Abstract Background Ustekinumab (UST) has become a pivotal treatment option for Crohn’s disease (CD). Despite its clinical benefits, up to 50% of patients experience primary non-response (PNR) in both clinical trials and real-world cohorts. Currently, there is no method to predict primary non-response to UST prior to treatment, and the molecular basis of therapeutic heterogeneity remains unclear. We aim to identify predictors of UST therapeutic response and to explore key regulatory factors and mechanisms underlying varied treatment responses. Methods We prospectively enrolled 55 biologic-naïve CD patients undergoing UST induction therapy and 34 healthy controls. CD patients were classified as responders (R) and non-responders (PNR) based on treatment response at 16 weeks following clinical and endoscopic criteria. Integrated proteomic and untargeted metabolomic profiles were employed to identify predictive signatures of responders and primary non-responders at baseline. Single-cell RNA sequencing (scRNA-seq), multiplex immunofluorescence, and functional validation through in vivo murine models and macrophage experiments were performed to investigate regulatory mechanisms driving efficacy differences. Results Of 55 UST-treated CD patients, 30 achieved primary response, while 25 were primary non-responders. Plasma proteomic and metabolomic analyses identified four key biomarkers: elevated CXCL11, HLA-G, and MARCO, and reduced androsterone levels in PNR patients. Single-cell transcriptomics and immunostaining revealed an accumulation of CXCL11+ monocyte-like macrophage subset (MΦMoCXCL11) in intestinal mucosa of non-responders. Mechanistically, androsterone activated the androgen receptor (AR) to induce USP53, which deubiquitinates TAK1 and suppresses NF-κB–mediated inflammation in macrophages. In vivo study demonstrated that ADT markedly enhanced UST efficacy in DSS-induced colitis models. Additionally, we explored potential therapeutic strategies for UST non-responders to optimize personalized CD treatment, finding that switching to infliximab or upadacitinib yielded high rates of endoscopic and clinical improvement. Conclusion Our study developed a plasma-based biomarker panel to predict UST efficacy in CD patients. We also revealed a metabolic–immune axis, mediated by the androsterone–AR–USP53–TAK1 pathway—that regulates UST efficacy in CD. These findings support the development of personalized treatment strategies and suggest viable alternatives for UST non-responders. Conflict of interest: Dr. Wu, Lexi: No conflict of interest Huang, Lingjie: No conflict of interest Liu, Rongbei: No conflict of interest Cao, Qian: No conflict of interest
Wu et al. (Thu,) studied this question.