Purpose: Preoperative anti-vascular endothelial growth factor (VEGF) therapy in proliferative diabetic retinopathy (PDR) may exacerbate fibrovascular contraction, posing significant challenges. This study aims to characterize proteo-metabolomic changes induced by anti-VEGF therapy, validate key candidates on orthogonal analytical platforms, and provide functional evidence for the lead molecule. Methods: Aqueous humor from 25 patients with PDR with simple vitreous hemorrhage (VH; n = 13) or VH with tractional retinal detachment (VH + TRD; n = 12) underwent data-independent acquisition (DIA) proteomics and widely targeted metabolomics before and 7 days after intravitreal aflibercept. Molecules altered in both cross-sectional and longitudinal comparisons were classified as "aggravation-type" or "improvement-type." Aggravation-type molecules were validated by parallel reaction monitoring/selected reaction monitoring (PRM/SRM) in an independent cohort (cataract controls, n = 10; simple VH, n = 5; VH + TRD, n = 5), and GALNS was functionally assessed by siRNA knockdown in fibroblasts. Results: Intersection analysis identified 38 molecules (5 proteins and 33 metabolites): 35 improvement-type and 3 aggravation-type (CAST, GALNS, and 3-hydroxypropanoic acid 3-HPA). Independent validation using PRM proteomics and SRM metabolomics confirmed that only GALNS was robustly validated across both cross-sectional (P = 0.016) and longitudinal (P = 0.049) comparisons, whereas CAST and 3-HPA did not reach statistical significance in validation. GALNS exhibited a biphasic pattern-elevated in PDR relative to cataract controls yet depleted in VH + TRD and further declined post-treatment. GALNS knockdown in fibroblasts upregulated α-SMA and collagen I, accelerated migration, and enhanced contractility. Conclusions: This multi-omics study, reinforced by independent validation, reveals GALNS depletion as a potential pro-fibrotic molecule following anti-VEGF therapy in PDR, offering a promising candidate for perioperative risk stratification. CAST and 3-HPA remain exploratory candidates requiring validation in larger cohorts.
Huang et al. (Mon,) studied this question.