GUT MICROBIAL METABOLISM OF 5-ASA IS PROSPECTIVELY ASSOCIATED WITH TREATMENT FAILURE IN PATIENTS WITH IBD

Abstract

Listen

Translate article

Abstract BACKGROUND Variation in clinical response to 5-aminosalicylic acid (5-ASA) has been attributed in part to its inactivation by gut microbes. Recently, in the Inflammatory Bowel Disease (IBD) Multi’omics Database (IBDMDB), a multicenter year-long cohort of 100+ participants with IBD, we identified 12 gut microbial enzymes from two protein families that convert 5-ASA to N-acetyl 5-ASA, a compound that lacks anti-inflammatory effects. Within the IBDMDB, we then found that a subset of these enzymes was cross-sectionally linked with greater risk of treatment failure, defined by corticosteroid use. The aim of this study was to prospectively associate these drug-degrading gut microbial enzymes with treatment failure and to confirm in vitro conversion of 5-ASA by these enzymes. METHODS We examined the prospective association between gut microbial acetyltransferases with risk of 5-ASA treatment failure in the ongoing “Study of a Prospective Adult Research Cohort with IBD” (SPARC IBD). We included 208 participants who at study entry 1) gave a stool sample at baseline, 2) were on 5-ASA, and 3) were steroid-free. Fecal metagenomic data was processed by the same analysis pipeline (biobakery3) as in the IBDMDB. Exposure was defined identically as in the IBDMDB as metagenomic carriage of 3-4 drug-degrading acetyltransferases compared to 0-2 acetyltransferases. We calculated odds ratios (ORs) and 95% confidence intervals (CIs) of incident steroid use using multivariable generalized estimating equations, adjusting for age and sex. For biochemical characterization, the 12 enzymes identified in the IBDMDB were heterologously expressed by E. coli. A representative enzyme was selected from each protein family for purification, and then incubated with 5-ASA and acetyl CoA for 6 hr at 37C with 1 mM of each substrate and 50 μM enzyme. Using a custom LC–MS assay, we then detected N-acetyl 5-ASA production. RESULTS Over a median follow-up of 8 months, we identified 60 cases of corticosteroid use. Consistent with data from the IBDMDB, we found that metagenomic carriage of 3 or more microbial acetyltransferase genes in SPARC IBD (compared to 2 or fewer) was associated with treatment failure (OR 2.77, 95% CI 1.03-7.43). In a random effects meta-analysis of the two cohorts, we observed a three-fold increased risk of drug failure (OR 3.12, 95% CI 1.41-6.89) (Fig 1). In vitro biochemical experiments confirmed the ability of a thiolase and an acyl-CoA N-acetyltransferase to acetylate 5-ASA, generating product at levels consistent with >25% conversion (Fig 2). CONCLUSIONS We characterized two gut microbial protein families that are directly involved in 5-ASA metabolism and, in turn, are prospectively associated with 5-ASA treatment failure. These findings advance the possibility of microbiome-based personalized medicine for patients with IBD.