Changes in fecal metabolic and lipidomic features by anti-TNF treatment and prediction of clinical remission in patients with ulcerative colitis.

Background: Therapeutic targets for ulcerative colitis (UC) and prediction models of antitumor necrosis factor (TNF) therapy outcomes have not been fully reported. Objective: Investigate the characteristic metabolite and lipid profiles of fecal samples of UC patients before and after adalimumab treatment and develop a prediction model of clinical remission following adalimumab treatment. Design: Prospective, observational, multicenter study was conducted on moderate-to-severe UC patients (푛=116). Methods: Fecal samples were collected from UC patients at 8 and 56weeks of adalimumab treatment and from healthy controls (HC, 푛=37). Clinical remission was assessed using the Mayo score. Metabolomic and lipidomic analyses were performed using gas chromatography mass spectrometry and nano electrospray ionization mass spectrometry, respectively. Orthogonal partial least squares discriminant analysis was performed to establish a remission prediction model. Results: Fecal metabolites in UC patients markedly differed from those in HC at baseline and were changed similarly to those in HC during treatment; however, lipid profiles did not show these patterns. After treatment, the fecal characteristics of remitters (RM) were closer to those of HC than to those of non-remitters (NRM). At 8 and 56 weeks, amino acid levels in RM were lower than those in NRM and similar to those in HC. After 56weeks, levels of 3-hydroxybutyrate, lysine, and phenethylamine decreased, and dodecanoate level increased in RM similarly to those in HC. The prediction model of long-term remission in male patients based on lipid biomarkers showed a higher performance than clinical markers. Conclusion: Fecal metabolites in UC patients markedly differ from those in HC, and the levels in RM are changed similarly to those in HC after anti-TNF therapy. Moreover, 3-hydroxybutyrate, lysine, phenethylamine, and dodecanoate are suggested as potential therapeutic targets for UC. A prediction model of long-term remission based on lipid biomarkers may help implement personalized treatment. [ABSTRACT FROM AUTHOR]