Identification and Validation of Key Purine Metabolism-Related Genes in Ulcerative Colitis Using Bioinformatics and Machine Learning

Su Zhang,1,* Yifang Zhang,2,* Dongwei Du,2,* Yanling Zeng,3 Shengkai Zhang,1 Qinqin Wang,1 Wenjing Xue,1 Xiang Wen,4 Yi Lan,5 Weitao Hu2 1Department of Rheumatology, the Nanping First Affiliated Hospital of Fujian Medical University, Nanping, People’s Republic of China; 2Department of Gastroenterology, the Second Affiliated Hospital of Fujian Medical University, Quanzhou, People’s Republic of China; 3Department of Hematology, the Nanping First Affiliated Hospital of Fujian Medical University, Nanping, People’s Republic of China; 4Department of Pathology, the Nanping First Affiliated Hospital of Fujian Medical University, Nanping, People’s Republic of China; 5Department of General Practice, the Nanping First Affiliated Hospital of Fujian Medical University, Nanping, People’s Republic of China*These authors contributed equally to this workCorrespondence: Yi Lan, Email [email protected] Weitao Hu, Email [email protected]: Ulcerative colitis (UC) is a common inflammatory bowel disease with a complex pathogenesis that makes diagnosis and treatment difficult. Purine metabolism is closely related to many diseases, and its specific mechanism of action in UC remains unclear. The aim of this study was to find the relevant biomarkers of purine metabolism in UC.Methods: UC-related datasets downloaded from the Gene Expression Omnibus (GEO) database were used to screen for differentially expressed genes (DEGs). Weighted gene co-expression network analysis (WGCNA) was then performed to identify key module genes in UC. Then, further differentially expressed purine metabolism-related genes in UC were identified and defined as UCDE-PMRGs. Subsequently, functional enrichment of UCDE-PMRGs was performed. Next, three machine learning algorithms screened the key UCDE-PMRGs and further validated them in a separate validation cohort. We also utilized single-cell sequencing data to analyze the cellular distribution of key UCDE-PMRGs in the UC. Finally, the expression of key genes was validated in clinical samples, in vitro and in vivo experiments.Results: A total of 2133 DEGs and 9 UCDE-PMRGs were identified in UC. Machine learning was employed to identify the key UCDE-PMRG (PDE4B). PDE4B was significantly associated with immune infiltrating cells. Additionally, clinical samples validated that PDE4B is highly expressed in UC and positively correlated with disease activity. Furthermore, inhibiting PDE4B expression promotes intestinal epithelial barrier repair and alleviates symptoms in UC mice.Conclusion: PDE4B is a good biomarker related to purine metabolism in UC. Inhibiting PDE4B expression helps alleviate UC symptoms, providing a new approach to the pathogenesis and treatment of UC.Keywords: ulcerative colitis, purine metabolism, immune infiltration, PDE4B, machine learning, bioinformatics