P852 Metataxonomic and metabolomic profiling to predict response to anti-TNF therapy in Crohn’s Disease

Abstract

Abstract Background The composition of the gut microbiota is associated with response to treatments in inflammatory bowel disease1. We sought to identify gut microbial and metabolomic biomarkers associated with response to anti-TNF therapy in Crohn’s disease. Methods Faecal and serum samples collected at baseline and week 14 of anti-TNF treatment in the "Predictors of anti-TNF treatment failure" (PANTS) study were used2. Response to treatment (defined clinically and biochemically) was determined at week 142. Faecal samples were processed by GA-map®, with targeted 16S rRNA PCR sequencing. Serum samples were analysed with targeted liquid chromatography tandem mass spectrometry for bile acids and tryptophan metabolites. Multivariate and univariate analyses, as well as Spearman’s rank test, were performed. Results Two-hundred and twenty-four patients with Crohn’s disease commencing anti-TNF therapy were included (54.1% anti-TNF responders, 48.6% female, median age 37.2 (26.5-50.0) years, 56.9% receiving immunomodulator-based treatment). A significantly higher proportion of non-responders received corticosteroid treatment compared to responders at baseline (p=0.0009). Sixty-nine bacterial taxa were identified. Significantly lower abundances of Clostridium difficile and Mycoplasma were identified in responders relative to non-responders at baseline (q-value=0.046 and q-value=0.034, respectively). No significant differences in the abundances of Faecalibacterium prausnitzii, Shigella or Escherichia Coli were found (all q-values > 0.05). Eighteen serum bile acids and thirteen tryptophan metabolites were identified. Significantly higher serum abundances of cholic acid, tryptophan and 3OH-kynurenine were identified in responders at baseline (q-values=0.049, 0.033, and 0.033, respectively). Exploration of microbial–metabolite associations revealed negative correlations between secondary bile acids and several bile salt hydrolase producing bacteria, such as Bifidobacterium (r ≤ -0.50). Positive correlations were also identified between xanthurenic acid and Bifidobacterium and Roseburia intestinalis (r ≥ 0.50). Increased abundances of short-chain fatty acid producing bacteria, such as Bifidobacterium and Bacteroides vulgatus, were noted in responders at week 14. Conclusion We have identified baseline differences in the faecal microbiota and serum metabolome of anti-TNF responders and non-responders. These findings require validation in larger cohorts to determine their robustness for potential use in treatment personalisation.