Refinement of genetic risk assessment for type 1 diabetes in relatives and in the general population: composition of homogeneous study groups to evaluate interventions aiming at the prevention of diabetes.

Type 1 diabetes is presently an incurable disorder whose chronic invalidating complications cannot be prevented by insulin treatment alone. The absolute risk to develop type 1 diabetes before age 40 years in Belgium is about 0.4%. The chronic complications reduce the quality of life and life expectancy of the patient, and cause both psychosocial and financial problems . The majority of type 1 diabetes is the consequence of an interaction between partly identified genetic susceptibility factors and largely unknown environmental factors that initiate an immune-mediated attack on the pancreatic beta-cells. Determination of beta-cell-specific auto-antibodies allows to identify this immune reaction in the preclinical phase and constitutes the cornerstone of current prediction. Prevention or delay of clinical disease onset will have a beneficial effect on chronic complications. Presently, several non-toxic pharmacological interventions (e.g. prophylactic administration of insulin, nicotinamide) can prevent type 1 diabetes in animal models. These preventive interventions are currently under investigation in antibody-positive relatives from diabetic patients; effective interventions should ultimately be tested in the general population. However, in view of the excess of sporadic type 1 diabetes cases (85%) compared to familial cases, any predictive or preventive measurement should ultimately be applicable in the general population. However, type 1 diabetes displays a marked heterogeneity in terms of progression rate to beta cell-destruction during the preclinical phase. Although many antibody-positive individuals will develop the disease at very young age, the number of antibody-positive children is much higher than the number of patients who will develop the disease rapidly. Type 1 diabetes occurs most often in adulthood with, in general, a less severe phenotype (lower prevalence of ketosis, lower insulin requirements, larger residual betacell mass as refelected by higher residual C-peptide titers), and with lower titers of diabetes-specific auto-antibodies (directed against islets antigens, ICA; insulin, IAA; and IA-2 protein, IA-2-A). Som marker-positive individuals will develop hyperglycemia and clinical diabetes slowly or not at all, even without any preventive measurement. This heterogeneity implies that, in order to reveal any significant effect, large numbers of study subjects have to take part in prevention trials. As a consequence, some of these subjects would be needlessly exposed to preventive medication.

Several data suggest the involvement of HLA class II alleles on the modulation of the progression of beta-cell destruction. Combination of immune and genetic markers should allow the selection for prevention trials of study groups carrying a more homogeneously risk to develop diabetes. The practical implementation of this principle requires a better knowledge of the genetic determinants (± 15 loci) of diabetes initiation and progression. However, the predictive value of these loci for the future development of diabetes is rather weak. The most important loci are HLA class II genes on chromosome 6 (mainly HLA-DQ and, to a lesser extent HLA-DR genes) that represent together ~ 50% of the genetically determined risk; other genes from the HLA regio (class I and class III) and several non-HLA genes, including the well-characterised insulin promoter region (5' INS). Several of these genetic markers seem to be associated with a younger age at clinical onset (HLA-DQA1*-DQB1* 0301-0302/0501-0201 [class II], MICA [class I], with the male excess of diabetic patients (possible involvement of an X-linked locus), or with a lower residual beta-cell capacity at diagnosis (HLA-A24 [class I]). The associations between the MICA and HLA-A genes have not yet been studied sufficiently in HLA-DQ-identical study populations. However, in order to identify their role in the etiology of type 1 diabetes and in the progression rate of beta-cell destruction, the interactions of these genes or gene products (e.g. between DQ and 5'INS) together with the strong linkage disequilibrium between HLA loci (e.g. DQB1*03 and DRB1*04 [DR4]) must be taken into account. The assessment of the weak individual contributions of these additional HLA and non-HLA loci to the overall genetic risk is complicated by the relatively strong HLA-DQ genetic background. The prevalenve of the most specific genetic marker (heterozygote HLA-DQA1*-DQB1* genotype 0301-0302/0501-0201) is not higher than 2% in the general population compared to 30% in type 1 diabetic patients. The sole presence of this marker confers an absolute risk to develop type 1 diabetes not higher than 5-10%. Refinement of the genetic risk assessment by better knowledge of the interactions between different genetic risk factors and between genetic risk factors and immune markers at diagnosis and in the preclinical phase might improve the selection of first degree relatives for inclusion in prevention trials. The selection of at risk individuals from the general population might also benefit from this approach since initial genetic screening and subsequent monitoring for the appearance of antibodies is presently considered as the most appropriate strategy for selection of study subjects to be included in prevention trials for sporadic cases.