A new molecular approach to the diagnosis of acute rejection

Abstract

Renal biopsy is the gold standard for detection of rejection in kidney transplant recipients but is not considered until evidence of renal dysfunction is apparent. Now, Suthanthiran and colleagues suggest that mRNA levels in urinary cells from these patients might be diagnostic and prognostic of acute cellular rejection.

Accurate diagnosis of acute or impending allograft rejection is an obstacle to successful long-term kidney transplantation. The time-honoured gold standard for detection of rejection is the renal transplant biopsy. In current practice, however, biopsy is not considered until clear evidence of substantial renal transplant dysfunction is apparent. Hence, the diagnosis of rejection lags behind the advent of obvious renal damage and the inevitable accumulation of donor-specific memory cells that are refractory to immunosuppressive drugs. Transplant biopsies are associated with inconveniences, costs, time delays, and inherent risks. Moreover, only a very small part of the transplanted kidney is sampled. We have seen instances where a first biopsy core shows a normal kidney and a second core shows vigorous rejection. A biopsy sample also provides a largely descriptive indication rather than a functional or molecular interpretation of the recipient antiallograft response. For example, although T cell interstitial infiltrate within the transplanted kidney may indicate cellular rejection, in certain experimental situations these cells may have a beneficial regulatory phenotype. In analyses of 0 h transplant biopsy samples, a molecular approach can often identify processes of diagnostic and prognostic importance that are not evident using microscopy.¹ Now, new data from Suthanthiran and colleagues suggest that mRNA levels in urinary cells from renal transplant recipients could provide a diagnostic signature of acute cellular rejection.

Development of a noninvasive, easy-to-perform biomarker-based assay that gives a more global account of processes within the allograft than does a biopsy sample is a reasonable goal. This approach could enable detection of the rejection process before the development of full-blown, clinically evident rejection. With the advent of the sensitive and remarkably reproducible reverse transcriptase PCR, a molecular signature of acute allograft rejection can be obtained from transplant biopsy samples in experimental models and in renal transplant recipients. As donor-specific cytotoxic T cells infiltrate rejecting transplants,² amplified gene expression of cytotoxic T cell effector molecules expressed by these cells, including granzyme B, perforin and Fas ligand, is evident in biopsy samples from human kidney transplants undergoing rejection.³ In addition, transplants undergoing rejection manifest amplified expression of an invariant portion of the T cell receptor gene as marker of T cell infiltration.⁴ At the time of rejection, amplified expression of cytotoxic T cell effector genes were detected in the circulating blood of renal transplant recipients who had not received inductive lympholytic agents, although false positives were evident in those with active cytomegalovirus infection.⁵

Suthanthiran and colleagues cleverly reasoned that mononuclear leukocytes in the urinary sediment of renal transplant patients—especially those under attack by the host immune system—are rich in cells that have trafficked through the transplanted kidney. Hence analysis of gene expression in serial samples of urinary sediment cells may yield valuable clues as to the nature of the allograft response, even in patients receiving induction therapy with agents that produce lymphopenia. In a series of well-designed and executed studies, Suthanthiran and colleagues previously showed that in urinary sediment cells, the expression of cytotoxic T cell genes (including genes that encode granzyme B, perforin, granulysin, the universal T cell marker CD3ε, IP-10 [a gene induced in monocytes stimulated with interferon-γ] and the regulatory T cell transcription factor FOXP3) correlate with rejection.^6–8

In a new prospective study of 4,300 urine samples (of which 3,559 passed quality standards) from 485 kidney transplant recipients, Suthanthiran and collaborators demonstrate that a ‘parsimonious’ panel of 18S ribosomal mRNA levels plus CD3ε mRNA and IP-10 mRNA quantitation in urinary sediment cells is a sensitive and accurate marker for transplant rejection (P <0.001) and can even predict rejection.⁹ Gene expression of granzyme B and perforin were also excellent correlates of rejection. The study demonstrates the considerable promise of this molecular approach to the diagnosis of acute rejection as well as the challenges faced when using this approach as a solitary alternative to renal biopsy for the management of renal transplant recipients. The researchers found that their panel discriminates acute cellular rejection from acute antibody-mediated rejection and urinary tract infection. However, the test was more sensitive for the diagnosis of rejection in patients receiving anti-IL-2 receptor antibodies than in those receiving T cell depleting antibodies. This finding indicates that the sensitivity of molecular signatures for the diagnosis of rejection varies with treatment protocols. Suthanthiran et al. analysed receiver-operating-characteristic (ROC) curves and presented a set point for illustrative purposes that showed high specificity and sensitivity for the diagnosis of acute cellular rejection. Using arbitrary parameters derived from the ROC curves, the probability analysis may suggest that the test is not precise enough to forego renal biopsies in many patients. However, if higher specificity, albeit at the inherent loss of sensitivity, is sought, the ROC curves can be used at set points that identify acute cellular rejection with specificity sufficient for physicians to consider foregoing biopsy in some patients.

Unsurprisingly, the study also indicates that processes such as BK polyoma virus infection can to some extent mimic the rejection signature. In other words, the molecular signature describes a molecular pattern of immunity but not the provocative antigen. Hence, increased specificity can almost certainly be obtained without loss of sensitivity if analysis of BK polyoma viral load is added to the transcriptional profile. Little doubt exists that expression of other genes (some not as yet identified) in these urine sediment samples can add to the sensitivity and perhaps specificity of the panel. Following resolution of the lymphopenia that is caused by inductive biologic therapies, monitoring of the molecular signature in blood samples, which can be more conveniently obtained and handled than urine samples, may also prove useful.

Molecular approaches that quantitate protein, rather than gene expression, are attractive in concept and may add to or even replace analysis of gene expression as a long-sought alternative to biopsy as the gold standard for the diagnosis of rejection.¹⁰ The validity of these molecular approaches for the prediction or diagnosis of rejection await refinement and validation in clinical trials.