Characterization of diabetes in early adulthood

Original article:
Relation between disease phenotype and HLA-DQ genotype in diabetic patients diagnosed in early adulthood. Weets I, Siraux V, Daubresse J-C, De Leeuw IH, Féry F, Keymeulen B, Krzentowski G, Letiexhe M, Mathieu C, Nobels F, Rottiers R, Scheen A, Van Gaal L, Schuit FC, Van der Auwera B, Rui M,De Pauw P, Kaufman L, Gorus FK, for the Belgian Diabetes Registry. J Clin Endocrinol Metab 2002; 87(6): 2597–605.

Summary
A population of 1584 Belgian patients with recent-onset diabetes, aged 15–39 years, was characterized in relation to islet autoantibodies and HLA-DQ genotypes associated with type 1 diabetes. Islet antibody-positive patients (n = 1198) compared with islet antibody-negative patients (n = 386) had distinct clinical, metabolic and genetic features of type 1 diabetes. The islet antibody-positive group was younger and more symptomatic at diagnosis, had lower BMI and random C-peptide, higher glycemia and insulin requirement, ketonuria, and a higher prevalence of HLA-DQ and 5' insulin gene susceptibility genotypes. Within the antibody-positive population, these characteristics did not vary according to HLA-DQ susceptibility genotypes. In contrast, in the antibody-negative group, HLA-DQ susceptibility genotypes were associated with more clinical features of type 1 diabetes. The authors conclude that the association of susceptibility HLA-DQ genotype and clinical features of type 1 diabetes in antibody-negative patients supports the view that a subgroup of seronegative patients may have an immune-mediated disease.

Comment
Type 1 diabetes is an autoimmune disease characterized by circulating islet autoantibodies and an increased frequency of specific HLA genes, represented by the highest risk haplotype DR3,4; DQ2,8. Islet autoantibodies include islet cell antibody detected on frozen sections of pancreas by indirect immunofluorescence and antibodies specifically to insulin, glutamic acid decarboxylase and tyrosine phosphatase insulinoma-like antigen IA-2. HLA genes encode proteins that bind and present antigenic peptides on the cell surface to T cells. Therefore, the association of specific HLA proteins with type 1 diabetes implies that they present autoantigen peptides to activate T cells that are ‘diabetogenic’ or to inhibit T cells that are anti-‘diabetogenic’.
b-Cell destruction in type 1 diabetes is believed to be T cell-mediated. This viewpoint is reinforced by a recent case report from Martin et al. [1] of a boy with X-linked agammaglobulinemia who developed type 1 diabetes. Although islet autoantibodies are specific markers, up to 10% of Caucasoid patients presenting with the classical features of type 1 diabetes have no detectable islet autoantibodies. Antibody-negative, insulinopenic patients presenting with acute diabetes in adulthood have also been well documented in Japan [2].
Weets et al. analysed data from a large group of patients and determined that the HLA-DQ susceptibility haplotypes were over-represented in islet autoantibody-negative patients with clinical features of type 1 diabetes. Their suggestion that this indicates an immune-mediated process is reasonable, given the central role of HLA proteins in mediating immune responses. These islet autoantibody-negative patients may have pathogenic anti-islet T cell responses but only transient or no detectable antibody responses. Alternatively, they may have as yet unknown antigen specificities not revealed by current islet cell antibody assays.
The higher frequency of HLA-DQA1-DQB1 risk haplotypes (0301–0302 and/or 0501–0201) in antibody-positive patients reported by Weets et al. is consistent with their known association with classical type 1 diabetes. Nevertheless, 43% of antibody-positive patients lacked HLA-DQ susceptibility haplotypes. HLA risk is attributable not only to HLA-DQ status but also to DR status independent of linkage disequilibrium with DQ [3], which was not analysed. Given the weak association between HLA-DQ haplotype status and clinical features of type 1 diabetes in their antibody-positive subjects, Weets et al. postulate that HLA-DQ may not determine disease severity once b-cell autoimmunity is initiated. This could be age-related, because their study group included older patients. The reported association between high-risk HLA alleles and an earlier age at clinical diagnosis [4, 5] (potentially more aggressive autoimmune-mediated
b-cell destruction) appears not to hold after about age 20. Consistent with this, they found that antibody-positive patients developed diabetes overall at a younger age regardless of HLA-DQ haplotype, whereas in the absence of islet antibodies, patients with HLA-DQ risk haplotypes had an earlier onset of disease.
Weets et al. cite evidence that a majority of cases of type 1 diabetes present after the age of 15. However, distinguishing type 1 from type 2 diabetes, particularly in early adulthood, is difficult in the absence of immune markers. There is no argument about the need for additional markers of type 1 diabetes, apart from islet antibodies, to define clinical forms of the disease. The obvious markers are islet antigen-specific T cells. Unfortunately, however, reliable assays for these, from the only accessible human tissue (blood), are not yet a reality. New genetic markers would also be useful to define subtypes of antibody-negative patients and predict disease progression and time to insulin requirement. This study serves to remind us that many pieces of the pathogenetic puzzle of type 1 diabetes are still missing.