C. Imai

Whole Blood Genomic Biomarkers of Acute Cardiac Allograft Rejection

BACKGROUND Significant progress has been made in cardiac transplantation over the past 30 years; however, the means for detection of acute cardiac allograft rejection remains in need of improvement. At present, the endomyocardial biopsy, an invasive and inconvenient procedure for patients, is required for the surveillance and diagnosis of acute cardiac allograft rejection. In the Biomarkers in Transplantation initiative, we investigated gene expression profiles in peripheral blood of cardiac transplant subjects as potential biomarkers for diagnosis of allograft rejection. METHODS Whole blood samples were obtained from 28 cardiac transplant subjects who consented to the study. Serial samples were collected from pre-transplant through 3 years post-transplant according to the standard protocol. Temporally correspondent biopsies were also collected, reviewed in a blinded manner, and graded according to current ISHLT guidelines. Blood samples were analyzed using Affymetrix microarrays. Genomic profiles were compared in subjects with acute rejection (AR; ISHLT Grade > or =2R) and no rejection (NR; Grade 0R). Biomarker panel genes were identified using linear discriminant analysis. RESULTS We found 1,295 differentially expressed probe-sets between AR and NR samples and developed a 12-gene biomarker panel that classifies our internal validation samples with 83% sensitivity and 100% specificity. CONCLUSIONS Based on our current results, we believe whole blood genomic biomarkers hold great potential in the diagnosis of acute cardiac allograft rejection. A prospective, Canada-wide trial will be conducted shortly to further evaluate the classifier panel in diverse patients and a range of clinical programs.

Plasma Protein Biosignatures for Detection of Cardiac Allograft Vasculopathy

BACKGROUND Coronary angiography remains the most widely used tool for routine screening and diagnosis of cardiac allograft vasculopathy (CAV), a major pathologic process that develops in 50% of cardiac transplant recipients beyond the first year after transplant. Given the invasiveness, expense, discomfort, and risk of complications associated with angiography, a minimally invasive alternative that is sensitive and specific would be highly desirable for monitoring CAV in patients. METHODS Plasma proteomic analysis using isobaric tags for relative and absolute quantitation-matrix-assisted laser desorption ionization double time-of-flight mass spectrometry was carried out on samples from 40 cardiac transplant patients (10 CAV, 9 non-significant CAV, 21 possible CAV). Presence of CAV was defined as left anterior descending artery diameter stenosis ≥ 40% by digital angiography and quantitatively measured by blinded expert appraisal. Moderated t-test robust-linear models for microarray data were used to identify biomarkers that are significantly differentially expressed between patient samples with CAV and with non-significant CAV. A proteomic panel for diagnosis of CAV was generated using the Elastic Net classification method. RESULTS We identified an 18-plasma protein biomarker classifier panel that was able to classify and differentiate patients with angiographically significant CAV from those without significant CAV, with an 80% sensitivity and 89% specificity, while providing insight into the possible underlying immune and non-alloimmune contributory mechanisms of CAV. CONCLUSION Our results support of the potential utility of proteomic biomarker panels as a minimally invasive means to identify patients with significant, angiographically detectable coronary artery stenosis in the cardiac allograft, in the context of post-cardiac transplantation monitoring and screening for CAV. The potential biologic significance of the biomarkers identified may also help improve our understanding of CAV pathophysiology.

Predicting acute cardiac rejection from donor heart and pre-transplant recipient blood gene expression

BACKGROUND Acute rejection in cardiac transplant patients remains a contributory factor to limited survival of implanted hearts. Currently, there are no biomarkers in clinical use that can predict, at the time of transplantation, the likelihood of post-transplant acute cellular rejection. Such a development would be of great value in personalizing immunosuppressive treatment. METHODS Recipient age, donor age, cold ischemic time, warm ischemic time, panel-reactive antibody, gender mismatch, blood type mismatch and human leukocyte antigens (HLA-A, -B and -DR) mismatch between recipients and donors were tested in 53 heart transplant patients for their power to predict post-transplant acute cellular rejection. Donor transplant biopsy and recipient pre-transplant blood were also examined for the presence of genomic biomarkers in 7 rejection and 11 non-rejection patients, using non-targeted data mining techniques. RESULTS The biomarker based on the 8 clinical variables had an area under the receiver operating characteristic curve (AUC) of 0.53. The pre-transplant recipient blood gene-based panel did not yield better performance, but the donor heart tissue gene-based panel had an AUC = 0.78. A combination of 25 probe sets from the transplant donor biopsy and 18 probe sets from the pre-transplant recipient whole blood had an AUC = 0.90. Biologic pathways implicated include VEGF- and EGFR-signaling, and MAPK. CONCLUSIONS Based on this study, the best predictive biomarker panel contains genes from recipient whole blood and donor myocardial tissue. This panel provides clinically relevant prediction power and, if validated, may personalize immunosuppressive treatment and rejection monitoring.