Developing and validating a computational model of the gut microbiota–mucosa interactions to replace and reduce animal experiments

Foster, Timothy Roger ORCID: 0000-0001-8559-0983 (2023). Developing and validating a computational model of the gut microbiota–mucosa interactions to replace and reduce animal experiments. University of Birmingham. Ph.D.

Preview

Foster2023PhD.pdf
Text - Accepted Version
Available under License All rights reserved.
Download (5MB) | Preview

Abstract

The human gut is teeming with microbial life. Within the most densely populated part of the gut - the colon - there are an estimated 1 trillion microbial cells. These microbes - the gut microbiota - have profound impacts on our health from the moment of birth, all the way through our lives. Some provide epithelial cells with nutrition, by breaking down dietary fibres, some may be essential in the development of the immune system, and some microbes can provide protection against the invasion of the gut by harmful pathogens. Understanding such a complex and diverse system calls for the development of models that can represent key features of the system, while simplifying it, to make generalisable conclusions. The most widely utilised model of the human gut is the mouse gut, which is used in a great many studies on the gut microbiota. In addition, various in vitro models have been developed, which typically represent the gut using chemostats or “gut-on-a-chip” systems that may include epithelial cells. Several population-level in silico models of the gut microbiota have also been developed, based on systems of differential equations. However, agent-based modelling - a modelling approach which allows researchers to investigate how local interactions between cells can influence population-level outcomes - has not been utilised extensively to investigate the gut microbiota.

In this thesis, I will present a new modelling platform - eGUT - an agent-based modelling platform for modelling the gut microbiota. This model focusses on the interactions between microbial cells and the host epithelium at the mucosal surface, by explicitly modelling the spatial detail of the mucosal region. This mucosal compartment can be connected to other regions, such as the gut lumen and the host’s circulation system, in order to model the interactions between these regions, such as the exchange of chemicals and microbial cells and the flow of digesta through the gut. Partial differential equation solvers are used to model the reactions and diffusion of the various chemicals within the system, allowing interactions such as cross-feeding, competition and other metabolic interactions to arise naturally, and to explore how niches may be created or partitioned, and how the various behaviours of gut microbes and the host epithelium affect the fate of particular gut microbes.

In order to ensure that the eGUT platform enables a physically and biologically realistic representation of reality, it has been subjected to various tests and validation experiments. These include test of numerical algorithms that ensure the models of solute diffusion and consumption work as expected, as well as more complex tests, such as benchmarking eGUT against other in silico models able to simulate biofilms, and conducting validation experiments using MIMic, a chemostat-based in vitro model of the gut microbiota.

eGUT now passes all numerical tests without issue, showing that its basic models of reaction and diffusion are functioning correctly. It is also in good agreement with a range of other in silico models when running the benchmark biofilm model, given the variation between models and their fundamentally different approaches. However, the results of a colonisation resistance study conducted both in MIMic and eGUT show different results. It is unclear whether this is the result of problems with the assumptions made in the modelling system, or imprecision in the values of estimated growth parameters for the species used in the model.

Although some issues remain to be further investigated, eGUT has the potential to be a valuable resource to gut scientists wanting to model the interactions between members of the microbiota and their host. eGUT may act as a convenient testing ground for hypotheses and questions regarding population dynamics in the gut, or for the development of probiotics and/or prebiotics. This may lead to the reduction or possibly the replacement of the use of animal models, including mice, in gut research laboratories.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):