Dietary carbohydrates nourish human gut bacterial communities (microbiota) that resist pathogens, metabolize drugs, and train the immune system. Short-chain fatty acids, which are end products of microbial polysaccharide fermentation, are also crucial metabolic precursors and energy sources for human colon cells. The potential health benefits of dietary carbohydrates that stimulate the growth and activity of intestinal bacteria (prebiotics) have led millions of Americans to consume these compounds annually. Yet, the effects of prebiotics on gut microbiota and their fermentation are known to vary substantially between individuals. Our objective here is to understand why and how prebiotics should be tailored to individuals and their gut microbiota. To address advance prebiotic research, we have developed innovative new tools: a microfluidic technique for creating and assaying millions of individual bacterial cultures; Bayesian state-space models for longitudinal microbiota data; and, an artificial human intestine that we can sample and manipulate with arbitrary frequency. We propose combining these new methods to test our central hypothesis that the impact of prebiotic treatments can be maximized by personalization to individuals and their microbiota. Our proposal has three specific aims: 1) Use our microfluidic culture techniques to measure how different carbohydrate compounds affect the growth and metabolism of thousands of distinct human gut bacterial species. By identifying which bacterial species are directly stimulated by prebiotics, we can begin to understand how these treatments reshape each individuals’ gut microbiota. 2) Develop a probabilistic state-space model of microbial community dynamics and apply it our existing datasets tracking human diet, gut microbiota, and short chain fatty acid levels over time. The resulting model will pinpoint interactions between bacterial species growth, microbial fermentation, and subject diet that influence response to prebiotic treatments. 3) Use our artificial human gut models to carry out prebiotic trials on human gut microbiota with doses that are either fixed or periodically updated based on changes to microbiota structure and function. The resulting data will establish how initial microbiota shifts caused by prebiotic treatment affect later dose responses, as well as assess whether eliminating differences in host compliance and response affect variations in prebiotic impact. Ultimately, the proposed aims are expected to provide the first systematic study on the underlying mechanisms driving individualized responses to prebiotic treatments and help establish a new research field focused on personalized prebiotic treatments. The computational and experimental techniques developed here will also serve as a preclinical platform that could directly translate potential prebiotics to human clinical trials. More broadly, these techniques are designed to be generalizable and could thus be used for the rational optimization of other microbiota functions, including drug or hormonal metabolism within hosts, or even toxin bioremediation or nutrient production in the environment.
This work is sponsored by NIH award 5R01DK116187.