The metabolic consequences of type 1 diabetes result from progressive beta cell deficiency due to loss of functional beta cells. This section reviews the evidence that beta cell destruction is mediated by the immune system, and the associated histopathological changes in the pancreatic islets. The immunological mechanisms involved include humoral and cellular immune responses directed against beta cell constituents, with a further potential contribution from the innate immune system. Possible models of immune-mediated beta cell destruction are outlined, and the role of animal models of type 1 diabetes is described.
Aetiology and pathogenesis
The aetiology ('cause') of type 1 diabetes is unknown, but a good deal is known about its pathogenesis (the way in which it develops). As with other complex diseases, the outcome is determined by the interplay of multiple genes and (most likely) multiple environmental determinants, together with an element of happenstance. The risk of developing diabetes is strongly influenced by genes affecting immune function, particularly the HLA system, but other factors are involved. The best evidence for this comes from the study of identical twins: if one twin develops diabetes in childhood, the other (who has identical genes) has no more than a one in three chance of developing the disease. Prospective studies in human populations reveal that circulating autoantibodies directed against the islets typically appear in the first 5 years of life, and may be present for many years (sometimes 20 years or more) before the disease develops. A similar latent period is seen with coeliac disease, and implies that some sort of regulatory balance has been achieved within the immune system, and subsequently lost. This opens the door to the possibility of 're-education' of the immune system by various forms of immune intervention.
Relative insulin deficiency is found in both types of diabetes, but marked insulin deficiency with normal or near-normal insulin sensitivity is the hallmark of type 1 diabetes. Even low rates of insulin secretion are sufficient to inhibit ketogenesis in the liver, which is why diabetic ketoacidosis is characteristic of type 1 but not type 2 diabetes. Prospective studies of children at high risk of type 1 diabetes have shown that beta cell function declines over many months or years before clinical onset. The earliest clinical markers are loss of the first-phase insulin response (FPIR) to intravenous glucose, and the progressive development of glucose intolerance. Acute symptoms of thirst, polyuria and weight loss appear to develop suddenly (over days or weeks) against this background, sometimes precipitated by stress or intercurrent illness. Earlier investigators considered that some 90% of beta cells had been destroyed by the time of clinical diagnosis, but the beta cell reserve is now considered to be much greater than this. This could help to explain the 'remission' or 'honeymoon' phase sometimes seen following diagnosis, and provides much of the rationale for trials of immunotherapy in newly diagnosed patients.
Evidence for an immune pathogenesis
InsulitisAlthough autoimmunity may not cause type 1 diabetes, autoimmune processes undoubtedly mediate many of its pathological consequences. Evidence for this includes the overlap between type 1 diabetes and other HLA-associated autoimmmune conditions, the presence of insulitis (lymphocytic infiltration of the islets) in newly diagnosed children, evidence of humoral and cellular immunity directed against islet constituents, the observation of passive transfer of diabetes from marrow donors to recipients, the re-enactment of autoimmunity in twin-to-twin transplants, and the response to immunosuppressive therapy in the newly diagnosed. Animal models of immune-mediated diabetes such as the non-obese diabetic (NOD) mouse lend further support to the autoimmune hypothesis.
Post mortem studies of the pancreatic islets from patients with long-term type 1 diabetes show that they contain little or no insulin, implying that most functioning beta cells have been destroyed. Scattered functional beta cells may nonetheless be found even after 50 years. Other islet cells are unaffected. Studies of the pancreas in those who died soon after clinical presentation of diabetes show lymphocytic infiltration ('insulitis') consistent with an immune-mediated process, but these changes are less commonly present in older individuals.
The first evidence of circulating antibodies directed against the islets came from the demonstration of islet cell antibodies (ICA) by indirect immunofluorescence in 1974. ICA staining is a composite of antibodies directed against a variety of islet-associated molecular entities, and was superseded by specific assays as these entities came to be identified. Insulin autoantibodies (IAA) were reported in 1984, followed by antibodies directed against glutamic acid decarboxylase (GADA), islet antigen-2 (IA-2A) and the zinc transporter ZnT8. All these autoantigens are related to the beta cell secretory apparatus. Islet autoantibodies can typically be detected within the first few years of life, but may appear later; the appearance of multiple antibody species indicates established islet autoimmunity and strongly predicts the subsequent onset of diabetes.
Type 1 diabetes appears to develop as the consequence of an imbalance between pathogenic and regulatory T lymphocytes. CD4+ T cells and CD8+ T cells appear to mediate beta cell destruction, and diabetes can be induced in animal models by transfer of these cells. CD8+ T lymphocytes are capable of destroying beta cells by release of granules containing granzyme or perforin, or via Fas-ligand interactions. Circulating T lymphocytes directed against a wide range of islet antigens have been identified, but do not as yet offer the degree of specificity or disease prediction provided by measurement of islet autoantibodies.
Animal models of autoimmune diabetes
Two animal models of spontaneous autoimmune diabetes have been widely used: the NOD mouse and the BioBreeding (BB) Wistar rat. These provide useful analogies with the human situation, but differ from it in a number of important respects.