Review Precision medicine in the management of type 2 diabetes Anna L Gloyn, Daniel J Drucker The study of type 2 diabetes has been driven by advances in human genetics, epigenetics, biomarkers, mechanistic Lancet Diabetes Endocrinol 2018 studies, and large clinical trials, enabling new insights into disease susceptibility, pathophysiology, progression, and Published Online development of complications. Simultaneously, several new drug classes with different mechanisms of action have April 23, 2018 been introduced over the past two decades, accompanied by data about cardiovascular safety and non-glycaemic http://dx.doi.org/10.1016/ S2213-8587(18)30052-4 outcomes. In this Review, we critically examine the progress and integration of this new science into clinical practice, Oxford Centre for Diabetes, and review opportunities for enabling the use of precision medicine in the diagnosis and treatment of type 2 diabetes. Endocrinology and We contrast the success in delivering personalised medicine for monogenic diabetes with the greater challenge of Metabolism, Radcliffe providing a precision medicine approach for type 2 diabetes, highlighting gaps, limitations, and areas requiring Department of Medicine, and further study. Wellcome Centre for Human Genetics, University of Oxford, Oxford, UK Introduction a marked sensitivity to sulfonylureas, a finding that was (Prof A L Gloyn DPhil); NIHR There has been much interest in our increased ability to elegantly substantiated in a randomised controlled trial Oxford Biomedical Research incorporate data from human genetics, along with that provided the first example of personalised medicine Centre, Churchill Hospital, Oxford, UK (Prof A L Gloyn); 3 lifestyle and environmental information, to individualise in diabetes. The precise molecular mechanism for and Department of Medicine, treatment decisions. Although not a novel concept, the sensitivity to this class of oral hypoglycaemic drugs in Lunenfeld-Tanenbaum technological advancements that have driven the omics patients with HNF1A mutations remains elusive but it Research Institute, Mount Sinai revolution have provided the impetus and expectation has been argued that closure of the ATP­sensitive Hospital, University of Toronto, Toronto, ON, Canada that a new dawn in precision medicine is approaching. In potassium (KATP) channel by sulfonylureas bypasses the (D J Drucker MD) this Review, we use the term precision medicine to major sites of β­cell dysfunction, which are upstream of Correspondence to: describe the stratification of patients into groups on the the KATP channel, thus reigniting insulin secretion. Dr Daniel J Drucker, Mount Sinai basis of molecular and genetic biomarkers and clinical Sulfonylureas are also the first­line treatment for indivi­ Hospital, Toronto, ON M5G 1X5, characteristics to optimise therapeutic outcomes. The duals with neonatal diabetes due to activating mutations Canada [email protected] scope of precision medicine has appropriately been in the KCNJ11 and ABCC8 genes, which prevent the KATP broadened to not only include genetics but also channels from closing in response to ATP generated by environmental and lifestyle factors, with recommen­ glycolysis.4–7 Closure of the channel by an ATP­ dations targeting groups rather than individual patients. independent mechanism circumvents the cause of the In this Review, we examine the evidence for precision β­cell dysfunction, restoring insulin secretion in these medicine in diabetes, focusing on its success in individuals and, remarkably, improving their glycaemic monogenic diabetes, the emerging evidence for type 2 control. diabetes, and identifying challenges for enabling Personalised medicine in monogenic diabetes extends improved individualised treatment recommendations in beyond treatment response: for example, individuals with the future. loss­of­function mutations in the GCK gene have stable fasting hyperglycaemia and are unlikely to develop Precision medicine in monogenic versus diabetic complications.8 For neonatal diabetes, non­ polygenic diabetes selective genetic testing now provides information about The past 10 years have seen unprecedented progress in the future development of additional clinical features, our understanding of the genetic basis of diabetes. With disease management, and likelihood of disease remission.9 this progress has come an expectation among the clinical To date, however, genetic discovery efforts for community that our ability to diagnose, characterise, and type 2 diabetes have had limited translational benefits. treat patients with diabetes will be transformed. Given Reconciling the different bench­to­bedside trajectories of the well documented success in rapid translation of gene monogenic versus complex forms of diabetes can be discovery efforts for monogenic varieties of diabetes into achieved by appreciating the differences in the genetic changes in clinical care pathways, this belief is under­ contributions to the development of diabetes, which standable. Up to 3% of cases of diabetes diagnosed in makes the eventual classification of type 2 diabetes into children have a monogenic basis, with the most common multiple discrete subtypes less likely than into monogenic cause being mutations in the transcription factor gene forms of diabetes (figure). As researchers have tried to HNF1A.1 Individuals with rare penetrant mutations in understand how genetics might be used to support HNF1A typically present before age 25 years with non­ precision medicine, several theories have been proposed insulin­dependent diabetes, are slim, and have a strong for complex diseases such as type 2 diabetes and coronary family history of diabetes. Clinical investigation usually artery disease. Two untested models10,11 that have garnered reveals that such individuals are autoantibody negative interest propose that patients with type 2 diabetes or but C­peptide positive. An early case study2 showed that coronary artery disease are likely to have a diverse set of individuals with diabetes due to an HNF1A mutation had overlapping mechanisms for their inability to regulate www.thelancet.com/diabetes-endocrinology Published online April 23, 2018 http://dx.doi.org/10.1016/S2213-8587(18)30052-4 1 Review A B Single gene disorder Complex trait Examples of monogenic subtypes of diabetes GCK HNFA KCNJ PPARG/LMNA Glucose sensing Insulin secretion Channel disorder Ectopic fat Examples of treatment No treatment Low dose High dose Metformin sulfonylurea sulfonylurea Insulin sensitisers Sulfonylureas Incretin therapies Figure: Precision medicine approaches for the treatment of type 2 diabetes (A) Treatment of patients with selected monogenic subtypes of diabetes. (B) Individuals with complex genetic determinants of type 2 diabetes. Monogenic diabetes subtypes have specific treatments whereas complex trait type 2 diabetes is a sliding scale making individualised treatment more difficult. Many subtypes of monogenic diabetes can be treated with specific pharmcological agents, whereas individuals with type 2 diabetes have a heterogeneous phenotype, with different degrees of abnormalities in multiple pathways. their blood glucose or lipid concentrations, respectively, in about 9000 newly diagnosed patients with diabetes, and that these combinations will be difficult to tease the researchers were able to identify five subtypes of apart into discrete subtypes. Both models propose that diabetes based on clinical variables (autoantibodies, age what matters is where an individual is on a series of at diagnosis, BMI, HbA1c, and estimates of β­cell function hypothetical sliding scales of intermediate phenotypes and insulin resistance), which were replicated in for particular pathophysiological processes (eg, β­cell independent cohorts. In support of the presence of mass, insulin sensitivity, and incretin response in different pathophysiological features, the clustering of patients with type 2 diabetes, or LDL cholesterol con­ established genetic associations differed between these centrations in patients with coronary artery disease).10,11 If subtypes. In line with different degrees of shared patho­ we can understand the composition of the pallette or physiology, the risks of complications differed across the blend and where individuals are located on these sliding subtypes. Notably, diabetic retinopathy was identified scales then we can make more tailored treatment choices, earlier in patients with relative insulin deficiency, particularly when identifying patients who are unlikely to whereas patients with insulin­resistant diabetes showed respond to a particular therapy. For example, people with an increased risk of developing diabetic kidney disease.14 diabetes largely due to loss of β­cell mass are unlikely to These findings, if replicated in ethnically diverse respond to drugs that increase insulin secretion. There is populations across a range of age groups, might be of already evidence for this hypothesis from studies potential value for clinical trial enrolment or for early showing the reduced efficacy of glucagon­like peptide­1 treatment stratification, perhaps with drugs such as (GLP­1) receptor agonists in insulin­treated patients with sodium­glucose cotransporter 2 (SGLT2) inhibitors, low C­peptide concentrations and autoantibodies.12 which show potential for modification of renal out­ Encouraging evidence for the clinical utility of genetic comes.15 Future studies will now need to establish risk profiling has emerged from efforts to understand whether