Full text 2026

m5C: Novel Diagnostic and Drug Repurposing Targets for Nonalcoholic Steatohepatitis

Chen S, Duan R, Qiu J, et al.

Full text

Loading PDF… Expand reader Download

Abstract

<h4>Background</h4>Epigenetic medication, such as RNA 5-methylcytosine (m5C), is well-recognized as a key regulator in hepatic metabolism and immune responses. However, m5C regulatory mechanisms in NASH pathogenesis have not yet been clearly elucidated.<h4>Methods</h4>By utilizing three bulk profiles of NASH patients acquired from GEO and integrative bioinformatic pipelines, such as Limma framework, consensus clustering, and machine learning, we first identified m5C-related molecular subgroups and hub genes for NASH patients. Besides, diagnostic performance and biological characteristics of m5C-related hub gene were estimated at bulk level. Indeed, the heterogeneity of m5C-related hub gene for NASH patients was deciphered in single-cell transcriptomic profiles at temporal and spatial manners, especially in artificial intelligence (AI)-driven virtual cells. Furthermore, potential therapeutic agents targeting m5C-associated hub genes for the treatment of NASH were enriched by AI-driven drug enrichment framework (DrugReflector) based on NASH bulk profile and then validated by molecular docking. Finally, in vitro studies quantified the expression of m5C-associated hub genes compared to normal control.<h4>Results</h4>m5C can divide NASH patients into two various consensus groups with different molecular and immune patterns. Furthermore, ERCC2 and FOXC2 can be considered two upregulated m5C-associated hub genes involved in NASH pathogenesis, which were mainly distributed at cholangiocyte. BRD-K93672499 can be considered a multitarget therapeutic strategy targeting ERCC2 and FOXC2 for the treatment of NASH.<h4>Conclusion</h4>Our study first deciphered the m5C in predictive and therapeutic potential for NASH patients, which gains more insight into their personalized and precision medicine.

Keywords

Nonalcoholic steatohepatitis Machine Learning Drug Repurposing M5c Multiomics