Kidney transplantation is the standard renal replacement therapy. Despite the requirement of immunosuppressants to suppress the immunological reaction against alloimmunity, kidney transplantation improves the quality of life and life expectancy of patients with end-stage kidney disease, compared to dialysis therapies. Although the outcome of kidney transplantation has been improving following the development of immunosuppressants and increased understanding of proper management for graft rejection, kidney grafts tend to lose their function due to allograft rejections and injuries such as recurrence of original disease, toxicity of immunosuppressants, development of metabolic disorders, and glomerular overload [2]. Improving survival of kidney grafts is thus a challenge and an unsolved issue.
The diagnosis of graft injury (including graft rejection) relies on clinical manifestations such as decrease in urine volume and fever, urine and blood analyses, blood chemical analysis, and radiographic evaluation such as ultrasonography, computed tomography or radioisotope imaging. However, the definitive diagnosis is still made by graft pathology [3, 4]. A graft biopsy can generally be performed safely; however, there are risks for patients in the early period after kidney transplantation and for those taking antithrombotic agents because of the slight invasiveness. Moreover, the final diagnosis takes several days. Thus, there is need for a non-invasive biomarker assay that can yield a correct diagnosis of graft injuries with comparable performance to histology. The assay of chemical substances such as neutrophil gelatinase-associated lipocalin [6, 18] and liver-type fatty acid-binding protein [19], which are enhanced in tubular injury has been proposed; however, a diagnostic modality that can detect several types of graft injury is ideal. We focused on exosomes in patient’s urine in this study. Exosomes are microvesicles discharged from cells; they include cell membrane components, protein, DNA, mRNA, and miRNA. Exosomes have been focused on as an information source since the late 1990’s [20]. The intercellular signal from renal injury and lymphocytes is included. Strictly, the sizes of exosomes and microvesicles are 50–100 nm and 100–1000 nm, respectively, but they are often collectively referred to as EVs [20]. EVs are located in blood or fluid such as bile and ascites [13] and can be recovered from any part of body. They are an ideal biomarker source for investigating kidney or urinary tract disease [5, 22, 23] because EVs from urine can be recovered non-invasively. The efficacy of EV evaluation by RT-PCR in nephritis [11] diabetic kidney disease [12], and bladder cancer [10] has been proven. Furthermore, EVs are generally retrieved by an ultracentrifugation method; however, this procedure is complicated and yields limited measurable samples [23]. As an alternative, we explored the seamless assay system for recovery of EV, extraction of mRNA, and generation of cDNA, establishing a protocol for rapid management of multiple samples [9]. A critical step during mRNA assay is preventing damage by RNase among urine sample contaminated in recovery or storage. Thus far, we recovered urine sample by way of ordinal sample handling for urinalysis and consecutive freezing preservation in a few hours, yielding RNA that was successfully measured. This may be because EVs are covered with cellular lipid membrane; RNA is thus resistant to temperature changes and RNase, preventing its degradation. Consequently, EVs are an ideal source of information [20].
We evaluated KGI using measurement of mRNA obtained from urine EVs. We previously introduced the usefulness of a single gene, ANXA1, in the detection of graft injury in a single center analysis of kidney injury model [14]. Subsequently, a nation-wide survey including the search of candidate gene by next generation sequencer (NGS) was developed to verify this result.
Here, 39 candidate genes, which were selected based on our preparation study, were analyzed by qPCR among 127 cases. CXCL9, CXCL10, SPDEF, SPNS2, and UMOD showed statistical differences between some graft rejection types. Among these, CXCL9/CXCL10 and UMOD were shown to be significant biomarkers in TCMR, as their expression showed robust enhancements in samples from TCMR; in contrast, there was no increase in the expression of these genes in samples from patients with antibody-mediated rejection. Previous literature stated that the chemokines—CXCL9 and CXCL10—were significant biomarkers for detecting allograft rejection in animal models and a clinical multiple-institute study. Our present study clearly supports these results [7, 8, 15, 16]. In this study, the detection of TCMR by single genes other than CXCL9 or CXCL10 was difficult; however, a combination of multiple candidate mRNA generated reliable diagnostic formula and became the promising biomarker instead of graft biopsy and pathology in the diagnosis of KGI. For example, we also determined that UMOD can be an alternative biomarker for TCMR detection. UMOD is a gene encoding uromodulin, also called Tamm-Horsfall protein. Uromodulin, a kidney specific protein located in the medullary thick ascending limb of loop, is reportedly a predictor of tissue injury in patients with anti-neutrophil cytoplasmic antibody-related nephritis24). Moreover, UMOD expression in urine EVs is a predictive biomarker of the development of diabetic kidney disease in patients with type 2 diabetes [12].
B4GALT1 expression was increased in cABMR but decreased in cCNIT. Both KGIs induce gradual arteriole stenosis and consecutive tissue injury as a result of chronic ischemic changes. B4GALT4 is a promising gene biomarker for distinguishing those two events and has critical significance given the contrary management of immunosuppressants dosing for these conditions. B4GALT1 is a gene encoding glycosyltransferase and influences B cell activation [25], and has been used as a predictive marker for disease progression and prognosis in malignancy26). The relationship between B4GALT4 and kidney injuries has not been well studied. In the present study, SPNS2 was also nominated as a biomarker gene and has similar expression patterns to B4GALT1. SPNS2 plays a role in anti-fibrotic and anti-inflammatory processes in human kidney gene tissue [27].
SLC12A1 was identified as a candidate marker for reflecting the severity of IFTA by qPCR analysis. NKCC exists on the cell surface and has two variants, NKCC1 and NKCC2; NKCC2 is expressed only in kidney tissue and encoded by SLC12A1 [28]. However, the role of NKCC in graft fibrosis and tubular atrophy is not understood.
Finally, POTEM and HAVCR1 were nominated as the candidate biomarkers for the detection of BChS, which is supposed to correlate graft loss. HAVCR1, also called TIM-1, is a known biomarker of kidney injury and has been proven to be a candidate marker for chronic KGI. While POTEM was also identified as a candidate biomarker, further study is needed regarding its mechanism of involvement in the progression of chronic graft damage.