Association of the lipidome with type 1 diabetes and the mediated effect of metabolites: A Mendelian randomization study

Abstract

The effects of lipidome and their metabolites on type 1 diabetes (T1D) have not been entirely elucidated, and this study aimed to explore the causal effects of lipidome on T1D and the mediated effects of metabolites by Mendelian randomization (MR). Data on lipidome, metabolites, and T1D were obtained from the genome-wide association study, and single nucleotide polymorphisms were screened according to the basic assumptions of MR. Subsequently, inverse variance weighted was used to analyze the causal relationship between lipidome and T1D, as well as the mediated effect of metabolites. Finally, the horizontal pleiotropy, heterogeneity, and robustness of the results were assessed by MR-Egger intercept, Cochran Q, and leave-one-out sensitivity analysis, respectively. The MR analysis revealed that phosphatidylcholine (PC) (O-16:0_20:4) reduced the genetic susceptibility to T1D by increasing myristoyl dihydrosphingomyelin levels (d18:0/14:0) (mediated proportion: 39.10%, mediated effect: -0.021, 95% confidence interval [CI] -0.037 to -0.005, P = .011) and docosahexaenoylcholine levels (mediated proportion: 31.80%, mediated effect: -0.017, 95% CI -0.032 to -0.002, P = .027). Additionally, PC (16:1_20:4) reduced the genetic susceptibility to T1D by increasing myristoyl dihydrosphingomyelin (d18:0/14:0) levels (mediated proportion: 64.30%, mediated effect: -0.021, 95% CI -0.039 to -0.003, P = .024). MR-Egger intercept showed no horizontal pleiotropy in these results (P ≥ .05). Cochran Q demonstrated no significant heterogeneity in the MR results (P ≥ .05). Sensitivity analysis indicated that all results were robust. Our findings revealed pathways by which PC (16:1_20:4) reduced the risk of T1D by increasing myristoyl dihydrosphingomyelin and docosahexaenoylcholine levels, as well as PC (O-16:0_20:4) reduced the risk of T1D by increasing myristoyl dihydrosphingomyelin (d18:0/14:0) levels.

Keywords: Mendelian randomization; lipidome; mediated analysis; metabolites; type 1 diabetes.

Figure 1.

MR design diagrams of lipidome – metabolites – T1D. The a is the effect of lipidome on metabolites; the b is the effect of metabolites on T1D; the c is the total effect of T1D; the c′ is the direct effect of lipidome on T1D; the d is the total effect of T1D on lipidome. The algorithm for the mediated effect is a × b. The algorithm for the mediated proportion is (a × b)/c. MR = Mendelian randomization, T1D = type 1 diabetes.

Figure 2.

Forest plot of MR analysis for lipidome – metabolites – T1D. MR = Mendelian randomization, T1D = type 1 diabetes.

Figure 3.

Schematic diagram of the metabolites mediation effect. (A) The mediated effect of myristoyl dihydrosphingomyelin (d18:0/14:0) levels between phosphatidylcholine (16:1_20:4) levels and T1D; (B) the mediated effect of myristoyl dihydrosphingomyelin (d18:0/14:0) levels between phosphatidylcholine (O-16:0_20:4) levels and T1D; (C) the mediated effect of docosahexaenoylcholine levels between phosphatidylcholine (O-16:0_20:4) levels and T1D. The black arrow indicates that the exposure promoted the outcome, while the red arrow indicates that the exposure suppressed the outcome. T1D = type 1 diabetes.