Genomic and Proteomic Biomarkers of Cardiac Allograft Vasculopathy

Long-term survival of cardiac transplant recipients remains a significant hurdle largely due to cardiac allograft vasculopathy (CAV) as an expression of chronic rejection (CR). Currently, coronary angiography and intravascular ultrasound, invasive and expensive modalities, are considered the standard for diagnosis of CAV. In the current study, we combined assessment of blood mRNA and proteins to identify potential biomarkers for diagnosis of CAV. Whole blood Affymetrix microarrays and plasma iTRAQ-MALDI-TOF/TOF proteomic analyses were carried out on 25 cardiac transplant patient samples. CAV diagnosis was made based on blinded quantitative analysis of angiograms, intravascular ultrasound images and clinical chart review. The genomics and proteomics data were assessed by moderated t-tests and discriminant analysis. Two panels of 10 differentially expressed genes (FDR<5%), and 10 proteins with differential relative levels (p-values<0.025) between CAV and non-CAV samples were independently identified. Classification of new test samples revealed that the 10 genes together can discriminate between CAV and non-CAV samples with 83% sensitivity and specificity. Classification of the same samples using the 10 PGC’s dropped the sensitivity to 67%. A combinatorial panel derived from the genomic and proteomic panels provided improvement in performance, at 100% sensitivity and 83% specificity. Leave-one-out cross validation showed similar results. Genomic, proteomic and combinatorial blood-based biomarkers are promising as potential minimally-invasive, sensitive and specific modalities for the diagnosis of CAV. Classifier panels generated from current work are being evaluated in a prospective, multi-site observational trial. This study, approved by the Providence Health Care, the Vancouver Coastal Health and the UBC Research Ethics Boards, was carried out under the Biomarkers in Transplantation initiative on subjects who received a cardiac transplant and signed consent forms. A total of 25 CAV and non-CAV blood samples, collected between year one and year five post-transplantation from 17 patients were selected for analyses. Thirteen subjects with ‘extreme phenotypes’ were selected for the biomarker discovery analysis, resulting in a training cohort of 7 CAV and 6 non-CAV samples, collected mostly at one year post-transplant. The remaining samples (6 CAV and 6 non-CAV) were subsequently utilized as the test cohort. Diagnosis of CAV was made based on chart reviews (i.e., angiography and/or IVUS measurements) at timepoints corresponding to the blood sample collection dates.