University of Southern California Dept of Biomedical Engineering
"Metabolomics and Mechanistic Kinetic Modeling Reveal Mechanisms Driving Intracellular Metabolism and Insulin Secretion in Pancreatic Beta Cells"
Pancreatic beta cells maintain blood glucose levels within a healthy range by producing insulin. Insulin production is heavily dependent on the intracellular metabolic reactions carried out by the cell, and diseases, such as Type 2 Diabetes, occur when that metabolism functions poorly. In order to better treat diabetes, we must understand glucose-stimulated insulin secretion in beta cells. Systems-focused computational modeling of metabolic processes can provide quantitative insight into the mechanisms driving insulin production in varied extracellular conditions, informing future research into novel treatments for diabetes.
We developed a kinetic, ordinary differential equation model of PBC intracellular metabolism. The model includes glycolysis, glutaminolysis, the TCA cycle, the pentose phosphate pathway, and the aldose-reductase pathway. We linked metabolism to insulin production using partial least-squares regression. We performed a global sensitivity analysis to determine the kinetic parameters that significantly influence predicted metabolite levels. We trained the model by fitting its reaction velocities (Vmax parameters) to mass spectrometry-based metabolomics measurements of metabolite levels following 5- and 30-minute stimulation of the INS-1E cell line with varied concentrations of glucose. We applied the kinetic model to simulate clinically-relevant metabolic perturbations.
The sensitivity analyses identified influential metabolic reactions. At both time points, Vmax values for the glucose transporter (GLUT2) reaction and the Glucokinase reaction were impactful on predicted metabolite levels. The results make sense, given both the primary role of beta cells (to import glucose into the cell to produce insulin) and published results showing GK to be a key regulator in overall PBC activity. At only the 30-minute time point, the Lactate Dehydrogenase, Monocarboxylate Cotransporter (MCT), and Aldose Reductase reaction velocities were found to be significantly impactful. The results suggest possible mechanisms by which extended treatment with glucose causes cells to adjust their metabolism to avoid glucotoxicity. At the 5-minute time point, the Triose-phosphate isomerase reaction velocity was influential, which may support the experimentally-seen rapid equilibration of glyceraldehyde-3-phosphate and dihydroxyacetone phosphate. Using the fitted model, we simulated 95% knockdown and upregulation of the GLUT2, MCT, pyruvate-hydrogen shuttle, and glucose-6-phosphate dehydrogenase reactions, to understand the effect of adjusted glucose import, lactate excretion, TCA Cycle flux, and PPP flux, respectively. Though the network as a whole is robust to many changes, the model predicts that controlling flux into the TCA Cycle had substantial effects at the 5-minute time point, suggesting that targeting TCA Cycle reactions may improve insulin production.