Identification and selective inhibition of patient-specific pathobionts in inflammatory bowel disease

Background The microbiome has been linked to various chronic diseases. However, the congruency of proposed bacterial taxa between studies is low and identifying inflammation-driving pathobionts for personalised therapy is hampered by a complex microbiota, heterogeneous patient risk factors and technical challenges to sort and maintain viable bacteria. Objective Development of an approach that enables patient-specific identification of bacterial triggers of inflammation and their selective elimination. Design Faecal samples were collected from individual inflammatory bowel disease (IBD) patients during inflammation and remission and transferred into germ-free (GF) mice. Comparing microbiomes collected at different times from the same individual uncoupled the inflammatory capacity of the microbiome from other factors, such as genetic or environmental risk factors. Results We show that inflammatory states of patients with IBD were recapitulated on faecal microbiota transplantation and identified putative disease-driving bacterial taxa by sorting and culture of immunoglobulin-coated intestinal bacteria. Colonisation of GF mice identified Sarcina perfringens and Hathewaya massiliensis as pathobionts promoting inflammation. Mechanistically, innate immune cells required TLR1/2-MyD88 signalling to sense both pathobionts. Compared with healthy individuals, S. perfringens and H. massiliensis were enriched in ~25%–30% and ~5% of patients with IBD, respectively, across diverse populations. S. perfringens abundance was also correlated with mucosal inflammatory transcripts. Both pathobionts increased IL-8 production or reduced ZO-1 expression in IBD patient-derived organoids. Screening >1500-approved compounds revealed five molecules, including oxolinic acid, that selectively inhibited pathobiont growth and protected from intestinal inflammation. Conclusions Collectively, we developed an approach for identifying disease-driving pathobionts and selecting compounds for targeted therapy.