Identification of novel biomarkers linked to M1 macrophage infiltration in the diagnosis of inflammatory bowel diseases

Xiang-Guang Li; Huantao Li; Ding Luo; Jie Li; Yu Wang; Jie Zheng; Junxia Zheng; Jie Song; Chuqiao Xiao

doi:10.1016/j.intimp.2025.115138

Identification of novel biomarkers linked to M1 macrophage infiltration in the diagnosis of inflammatory bowel diseases

Int Immunopharmacol. 2025 Jun 28:162:115138. doi: 10.1016/j.intimp.2025.115138. Online ahead of print.

Authors

Xiang-Guang Li¹, Huantao Li¹, Ding Luo², Jie Li¹, Yu Wang³, Jie Zheng¹, Junxia Zheng¹, Jie Song⁴, Chuqiao Xiao⁵

Affiliations

¹ Department of Pharmaceutical Engineering, School of Biomedical and Pharmaceutical Sciences, Guangdong University of Technology, Guangzhou 510006, China.
² The Second Affiliated Hospital of Guangzhou University of Chinese Medicine, Guangzhou 510120, China.
³ Department of Pharmaceutical Engineering, School of Biomedical and Pharmaceutical Sciences, Guangdong University of Technology, Guangzhou 510006, China; Center of Digestive Endoscopy, The Affiliated Guangdong Second Provincial General Hospital of Jinan University, Guangzhou 510310, China.
⁴ Center of Digestive Endoscopy, The Affiliated Guangdong Second Provincial General Hospital of Jinan University, Guangzhou 510310, China. Electronic address: Songj29@mail2.sysu.edu.cn.
⁵ Department of Pharmaceutical Engineering, School of Biomedical and Pharmaceutical Sciences, Guangdong University of Technology, Guangzhou 510006, China. Electronic address: cqxiao@gdut.edu.cn.

PMID: 40582145
DOI: 10.1016/j.intimp.2025.115138

Abstract

Inflammatory bowel disease (IBD) often lacks a definitive diagnostic standard, leading to diagnoses through exclusion. This study aimed to create a predictive model for IBD using bioinformatics and deep learning while identifying potential biomarkers to improve management outcomes. Analyses indicated that differentially expressed genes in IBD patients were associated with cytokine activity and pro-inflammatory pathways. A significant correlation was found between M1 macrophage infiltration and IBD pathogenesis. Through a Weighted Gene Correlation Network Analysis, we identified 31 genes related to M1 macrophages and 18 metatranscriptomic features linked to IBD. Integrating these, we developed a two-class predictive model, demonstrating the effectiveness of deep learning techniques, particularly neural networks in IBD diagnostics. Using the SHAP algorithm, we highlighted 10 critical features, primarily host genetic variants, that illustrate the interaction between gut microbiota and host genetics in IBD. Furthermore, our findings emphasize the pivotal role of M1 macrophage-associated CXCL10 in IBD, suggesting its potential as a promising biomarker and therapeutic target in managing IBD. We envision a future where global analyses of IBD data can be enhanced through artificial intelligence, yielding more profound insights into the etiological factors driving IBD pathology.

Keywords: Artificial intelligence; CXCL10; Inflammatory bowel disease; Interpretable prediction model; M1 macrophage.