Gene Expression Profiling for the Identification and Classification of Antibody-Mediated Heart Rejection
Background—Antibody-mediated rejection (AMR) contributes to heart allograft loss. However, an important knowledge gap remains regarding the pathophysiology of AMR and how detection of immune activity, injury degree and stage could be improved by intragraft gene expression profiling.
Methods—We prospectively monitored 617 heart transplant recipients referred from four French transplant centers (January 1st 2006 - January 1st 2011) for AMR. We compared patients with AMR (n=55) to a matched control group of 55 patients without AMR. We characterized all patients using histopathology (ISHLT 2013), immunostaining, and circulating anti-HLA donor-specific antibodies (DSA) at the time of biopsy, together with systematic gene expression assessments of the allograft tissue using microarrays. Effector cells were evaluated using in vitro human cell cultures. We studied a validation cohort of 98 heart recipients transplanted in Edmonton, Canada including 27 cases of AMR and 71 controls.
Results—A total of 240 heart transplant endomyocardial biopsies were assessed. AMR showed a distinct pattern of injury characterized by endothelial activation with microcirculatory inflammation by monocytes / macrophages and natural killer (NK)-cells. We also observed selective changes in endothelial/angiogenesis and NK cell transcripts, including CD16A signaling and select interferon gamma (IFNG)-inducible genes. The AMR selective gene sets accurately discriminated patients with AMR from those without and included NK transcripts (AUC=0.87), endothelial activation transcripts (AUC=0.80), macrophage transcripts (AUC=0.86), and IFNG transcripts (AUC=0.84, p<0.0001 for all comparisons). These 4 gene sets showed increased expression with increasing AMR ISHLT grade (p<0.001) and association with DSA levels. The unsupervised PCA analysis demonstrated a high proportion of molecularly inactive pAMRI+ while there was significant molecular overlap between pAMR1H+ and full-blown pAMR2-3 cases. Endothelial activation transcripts, IFNG and NK transcripts showed association with chronic allograft vasculopathy. The molecular architecture and selective AMR transcripts together with gene set discrimination capacity for AMR identified in the discovery set, were reproduced in the validation cohort.
Conclusions—Tissue-based measurements of specific pathogenesis-based transcripts reflecting NK burden, endothelial activation, macrophage burden and IFNG effects accurately classify AMR and correlate with degree of injury and disease activity. This study illustrates the clinical potential of a tissue-based analysis of gene transcripts to refine diagnosis of heart transplant rejection.
- antibody-mediated rejection
- HLA antibody
- microarray biomarker
- translational research
- genetics, microarrays
- Received April 8, 2016.
- Revision received December 30, 2016.
- Accepted January 23, 2017.