Dan Jian, Jian Chen, Jinping Yuan, Kunwar Namrata, Dan Su, Bingxue Bai

Department of Dermatology, The Second Affiliated Hospital of Harbin Medical University, Harbin, 150000, People’s Republic of China

Correspondence: Bingxue Bai, Department of Dermatology, The Second Affiliated Hospital of Harbin Medical University, 246 Xuefu Road, Nangang, Harbin, 150000, People’s Republic of China, Tel +86-15114517408, Email baibingxue@hrbmu.edu.cn

Background: Atopic dermatitis (AD) and ulcerative colitis (UC) are increasingly prevalent, and growing evidence suggests a potential link between them. However, the shared pathogenesis remains unclear. This study aims to identify the hub genes and immunological features underlying the connection between AD and UC.
Methods: In this study, we downloaded the GSE121212 and GSE75214 datasets from the GEO database. Differentially expressed genes (DEGs) were identified through DESeq2 and limma R packages, revealing enrichment in immune-related pathways such as neutrophil migration and chemokine signaling. Subsequently, weighted gene co-expression network analysis (WGCNA) and protein-protein interaction (PPI) network analysis were performed to identify common candidate genes. Three machine learning algorithms were employed to select the hub gene, and single-cell RNA sequencing was used to validate the findings.
Results: Seven common candidate genes (CXCL1, CCL20, CXCL2, ZC3H12A, PI3, CXCL3, LCN2) were identified, showing significant expression differences in both diseases and associations with immune cell infiltration. Among them, PI3 emerged as the hub gene with strong diagnostic potential (AUC > 0.95) based on three machine learning models. Single-cell RNA sequencing supported high PI3 expression in UC intestinal epithelial cells and AD keratinocytes, with correlations to certain CCL and CXCL chemokines. These chemokines play overlapping roles in recruiting M1 macrophages, CD4 T cells, and neutrophils, thereby regulating inflammation in both AD and UC.
Conclusion: Our study uncovers shared immune cell recruitment mechanisms in both diseases, suggesting that CCR1 may serve as a potential common target. Additionally, PI3 is identified as a potential biomarker for both AD and UC, providing new insights into the mechanisms and potential connections between these diseases.

Keywords: biomarkers, machine learning, immune recruitment, single-cell RNA sequencing, neutrophil