Mucosal Gene Expression Signatures Associated with Response to Tofacitinib and Upadacitinib in Ulcerative Colitis

Background/Introduction: Response to Janus kinase (JAK) inhibitors in ulcerative colitis (UC) varies, and reliable pretreatment molecular markers predicting therapeutic response are lacking. We investigated mucosal gene-expression profiles before and after treatment with tofacitinib or upadacitinib to identify baseline molecular signatures associated with clinical response and treatment-related changes. Methods: Paired baseline and follow-up mucosal biopsies were obtained from 57 patients with UC receiving upadacitinib (n = 32) or tofacitinib (n = 25), alongside biopsies from 11 healthy controls. Gene expression was assessed using a targeted 84-gene RT2 Profiler PCR Array. Gene-level and biologically informed module-level analyses were performed with Benjamini–Hochberg correction, while exploratory lasso logistic-regression models were evaluated using repeated nested cross-validation. Results: At baseline, 42 of 84 genes differed significantly between subsequent responders and non-responders [False Discovery Rate (FDR) < 0.05]. The strongest response-associated genes included IL13RA1, IL12B, IL18R1, IL6, BCL6, and STAT1, whereas IL25 and SPP1 showed lower expression in responders. Ten of 11 expression modules were higher in responders. Longitudinal analyses revealed treatment-associated molecular changes in both responders and non-responders. Tofacitinib responders exhibited significant reductions in several cytokine- and JAK/STAT-related genes and modules, whereas fewer changes reached statistical significance among upadacitinib responders. The combined gene-level prediction model achieved an internally cross-validated area under the curve (AUC) of 1.000 using participant-level predictions averaged across 50 repetitions, compared with 0.889 and 1.000 for the upadacitinib- and tofacitinib-specific models, respectively. Conclusions: These findings suggest that baseline mucosal gene-expression signatures are strongly associated with subsequent response to JAK inhibition. However, the small cohort and lack of external validation limit their immediate clinical applicability. Notably, molecular changes observed in non-responders indicate that pharmacodynamic pathway modulation does not necessarily translate into clinical improvement.