Non-invasive screening for sub-clinical antibody-mediated rejection as a new tool for indication of kidney allograft biopsy

Abstract

Antibody-mediated rejection (ABMR) is nowadays the major probdiagnosis) and inspected arbitrarily chosen thresholds, which could lem for long-term kidney allograft survival. ABMR frequently takes a continuous and fluctuating course. It often occurs sub-clinically, and abnormal serum creatinine levels or the detection of proteinuria is only detectable at stages with already advanced graft injury. Frequent follow-up controls to uncover the rise of known pre-transplant donorspecific antibodies (DSA) or their de novo formation, that might trigger diagnostic biopsies for detection of earlier stages of graft injury, or the implementation of protocol biopsies are not always possible in transplant centres. However, a timely treatment of ABMRwould enhance significantly graft outcome [1–4]. Efforts to develop better measures that allow early detection of rejection before any clinically obvious, irreversible damage has accumulated and range from protocol biopsies to less invasive approaches such asmetabolomics, proteomics or mRNA profiling of single or multiple markers in urine or serum. These latter techniques aim at detecting active injury, due to acute rejection, T-cellmediated rejection (TCMR) in particular, and, to a lesser extent, ABMR [5–8]. Although innovative, none of these approaches has been widely adopted for routine clinical procedures, mainly due to the lack of confirmation of these often small to medium-sized studies. Owing to these nonetheless encouraging data, Van Loon et al. sought to assess the transcriptome from peripheral blood prospectively and, if possible, renal allograft tissue, in 630 renal allograft recipients from multiple European centres [9]. They split the study into three distinct parts: into a discovery, a derivation and a validation phase, and correlated blood mRNA levels to clinical, serological, and histological data. The authors first tested blood and tissue samples in a genome-wide expression assay on anRNAmicroarray. They computedwithfive different statistical approaches a multivariate score for each transcript, which could separate four pre-defined and centrally confirmed diagnostic situations (no rejection, pure ABMR, pure TCMR and mixed rejection but excluding glomerulonephritis, BK-virus nephropathy and unclear