Examination of hydrogen cross-feeders using a colonic microbiota model

Nick W Smith,Warren C Mcnabb,Eric Altermann,Nicole C Roy,Paul R Shorten

doi:10.1186/s12859-020-03923-6

Nick W Smith, Warren C Mcnabb + Show 3 more

Open Access

https://doi.org/10.1186/s12859-020-03923-6

Copy DOI

Abstract

BackgroundHydrogen cross-feeding microbes form a functionally important subset of the human colonic microbiota. The three major hydrogenotrophic functional groups of the colon: sulphate-reducing bacteria (SRB), methanogens and reductive acetogens, have been linked to wide ranging impacts on host physiology, health and wellbeing.ResultsAn existing mathematical model for microbial community growth and metabolism was combined with models for each of the three hydrogenotrophic functional groups. The model was further developed for application to the colonic environment via inclusion of responsive pH, host metabolite absorption and the inclusion of host mucins. Predictions of the model, using two existing metabolic parameter sets, were compared to experimental faecal culture datasets. Model accuracy varied between experiments and measured variables and was most successful in predicting the growth of high relative abundance functional groups, such as the Bacteroides, and short chain fatty acid (SCFA) production. Two versions of the colonic model were developed: one representing the colon with sequential compartments and one utilising a continuous spatial representation. When applied to the colonic environment, the model predicted pH dynamics within the ranges measured in vivo and SCFA ratios comparable to those in the literature. The continuous version of the model simulated relative abundances of microbial functional groups comparable to measured values, but predictions were sensitive to the metabolic parameter values used for each functional group. Sulphate availability was found to strongly influence hydrogenotroph activity in the continuous version of the model, correlating positively with SRB and sulphide concentration and negatively with methanogen concentration, but had no effect in the compartmentalised model version.ConclusionsAlthough the model predictions compared well to only some experimental measurements, the important features of the colon environment included make it a novel and useful contribution to modelling the colonic microbiota.

Highlights

Hydrogen cross-feeding microbes form a functionally important subset of the human colonic microbiota
The pH and substrate preferences of each Microbial functional group (MFG) are included, leading to 243 model parameters in total. These parameters were set based on existing knowledge of the MFGs, rather than parameterisation of microPop with faecal culture data
The major differences between the two parameter sets are in the viable and optimal pH range of each MFG, which were altered by as much as 0.95 pH units between parameter sets, and in the maximum growth rates for non-starch polysaccharide (NSP) and resistant starch of all MFGs able to metabolise these substrates. These are the most abundant substrates included in the model, substantial changes in model predictions between the two parameter sets were anticipated

Summary

Introduction

Hydrogen cross-feeding microbes form a functionally important subset of the human colonic microbiota. Efforts to study the colonic microbiota have included observational and interventional studies, coupled with in vitro, animal, and computational models This last technique, dependent on experimental data for model validation, has appeal as a fast, cheap and high-throughput method, leading to the creation of several mathematical models predicting colonic microbial dynamics in recent years Hydrogen is produced through various microbial metabolic pathways involved in the degradation of carbohydrates in the colon, creating a niche for microbes that can cross-feed on hydrogen [5] These hydrogenotrophs have demonstrated and hypothesised impacts on both the microbiota and the host, including increasing the rate of carbohydrate fermentation of saccharolytic bacteria in vivo, with associated increases in host adiposity [6], and links to negative health outcomes such as Irritable Bowel Syndrome [7], Inflammatory Bowel Disease [8] and colorectal cancer development [9]. Modelling could provide useful insight into the dynamics of these taxa in vivo, contributing to our understanding of these microbes in human health, nutrition and wellbeing

Methods

Results

Discussion

Conclusion