Integrative Multi-Omics and Machine Learning Reveal Senescent Fibroblast-Associated ADAMTS2 as a Critical Indicator of Intestinal Inflammation in IBD.

Although cellular senescence contributes to tissue dysfunction in inflammatory bowel disease (IBD), the heterogeneity of senescence at the single-cell resolution remains poorly defined. In this study, we systematically characterized the senescence landscape within the IBD microenvironment, with particular emphasis on fibroblast subpopulations and their systemic consequences. Single-cell RNA sequencing (scRNA-seq) dataset was applied to profile senescence-associated feature heterogeneity. Fibroblasts were classified into high-senescence (HS-Fibro) and low-senescence (LS-Fibro) groups according to senescence scores. HS-Fibro-specific gene signatures were identified and functionally annotated. Intercellular communication networks were constructed using CellChat to delineate the HS-Fibro interactome atlas, with emphasis on identifying cellular targets responsive to senescence-associated secretory phenotype (SASP) signals. Integrated microarray datasets were employed to elucidate the systemic impact of HS-Fibro in IBD. Through the integration of differential expression analysis, Weighted Gene Co-expression Network Analysis (WGCNA), and an ensemble of 113 machine learning algorithms, seven hub genes implicated in IBD progression were systematically identified and prioritized. In the IBD microenvironment, fibroblasts exhibited enhanced senescence activity. HS-Fibro displayed potent pro-inflammatory properties and enhanced communicative capacity. IBD patients with high HS-Fibro infiltration showed sustained immune-inflammatory activity. Systematic bioinformatic analysis and immunohistochemistry validation identified ADAMTS2 as a robust diagnostic indicator for IBD, with substantially elevated expression in patients relative to controls. The human colon fibroblast cell line CCD-18Co was treated with H2O2 to establish a cellular senescence model, demonstrating marked upregulation of ADAMTS2, SASP markers, and fibrosis-related genes. Moreover, siRNA-mediated knockdown of ADAMTS2 resulted in obvious attenuation of these SASP and fibrosis-associated genes.