Integrated Single-Cell Transcriptomics and Metabolomics Reveal Potential Associations Between Monascus pilosus Fermentation Metabolites and Diabetes-Related Pancreatic Transcriptional Networks

Diabetes mellitus is a global metabolic disease characterized by immune–metabolic dysregulation, yet its pathogenesis remains unclear. Natural products from the medicinal and edible fungus Monascus show hypoglycemic activity, but clinical translation is hindered by a lack of systematic understanding of cellular heterogeneity and molecular targets. We integrated public single-cell transcriptome data (GSE101207) and 21 in-house untargeted metabolomics datasets for Monascus. Single-cell analysis, metabolic communication inference, differential metabolome intersection, three machine learning algorithms (Lasso regression, support vector machine–recursive feature elimination, and random forest), network pharmacology, immune cell infiltration assessment, and virtual gene knockout analyses were performed. Pancreatic ductal cells exhibited the highest disease-association score with diabetes and were predicted to be highly involved in metabolite-mediated crosstalk. Four candidate overlapping metabolites and three candidate genes—hypoxanthine phosphoribosyltransferase 1 (HPRT1), mannosidase beta (MANBA), and ornithine decarboxylase 1 (ODC1)—were identified. HPRT1 was predicted to be associated with the IL-17 and TGF-β pathways; MANBA was predicted to be related to ECM–receptor interactions and TNF signaling; ODC1 was predicted to be linked to the NF-κB and NOD-like receptor pathways. These genes showed exploratory associations with immune subsets, with numerically higher CIBERSORT-estimated proportions of memory B cells and CD8+ T cells in the diabetic group, requiring further validation. Virtual gene knockout predicted that perturbing any of these genes reshaped the inflammation- and metabolism-related transcriptional network. This multidimensional framework suggests potential associations between Monascus fermentation-derived metabolites and diabetes-associated pancreatic transcriptional networks involving PDCs and the HPRT1/MANBA/ODC1 gene set. In vitro qRT-PCR in mouse pancreatic ductal epithelial cells partially supported the predicted upregulation of HPRT1 and ODC1 under diabetic-like conditions and their reduction after Monascus fermentation metabolite treatment. The study provides a hypothesis-generating framework for future experimental validation and a new model for elucidating natural product mechanisms.