Novel insights into the mechanisms of hard exudate in diabetic retinopathy: Findings of serum lipidomic and metabolomics profiling.

Objective: Retinal hard exudates (HEs) result from lipoproteins leaking from capillaries into extracellular retinal space, and are related to decreased visual acuity in diabetic retinopathy (DR). This study aims to identify differential serum lipids and metabolites associated with HEs.

Materials And Methods: A cross-sectional study was conducted Jul 2017 ∼ Mar 2021. We assessed the amount of HEs using standard ETDRS photographs for comparison. HEs severity was rated as "no or questionable", "moderate" or "severe". Serum samples were processed via high coverage pseudotargeted lipidomics analysis, and untargeted liquid chromatography coupled with time-of-flight mass spectrometry for metabolomics study, respectively. Weighted gene co-expression network analyses, partial least squares-discriminant analysis, and multi-receiver operating characteristic analysis were applied.

Results: A total of 167 patients were included. Discovery group: 116 eyes (116 patients). Validation group: 51 eyes (51 patients). 888 lipids were detected and divided into 18 modules (MEs), ME1 ∼ ME18. Lipids in ME1 significantly increased in patients with HEs in DR (NPDR and PDR combined), NPDR, and PDR, respectively. ME1 enriched to triglycerides (29%), ceramides (17%), and -acylethanolamines (15%). A combined model of 20 lipids was the best to discriminate HEs, area under curve = 0.804, 95% confidence interval = 0.674-0.916. For metabolomics analysis, 19 metabolites and 13 pathways associated with HEs were identified. Taurine and hypotaurine metabolism, cysteine and methionine metabolism were closely related to HEs ( < 0.01).

Conclusions: The lipids and metabolites identified may serve as prediction biomarkers in the early stage of HEs in DR.