Nutritional Systems Biology of Fat : integration and modeling of transcriptomics datasets related to lipid homeostasis

Fatty acids, in the form of triglycerides, are the main constituent of the class of dietary lipids. They not only serve as a source of energy but can also act as potent regulators of gene transcription. It is well accepted that an energy rich diet characterized by high intakes of dietary fat is linked to the dramatic increase in the prevalence of obesity in both developed and developing countries in the last several decades. Obese individuals are at increased risk of developing the metabolic syndrome, a cluster of metabolic abnormalities that ultimately increase the risk of developing vascular diseases and type 2 diabetes. Many studies have been performed to uncover the role of fatty acids on gene expression in different organs, but integrative studies in different organs over time driven by high throughput data are lacking. Therefore, we first aimed to develop integrative approaches on the level of individual genes but also pathways using genome-wide transcriptomics datasets of mouse liver and small intestine that are related to fatty acid sensing transcription factor peroxisome proliferator activated receptor alpha (PPARα). We also aimed to uncover the behavior of PPARαtarget genes and their corresponding biological functions in a short time series experiment, and integrated and modeled the influence of different levels of dietary fat and the time dependency on transcriptomics datasets obtained from several organs by developing system level approaches.

We developed an integrative statistical approach that properly adjusted for multiple testing while integrating data from two experiments, and was driven by biological inference. By quantifying pathway activities in different mouse tissues over time and subsequent integration by partial least squares path model, we found that the induced pathways at early time points are the main drivers for the induced pathways at late time points. In addition, using a time course microarray study of rat hepatocytes, we found that most of the PPARα target genes at early stage are involved in lipid metabolism-related processes and their expression level could be modeled using a quadratic regression function. In this study, we also found that the transcription factorsNR2F, CREB, EREF and RXR might work together with PPARα in the regulation of genes involved in lipid metabolism. By integrating time and dose dependent gene expression data of mouse liver and white adipose tissue (WAT), we found a set of time-dose dependent genes in liver and WAT including potential signaling proteinssecreted from WAT that may induce metabolic changes in liver, thereby contributing to the pathogenesis of obesity.

Taken together, in this thesis integrative statistical approaches are presented that were applied to a variety of datasets related to metabolism of fatty acids. Results that were obtained provide a better understanding of the function of the fatty acid-sensor PPARa, and identified a set of secreted proteins that may be important for organ cross talk during the development of diet induced obesity.