Abstract Background Limited evidence is available about predictors of response to specific medications for ulcerative colitis. Variants of OCTN1, an organic cation transporter, could modulate the inflammatory response and potentially predict individual response to immune-modulatory therapy. Methods We first assessed the ability of OCTN1 to differentially modulate IL1β secretion under different stimuli in a human leukemia monocytic cell line (THP-1) and in human peripheral blood mononuclear cells. We therefore prospectively enrolled a cohort of UC patients starting either anti-TNFα or vedolizumab, determined their OCTN1 genotype and evaluated their response to therapy after six months, as well as steroid free therapy persistence after 2 years. Finally, we performed an integrated machine learning based analysis including serum cytokines and gut microbiota analysis. Results OCTN1-deficient THP1 cells showed reduced IL-1β secretion in response to peptidoglycan (PGN) and Fusobacterium nucleatum (Fn), whilst primary monocytes bearing OCTN1 503F variant displayed an increased production of IL-1β in response to PGN and live bacteria. We therefore enrolled 90 patients, of which 50 starting anti-TNFα and 40 vedolizumab. At the multivariate analysis, a significant reduction of clinical response after six months was observed in patients with 503F variant treated with vedolizumab. The same trend was observed for steroid free therapy persistence after two years. The machine learning analysis revealed that changes in bacterial composition, notably the relative abundance of Gemellaceae, followed by Lachnospiracea, were significant predictors of lack of two-year steroid-free therapy persistence. Conclusion Mutated OCNT1 genotype (503F) could favor an increased IL1β secretion from human leukocytes, thus suggesting a worse disease course and lower response to vedolizumab as compared to 503L. ovel machine learning data analysis provides further insights into such topic.
Read full abstract