Identifying low-grade cellular rejection after heart transplantation in children by using gene expression profiling

Endomyocardial biopsy (EMB) remains the gold standard for detecting rejection after heart transplantation but is costly and invasive. This study aims to distinguish no rejection (0R) from low-grade rejection (1R/2R) after heart transplantation in children by using global gene expression profiling in blood. A total of 106 blood samples with corresponding EMB from 18 children who underwent heart transplantation from 2011 to 2014 were analyzed (18 baseline/pretransplantation samples, 88 EMB samples). Corresponding rejection grades for each blood sample were 0R in 39% (34/88), 1R in 51% (45/88), and 2R in 10% (9/88). mRNA from each sample was sequenced. Differential expression analysis was performed at the gene level. A k-nearest neighbor (kNN) analysis was applied to the most differentially expressed (DE) genes to identify rejection after transplantation. Mean age at transplantation was 10.0 ± 5.4 yr. Expression of B cell and T cell receptor sequences was used to measure the effect of posttransplantation immunosuppression. Follow-up samples had lower levels of immunoglobulin gene families compared with pretransplantation ( P < 3E-5) (lower numbers of activated B cells). T cell receptor alpha and beta gene families had decreased expression in 0R samples compared with pretransplantation ( P < 4E-5) but recovered to near baseline levels in 1R/2R samples. kNN using the most DE gene (MKS1) and k = 9 nearest neighbors correctly identified 83% (73/88) of 1R/2R compared with 0R by leave-one-out cross validation. Using a genomic approach we can distinguish low-grade cellular allograft rejection (1R/2R) from no rejection (0R) after heart transplantation in children despite a wide age range.