Transcriptional responses of insulin resistance and mechanisms of drug action

Insulin resistance is the defining feature of the metabolic syndrome and the primary defect leading to type 2 diabetes. Insulin resistance is a pathological state in which the normal physiological function of insulin is impaired in target tissues including skeletal muscle, adipose tissue, and liver. Thiazolidinediones, synthetic ligands of the PPAR[gamma] nuclear receptor, are clinically potent insulin sensitizing drugs; however, not all individuals respond to treatment. Although PPAR[gamma] -mediated gene regulation is the predominant mode of thiazolidinedione-enhanced insulin sensitivity, the exact mechanisms by which PPARgamma activation leads to insulin sensitization or by which insulin sensitization is prevented are poorly understood. Addressing these expansive but fundamental questions concerning multi- tissue insulin resistance and PPAR[gamma] ligand-mediated insulin sensitization necessitates a comprehensive system- wide approach. The advent and maturation of the gene expression microarray allows for high-throughput transcript measurements. The massively parallel nature of microarrays, though, requires robust statistical analysis in order to obtain sensitive, accurate, and reliable results. With these tools, we characterize human subjects whom range in insulin sensitivity and relate their clinical phenotypes with functional pathway alterations. Multi-tissue gene expression profiles are further interrogated for distinctions between thiazolidinedione treatment responder and non-responder subjects. We further characterize the insulin sensitivity and multi-tissue gene expression profiles of lean and insulin resistant, obese Zucker rats treated with one of four PPAR[gamma] ligands, in order to identify functional pathways which are necessary for effective insulin sensitization.A robust statistical approach is formulated, through which sensitivity analysis of leading microarray statistical testers on various controlled datasets demonstrates that a global variance modeling methodology is advantageous towards increased sensitivity for the detection of true differential expression. With this devised approach, we identify pathway alterations in multiple tissues which are characteristic of insulin resistance, TZD-induced insulin sensitization, and potential TZD responsiveness. We also identify transcriptional biomarkers of insulin resistance and response signatures of PPAR[gamma] ligand treatment. These new insights suggest new targets for the therapy of type 2 diabetes, and measures to decrease risk for existing treatment