PersonalizedClostridioides difficileengraftment risk prediction and probiotic therapy assessment in the human gut

Abstract

Clostridioides difficile colonizes up to 30-40% of community-dwelling adults without causing disease. C. difficile infections (CDIs) are the leading cause of antibiotic-associated diarrhea in the U.S. and typically develop in individuals following disruption of the gut microbiota due to antibiotic or chemotherapy treatments. Further treatment of CDI with antibiotics is not always effective and can lead to antibiotic resistance and recurrent infections (rCDI). The most effective treatment for rCDI is the reestablishment of an intact microbiota via fecal microbiota transplants (FMTs). However, the success of FMTs has been difficult to generalize because the microbial interactions that prevent engraftment and facilitate the successful clearance of C. difficile are still only partially understood. Here we show how microbial community-scale metabolic models (MCMMs) accurately predicted known instances of C. difficile colonization susceptibility or resistance. MCMMs provide detailed mechanistic insights into the ecological interactions that govern C. difficile engraftment, like cross-feeding or competition involving metabolites like succinate, trehalose, and ornithine, which differ from person to person. Indeed, three distinct C. difficile metabolic niches emerge from our MCMMs, two associated with positive growth rates and one that represents non-growth, which are consistently observed across 14,862 individuals from four independent cohorts. Finally, we show how MCMMs can predict personalized engraftment and C. difficile growth suppression for a probiotic cocktail (VE303) designed to replace FMTs for the treatment rCDI. Overall, this powerful modeling approach predicts personalized C. difficile engraftment risk and can be leveraged to assess probiotic treatment efficacy. MCMMs could be extended to better understand personalized engraftment of other opportunistic bacterial pathogens, beneficial probiotic organisms, or more complex microbial consortia.

Integrating biological knowledge for mechanistic inference in the host-associated microbiome

“…Moreover, GSMNs from multiple organisms can be aggregated in order to simulate entire microbial communities with tools such as MICOM ( Swainston et al, 2017 ). Recently, MICOM was used to predict the risk of Clostridium difficile infection, the leading cause of antibiotic associated diarrhea, based on the metabolic strategies of C. difficile in different host microbiome and diet contexts ( Carr et al, 2023 ). GSMNs can therefore serve as a blueprint for the suite of metabolic transformations possible and facilitate the understanding of the metabolic potential of a given microbial community ( Mendoza et al, 2019 ; Esvap and Ulgen, 2021 ; Passi et al, 2021 ).…”

Top-down and bottom-up microbiome engineering approaches to enable biomanufacturing from waste biomass

“…This approach has been proposed and applied for several decades, and, while successful in some cases, a common issue across contexts is the loss of the added organism over time ( Mikesell & Boyd, 1988 ; Albright et al., 2022 ; Shan et al., 2023 ). Adding synthetic consortia with functional features of interest could potentially improve their retention, but rational rules for their design can be challenging to identify (Carr et al., 2023 ; Louie et al., 2023 ). Two lignocellulolytic microbial consortia, CL and YL, that were enriched from the rumen content of Tibetan yak and Holstein cow, respectively, were augmented with Lactobacillus plantarum (LP) to enhance the fermentation of alfalfa silage.…”