Histologic and Molecular Patterns in Responders and Non-responders With Chronic-Active Antibody-Mediated Rejection in Kidney Transplants

Abstract

Introduction: There is no proven therapy for chronic-active antibody-mediated rejection (caABMR), the major cause of late kidney allograft failure. Histological and molecular patterns associated with possible therapy responsiveness are not known.

Methods: Based on rigorous selection criteria this single center, retrospective study identified 16 out of 1027 consecutive kidney transplant biopsies taken between 2008 and 2016 with pure, unquestionable caABMR, without other pathologic features. The change in estimated GFR pre- and post-biopsy/treatment were utilized to differentiate subjects into responders and non-responders. Gene sets reflecting active immune processes of caABMR were defined a priori, including endothelial, inflammatory, cellular, interferon gamma (IFNg) and calcineurin inhibitor (CNI) related-genes based on the literature. Transcript measurements were performed in RNA extracted from stored, formalin-fixed, paraffin-embedded (FFPE) samples using NanoString™ technology. Histology and gene expression patterns of responders and non-responders were compared.

Results: A reductionist approach applying very tight criteria to identify caABMR and treatment response excluded the vast majority of clinical ABMR cases. Only 16 out of 139 cases with a written diagnosis of chronic rejection fulfilled the caABMR criteria. Histological associations with therapy response included a lower peritubular capillaritis score (p = 0.028) along with less glomerulitis. In contrast, no single gene discriminated responders from non-responders. Activated genes associated with NK cells and endothelial cells suggested lack of treatment response.

Conclusion: In caABMR active microvascular injury, in particular peritubular capillaritis, differentiates treatment responders from non-responders. Transcriptome changes in NK cell and endothelial cell associated genes may further help to identify treatment response. Future prospective studies will be needed which include more subjects, who receive standardized treatment protocols to identify biomarkers for treatment response.

Clinical trial registration: [ClinicalTrials.gov], identifier [NCT03430414].

Keywords: Banff classification; chronic-active ABMR; eGFR slope; kidney transplantation; therapy response; transcriptome.

Conflict of interest statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2022 Sazpinar, Gaspert, Sidler, Rechsteiner and Mueller.

Figures

FIGURE 1

Algorithm of the biopsy selection process. cABMR, chronic antibody mediated rejection; aABMR, active antibody mediated rejection; IF/TA, interstitial fibrosis and tubular atrophy; ICT, information and communication technology; caABMR, chronic-active antibody mediated rejection fulfilling Banff 2017 criteria; ptc, peritubular capillaritis; g, glomerulitis; cg, transplant glomerulopathy; C4d, complement split product.

FIGURE 2

Kidney function before and after biopsy in treatment responders versus non-responders. (A) eGFR slopes in 10 therapy responders, (B) in 6 non-responders. Every line represents one patient. The linearized slopes were calculated by linear regression of raw data before and after biopsy and plotting of the two curves ypre–biopsy = mpre–biopsy *x + qpre–biopsy and y = mpost–biopsy*x + qpost–biopsy where qpre–biopsy = qpost–biopsy = measured eGFR at biopsy day.

FIGURE 3

Overview of Banff scores of responders versus non-responders. t, tubulitis; i, interstitial inflammation; ti, total cortical inflammation; ptc, peritubular capillaritis; v, vasculitis; cv, arterial fibrous intimal thickening; g, glomerulitis; cg, transplant glomerulopathy; mm, mesangial matrix expansion; ci, interstitial fibrosis; ct, tubular atrophy; ah, arteriolar hyalinosis; aah, hyaline arteriolar thickening; C4d, complement split product; IFTA, interstitial fibrosis–tubular atrophy.

FIGURE 4

Clustering of histopathology in the caABMR biopsies. (A) Principal component analysis indicating distribution and correlation between Banff scores. (B) Unsupervised hierarchical clustering of the Banff characteristics based on the samples and (C) of the biopsies based on 8 Banff features. Due to high Euclidian distances 3 samples did not include in the analysis.

FIGURE 5

Clustering of transcriptomes in the caABMR biopsies. (A) Unsupervised hierarchical clustering of the 44 gene transcripts. The dendrogram shows the clustering of the 44 genes according to similarities and dissimilarities. (B) Dendrogram of distribution of responders and non-responders based on gene expression.

FIGURE 6

Heatmap analysis of gene expression. Columns represent the 16 different samples, rows the 44 analyzed genes. High gene expression is shown in green, low gene expression in red. Similarity metric was calculated with uncentered Pearson correlation for patients/biopsies and genes; average linkage was the used linkage method.