氨基酸与2型糖尿病因果关系的孟德尔随机化分析

A mendelian randomization analysis on the causal associations between amino acids and type 2 diabetes

目的 探讨氨基酸与2型糖尿病(T2D)的双向因果关联。方法 利用氨基酸和T2D全基因组关联研究汇总数据,通过逆方差加权法、加权中位数法、MR-Egger回归法、MR稳健调整轮廓得分法、MR多效性残差及异质性检测法(MR-PRESSO)进行两样本双向孟德尔随机化(MR)分析,并采用Cochran's Q检验、MR-Egger回归截距项检验、MR-PRESSO全局检验评估结果的异质性和水平多效性。结果 逆方差加权法结果显示,亮氨酸(OR=1.239,95%CI:1.078~1.424,P=0.002 5)、缬氨酸(OR=1.222,95%CI:1.071~1.394,P=0.002 9)会增加T2D发病风险,甘氨酸(OR=0.885,95%CI:0.831~0.942,P=0.000 1)则降低T2D发病风险。反向MR分析结果显示,T2D分别与亮氨酸、异亮氨酸、缬氨酸、丙氨酸、苯丙氨酸、酪氨酸的增加和甘氨酸的降低有因果关联。多种MR方法的结果基本一致,且敏感性分析未发现潜在的多效性。结论 较高水平的亮氨酸、缬氨酸及较低水平的甘氨酸与T2D发病风险具有双向因果关联,T2D对异亮氨酸、丙氨酸、苯丙氨酸、酪氨酸有正向因果效应。

Objective To investigate the bidirectional causal association between amino acids and type 2 diabetes(T2D).Methods Data of genome-wide association study on amino acids and T2D were collected.Inverse variance weighted method,weighted median method,MR-Egger regression method,MR robust adjusted profile score method and MR pleiotropy residual sum and outlier method(MR-PRESSO)were adopted for two-sample bidirectional mende-lian randomization(MR)analysis.Cochran's Q test,MR-Egger intercept test,and MR-PRESSO global test were used to evaluate the heterogeneity and pleiotropy of the results.Results The results of inverse variance weighted method showed that leucine(OR=1.239,95%CI:1.078-1.424,P=0.0025)and valine(OR=1.222,95%CI:1.071-1.394,P=0.0029)increased the risk of T2D,while glycine(OR=0.885,95%CI:0.831-0.942,P=0.0001)decreased the risk.The results of reverse MR analysis showed that the risk of T2D was causally associated with increased leucine,isoleucine,valine,alanine,phenylalanine and tyrosine,and decreased glycine.The results of multiple MR methods sug-gested robustness of causal associations and no potential pleiotropy was found in the sensitivity analysis.Conclusion Higher levels of leucine and valine and lower level of glycine have bidirectional causal associations with T2D risk.T2D has positive causal effects on isoleucine,alanine,phenylalanine and tyrosine.