DYNAMIC MODELING IN DIABETES: FROM WHOLE BODY TO GENES

Abstract

Diabetes is one of the major chronic diseases such that, together with its complications it can account for more than 10% of national healthcare expenditure. Mathematical modeling can enhance understanding of this disease in quantitative terms and is becoming an increasingly important aid in diagnosis, prognosis and in the planning of therapy. Mathematical modeling in relation to carbohydrate metabolism and diabetes has a long history stretching back some 45 years. Initially modeling has focused on the dynamics of glucose and insulin and their interactions, principally at the whole body and organ levels. However, over recent years the scope of mathematical modeling in relation to diabetes has seen dramatic expansion such that it is now being applied across the spectrum from populations of patients (public health) to the molecular level. This paper will explore recent developments of mathematical modeling in our laboratory across this ever increasing spectrum. Ingredients will include models to assess, at whole body, the efficacy of homeostatic control and system fluxes and, at organ level, unit processes in skeletal muscle, a key target tissue. To do so both whole body as well as regional tracer experiments, these last employing Positron Emission Tomography, will be discussed not only to understand the physiology but also the pathophysiology of glucose metabolism, like obesity and diabetes. Microarray technology offers an important tool to understand how genes change expression and interact as a consequence of external/internal stimuli. Dynamic stimulus/response experiments can provide time series expression data from which regulatory networks can be obtained by reverse engineering, and this is illustrated for insulin stimulation of muscle rat cells. Recent technological advances in diabetes include more reliable subcutaneous glucose sensors: interpretation and clinical use of continuous glucose monitoring time series data can be powered by dynamic modeling, in particular we show how critical hypoglycemic events can be predicted ahead in time. Finally, the importance of dynamic modeling in an important diabetes health care problem is discussed by showing its use in conjunction with gait analysis for preventing diabetic foot complications.