Integrative single-cell and bulk transcriptomics with Mendelian randomization reveals NAD-related biomarkers in diabetic retinopathy: Insights and experimental validation

Abstract

Background: Metabolic dysregulation involving nicotinamide adenine dinucleotide (NAD) has been implicated in the pathogenesis of various diseases; however, its role in diabetic retinopathy (DR) remains poorly understood. This study aimed to identify NAD-associated biomarkers and elucidate their molecular mechanisms in DR.

Methods: DR-related data were retrieved from public databases. NAD-associated biomarkers were identified through differential expression analysis, NAD-related gene (NAD-RG) score comparison, Mendelian randomization (MR) analysis, and experimental validation. Functional enrichment and immune microenvironment analyses were conducted to explore the mechanisms underlying biomarker involvement in DR. Single-cell RNA sequencing (scRNA-seq) was employed to identify key cell types, and pseudo-temporal trajectory analysis was performed to assess dynamic biomarker expression during cellular differentiation. Reverse transcription quantitative PCR (RT-qPCR) was used to validate biomarker expression levels.

Results: CCDC88A and GPD2 were identified as DR potential biomarkers and risk factors for disease progression. These genes were co-enriched in key pathways, including oxidative phosphorylation. CCDC88A exhibited the strongest positive correlation with monocytes (cor = 0.35, p < 0.05). scRNA-seq analysis highlighted microglia as key cells in DR. During microglia differentiation, CCDC88A expression initially decreased before rising, while GPD2 expression increased initially, declined, and then slowly increased in later stages. RT-qPCR confirmed significantly elevated expression of CCDC88A and GPD2 in the DR group (P < 0.05).

Conclusion: This study integrates bulk and single-cell transcriptomic analyses to identify CCDC88A and GPD2 as DR potential biomarkers and microglia as key cellular players, providing new insights into DR pathogenesis and potential diagnostic strategies.

Keywords: Biomarkers; Diabetic retinopathy; Mendelian randomization; Nicotinamide adenine dinucleotide; Single-cell sequencing analysis.