Integrative pathway dissection of molecular mechanisms of moxLDL-induced vascular smooth muscle phenotype transformation

Abstract

Background: Atherosclerosis (AT) is a chronic inflammatory disease characterized by the accumulation of inflammatory cells, lipoproteins and fibrous tissue in the walls of arteries. AT is the primary cause of heart attacks and stroke and is the leading cause of death in Western countries. To date, the pathogenesis of AT is not well-defined. Studies have shown that the dedifferentiation of contractile and quiescent vascular smooth muscle cells (SMC) to the proliferative, migratory and synthetic phenotype in the intima is pivotal for the onset and progression of AT. To further delineate the mechanisms underlying the pathogenesis of AT, we analyzed the early molecular pathways and networks involved in the SMC phenotype transformation.

Methods: Quiescent human coronary artery SMCs were treated with minimally-oxidized LDL (moxLDL), for 3 hours and 21 hours, respectively. Transcriptomic data was generated for both time-points using microarrays and was subjected to pathway analysis using Gene Set Enrichment Analysis, GeneMANIA and Ingenuity software tools. Gene expression heat maps and pathways enriched in differentially expressed genes were compared to identify functional biological themes to elucidate early and late molecular mechanisms of moxLDL-induced SMC dedifferentiation.

Results: Differentially expressed genes were found to be enriched in cholesterol biosynthesis, inflammatory cytokines, chemokines, growth factors, cell cycle control and myogenic contraction themes. These pathways are consistent with inflammatory responses, cell proliferation, migration and ECM production, which are characteristic of SMC dedifferentiation. Furthermore, up-regulation of cholesterol synthesis and dysregulation of cholesterol metabolism was observed in moxLDL-induced SMC. These observations are consistent with the accumulation of cholesterol and oxidized cholesterol esters, which induce proinflammatory reactions during atherogenesis. Our data implicate for the first time IL12, IFN-α, HGF, CSF3, and VEGF signaling in SMC phenotype transformation. GPCR signaling, HBP1 (repressor of cyclin D1 and CDKN1B), and ID2 and ZEB1 transcriptional regulators were also found to have important roles in SMC dedifferentiation. Several microRNAs were observed to regulate the SMC phenotype transformation via an interaction with IFN-γ pathway. Also, several "nexus" genes in complex networks, including components of the multi-subunit enzyme complex involved in the terminal stages of cholesterol synthesis, microRNAs (miR-203, miR-511, miR-590-3p, miR-346*/miR- 1207-5p/miR-4763-3p), GPCR proteins (GPR1, GPR64, GPRC5A, GPR171, GPR176, GPR32, GPR25, GPR124) and signal transduction pathways, were found to be regulated.

Conclusions: The systems biology analysis of the in vitro model of moxLDL-induced VSMC phenotype transformation was associated with the regulation of several genes not previously implicated in SMC phenotype transformation. The identification of these potential candidate genes enable hypothesis generation and in vivo functional experimentation (such as gain and loss-of-function studies) to establish causality with the process of SMC phenotype transformation and atherogenesis.

Figure 1

Enrichment map for moxLDL, p=0.005, q=0.1; Nodes represent gene sets that are enriched at the top or bottom of the ranking of differentially expressed genes, as determined by GSEA, where the node size corresponds to number of genes in the set. Edges indicate overlap between gene sets, where the thickness indicates the size of the overlap. Red indicates up-regulation, blue indicates down-regulation. The centre of the nodes corresponds to the early time point (3h), whereas the border of the node corresponds to the late time point (21h).

Figure 2

‘Endopeptidase Inhibition’ theme heatmap of representative SMC differential gene expression patterns induced by treatment with moxLDL at 3h and 21 h.

Figure 3

Canonical Pathways from IPA. Bars correspond to the top 20 Canonical Pathways that surpassed the Ingenuity statistical threshold (orange squares) using the Fisher’s exact test in the 3h (3A) and 21h (3B) moxLDL experiments, respectively.

Figure 4

Cholesterol biosynthesis theme analysis. (A) Cholesterol biosynthesis theme from enrichment map. (B) Cholesterol biosynthesis-associated heatmap. (C) Network of interactions among moxLDL-SMC cholesterol metabolism related genes, as retrieved by the GeneMANIA website, colored by gene expression at 3h (node center) and 21h (node border). Red indicates up-regulation, blue indicates down-regulation and white indicates no differential expression or no expression data available. Circles represent genes and connecting lines represent interactions between genes. The network was generated using a GeneMANIA query of 43 cholesterol synthesis related genes (large circles) differentially expressed at either 3h or 21h. GeneMANIA retrieved known and predicted interactions between these genes and added extra genes (smaller white circles) that are strongly connected to query genes (we used the default setting of 20 additional connecting genes). Light blue lines indicate pathway interactions from the Reactome pathway database, dark blue lines indicate experimentally determine physical interactions, from various protein interaction databases included in GeneMANIA, and brown lines indicate predicted interactions, mostly from the I2D database of protein interactions predicted from experimentally determined physical interactions in other species. GeneMANIA advanced settings were used to search only physical, pathway and the default set of predicted networks. We excluded co-expression, co-localization and genetic interaction networks from the search to focus the analysis on higher confidence physical and pathway interactions.

Figure 5

Cytokine and growth factor theme analysis. IPA networks from 3h (A, B, D) and 21h (C) experiments, involving cytokines and growth factors as molecular hubs. Genes/proteins are illustrated as nodes and molecular relationships as connecting lines between two nodes (direct relationships as normal lines; indirect relationships as dashed lines). Molecular relationships are supported by at least one literature reference, or by canonical information stored in the IPKB. Grey nodes represent genes of interest, while white nodes represent hubs that were added by the IPA algorithm to connect a set of genes of interest.

Figure 6

GPCR signaling theme analysis. (A) GPCR theme from enrichment map. (B) GPCR - associated heatmap. (C)GPCR-associated IPA network (3h) (D) GPCR-associated IPA network (21h). For color coding interpretation, please refer to Figures 1 &5.

Figure 7

Cell adhesion theme analysis. (A) cell-to-cell junction theme from enrichment map. (B) Cell- to-cell junction-associated heatmap. For color coding interpretation, please refer to Figure 1.

Figure 8

Cell-cycle control theme analysis. (A) Cell-cycle theme from enrichment map. (B) Cell-cycle- associated heatmap. (C) Cell-cycle-associated IPA network (21h). For color coding interpretation, please refer to Figures 1 &5.

Figure 9

Myogenic contraction-associated IPA network. For color coding interpretation, please refer to Figure 5.