Aims/hypothesis: Juvenile-onset, insulin-dependent diabetes is associated with islet cell antibodies and with specific "high-risk" HLA-DRB1 and HLA-DQB1 genotypes. Patients with Type II (non-insulin-dependent) diabetes mellitus can have islet-related antibodies, but the genotypic associations at different ages of onset have not been evaluated. Our aim was to determine (i) the prevalence of DRB1 and DQB1 genotypes in patients at diagnosis of Type II diabetes at different ages from 25 to 65 years compared with the general population, and (ii) whether the presence of islet cell antibodies (ICA) or glutamic acid decarboxylase antibodies (GADA) or both by age is associated with different DRB1 and DQB1 genotypes., Methods: The antibodies to islet cells and those to glutamic acid decarboxylase were measured in 1712 white Caucasian diabetic subjects at diagnosis of diabetes and they were genotyped for HLA DRB1*03 and DRB1*04 and the high-risk DRB1*04-DQB1* 0302 haplotype. To assess over-representation of high-risk alleles for Type I (insulin-dependent) diabetes mellitus, the prevalence of high-risk alleles in diabetic patients was expressed relative to the prevalence of low-risk alleles, non-DR3/non-DR4, that provided a reference denominator in both the diabetic patients and in 200 non-diabetic control subjects. The prevalence of ICA or GADA or both in patients with different HLA genotypes was assessed in those diagnosed in four age groups, 25-34 years, 35-44 years, 45-54 years and 55-65 years., Results: In Type II diabetic patients presenting at ages 25-34, 35-44 and 45-54 years, there was an increased prevalence of DR3/DR4 compared with the general population with approximately 6.5-fold, 2.9-fold, 2.1-fold over-representation, respectively (p < 0.0001, < 0.01, < 0.05) but this was not found in those aged 55-65 years old. In the group aged 25-34 years, 32 % of patients with ICA or GADA or both had DRB1*03/DRB1*04-DQB1*0302 compared with 10% in those aged 55-65 years and expected 3% prevalence. Conversely, only 14% of those aged 25-34 years with antibodies had non-DR3/non-DR4, compared with 35 % in those aged 55-65 years. There was thus pronounced age heterogeneity in DRB1 and DQB1 predisposition to Type II diabetes. The antibodies displaced DRB1 or DQB1 genotypes in the multivariate model for requiring insulin therapy by 6 years of follow-up., Conclusion/hypothesis: The age of presentation of Type I diabetes in adulthood was in part dependent on the DRB1/DQB1 genotype. Islet cell antibodies and glutamic acid decarboxylase antibodies were strongly associated with DRB1*03/DRB1*04-DQB1*0302 in early adulthood but showed little relation with specific HLA genotypes after the age of 55 years.

Analysis of the geographical variation of risk for a disease is a key issue in descriptive epidemiology and may provide useful suggestions for planning further studies to identify the underlying causes. We adopted a Bayesian approach to investigate the geographical distribution of insulin-dependent diabetes mellitus (IDDM) incidence rate across Sardinia. Data on incidence of IDDM in children aged under 15 years (619 IDDM cases) in Sardinia was obtained by the Sardinian Eurodiab ACE register. The overall completeness of ascertainment was: 91.3%. The average yearly standardized incidence rate for the years 1989-1994 was 33.24 per 100000 (95% C.I. 30.60, 35.88), which is the second highest in Europe after Finland. Sex and age-specific risks were higher in males than in females. Considering the variation of IDDM risk according to the age at diagnosis, the risk profile increased up to the 13th year of age for both sexes, being steeper in males. The degree of geographical variation in IDDM risk was small with a slight difference between the highest and the lowest standardized rate across the map. Indeed, even the municipalities at lowest risk in Sardinia showed a risk higher than most European countries. The Sardinian population is genetically atypical, characterized by genetic homogeneity and marked susceptibility to autoimmune diseases. Our finding of a small geographical variation within the island coupled with a marked temporal trend previously observed in data on military conscripts could be interpreted as evidence of a relatively recent environmental aetiological factor that was uniformly distributed across the island and had its effect in a genetically predisposed population.

Identification of islet autoantigens offers the possibility that antibody tests other than islet cell antibodies may be used for assessing risk of insulin-dependent diabetes mellitus (IDDM). The aim of this study was to determine the combination of islet autoantibody markers that could identify most future cases of IDDM. Islet cell antibodies, antibodies to glutamic acid decarboxylase (GAD)65, 37,000/40,000 M(r) islet tryptic fragments, carboxypeptidase-H, and islet cell autoantigen (ICA)69 were measured in sera from 100 newly-diagnosed IDDM patients, 27 individuals prior to onset of IDDM, and 83 control subjects. Islet cell antibodies were detected in 88% of IDDM patients and 81% with pre-IDDM, GAD65 antibodies in 70% of IDDM patients and 89% with pre-IDDM, and antibodies to 37,000/40,000 M(r) islet tryptic fragments in 54% of IDDM patients and in 48% with pre-IDDM. The latter were found only in conjunction with islet cell antibodies and were more frequent in young onset cases. All 20 IDDM patients and the 3 pre-IDDM subjects who had islet cell antibodies without GAD65 antibodies had antibodies to 37,000/40,000 M(r) islet tryptic fragments, and all but one had disease onset before age 15 years. No sera strongly immunoprecipitated in vitro translated ICA69 or carboxypeptidase-H; 4% of patients had anti-ICA69 and 11% anti-carboxypeptidase-H levels above those of the control subjects. The findings suggest that none of the single antibody specificities are as sensitive as islet cell antibodies, but that a combination of GAD65 antibodies and antibodies to 37,000/40,000 M(r) islet tryptic fragments has the potential to identify more than 90% of future cases of IDDM.(ABSTRACT TRUNCATED AT 250 WORDS)

Finland and Sweden have the highest incidence of insulin-dependent diabetes in children in the world, about 3-4 times that of countries in the Mediterranean area, with the exception of Sardinia. We have collected information from several European clinics and from Pittsburgh, USA, in order to find out whether this difference incidence is associated with corresponding differences of the disease pattern. Patients in Finland or Sweden ('North') and Pittsburgh were younger (< 10 years old) at diagnosis compared with those in the other clinics in Europe (P < 0.05 versus P < 0.02). In the North, boys were in excess (58%) in contrast to France (40%) and Pittsburgh (46%). Patients in the North had a shorter duration of symptoms (< 8 days; P < 0.001) and higher blood glucose (> 20 mmol/l; P < 0.05) than those attending the other European clinics. Irrespective of age, there were more ICA-positive patients in the North (94%) than in Berlin-Vienna (67%; P < 0.01) or in France (70%; P < 0.01). There was a tendency for non-diabetic parents and siblings in the North to have lower C-peptide values (< 0.26 pmol/ml) at the time of diagnosis of the proband and to be ICA-positive more often than relatives in the other European clinics. The seasonal variation of diagnosis showed no obvious geographical differences, with recorded diagnosis always lowest during the summer. We conclude that certain factors seem to cause not only a high incidence of diabetes in children in Finland and Sweden but perhaps also a more aggressive early disease process.

We studied metabolic progression to IDDM in a cohort of adults who are ICA-positive and have associated autoimmune endocrine disease or circulating organ-specific autoantibodies (the Polyendocrine Study). Of the 186 individuals recruited 27 developed overt diabetes after a median follow-up of 4.5 years (range 0.4-12). Of these, eight patients did not require insulin treatment until at least 6 months after clinical diagnosis, with an interval of 1.8 years (1.2-5.7). An IVGTT was performed in 38 subjects and 23 had sequential studies. Of the initial 38 subjects six developed diabetes and only three showed a loss of FPIR to glucose (below the first percentile of a normal control group) before clinical onset of the disease. An additional three subjects showed a loss of the FPIR, and all still have normal glucose tolerance after median follow-up of 28 months (22-95). A "whole" or "mixed" pattern of islet cell staining was found in five of the six patients who developed diabetes and antibodies against an islet 37 k-antigen were detectable in four patients, all of whom required insulin soon after diagnosis. A beta-cell "selective" ICA staining pattern was seen in 14 of 17 subjects who did not develop diabetes and the "mixed" pattern in only three. None of this group had detectable 37 k-antibodies. We conclude that metabolic deterioration is slow in polyendocrine patients, and that the IVGTT has less prognostic significance in this group than in first degree relatives of patients with IDDM.(ABSTRACT TRUNCATED AT 250 WORDS)

In order to obtain new and more detailed information about temporal trends and geographic distribution of Type 1 (insulin-dependent) diabetes mellitus in Sardinia, we screened a series of birth cohorts (1936-1973) of all male army conscripts aged 18-19 years, filed in the Sardinian Conscript Register where Type 1 diabetes is a cause of rejection. A total of 678 diabetic subjects, born and permanently residing in Sardinia, was identified. The point prevalence (x 1000) at the age of 20 years in the birth cohorts ranged from values close to zero for the first ten cohorts (1936-1945) up to a maximum of 3.08 (95% confidence limits 2.28-4.08) for the 1966 cohort and continued high thereafter although an apparent decrease was observed from the early 1970s birth cohorts. Type 1 diabetes was distributed throughout the four provinces of Sardinia with no particularly significant heterogeneity; however, in accordance with the geographical distribution of diabetes cases of the Eurodiab Ace survey (1989-1990), the highest prevalence of the disease was observed in the Cagliari and Oristano provinces, followed by Nuoro and Sassari. These data suggest a gradually increasing trend of male Type 1 diabetes prevalence in Sardinia with a 29-fold increase between the late 1930s and the late 1960s birth cohorts. This seems to confirm the high incidence of Type 1 diabetes in the 0-14 and 0-29 year age groups recently reported among Sardinians during the Eurodiab Ace collaborative multicentre study.

The cytoplasmic islet cell antibody patterns of sera from islet cell antibody positive non-diabetic and diabetic endocrine autoimmune patients, and newly-diagnosed Type 1 (insulin-dependent) diabetic patients were characterised using four layer immunofluorescence with monoclonal anti-proinsulin or anti-glucagon antibodies. Two distinct islet cell antibody types were identified. One gave a diffuse cytoplasmic staining in both Beta and Alpha cells ('whole' islet pattern), and was not affected by pre-incubation with rat brain homogenate. The other had a granular appearance with staining restricted predominantly to Beta cells ('selective' islet pattern) and was completely inhibited by pre-incubation with rat brain homogenate. Some sera appeared to have a 'mixed' islet pattern, in which glucagon-positive cells gave a weaker cytoplasmic staining than proinsulin-positive cells. The granular 'selective' pattern was found in sera from 19 (79%) of 24 non-diabetic endocrine autoimmune patients, in two (22%) endocrine autoimmune patients who developed Type 1 diabetes (p less than 0.0001 vs non-diabetic endocrine autoimmune patients), and in none of 19 newly-diagnosed diabetic patients. The 'whole' islet pattern was found only in sera from patients who had, or who subsequently progressed to, Type 1 diabetes. This study has identified a novel islet cell antibody specificity and demonstrates that in islet cell antibody positive endocrine autoimmune patients, only islet cell antibodies which stain both Beta and Alpha cells are associated with progression to Type 1 diabetes.

Alterations of lymphocyte subsets have been recently reported in pre-type 1 diabetes but the relation with other immunological markers, in particular islet cell antibodies (ICA), is still unknown. In the present study, we have investigated prospectively changes of lymphocyte subsets in 86 first-degree relatives of patients affected by type 1 diabetes and correlated such modifications with ICA titres. Among individuals with persistent ICA, 8 had ICA titres of more than 20 JDF units, 14 had ICA titres between 5 and 20 JDF units and 64 had ICA titres between 0 and 5 JDF units. First-degree relatives with ICA titres of more than 20 JDF units had significantly decreased proportions of CD3 cells. This reduction was predominantly in the CD4 subset, giving rise to a decreased CD4/CD8 lymphocyte ratio. Those with ICA titres between 5 and 20 JDF units showed abnormalities in both CD3 and CD4 lymphocytes, but not in CD4/CD8 lymphocyte ratio. Further characterization of the CD4 cell subset was performed using three other monoclonal antibodies, CD45RO (UCHL1), CD45RA and CD29, phenotyping memory T-cells, the inducer cells of suppressor function and helper-inducer cells, respectively. The proportions of total CD45RO and CD45RA were not significantly different among first-degree relative with distinct ICA titres in a cross-sectional study, whereas a trend towards a reduced proportion of CD4/CD45RA cells was observed. The longitudinal study demonstrated that individuals potentially susceptible to the development of type 1 diabetes and who possess high titres of ICA have impairment of CD4/CD8 lymphocyte ratio, mainly due to a reduction in the CD4 subset.(ABSTRACT TRUNCATED AT 250 WORDS)

Basal insulin secretion was compared in nine islet-cell antibody positive, non-diabetic first-degree relatives of children with Type 1 (insulin-dependent) diabetes mellitus and nine normal control subjects matched for age, sex and weight. Acute insulin responses to a 25 g intravenous glucose tolerance test were similar in the two groups (243 (198-229) vs 329 (285-380) mU.l-1 x 10 min-1, mean (+/- SE), p = 0.25). Fasting plasma insulin was assayed in venous samples taken at one min intervals for 2 h. Time series analysis was used to demonstrate oscillatory patterns in plasma insulin. Autocorrelation showed that regular oscillatory activity was generally absent in the islet-cell antibody-positive group, whereas a regular 13 min cycle was shown in control subjects (p less than 0.0001). Fourier transformation did, however, show a 13 min spectral peak in the islet-cell antibody positive group, consistent with intermittent pulsatility. We conclude that overall oscillatory patterns of basal insulin secretion are altered in islet-cell antibody positive subjects even when the acute insulin response is within the normal range.

In this study we report for the first time, the molecular analysis of HLA-DR and -DQ gene frequencies in a large cohort of well characterized type 1 (insulin-dependent) diabetes mellitus (IDDM) patients (n = 72), and ethnically matched controls (n = 59) collected in sub-Saharan Africa. High molecular mass DNA was prepared and analyzed in Southern blots with DRB1, DQA1, and DQB1 probes. By identifying DR and DQ allele-specific restriction fragment length polymorphisms (RFLPs), we have shown a strong positive association between IDDM and the Asp 57- DQB1 allele *0201 (DQw2). A rare DR4, DQw2 haplotype was also identified at high frequency in the IDDM cohort. We can now confirm that the association between Asp 57-DQB1 alleles and IDDM, previously reported in ethnically diverse cohorts collected in Western Europe, North America, and South Asia, is also present in an IDDM cohort collected in Africa.

Forty-one assays were analysed at the 3rd International Workshop on the standardisation of islet cell antibodies. Analysis of precision demonstrated assays consistently detecting blind duplicates within one doubling dilution and capable of discriminating one doubling dilution differences in islet cell antibody concentration. Some assays, however, reported duplicates discrepantly by more than seven doubling dilutions, and consequently could not distinguish even large quantities of islet cell antibodies. Precision was best in assays from laboratories which had participated in all three Standardisation Workshops and was not dependent upon methodology. The use of the Juvenile, Diabetes Foundation reference islet cell antibody standard and standard curves reduced the scatter of results, and was best amongst assays with better precision. Twenty-seven assays reported all ten blood donor sera as negative. However, 14 assays did not, and specificity (negativity in health) was less than 50% in three assays. Low specificity was strongly associated with poor precision. The detection limit of assays ranged from less than 5 to 50 JDF units and was partially dependent upon methodology. Assays incorporating extended incubation had the lowest detection limits without a decrease in the specificity of the ten blood donor sera. Precise quantification is fundamental for the standardisation and comparability of islet cell antibodies. Precise quantitative assays have been identified and reference standards and common units established.

The observations emerged from the pancreatic transplant experiments in identical twins indicate that Type I diabetic patients maintain memory cytotoxic T cells for several years and these can be "re-awakened" when Class I identical beta cells are re-introduced into the diabetic melieu. In addition, these data have shown that these lymphocytes can (in a matter of a few weeks) cause rapid irreversible decompensation of beta cell function. From this, an important lesson is learnt: cytotoxic T cells, when generated in sufficient number against beta cells do not leave much "breathing space" for these cells. This has important implications for explaining the long latency period preceding the acute onset of Type I diabetes. ICA, when produced, seem unable to cause gross damage to beta cells. They persist for several years in the blood, but beta cell function remains apparently unaltered. It is only when cytotoxic T cells, with fine specificity for beta cells are generated that the 'killing cycle' is completed. Whether these cells are present all the time, and kept under tight control by active suppressor mechanisms or whether they appear only after an environmental trigger (e.g., retroviruses) is unknown. If the former is the case, this would give strong support to the suggested important role of suppressor T cells in the pathogenetic circuit. Some evidence for this has been produced (rev. in [116]) but, obviously, it requires confirmation (see debate which followed [116]). If, on the other hand, the latter is experimentally confirmed, one can return to the theory that cytotoxic T cells acquire the characteristics of autoreactivity by expressing receptors on their surface in a configuration which enables combination with self-autoantigens (rev. in [45]). In summary, the study of the etiopathogenesis of Type I diabetes and human autoimmunity in general has attracted a great deal of interest among immunologists. It is only by further dissecting the various limbs of the undesirable immune response against beta cells and, by trying to formulate novel hypotheses, sometimes against accepted dogmas [32], that the complete picture will be finally disclosed. At this stage it will be possible to design effective therapy trials, so that Type I diabetes and other related autoimmune disorders ultimately may be prevented.

To determine the value of islet-cell antibodies, both complement-fixing and non-complement-fixing, in predicting the later development of Type 1 (insulin-dependent) diabetes, we studied different groups of identical twins. Twelve twins have developed diabetes and 11 of these had non-complement-fixing islet-cell antibodies before diagnosis, and eight out of nine tested had complement-fixing islet-cell antibodies. Of the twins who have remained non-diabetic for many years and are now unlikely to develop diabetes, twelve have had non-complement-fixing islet-cell antibodies at some stage but only four have ever had complement-fixing antibodies. In 29 non-diabetic co-twins tested within 5 years of the diagnosis of diabetes in the affected twin the presence of islet-cell antibodies, especially complement-fixing, predicted the progression to frank diabetes with a high specificity (100%), sensitivity (88%) and predictive value (100%). In pairs remaining discordant the antibodies were found more frequently in the diabetic than the non-diabetic twin. We conclude that the presence of islet-cell antibodies is not genetically determined and can occur without progression to diabetes. However, the presence of islet-cell antibodies, especially complement-fixing, in non-diabetic twins tested soon after the diagnosis of their co-twin, indicates a high risk for the development of diabetes.

A prospective study of lymphocyte subsets has been carried out for 18 months in 58 healthy first-degree relatives of Type 1 (insulin-dependent) probands. Subjects selected for presence or absence of islet cell antibodies included 10 with complement-fixing islet cell antibodies, 10 with conventional islet cell antibodies and 38 without islet cell antibodies. Immunoregulatory and effector lymphocytes subsets, and in particular activated T-cells, were investigated using a panel of monoclonal antibodies. The results showed no significant changes in total T, helper, suppressor/cytotoxic cell or K/NK cells. Activated T-cells were observed at least once in 22 subjects using the 4F2 monoclonal antibody and in 11 using the Tac antibody. Seven subjects had 4F2-positive cells on repeated occasions and one twice showed Tac-positive cells. Fluctuations and/or loss of islet cell antibodies were observed during follow-up. There was no correlation between presence of activated T-cells and either islet cell antibody status of HLA haplotype sharing with the diabetic proband. On the other hand, a significant correlation was observed between HLA-DR3 positivity of subjects and the occurrence of activated T-cells (both 4F2-positive and Tac-positive). We conclude that subjects with HLA-DR3 may be especially prone to T-cell activation. As none of the 'high risk' individuals developed diabetes in the course of follow-up, the relevance of these observations in the pathogenesis of Type 1 diabetes needs more prolonged investigation.

The case of a female patient with fasting hypoglycaemia before the development of Type 1 (insulin-dependent) diabetes mellitus is reported. She presented with primary hypothyroidism, partial hypopituitarism, adrenal insufficiency and glucagon deficiency. Thyroid microsomal and gastric parietal cell antibodies were detected as well as HLA-B8, whereas islet cell antibodies were not demonstrable, even 2 years after the onset of diabetes. Plasma chromatography revealed true pancreatic glucagon (IRG3500) close to undetectable in basal samples with a questionable increase from 3 to 18 pg/ml during insulin-induced hypoglycaemia. After an overnight fast, moderate hyperaminoacidaemia was found with elevations of alanine, glycine, serine, arginine and ornithine as seen in pancreatectomized patients. It is suggested that the deficient glucagon secretion in this patient might, at least in part, have been the cause of fasting hypoglycaemia and the failure of glucose recovery following insulin-induced hypoglycaemia. Possible, the A cell deficiency was part of the polyglandular failure syndrome in this patient.

Over a period of 5 years, lymphocyte subpopulations and their markers of activation were studied prospectively in 56 first degree relatives of Type 1 (insulin-dependent) diabetic probands. Lymphocytes were phenotyped using a panel of monoclonal antibodies which recognise CD3, CD4, CD8 lymphocytes, K/NK cells, HLA Class II products and IL-2 receptors (IL-2r). Twenty-six subjects were negative for islet cell antibody (ICA), 18 had complement fixing ICA (CF-ICA) and 12 only conventional ICA (ICA-IgG). The total number of observations (blood samples collected) was 386. Overall, changes in T cell data were observed in the three groups of first degree relatives compared to 70 normal subjects without a family history of diabetes. Six individuals developed Type 1 diabetes in the course of the study. They all possessed CF-ICA and five out of six showed a persistent reduction (less than 1.5) of the CD4/CD8 lymphocyte ratio before the clinical onset of the disease. Activated lymphocytes were found on two occasions in two of these subjects. We conclude that imbalance of lymphocyte immunoregulatory subsets is present before the onset of Type 1 diabetes in susceptible individuals; the persistence of a reduced CD4/CD8 lymphocyte ratio may reflect the ongoing process leading to B-cell destruction.

Residual insulin secretion and islet-cell antibodies were studied in 399 insulin-dependent diabetics with age at onset of between 10--19.9 years (248 patients) or 30--39.9 years (151 patients). We found the prevalence of islet-cell antibodies to be independent of residual beta-cell function as measured by serum C-peptide and age at onset. The cause and role of the persistence of islet-cell antibodies in insulin-dependent diabetics remain obscure.

Sixteen pancreases from 11-24 week old human fetuses were cultured for up to 11 days to investigate islet cell surface antibodies. Hormonal content and presence of cytoplasmic autoantigen were assessed by immunofluorescence with specific antihormone sera and high titre cytoplasmic islet cell antibody positive sera. Viable islet cells cultured on coverslips were tested with 21 islet cell antibody positive sera from Type 1 (insulin-dependent) diabetics, one islet cell antibody positive serum from a non-diabetic and four normal control sera. Surface binding immunoglobulins were detected by indirect immunofluorescence in nine out of 11 newly diagnosed Type 1 diabetics and in two out of ten longstanding diabetics with another coexistent autoimmune endocrinopathy. The four-layer double immunofluorescence technique showed that the surface antibody stained insulin secreting cells, but owing to rarity of A and D cells in the fetal cultures it has not yet been possible to exclude the reactivity of islet cell surface antibodies with glucagon or somatostatin cells.

81 patients with spontaneously acquired myasthenia gravis (MG) were investigated for the presence of autoimmune (AI) diseases and their sera were tested for a range of organ-specific autoantibodies. 77 of the patients were HLA-phenotyped. Antibody titres to acetylcholine receptors (AChR) were higher in non-thymomatous patients who possessed HLA-B8 (p less than 0.05) and/or -DR3 (p less than 0.05) as compared to patients lacking these HLA antigens. 3 out of 20 (15%) patients with ocular MG, 7/23 (30%) with generalized MG of early onset, 11/23 (48%) generalized MG of late onset and 5/14 (35%) patients with thymoma had either overt AI diseases or significant titres of organ-specific autoantibodies suggesting subclinical AI disease. In ocular MG, low titres and an infrequent finding of antibodies to AChR (32%) as well as the low prevalence of associated autoantibodies and AI diseases indicate that this subgroup of MG consists of patients with restricted AI reactivity. HLA-B8 and -DR3 were present in all the patients with associated AI disorders in the young onset group but in none of the patients with old age of onset. In the young group, 6 out of 7 patients with associated AI conditions were women whereas the sex ratio was about equal in the older cases in both, patients with and without associated AI diseases or autoantibodies. We conclude from these observations that ageing provides conditions that allow the breakdown of self tolerance. The simultaneous presence of HLA B8, DR3 and female sex provide important additional factors for early expression of MG.

Antibody dependent cell mediated cytotoxicity in relation to the levels of circulating killer cells was investigated in 16 newly diagnosed classical insulin dependent (Type 1) diabetics, 11 islet cell antibody positive non diabetic children with at least one HLA haplotype in common with their diabetic sibling, and in 15 normal controls. Antibody dependent cell mediated cytotoxicity was evaluated using, as target, 51Cr labelled human 0+ erythrocytes sensitised with an anti-CD antiserum. Killer cells were measured by the low affinity E-rosetting cell technique. Increased killer cell levels (greater than normal mean + 2SD) were accompanied by a significant enhancement in antibody dependent cell mediated cytotoxicity both in newly diagnosed diabetics (p less than 0.05) and in unaffected siblings (p less than 0.01). These preliminary results indicate that raised antibody dependent cell mediated cytotoxicity is a feature of insulin dependent diabetes at diagnosis and suggest that active B cell damage might be occurring some time before the onset of clinical symptoms.

Pancreatic islet cell antibodies and 12 other autoantibodies were measured at the time of diabetes diagnosis in 46 Pima Indians, aged 17--47 years, and in 46 age-sex matched non-diabetic controls. Islet cell antibodies were found in only two diabetics, aged 20 and 25, compared with none of 46 controls. Neither of the subjects with islet cell antibodies had other autoantibodies. At least one type of autoantibody was found in 14 (30%) of the diabetics and in 14 controls, but none was significantly associated with diabetes. This study indicates that diabetes in the Pima Indians, even those with an onset below 25 years of age, is almost entirely of type II, in that the disease is not associated with islet cell antibodies, ketoacidosis, or insulin dependence.

A polymorphic DNA sequence in the 5'-flanking region of the human insulin gene was studied in relation to Type 1 (insulin-dependent) diabetes. In 141 Caucasoid subjects analysed by Southern blot hybridisation techniques, two major DNA insertions were observed: a Class 1 allele or a Class 3 allele. The Class 2 allele was not observed in this group of subjects. Genotype frequencies in a control population (n = 88) were: homozygous 1/1, 42%; heterozygous 1/3, 50%; and homozygous 3/3, 8%. Corresponding genotype frequencies in 53 Type 1 diabetic patients were 79%, 21% and 0%, respectively (p less than 0.0005 from chi 2 test). This confirms prevalence data reported by Bell et al. [16]. There appeared to be no coinheritance with HLA-DR3/DR4 related antigens, nor with autoimmune features. Analysis of 17 Type 1 diabetic pedigrees including 34 diabetic and 69 non-diabetic subjects did not demonstrate genetic linkage of these DNA inserts with diabetes, using an autosomal recessive, single locus model of inheritance.

The inappropriate expression of HLA Class II molecules by the target cells of endocrine autoimmune diseases is a recent observation that has been intensively studied in thyroid autoimmunity and type I diabetes mellitus. In vitro studies have shown that interferon-gamma can induce Class II expression, either alone, as in thyrocytes, or in combination with other mediators like tumour necrosis factor or lymphotoxin, as in islet cells, pointing to possible mechanisms operating in vivo. Endocrine cells expressing Class II molecules are able to present their autoantigens to helper T cells, thus possibly inducing the autoimmune process. However, until now it is still unclear if the expression of Class II molecules by the target cells is the primary immune phenomenon, which might possibly be triggered by a latent viral infection of the endocrine cell. Alternatively, it might be a secondary response in an ongoing autoimmune process. Particularly data obtained in the diabetic pancreas favour the first possibility, but only progress in our understanding of the role of HLA antigens in immunoregulation will make it possible to interpret the phenomenon properly.

Type 1 (insulin-dependent) diabetes mellitus, like some other autoimmune diseases, is linked to certain alleles coded by genes in the HLA-D region. Sequence analysis of DQ beta chains indicates that aspartic acid at codon 57 confers resistance to the development of Type 1 diabetes. However, a full explanation for the HLA-association of Type 1 diabetes, particularly the increased susceptibility of DR3/4 heterozygotes is still awaited. The localisation of tumour necrosis factor genes on the short arm of chromosome 6 between HLA-B and the complement genes (Class III) prompted us to investigate a possible polymorphism of TNF-alpha at the genomic level in relation to Type 1 diabetes susceptibility. A dialleleic TNF-alpha restriction fragment length polymorphism was found with Ncol and its segregation with HLA-haplotypes analysed in diabetic families. We describe here a strong linkage of TNF-alpha alleles with certain DR haplotypes. For example, the common extended haplotype HLA A1-B8-DR3 was almost exclusively associated with the 5.5 kb TNF-alpha allele whereas Bw62-DR4 with the 10.5 kb allele. Thus both alleles segregate to diabetic patients. DR matched haplotypes of affected family members differed significantly from those of the non-affected at the TNF alpha locus. All affected sibling pairs in 11 multiplex affected families were identical for TNF-alpha alleles, even if they were only haploidentical for HLA-B-DR haplotypes. In addition, heterozygosity for the TNF-alpha alleles was significantly more frequent in the patients. This tight linkage of TNF-alpha alleles with some extended haplotypes could help to explain the HLA-association of Type 1 diabetes as well as some other autoimmune diseases.

The pattern of insulin secretion was studied in 107 normal individuals aged 5 to 50 years. Intravenous glucose tolerance tests were performed on 64 islet-cell antibody negative siblings of diabetic children and on 43 normal adults. Puberty was staged using Tanner's criteria and subjects were grouped as follows: I - stage 1 (n = 22), II - stages 2 and 3 (n = 18), III - stages 4 and 5 (n = 20), IV - adults greater than 17 years (n = 47). Basal and stimulated (incremental 0-10 and 10-60 min areas) insulin responses rose throughout puberty (Groups I-III), declined following puberty until the third decade (Groups III and IV) and then appeared constant thereafter. Insulin levels in the 17.6-22.5 year group were lower than in the 12.6-15 year group (p less than 0.01). Fasting insulin to glucose ratios and incremental 0-60 min insulin to glucose area ratios produced a similar age-related pattern indicating that changes in insulin levels were independent of glucose concentrations. Gender did not affect these changes and multiple regression analysis showed that HLA haplotype sharing did not influence insulin responses in siblings of diabetic patients. Age and pubertal status must be carefully considered when interpreting intravenous glucose tolerance tests from patients suspected of having early abnormalities of carbohydrate metabolism.

HLA-DQ alpha and HLA-DX alpha gene polymorphisms were analyzed by Southern blot techniques in 78 Caucasoid insulin-dependent diabetes mellitus (IDDM) subjects and 55 control subjects. Five restriction fragment length polymorphisms of the HLA-DQ alpha gene correlated with HLA-DR typing. Two allelic DX alpha-related gene fragments, of 2.1 kb (U) and 1.9 kb (L) in size were identified. Genotype frequencies in the IDDM group for UU, UL, and LL were 54%, 38.5%, and 7.5%, respectively, whereas the corresponding frequencies in the control group were 24%, 40%, and 36% (P less than 0.00005 for differences in genotype frequencies). The U allele was associated particularly with IDDM patients who were DR3, with healthy controls who were DR3, as well as with IDDM patients who were not DR3. Thus, if this DX alpha U allele is not the DR3-associated IDDM susceptibility gene, it is the closest marker hitherto studied.

IgG and IgM class insulin autoantibodies were measured by an enzyme-linked immunosorbent assay in sera from members of the Barts-Windsor-Middlesex prospective family study for Type 1 (insulin-dependent) diabetes. One hundred and twelve individuals from 28 families were selected for study on the basis of a clearly defined islet cell antibody status. IgG insulin autoantibodies were found to be significantly associated with islet cell antibody positive (n = 30) versus islet cell antibody negative (n = 57) first degree family relatives (p = 0.002), with increased significance (p = 0.0003) if complement-fixing (CF)-islet cell antibody individuals (n = 20) only were considered. In addition, a significant association of IgG insulin autoantibodies with subsequent development of diabetes was observed within the CF-islet cell antibody positive group (p less than 0.0003). No such associations were found for IgM insulin autoantibodies, but a higher prevalence of these autoantibodies was observed in islet cell antibody negative first degree relatives (n = 57) compared with a control group of 73 Blood Bank donors (p = 0.00007), and they were significantly associated with siblings (n = 48) rather than parents (n = 39), (p = 0.001). We conclude that the presence of IgG insulin autoantibodies and CF-islet cell antibodies confer more risk for future development of diabetes than the presence of either marker alone.