Diabetic Complications and Genetic Variants in the Million Veterans Program

糖尿病(DM)并发症是其发病率、死亡率和费用的主要原因。MVP009有 对潜在的遗传学有深入的了解。由于糖尿病护理没有利用在 遗传学，我们建议使用遗传学来支持临床翻译和机械发现。 在MVP009中，我们在全基因组关联研究中利用了高度特异的表型。 (I)射血分数保留或减少的心力衰竭(HF)；(Ii)低血糖--严重(急诊) (3)肾脏疾病；(4)眼科疾病。我们还发现，尽管是典型的， “青少年起病的1型糖尿病(T1D)”不包括服兵役，至少10%的退伍军人被认为患有T2D MVP中可能有“成人起病的T1D”--很大程度上无法识别。我们现在建议扩大这些研究结果。 与糖尿病中的精准医学共识报告一致，我们的目标是在 (一)诊断(遗传T1D与T2D)，(二)治疗[结合遗传和传统风险因素(RF)]，和(三) 预后(表观基因组对并发症的贡献)--结合遗传学，以便提供更多的护理 准确和个性化，并确定可能导致新治疗方法发现的机制。 目的1：评估T1D和T2D基因负荷对临床特征的影响 以及推定患有T2D的人的疾病轨迹。我们将用T1D对多种族遗传风险进行建模 和T2D多基因风险评分(PR，来自最近大型研究的多种族数据)；每个MVP老将 既有T1D也有T2D PR。结果将包括糖尿病事件和疾病轨迹：年龄和BMI 发病时出现胰岛素Rx、酮症酸中毒和低血糖。我们会评估该计划的效用，以 确定患有糖尿病的退伍军人，他们将从明确的T1D测试和/或早期使用胰岛素中受益。 目的2：评估遗传性/非遗传性RF对糖尿病发展的综合贡献 并发症。(SubAim A)确定RF和并发症之间的影响修正。遗传 交互分析将包括生活方式、人口统计和共病(例如，血压、糖化血红蛋白)，如 疾病位点和PR带来的并发症风险的修饰物。我们将使用这两种假设检验 已知基因座和PR的方法，以及假设生成方法(使用全基因组G×E 建模)以检查与糖尿病眼病(DED)、肾脏疾病(DKD)、心力衰竭和 低血糖和使用最先进的孟德尔随机化(MR)的因果联系，包括 多变量和中介先生(SubAim B)开发和测试预测模型。我们将使用摘要 来自MVP GWAS的统计数据和文献，以使用最近的“最佳实践”开发单独的PR 方法，用于DED、DKD、HF和低血糖，并用PRSS来阐明以前未知的 RFS。利用PRS、Phewas、来自SubAim(A)的信息、临床RF和治疗，我们将开发 基因组信息预测模型将在Emerge和更新的MVP参与者中进行评估。 目的3：确定糖尿病并发症的表观基因组标记和分子系统。 表观基因组变化调节基因表达，可以调节环境和生理效应，并具有 与T2D和相关的血糖特征有关。我们假设差异甲基化也将是 与糖尿病并发症相关。使用EPIC芯片(850,000个位点)的甲基化信息将是 适用于30,000名退伍军人，并可归因于其他退伍军人，允许表观基因组和多基因组 方法，如聚集分析和表观基因组范围的关联研究，以(I)确定与 糖尿病的并发症以及低血糖，以及(Ii)确定涉及的基因和途径。 影响：诊断遗传学、G×E、表观基因组学和预测模型都应该有所帮助 识别个体风险，并帮助个性化治疗以减少糖尿病并发症和 低血糖-并支持发现新的治疗方法来缓解潜在的过程。

Diabetes (DM) complications are the major cause of its morbidity, mortality, and costs. MVP009 has advanced understanding of the underlying genetics. Since DM care does not take advantage of progress in genetics, we propose to use genetics to support both clinical translation and mechanistic discovery. In MVP009, we utilized highly specific phenotypes in genome-wide association studies (GWAS) of (i) heart failure (HF) with preserved vs. reduced ejection fraction; (ii) hypoglycemia – severe (emergency visits) and incidental (outpatient visits); (iii) kidney disease; and (iv) eye disease. We also found that although typical, “juvenile-onset type 1 diabetes (T1D)” excludes military service, at least 10% of Veterans with presumed T2D in MVP may have “adult onset T1D” – largely unrecognized. We now propose to extend these findings. Consistent with the Precision Medicine in Diabetes Consensus Report, our Aims target precision in (i) diagnosis (genetic T1D vs. T2D), (ii) treatment [combining genetic with traditional risk factors (RF)], and (iii) prognosis (epigenomic contributions to complications) – to incorporate genetics so that care can be more accurate and individualized, and identify mechanisms that can lead to discovery of new treatments. Aim 1: Assess the contributions of T1D and T2D genetic loads to the clinical characteristics and disease trajectories of people presumed to have T2D. We will model multiethnic genetic risk with T1D and T2D polygenic risk scores (PRS, with multiethnic data from large recent studies); each MVP Veteran will have both a T1D and a T2D PRS. Outcomes will include incident DM, and the disease trajectory: age and BMI at onset, time to insulin Rx, and ketoacidosis and hypoglycemia. We will evaluate the utility of the PRS to identify Veterans with DM who would benefit from definitive T1D testing and/or early use of insulin. Aim 2: Assess the combined contributions of genetic/nongenetic RF to development of complications. (SubAim a) Identify effect modifications between RF and complications. Genetic interaction analyses will include lifestyle, demographics, and comorbidities (e.g., blood pressure, HbA1c), as modifiers of the risk of complications conferred by disease loci and PRS. We will use both hypothesis-testing approaches for known loci and PRS, and hypothesis-generating approaches (using genome-wide G×E modeling) to examine interactions associated with diabetic eye disease (DED), kidney disease (DKD), HF, and hypoglycemia, and causal associations using state-of-the-art Mendelian Randomization (MR), including multivariable and mediation MR. (SubAim b) Develop and test predictive models. We will use summary statistics from the MVP GWAS and the literature, to develop separate PRS using the “best practice” recent method, for DED, DKD, HF, and hypoglycemia, and PheWAS with the PRSs to elucidate previously unknown RFs. Utilizing the PRS, PheWAS, information from SubAim (a), clinical RF, and treatments, we will develop genome-informed predictive models that will be evaluated in eMERGE and more recent MVP participants. Aim 3: Identify epigenomic markers and molecular systems underlying DM complications. Epigenomic changes regulate gene expression, can mediate environmental and physiologic effects, and have been associated with T2D and related glycemic traits. We hypothesize that differential methylation will also be associated with DM complications. Methylation information using the EPIC chip (>850,000 sites) will be available on >30,000 Veterans, and can be imputed in other Veterans, allowing epigenomic and multi-omic methods such as aggregation analysis and epigenome-wide association studies to (i) identify associations with the complications of DM as well as hypoglycemia, and (ii) identify the genes and pathways involved. Impact: The genetics of diagnosis, G×E, epigenomics, and predictive models should both aid translation – to identify risk in individuals, and help personalize treatment to reduce DM complications and hypoglycemia – and support discovery of new therapies to mitigate the underlying processes.