Integrative Transcriptomics and Machine Learning Reveal the Association of CBX4 with Inflammation in Ulcerative Colitis as a Potential Epigenetic Regulator.

Background/Objectives: Epigenetic factors are increasingly recognized to contribute to the pathogenesis of intestinal diseases, yet the precise mechanisms through which these factors influence ulcerative colitis (UC) remain poorly understood. Methods: Transcriptome profiles pertaining to UC and genes associated with epigenetic factors (EFRGs) were retrieved from publicly accessible datasets. Candidate genes were ascertained through the intersection of differentially expressed genes (DEGs) and EFRGs. Key genes were screened through machine learning algorithms and validated via the Artificial Neural Network (ANN) model. Enrichment analysis and immune infiltration assays were conducted to elucidate the underlying mechanisms of these genes. The hub gene, CBX4 (Chromobox homolog 4), was validated through immunohistochemical analysis of both healthy controls and patients with UC, and the correlation was evaluated using UC-related clinical parameters. Additionally, CBX4 expression was knocked down in dextran sulphate sodium (DSS)-treated mice to examine its regulatory function. Unlike conventional broad-spectrum biomarker screens, this study specifically integrated epigenetic factor-related genes (EFRGs) with machine learning and experimental validation using both clinical samples and animal models. Results: SMARCB1, JAK2, CBX4, and PPARGC1A were identified as key genes, with SMARCB1, JAK2, and CBX4 being upregulated in the UC group, while PPARGC1A was significantly downregulated. The ANN model exhibited excellent diagnostic performance. Enrichment analysis revealed that the key genes were associated with pathways such as the "chemokine signaling pathway". Immune cell infiltration analysis results revealed marked differences in the abundances of 13 immune cell types between the UC and control groups, and there were notable associations between immune cell infiltration and key genes. Notably, CBX4 expression was elevated in both DSS-treated mice and patients with UC, showing positive correlations with clinical indicators of UC. Further in vivo experiments revealed that silencing CBX4 alleviated DSS-induced colon damage and inflammation. Conclusions: This study identifies four EFRG-related key genes (SMARCB1, JAK2, CBX4, PPARGC1A) in UC, suggesting that CBX4 may play a significant role as an epigenetic regulator. CBX4 is upregulated in UC intestinal tissues, and its knockdown mitigates DSS-induced colitis. These findings provide critical theoretical support for developing targeted therapies for UC.