Simulating and predicting glucose response in lean and obese mice

Diabetes mellitus is a serious and widespread disease currently affecting 25.8 million people in the United States. Its treatments including exercise, diet, hypoglycemic pills, and insulin often are not sufficient to mitigate the symptoms of the disease or prevent its progression over time. An artificial pancreas capable of continuously monitoring and adjusting blood glucose levels through closed-loop control thus maintaining optimal range of glucose levels at all times will be an ideal treatment for the disease. One of the largest hurdles in developing an artificial pancreas is the creation of an individualized algorithm capable of calculating insulin doses from blood glucose readings and controlling the pump to deliver the appropriate doses of insulin. As a step toward developing algorithms for artificial pancreas, we adopted a mathematical model by Lombarte et al. [3] to study the homeostasis of glucose and insulin in lean and obese mice. We administered glucose (1 g/kg) intra-peritoneally and measured blood glucose levels at various intervals for 120 min. We performed simulation using Arena
^TM
software based on the mathematical model and estimated the rate constants (8 parameters) for various terms in the differential equations using OPtQuest
^TM
. The simulated data fit accurately to the observed data for both lean and obese mice, validating the use of the mathematical model in mice at different stages of diabetes progression. Examination of the rate constants may provide insights into which parameters are mostly affected by the progression of diabetes and thus how tightly a control algorithm needs to be calibrated for these parameters to maintain proper control.