Transcriptomics of Chronic Active Antibody-Mediated Rejection of Human Kidney Allografts and Identification of Intragraft Overexpression of Natural Kill Cell Cytotoxicity Gene Set

The enigmatic natural killer (NK) cells mediate spontaneous cell-mediated cytotoxicity and antibody-dependent cell-mediated cytotoxicity via cell surface Fc receptors. The dual functionality of NK cells may enable their participation in chronic active antibody-mediated rejection (CA-ABMR), wherein evidence for complement activation is inconsistent. We RNA sequenced 57 human kidney allograft biopsies to determine whether NK cell cytotoxicity pathway gene set enrichment is an attribute of CA-ABMR. Among the 15,910 intragraft-expressed genes, 60 were uniquely overexpressed in CA-ABMR compared to active antibody-mediated rejection (active-ABMR) or acute T cell-mediated rejection (TCMR), versus no rejection (NR) biopsies. Cell type annotation showed enrichment for T cells and NK cells, and molecular pathways related to T cells and NK cells in CA-ABMR versus active-ABMR biopsies. NK cell cytotoxicity gene set enrichment in CA-ABMR than in ABMR biopsies, but not in TCMR, was confirmed by gene set variation analysis. Cellular deconvolution analysis divulged a higher proportion of NK cells in CA-ABMR compared to active-ABMR, but not in TCMR; immunohistochemistry of 138 consecutive clinically indicated allograft biopsies validated a higher proportion of CD56+ NK cells in CA-ABMR. Principal component analysis of deconvolved immune cell transcriptomes separated CA-ABMR and TCMR from active-ABMR and NR biopsies. NK cell cytotoxicity pathway gene set was found to be enriched in rejection compared to no rejection biopsies in two publicly available kidney allograft microarray datasets. Altogether, CA-ABMR is exemplified by the overexpression of the NK cell pathway, and, surprisingly, compared to active-ABMR, is exemplified by key gene sets that are similar to TCMR. Overall design: To study the NK cell cytotoxicity pathway transcriptome in caABMR biopsies (N=15), we selected three comparator groups: active antibody-mediated rejection (aABMR, N=7), acute T cell-mediated rejection (TCMR, N=17), and normal/no rejection (N=18). We excluded mixed rejection, BK virus nephropathy, and other diagnoses. For the purpose of this study, prior to the processing of the samples for RNA isolation, each biopsy was read again by both the transplant pathologists independent of each other and reported based on the Banff 2017 update of the Banff '97 classification (2, 3). Subsequently, prior to data analysis, we reviewed the biopsies again; all 57 biopsies still met the criteria for their respective diagnostic categories when reclassified using the Banff 2019 update (4). The biopsies we selected were as pristine as possible in terms of their diagnostic phenotype.