The Evolution of Type 1 Diabetes

Abstract

T YPE 1 DIABETES (T1D) OCCURS IN INDIVIDUALS WITH a genetic predisposition to the disease, predominantly from a human leukocyte antigen (HLA)related immunogenotype that accounts for approximately 60% of the genetic influence. In these individuals who are genetically at risk, an environmental trigger is thought to initiate an immune response targeting the insulin-secreting pancreatic islet cells. The initial immune response also may engender secondary and tertiary immune responses that contribute to the impairment of -cell function and destruction of cells. The rate of development of T1D varies, probably related to non-HLA genetic factors and additional environmental factors beyond the triggering exposure. The initial laboratory manifestation of -cell injury is seroconversion, the appearance of diabetes-related autoantibodies. These autoantibodies likely do not mediate -cell injury but rather are markers of such injury. With progressive impairment of -cell function, metabolic abnormalities become measurable, initially either as loss of early insulin response to intravenous glucose or as reduced -cell sensitivity to glucose resulting in decreased insulin secretion. Subsequently, glucose abnormalities manifest as dysglycemia, either impaired fasting glucose (plasma glucose concentration of 100125 mg/dL) impaired glucose tolerance (2-hour plasma glucose concentration of 140-199 mg/dL following an oral glucose test), or indeterminate glucose tolerance (plasma glucose concentration of 200 mg/dL at 30, 60, or 90 minutes following an oral glucose test). Eventually, overt diabetes presents with glucose concentrations meeting current criteria, fasting plasma glucose concentration of 126 mg/dL or 2-hour plasma glucose concentration of 200 mg/dL. At the point of development of overt diabetes there is still evidence of persistent -cell function shown through measurement of C-peptide. C-peptide is a component of the insulin precursor molecule proinsulin, which is secreted by cells on an equimolar basis as insulin but is not contained in commercial insulin preparations, thus allowing measurement of -cell function even in patients treated with insulin. After diagnosis of T1D, however, there is a progressive decline in C-peptide as -cell function becomes more impaired or absent. Nonetheless, even many years after diagnosis, some patients with T1D may have low detectable levels of C-peptide. Since the sequence of events leading to T1D involves immunogenotypes detectable at birth or in infancy, several groups have followed birth cohorts with the genetic risk. Working together, Ziegler and colleagues analyzing cohorts in Colorado, Finland, and Germany report in this issue of JAMA the risk of progression to T1D from the time of seroconversion by measuring detectable diabetes autoantibodies. Remarkably, the results are similar in all 3 cohorts. Among 585 children who developed 2 or more diabetes-related autoantibodies, nearly 70% (280 of 401 available for follow-up) had developed T1D within 10 years, and 84% (299 of 355 available for follow-up) had developed T1D within 15 years. This is the first reportof a longitudinal cohort followed frominfancy throughthedevelopmentofT1D.Because theparticipantswere recruited from both the general population (in Colorado and Finland) and offspring of parents with T1D (in Germany), the similar findings take on added significance, suggesting that the same sequence of events occurs in individuals with socalled sporadic T1D and in relatives of individuals with T1D. The findings in this report raise additional questions. The basis for the differential progression to T1D according to the type of single autoantibody and the pattern of multiple autoantibodies should be explored. Also, the nearly 2-fold higher prevalence of HLA DR3/DR4-DQ8 in participants with multiple autoantibodies (n=331) than in those with a single autoantibody, and the similar prevalence between participants with a single autoantibody and those with no autoantibodies need to be better understood. In addition, would the identification of those with higher titers provide even a greater likelihood of progression to T1D? Muchattention inrecentyearshas focusedonwhetherT1D can be prevented. To that end, studies have been conducted in individualswith thegenetic riskalone(primaryprevention) and in individuals who have autoantibodies (secondary prevention). Although no study has of yet convincingly demonstrateddelayorpreventionofT1D,someevidencesuggests this may be possible. Primary prevention studies have shown the potential todelay theappearanceof autoantibodiesbyvarying infant formula used at the time of weaning from breast milk, either by eliminating cow’s milk and instead using a formula with casein hydrolysate or by removing bovine insulin from the formula. It is unknown whether such actions will delay progression to T1D, but an international effort—the National InstituteofChildHealthandHumanDevelopment–sponsored Trial toReduceInsulin-DependentDiabetesMellitus intheGenetically At-Risk (TRIGR) study—is under way to test the casein hydrolysate formula, with results expected in 2017.