Type 2 Diabetes: Multiple Genes, Multiple Diseases

Current Diabetes Reports August 2019, 19:55 | Cite as Type 2 Diabetes: Multiple Genes, Multiple Diseases AuthorsAuthors and affiliations Miriam S. Udler Open Access Pathogenesis of Type 2 Diabetes and Insulin Resistance (M-E Patti, Section Editor) First Online: 10 July 2019 538 Downloads Part of the following topical collections:Topical Collection on Pathogenesis of Type 2 Diabetes and Insulin Resistance Abstract Purpose of Review Type 2 diabetes (T2D), which accounts for the vast majority of diabetes cases, is essentially a diagnosis of exclusion in current clinical practice. Therefore, it is not surprising that T2D is heterogenous in terms of patients’ clinical presentation, disease course, and response to treatment. This review summarizes published attempts to improve diabetes subclassification, with a particular focus on the role of genetics. Recent Findings A handful of diabetes subclassification schemas have been proposed using clinical data (patient characteristics and laboratory values), with some subgroups associated with distinct management trends or complication risks. However, phenotypically driven classifications suffer from dependencies on time of variable measurement and are not readily linked to disease mechanism. Germline genetic data, in contrast, are essentially unchanged over a person’s lifetime and rooted in mechanism. Clustering of T2D genetic loci has identified at least five groupings of loci representing mechanisms of disease that may aid in deconstructing heterogeneity of T2D, but further work is needed to determine clinical utility. Summary Exciting progress in subclassification of diabetes has demonstrated initial steps in deconstructing disease heterogeneity. Incorporation of genetics into classification schemas will require additional research but has the potential to improve our understanding and management of T2D, both as a single disease and as a part of an integrated metabolic disease network. KeywordsType 2 diabetes Subtypes Genetics Disease pathways Polygenic risk score  This article is part of the Topical Collection on Pathogenesis of Type 2 Diabetes and Insulin Resistance Introduction In current clinical practice, when a patient develops elevated blood glucose indicative of diabetes, the diagnostic process of determining the diabetes “type” typically involves initially assessing for causes other than type 2 diabetes (T2D). For example, detection of autoantibodies may point to type 1 diabetes (T1D) or latent autoimmune diabetes in adults (LADA) or the presence of glucocorticoids on the medication list might suggest glucocorticoid-induced hyperglycemia. If a specific reason for hyperglycemia is not identified, a patient will generally then be considered to have T2D. Indeed, in practice, T2D is a diagnosis of exclusion, yet it currently is estimated to account for approximately 90% of all cases of diabetes [1]. It is not surprising, therefore, that T2D is a highly heterogenous condition with patients varying considerably in clinical presentation and response to treatment [2]. The heterogeneity observed among patients with T2D likely reflects variable contributions from diverse genetic and environmental factors [3], and ongoing efforts have aimed to utilize clinical and molecular data to develop a rational and reproducible categorization of diabetes. The goal of such subclassification is not only to refine patient diagnosis but also to better inform clinical management, specifically as it relates to prevention of diabetes complications. This review will summarize the diabetes-subtyping schemas that have been proposed as tools for deconstructing the heterogeneity of disease, with a particular focus on the role of genetics and its potential to shape our understanding and management of T2D, both as a single disease and as a part of an integrated metabolic disease network. Evidence for Type 2 Diabetes Constituting “Multiple Diseases” The stereotypical phenotype of a patient with T2D is someone obese with evidence of insulin resistance; however, diabetologists know well that not all patients with T2D fit this mold. Likewise, not all patients with presumed T1D present with diabetic ketoacidosis (DKA) and have positive islet autoantibodies [1]. This recognition of disease heterogeneity has prompted several efforts to refine the classification of diabetes. In 2003, Maldonado et al. presented the “AB” scheme, which was proposed to categorize patients presenting with DKA [4], and provided a useful construct for illustrating the diversity of these patients, who traditionally would have been assumed to have T1D. In 103 patients of various ethnic backgrounds who were admitted to the hospital with DKA, the authors assessed for presence of islet autoantibodies (A+/−) and evidence of beta-cell functional reserve (B+/−). They found that 50% of the patients were A−B+, 22% A−B−, 17% A+B−, and 11% A+B+ [4]. With only 17% of patients displaying antibody positivity and reduced beta-cell functional reserve (typical of T1D), the majority of these patients did not fit with the classic phenotypic picture associated with DKA. Furthermore, the substantial representation beyond both A−B+ (typical of T2D) and A+B− (typical of T1D) clearly demonstrated that patients with diabetes developing DKA did not fit neatly into well-established disease categories and that different pathophysiologies underlay their diabetes [5]. Of additional note, individuals in the B+ and B− groups also differed significantly by age of onset, glycemic control, and duration of insulin dependence, suggesting that recognition of subtype had clinical implications [4]. The utility in capturing beta-cell function in classifying diabetes “types” was also supported by a “beta-cell centric classification schema” later proposed by Schwartz et al., which conceptualized at least 11 pathways causing beta-cell dysfunction, each of which the authors envisioned could be targeted using a tailored treatment strategy [2]. Diversity among clinical phenotypes leading to development of diabetes was also evidenced in the work of Hulman et al. analyzing multi-point oral glucose tolerance tests (OGTT) in 5861 individuals without diabetes for whom longitudinal data was available. While typically only the fasting and two-hour time points of the OGTT are considered for diagnosing diabetes in non-pregnant adults, this analysis incorporated a third 30-min time point and fit latent class mixed-effects models across the three time points to identify four distinct glucose trajectory patterns [6]. Of particular interest was a subgroup (group 3) comprising 13% of individuals who had non-elevated 2-h glucose values, but elevated 30-min values; after up to 13 years of follow-up, individuals in group 3 were found to have a fourfold increased risk of developing diabetes and almost twofold all-cause mortality risk compared with individuals with similarly low 2-h glucose values, but who had (...truncated)