PhD Thesis A nutritional systems biology approach to clarifying molecular pathology of metabolic syndrome, diabetes and transcriptomic effects of dietary fatty acid variation.


Introduction: Nutritional genomics is a field that seeks to clarify the molecular link between diet and health. With the current dramatic growth in obesity, the symptomatic precursors of diabetes and cardiovascular disease (referred to as metabolic syndrome) are important targets in nutritional genomic research as they are strongly modifiable by diet. The goal of this thesis is to use transcriptomic analysis to examine the effects of the pro-inflammatory IL-1 signalling pathway and pro-fibrotic TGF-ß1 pathway, as well as dietary intake of polyunsaturated fatty acids, as they relate to metabolic health. Methods: A variant of pathway analysis was used to identify pathways showing bi- directional changes in expression in response to dietary intervention. As an alternative to pathway analysis, a novel approach to metabolic network analysis was developed to discover co-expressed network paths linked to genes with diet- or genotype-sensitive expression. To identify treatment-sensitive genes with potential clinical relevance, gene expression patterns were jointly analyzed with plasma markers of metabolic health under a multivariate statistical framework. Results: In the context of high fat diet, genetic knockdown of the type I IL-1 receptor induced expression changes in transcriptional targets of STAT3, and up-regulation of diacylgycerol metabolism and tricarboxylic acid cycle. In a second mouse study, a treatment diet high in conjugated linoleic acid (CLA) had profound effects on hepatic gene expression; joint analysis with plasma markers identified a set of metabolic pathways linked to both dietary CLA and metabolic health. A network analysis of adipose tissue transcriptomes in humans with metabolic syndrome revealed correlation between dietary intake of n-3 PUFA correlated with expression of genes related to adipogenesis and energy metabolism, and highlighted the utility of a network-based approach to transcriptomic analysis. RNA-seq profiling of a kidney cell line treated with TGF-ß1 (a model of diabetic kidney disease) identified novel functional changes that were reiterated in human kidney fibrosis. Only a small fraction of these changes was identified in a parallel microarray analysis, emphasizing the sensitivity of RNA-seq. Discussion: Transcriptomics – in particular, RNA-seq - is a highly sensitive tool in nutritional genomic studies, though the high-throughput nature of this data type presents unique challenges. Multivariate analysis identified clinical relevance in many genes that may have been otherwise overlooked. Furthermore, network analysis showed promise as an analytical means to address emerging limitations of pathway analysis. Future work should prioritize extensions to multivariate methods to integrate genotype and additional high-throughput data types under a probabilistic statistical framework.

Paper Details


M. Morine