Abstract 1806: Discovery of Genes Expressed in Whole Blood which Identify Patients with Coronary Artery Disease
Background: Atherosclerosis is a complex process primarily occurring within the coronary vessel wall. We previously demonstrated that changes in RNA levels from peripheral blood mononuclear cell samples correlated with the extent of CAD. In the current study we sought to expand these observations to whole blood RNA from patients with angiographically defined CAD.
Methods: Total blood RNA was isolated from PAXgene tubes drawn from non-diabetic patients undergoing coronary angiography and participating in the CATHGEN registry, a biorepository paired with clinical and angiographic data in the Duke Databank for Cardiovascular Disease. Cases had ≥75% stenosis in ≥1 major coronary artery; controls had ≤ 25% luminal stenosis in any major coronary arteries. We performed Agilent whole genome microarrays on 92 subjects (54 cases, 38 controls), followed by RT-PCR validation on the same sample set.
Results: Cases were more often male (74% vs 30%, P<.01) and older (65 vs 51 years old, P<.01) than controls. Array analysis using sex as a covariate yielded 923 genes with >1.2 fold differential expression between cases and controls (p <0.05). A subset of 61 genes was selected for further testing by RT-PCR. In an unadjusted analysis, 30 of these genes were significant (p <0.05); in a robust linear model 20 of these genes retained significance after adjusting for sex and age. Inclusion of the 20 genes in a LASSO cross-validation model yielded an AUC of 0.88 (95% CI, 0.809 to 0.943), with a sensitivity of 65% and specificity of 86%. The significant genes were largely in pathways previously implicated in the development and progression of CAD, including nitric oxide generation, cysteine biosynthesis, chemokine signaling, lipid metabolism, extracellular matrix modification, and oxidative stress responses. In addition, database analysis suggests that a large fraction of these genes are predominantly expressed in the myelomonocytic lineage of leukocytes.
Conclusion: Whole blood gene expression can reflect the presence or absence of CAD and represents a possible tool for development of a diagnostic classifier to distinguish these patients.