The M3al Model; towards precision nutrition in overweight an obesity.

Background
An energy imbalance in overweight and obesity can induce disturbances in the metabolic crosstalk between the liver, skeletal muscle, and adipose tissue giving rise to the development of dyslipidemia, insulin resistance, ultimately a loss of glycaemic control. While specific nutritional interventions interventions have been shown to ameliorate the effects of these co-morbidities, and in some cases allow the recovery of a metabolically healthy state, the considerable inter-individual heterogeneity has greatly impeded the identification of effective intervention and treatment options for overweight/obesity. In this study we generate a novel physiology based computational model to quantify features of metabolic resilience from meal challenge test data, allowing for a personalised assessment of intervention response.
Methods
We construct the M3al Model, a novel physiology based computational model of postprandial glucose and lipid metabolism which can be personalised using post-meal time series of plasma glucose, insulin, triglyceride, and free-fatty acid concentrations. We applied the M3al Model to meal challenge test data from a large population of overweight/obese individuals (n=317) collected from five different intervention studies (caloric restriction, improved macronutrient quality, or a combination of both) identifying postprandial metabolic signatures insulin resistance and elevated liver fat.
Results
The M3al Model was capable of capturing the diverse individual responses to standardised meals. Moreover, the M3al Model is sufficiently robust to fit responses from mixed meals with diverse macro-nutrient compositions used in the different intervention studies included in this analysis. The M3al Model could be used to infer rates of fluxes which are not directly measured, with the model predicting an increase in VLDL secretion from the liver in individuals with increase insulin resistance, which is confirmed with MRS quantification of liver fat. The M3al Model also quantifies insulin resistance and rates of insulin secretion from the plasma trajectories of glucose, insulin, triglycerides, and free-fatty acids, thereby providing an individual assessment of metabolic resilience.
Conclusion
The use of personalisable in silico computational models such as the M3al Model can capture the physiologically relevant dynamic features of meal challenge test data, providing an objective and sensitive assessment of metabolic health for overweight and obese individuals supporting the provision of more targeted nutritional intervention.