Lipidome's Role in Diabetic Retinopathy Subtypes via Inflammation: Mendelian Randomization and Mediation Analysis.
Xiaotong Xu, Nianen Liu, Kaixuan Dong, Shuai Ouyang, Weihong Yu
Abstract
Open AccessPurpose: This study aims to explore the causal relationships between plasma lipidome, inflammatory cytokines, and diabetic retinopathy (DR) subtypes using a two-sample Mendelian randomization (MR) approach and mediation analysis. Design: Mendelian randomization study and mediation analysis. Subjects: Genome-wide association study (GWAS) data of 179 plasma lipid species and 91 inflammatory cytokines from the public GWAS database. Genome-wide association study data of DR and its subtypes from the FinnGen consortium. Methods: Primary causal estimates were derived via inverse-variance weighted method, complemented by 4 sensitivity methods (weighted median, MR-Egger, simple mode, weighted mode). Mediation analysis was performed to determine the extent to which inflammatory cytokines mediate the effects of lipid species on DR subtypes. The Cochran Q-test, MR-Egger intercept test, and leave-one-out were used for sensitivity analyses. Mendelian randomization-Egger regression and Mendelian Randomization Pleiotropy RESidual Sum and Outlier were used to detect potential directional pleiotropy. We employed a two-sample MR analysis to assess the causal effects of specific lipid species on the risk of DR and its subtypes. Genome-wide association study summary statistics for lipid species and inflammatory cytokines were used. Mediation analysis was performed to determine the extent to which inflammatory cytokines mediate the effects of lipid species on DR subtypes. Main Outcome Measures: Genetic causal associations between lipid species, inflammatory cytokines, and DR subtypes. Results: Mendelian randomization analyses showed that most of the identified plasma lipids were significantly protective against DR and its subtypes, mainly belonging to the phosphatidylcholine and phosphatidylethanolamine classes. Triacylglycerols have different roles in different severities of DR, and inflammatory cytokines had different causal effects on DR and its subtypes. Mediation analyses identified 10 specific inflammatory cytokine-mediated pathways of lipid species on DR, of which 7 had about 10% of the mediating effect, with inflammatory cytokines in the remaining 3 playing mediating effects opposite to the pathways. Conclusions: This study highlights the complex interplay between lipid metabolism and inflammation in the pathogenesis of DR. Specific lipid species protect against DR, with inflammatory cytokines mediating these effects, suggesting potential therapeutic targets for DR management. Financial Disclosures: The author has no/the authors have no proprietary or commercial interest in any materials discussed in this article.