This research addresses the mechanism of coupled oscillations in calcium ion and cellular ATP/ADP driven by mitochondria, which is directly related to the defect in pulsatile insulin secretion in diabetes.
The oscillation mechanism is studied at different scales, from an individual mitochondrion, to the coupling among multiple (100 to 200) mitochondria in a single cell, to the coupling among multiple (1,000 to 10,000) cells in a single pancreatic islet. After initially developing a rigorous mathematical modeling framework for multiscale modeling and simulation, the framework is applied to develop a model for the oscillation mechanism in a single mitochondrion. These models are then coupled together using spatial information to represent multiple mitochondria and the coupled oscillation in a single cell. Finally, massively parallel computation is used to simulate the coupled oscillation among multiple cells in a single pancreatic islet. This project rigorously tests the hypothesis that synchronization of the oscillations at all scales is necessary for the correct secretion of insulin.
Diabetes Mellitus is a leading cause of death in the developed world. Currently 25.30 percent of adults have a high at risk of becoming diabetic. The prevalence of diabetes has doubled in the United States in the last 30 years and is likely to continue to increase. In healthy individuals insulin is secreted into the bloodstream after a meal as a signal for muscles and fat to use the glucose produced from the digested food. This signaling process maintains low blood glucose levels. In diabetes this signaling process is disrupted. There are two types of diabetes. Type 1 diabetes or autoimmune diabetes is characterized by a destruction of the pancreatic beta cells, the cells which secrete insulin. Type 2 diabetes is characterized by a reduction in insulin secretion as well as an inability of tissues to take up and breakdown glucose to produce energy (insulin resistance). This project focuses on the problem of insulin secretion in the normally functioning pancreas and in type 2 diabetes. We developed a mathematical model for the coupled insulin oscillation system that contains a network of a thousand pancreatic beta cells and used this model to study how aging may increase the potential for type 2 diabetes. With the mathematical model for a network of 1,000 connected pancreatic beta cells, we tried to quantitatively measure the impact of weakened mitochondria (due to aging) to the insulin oscillation behavior. We discovered that if twenty percent of the pancreatic beta cells drop fifteen percent of its mitochondria strength, the overall oscillation will be seriously damaged and broken in a long run. Through detailed study of simpler networks, we confirmed the original finding that 'twenty percent' is a critical threshold for the ratio of malfunction cells. Our discovery suggests a possible connection between aging and type II diabetes. For a system with such a large size (a network of 1,000 cells contains more than 5,000 state variables that need to be simulated for a long time), simulation efficiency is critical. In this process we also worked on the development of efficient simulation methods and the corresponding software that help to model, simulate, analyze other biochemical systems. These algorithms improve accuracy and efficiency of numerical simulation by 5 to 10 folds. Algorithms developed in this research have been implemented in the software CoPASI, a widely used software package for simulation and analysis of biochemical networks, and is available to general public. Through this project, five graduate students and four undergraduate students received training in computational biology research. Among the four undergraduate students, three of them belong to underrepresented group (two female students and one disabled student). The research topics related to this project have been covered in the graduate course: Computational Cell Biology cross-listed in the departments of computer science and biological sciences. Part of the project material was also used in the undergraduate course 'Numerical Methods' .
