Comparative genome-wide transcriptional analysis of human left and right internal mammary arteries

Abstract

In coronary artery bypass grafting (CABG), the combined use of left and right internal mammary arteries (LIMA and RIMA) - collectively known as bilateral IMAs (BIMAs) provides a survival advantage over the use of LIMA alone. However, gene expression in RIMA has never been compared to that in LIMA. Here we report a genome-wide transcriptional analysis of BIMA to investigate the expression profiles of these conduits in patients undergoing CABG. As expected, in comparing the BIMAs to the aorta, we found differences in pathways and processes associated with atherosclerosis, inflammation, and cell signaling - pathways which provide biological support for the observation that BIMA grafts deliver long-term benefits to the patients and protect against continued atherosclerosis. These data support the widespread use of BIMAs as the preferred conduits in CABG.

Keywords: Cardiovascular disease; Coronary artery bypass grafting; Genomics; Transcriptome analysis.

Fig. 1

Statistical significance of canonical pathways with genes differentially expressed between the LIMA and RIMA based on the entire dataset of 43 IMA tissue samples (20 LIMA/23 RIMA). The yellow line represents the significance threshold based on permutation analysis.

Fig. 2

Statistical significance of canonical pathways with genes differentially expressed between the LIMA and RIMA based on the 28 paired samples derived from 14 patients using data derived from both IMA tissues.

Fig. 3

Inferred connectivity map centered on NF-κB for differentially expressed genes between LIMA and RIMA. Red indicates that the genes are up-regulated in LIMA relative to RIMA and green indicates they were down-regulated. Uncolored genes or complexes were not found to be differentially expressed in our original analysis but were found through the pathway analysis. Note that this network does not exhibit a coherent pattern of biologically-driven regulation but instead a more random distribution of genes and across the network.

Fig. 4

Statistical significance of (A) canonical pathways and (B) assigned biological functions associated with genes differentially expressed between aorta and IMA samples as determined using Ingenuity Pathway Assist. The height of bars reflects the p-value from Fisher’s exact test. Blue bars are associated with the 782 differentially expressed genes in all aorta vs. all IMA comparisons. Orange bars present results for the 653 differentially expressed genes from 14 trios. Cyan bars represent results for the 360 genes found to be differentially expressed in both sets. The annotated classes were sorted by the p-value of the overlapping gene set (cyan). Here the significance threshold, shown as a dashed line, was estimated by permutation testing. The canonical pathway “atherosclerosis signaling” was ranked as the most significant pathway. The vast majority of the biological processes found to be significant are associated with inflammation, including those associated with the term “cardiovascular disease.” Detailed information on these significant canonical pathways and functional classes and a list of their component genes is provided in the Supplementary materials.

Fig. 5

Inferred connectivity maps for differentially expressed genes in (A) the “inflammatory response” network and (B.) networks associated with “lipid metabolism” and “cellular movement” Red indicates that the genes are up-regulated in IMA relative to aorta and green indicates they were down-regulated. Uncolored genes or complexes were not found to be differentially expressed in our original analysis but were found through the pathway analysis.