Systems biology integration of miR-mRNA profiles identifies steroid refractoriness mechanisms and biomarker candidates in ulcerative colitis.

Approximately 60% of patients with ulcerative colitis (UC) show steroid resistance or dependence, underscoring the need for biomarkers predicting therapeutic response. Despite increasing availability of biologics, corticosteroids remain essential first-line therapy for moderate-to-severe flares in many settings. We applied an integrative systems biology and machine-learning framework to combine microRNA (miR) and mRNA expression profiles from matched rectal biopsies and plasma samples of UC patients treated with corticosteroids, aiming at exploring classification potential of such biological data. Transcriptomic mRNA and miR profiling was performed at baseline and after three days of therapy, and patients were classified as responders or non-responders after seven days. Differential expression results were embedded into a previously defined curated molecular network-based mathematical model of the mechanism of action (MoA) of glucocorticoid signaling over UC to enhance biological interpretability. miR-mRNA interactions were prioritized based on database support and relevance to glucocorticoid receptor and inflammatory signaling as defined in the molecular mathematical model. Key transcriptional co-regulators within this network, including NCOA3, CBP, NCOR1, and NRIP1, distinguished response groups, together with miRs such as miR-145-5p, miR-10b-5p, and miR-16-5p. Several miR candidates showed circulating-tissue consistency and conserved behavior in a TNBS-induced colitis mouse model, in terms of expression in response to corticoids and miR-mRNA inverse correlations. This study proposes a mechanistically grounded framework for corticosteroid response biomarker discovery; however, findings are exploratory and prospective validation in independent cohorts is required before clinical applicability can be established.