Developing models for the diagnosing of ulcerative colitis and prognosis of anti-TNF-α non-response based on neutrophil extracellular trap-associated genes.

BACKGROUND

Neutrophil extracellular traps (NET) play a pivotal role in the pathogenesis of ulcerative colitis (UC) and may contribute to the impaired response to anti-tumor necrosis factor alpha (TNF-α) therapies. However, the functional implications of NET-associated genes in UC remain poorly understood. This study aims to identify key NET-associated molecular signatures in UC, develop diagnostic models based on NET-related biomarkers, and construct predictive models for response to anti-TNF-α therapies (infliximab and golimumab).

METHODS

NET-associated genes were obtained from the Kyoto Encyclopedia of Genes and Genomes, whereas UC-related gene expression datasets were retrieved from the Gene Expression Omnibus. Unsupervised consensus clustering based on NET-related genes was used to stratify patients with UC into molecular subtypes. The CIBERSORT algorithm and gene set variation analysis were employed to characterize immune cell infiltration and biological pathway activity across clusters. Hub genes were identified using weighted gene co-expression network analysis and machine learning algorithms. Spearman correlation analyses were performed to assess associations between hub genes, immune cell infiltration, and clinical disease activity. A diagnostic model for UC and a prognostic model for anti-TNF-α treatment response were developed using hub genes identified through least absolute shrinkage and selection operator regression.

RESULTS

Based on 33 NET-associated genes, patients with UC were stratified into two distinct molecular clusters (C1 and C2). Cluster C1 exhibited a pronounced NET signature, characterized by significantly elevated neutrophil infiltration (p < 0.001) and activation of inflammatory signaling pathways, including IL-2/STAT5, TNF-α/NF-κB, and IL-6/JAK/STAT3. Notably, C1 was associated with a significantly higher rate of non-response to anti-TNF-α therapy (57.4% vs. 22.0% in C2, p = 0.003). A diagnostic model for UC was constructed using five hub genes (FCGR3B, IL1RN, CXCL8, S100A8, and S100A9) derived from C1. Moreover, a predictive model for anti-TNF-α non-responsiveness, based on two hub genes (FCGR3B and IL1RN), was developed using a golimumab dataset and validated in two independent infliximab datasets.

CONCLUSION

A distinct NET-associated cluster was identified among patients with UC, exhibiting non-responsiveness to anti-TNF-α treatment. Diagnostic and prognostic models based on NET-associated genes hold promise for guiding clinical treatment strategies.