An ecological framework to understand the efficacy of fecal microbiota transplantation

Commensal microorganisms in the human gut are key players in physiology and disease. While metagenomics and next-generation sequencing have advanced our understanding of microbiome composition and function, fundamental questions about the system's dynamics and resilience remain. This is especially crucial given the development of microbiome-modifying therapies like fecal microbiota transplantation (FMT). FMT involves introducing fecal material from a healthy donor into a recipient's gastrointestinal tract to treat various conditions, including recurrent *Clostridioides difficile* infection (rCDI), inflammatory bowel disease, irritable bowel syndrome, and other disorders. However, the high complexity of the gut microbiome raises safety concerns about untargeted interventions. This article aims to understand the ecological principles governing FMT's efficacy, focusing on rCDI, which currently has the strongest clinical evidence supporting FMT. *C. difficile*, an anaerobic bacterium, is transmitted through the fecal-oral route and causes CDI. Antibiotic treatment disrupts the gut microbiota, allowing *C. difficile* to colonize and produce toxins. Standard antibiotic therapy often fails, leading to rCDI, while FMT has demonstrated high cure rates for rCDI. Despite this success, many questions remain unanswered: What key ecological factors determine FMT success? Does FMT's efficacy vary between primary and recurrent CDI? Does donor-recipient compatibility matter? How can we design effective personalized probiotic cocktails to replace FMT? This study utilizes community ecology theory and network science to address these questions.

The literature extensively documents FMT's efficacy in treating rCDI, with numerous case reports and cohort studies demonstrating its success. Studies have also explored FMT's use in treating inflammatory bowel disease, irritable bowel syndrome, and various other conditions, both GI and non-GI. However, the literature lacks a cohesive ecological understanding of the mechanisms underlying FMT's success and failure. Existing research highlights individual microbial species or functions involved in restoration of the gut microbiome, but lacks a systems-level perspective. This research gap necessitates an ecological framework to understand the complex interactions and dynamics within the gut microbiome during FMT.

This study combines community ecology theory and network science to develop a theoretical framework for understanding FMT's ecological principles, using rCDI as a model. A generalized Lotka-Volterra (GLV) model is used to simulate the FMT process, modeling the gut microbiota of different hosts as local communities assembled from a global species pool. The model simulates the healthy, diseased, and post-FMT states of the recipient's gut microbiota. The efficacy of FMT is quantified using a recovery degree (η), representing the reduction in *C. difficile* abundance after FMT. The impact of host-dependent microbial dynamics, pre-FMT taxonomic diversity, and donor-recipient compatibility on FMT efficacy are investigated through extensive simulations. The simulations vary the level of host-dependency in microbial dynamics (from universal dynamics to highly host-dependent dynamics), the taxonomic diversity of the recipient's pre-FMT microbiota, and the compatibility between donor and recipient microbiomes. Real-world data from a clinical trial of FMT and a mouse model are analyzed to validate the theoretical predictions. Finally, an algorithm is developed for the rational design of personalized probiotic cocktails to decolonize *C. difficile*. This algorithm considers the network effect—the indirect influence of species interactions—and aims to identify minimal sets of effective species for specific patients.

The ecological modeling framework predicts several key factors influencing FMT efficacy. Simulations demonstrate that host-dependent microbial dynamics significantly reduce FMT success. FMT efficacy is negatively correlated with the recipient's pre-FMT taxonomic diversity—higher diversity reduces FMT success. Donor-recipient compatibility is also a crucial factor. While some 'super donors' are effective for all recipients, and some 'super recipients' respond to all donors, compatibility issues are pronounced for recipients with high pre-FMT diversity. Simple comparison of donor and recipient microbial compositions does not fully predict FMT efficacy. Net effects within the microbiome (rather than direct interactions) are key determinants of FMT success. Analysis of real data from a clinical trial shows that nonresponders have higher median taxonomic diversity than responders, supporting the simulation results. The study proposes an algorithm for designing personalized probiotic cocktails based on the global ecological network or the *C. difficile* ego-network. This algorithm accounts for indirect interactions among species and efficiently identifies minimal, effective probiotic cocktails specific to each patient's microbiota. Analysis of mouse data demonstrates the presence of the network effect, supporting the framework's core assumptions.

This study provides a novel ecological framework for understanding FMT's efficacy. The findings highlight the importance of considering host-dependent microbial dynamics, pre-FMT taxonomic diversity, and donor-recipient compatibility in optimizing FMT treatment. The algorithm for personalized probiotic cocktails offers a potential solution to the limitations of FMT, such as safety concerns and donor recruitment challenges. The focus on network effects emphasizes the complexity of the gut microbiome and the need for systems-level approaches in developing microbiota-based therapies. The results suggest that FMT's effectiveness in rCDI likely results from the relative universality of gut microbial dynamics. This study advances our understanding of microbial community dynamics and FMT's success, offering a promising path towards developing more effective and personalized treatments for rCDI and other microbiome-related diseases.

This research introduces an innovative ecological framework for understanding FMT efficacy in treating rCDI. The framework accurately predicts key factors affecting FMT success, including the level of host-dependent microbial dynamics, the pre-FMT taxonomic diversity of the recipient, and the degree of donor-recipient compatibility. Furthermore, the developed algorithm allows for the design of personalized probiotic cocktails that effectively decolonize *C. difficile*. Future research could focus on incorporating stochastic effects, detailed resource dynamics, functional data, dietary impacts, and more sophisticated population dynamics models into the framework to enhance its predictive power and generalizability. Experimental validation using animal models and clinical trials is crucial to further solidify the framework's applicability.

The current modeling framework simplifies several aspects of the human gut microbiome. Stochastic effects are considered negligible, and the model doesn't explicitly model resource dynamics or functional changes in the microbiome. The model also doesn't incorporate the impact of diet or drugs on microbial composition, nor the influence of higher-order interactions beyond pairwise interactions among species. The GLV model assumes linear functional response which might not represent all the complexities of ecological interactions. These simplifications may limit the model's precise predictive power. Further validation using experimental data is needed to fully assess the model's generalizability.