Gene networks and modules in atherosclerosis

Abstract

In this thesis we are using global gene expression profiles to unravel functional gene networks and modules. The focus is atherosclerosis, a disease with manifestations in the artery wall where deposits of lipids accumulate and trigger immune responses causing the development of plaques, which upon rupture can lead to a myocardial infarction or stroke. Atherosclerosis is a complex disease influenced by energy metabolism in multiple organs and by several genetic and environmental risk factors. To meet this complexity, we believe the most appropriate approach is to identify gene networks and modules in patients suffering coronary artery disease as well as a relevant mouse model with human-like dyslipidemia prone to atherosclerosis development.

First, we investigate structural properties of the regulatory gene network in yeast, integrating protein protein interactions with the transcription network resulting in an estimate the effective gene network underlying gene expression data. In this effective gene network, we show evidence of in-hubs and provide a method for predicting in-hubs directly from gene expression data.

In the second study, we used the Ldlr−/− Apob100/100 Mttpflox/flox Mx1-Cre mouse model to study atherosclerosis development and how this development is effected by plasma cholesterollowering. This mouse model has a lipid profile similar to human hyperlipidemia and develops atherosclerosis on a chow diet. Moreover, it contains a genetic switch (Mttpflox/flox Mx1-Cre) to turn off the VLDL synthesis in the liver and lowering plasma cholesterol by > 80%. Atherosclerotic lesions progressed slowly at first, then expanded rapidly, and plateaued after advanced lesions formed. Analysis of lesion expression profiles indicated lipid-poor macrophages accumulated prior the rapid expansion of the plaques. When macrophage concentration reached a critical point it was followed by a rapid expansion phase with accelerated foam-cell formation and inflammation, an interpretation also supported by lesion histology. A network of 8 cholesterol-responsive atherosclerosis genes was identified as central to the rapid expansion of the plaques.

Third, in the Stockholm Atherosclerosis Gene Expression (STAGE) study, including 124 well-characterized patients undergoing coronary artery bypass surgery, we measured and analyzed 278 expression profiles from the liver, skeletal muscle, mediastinal fat, and aortic lesion (atherosclerotic artery expression with unaffected arterial wall expression subtracted). Clustering of these gene expression profiles performed separately in each organ generated a total of 60 clusters. Two clusters, in aortic lesion (n = 49) and fat (n = 59), related to degree of atherosclerosis. Remarkably, in a validation cohort 27 genes were replicated in a cluster (n =55) also related to the degree of atherosclerosis. In all three clusters relating to atherosclerosis (i.e., the atherosclerosis module), genes in the transendothelial migration of leukocyte pathway (TEML) were overrepresented and the transcription co-factor LIM-domain binding 2 (LDB2) expressed in lesion macrophages and endothelial cells was identified as a potential regulator of this module.

In the last study, we first identified 2457 cholesterol-responsive genes in the atherosclerotic arterial wall by lowering plasma cholesterol at 10-weeks intervals during atherosclerosis development using the mouse model of Study II. To prioritize the most atherosclerosis-relevant genes among these 2457, we used a list of 1259 genes active during atherogenesis (Study II) together with three global gene networks generated from human atherosclerosis gene expression profiles in study III, public literature mining, and protein-protein interaction data. Using an integrative network approach to identify genes neighboring any of 68 atherosclerosis seed genes, we identified 35 cholesterol-responsive genes that were believed to be highly relevant to atherosclerosis.

Taken together, this thesis provides evidence that systems biological analysis of global gene expression profiles isolated from a wide range of biological specimens can be used to infer functional interactions of genes in modules or networks. The content and architecture of these modules and networks can be used to improve our understanding how complex disorders like atherosclerosis develop.