Integrative network toxicology and single-cell transcriptomics reveal TP53 as a key mediator of PCBs-induced microglial dysfunction in diabetic retinopathy

Diabetic retinopathy (DR), a leading cause of vision loss in diabetes, arises from intricate metabolic and environmental interactions. This study investigates how polychlorinated biphenyls (PCBs) contribute to DR pathogenesis. Network toxicology was employed to identify overlapping gene targets between PCBs and DR. Machine learning analyses subsequently refined these targets to four core genes: TP53, ESR1, EGR1, and HSPA5. Diagnostic modeling validated using human retinal transcriptomes demonstrated TP53's robust diagnostic accuracy, yielding area under the curve (AUC) values of 0.740 for non-proliferative DR (NPDR) and 0.920 for proliferative DR (PDR), with expression levels positively correlated with DR severity and ETDRS scores. Molecular docking confirmed strong binding affinities of toxic PCB congeners to TP53 and ESR1. Single-cell RNA Sequencing in a DR mouse model revealed enriched Trp53 expression in microglia, alongside microglial depletion and a pro-inflammatory shift. In vitro PCB138 exposure upregulated TP53 in high-glucose-cultured human microglial cells, promoting M1 polarization and cytokine secretion, effects that were attenuated upon pharmacological inhibition of p53 protein activity. These findings suggest that PCBs exacerbate DR through a TP53-driven pathway that promotes pro-inflammatory microglial activation, disrupting retinal homeostasis. TP53 emerges as a key biomarker and therapeutic target, highlighting the importance of reducing PCB exposure to mitigate DR progression.

Diabetic retinopathy; Machine learning; Microglia; Network toxicology; Polychlorinated biphenyls; TP53.