Improve understanding of gut-brain interaction
We modelled the fluid transit behaviour of the colon, integrated with existing models of the microbiome, to gain a deeper understanding of the role of microbiome-host interaction in disorders of gut-brain interaction.
Clinical need summary
Irritable bowel syndrome (IBS) is a disorder of gut-brain interaction associated with chronic pain in conjunction with altered bowel habits (e.g., diarrhoea or constipation). IBS remains a functional disorder as the underlying pathophysiology is poorly understood, so diagnosis relies on gastrointestinal symptom reporting. Many dietary and pharmacological interventions have been proposed for IBS symptom relief, but treatment response rates are low. As a result, patients with IBS will generally try several treatment options in a trial-and-error fashion, often without sustained success, leading to frustration and a loss of trust. The proposed pipeline would create a patient-specific lower gastrointestinal tract that could be used for testing treatment options in place of the patient.
Models
A multiscale, physics-based modelling framework of fluid handling in the distal colon has been developed, spanning protein to tissue levels. This framework provides a platform for mechanistically interpreting increasingly accessible omics datasets and offers a foundation for extension to other regions of the lower gastrointestinal tract. Developed in close collaboration with Platform 1, all models adhere to the modelling and annotation standards across the wider project, ensuring compatibility and reusability. As part of this effort, several novel bond-graph models were created to describe the key mechanisms underlying transport and microbiome-host interaction processes, including:
| Scale | Models | Key features/advances |
|---|---|---|
| Protein | Epithelial Na+ Channel (ENaC) | First thermodynamically consistent description of ENaC. Steady-state behaviour parameterised with patch clamp data using an SED-ML workflow developed by P1. Implemented in CellML. |
| Protein | Serotonin transporter (SERT) | First thermodynamically consistent description of SERT. Implemented in CellML. Model creation and parameterisation workflow defined in collaboration with P2. |
| Protein | G-protein coupled receptor | Introduced beta-arrestin, internalisation, recycling, and degradation pathways to the GPCR model developed by P1. |
| Subcellular processes | Post-translational modification template | An innovative phosphorylation sub-layer approach was created to capture phosphorylation states without duplicating the entire bond-graph model. Ubiquitination, a key trigger for degradation, was also defined. |
| Subcellular processes | Serotonin synthesis and degradation | Biochemical pathways implemented in CellML. |
| Subcellular processes | Carbonic anhydrase (H2O-HCO3- handling) | Biochemical pathway implemented in CellML. |
| Cell | Enterochromaffin cell | First mechanistic model of metabolite-sensitive serotonin secretion in enterochromaffin cells. |
| Cell | Enterocyte | Integrated ENaC model with existing bond graphs developed by P1. Implemented in CellML. |
| Tissue | Fluid transport FTU – distal colon (Full) | Detailed model of ion and water fluxes in the distal colon, based on an intestine-specific extension of the six-layer transport framework from P1. |
| Tissue | Fluid transport FTU – distal colon (MVP) | Reduced form that describes the contribution of Na+ transport to fluid movement in the colon. Implemented in CellML. |
| Tissue | Microbiome-host interaction serotonin handling | Description of microbiome-induced serotonin secretion, transport, and degradation dynamics. |
Workflow 1: Characterising metabolite-host interaction
We developed a bond-graph model describing serotonin handling associated with microbiome-host interaction. This model includes microbial metabolite sensing (via GPCRs), serotonin secretion (by enterochromaffin cells), and downstream uptake and degradation pathways. Model inputs were designed for compatibility with existing flux balance analysis microbiome models, such as MICOM, enabling future integration of microbial metabolism and its effect on host physiology.
Workflow 2: Enhanced IBS subtyping
Multi-omic cluster analysis has been performed using targeted metabolomics (faecal bile acid, faecal short-chain fatty acid, and plasma amino acid) and untargeted metabolomics (plasma and stool), metagenomics (stool), and symptom (gastrointestinal and psychological) information. Two key patient subgroups were identified with metabolic dysfunction and dysregulated ion transport. A journal article describing these findings has been drafted and is under peer review.
Workflow 3: Predicting treatment response
The modelling framework necessary to link omics data to predictions of fluid transport has been completed. Our collaborators have recently finalised the proteomic data analysis from the COMFORT study. However, due to when this data became available, integration into our framework was not possible. The next step is to develop a parameterisation pipeline that maps relative protein expression into absolute quantities within the fluid transport FTU. This capability will form the foundation of a patient-specific digital human that can be used for virtual screening of treatment options in future applications.