Dr. Muhammad J. Azam is a Transplant Nephrology fellow at Vanderbilt University Medical Center. He completed his Internal Medicine and Nephrology training from Health Education England in the United Kingdom. His areas of interest include chronic allograft nephropathy, BK nephropathy, and post-transplant GN recurrence.

---

Dr. Saed Shawar @kidneyshawar is a transplant nephrologist and an Assistant Professor of Medicine at Vanderbilt University Medical Center. He also serves as the Program Director of the kidney and pancreas transplant program. Actively involved and passionate about the education of medical students and fellows. His area of interest includes health-related quality of life in kidney transplant recipients, complement mediated kidney disease, and improving options for highly sensitized patients in kidney transplantation.

Kidney transplantation remains the gold standard treatment for end-stage kidney disease, significantly improving recipients’ quality of life and longevity. Despite advancements, kidney allograft rejection remains a significant challenge, threatening the long-term function and lifespan of the transplanted kidney. Prompt identification and management of rejection episodes are crucial for graft survival.

Traditional monitoring of kidney allograft function relies on following serum creatinine and proteinuria. While abnormalities can suggest potential rejection, a kidney biopsy,, the gold standard for diagnosis, is still necessary. However, biopsies carry inherent risks, such as bleeding, and suffer from inter-observer variability in readings. Additionally, limited sampling in biopsies can miss patchy rejection.  Also, creatinine and proteinuria are lagging indicators that are neither specific to the causes of kidney injury nor effective in predicting subclinical acute rejection (SubAR). Therefore, there is a pressing need for alternative non-invasive monitoring biomarkers.

Mrs. Silverstein, a two-time heart transplant recipient, wrote in a New York Times article before she passed away, “Gratitude is not permission for the status quo”. This highlights the necessity for advancements in managing long-term health issues post-transplantation. We need to move from protocol-driven to personalized precision medicine to tailor immunosuppressive regimens to individual needs, improving outcomes.

Detecting or predicting the “smoke” of SubAR before the “fire” of full-blown rejection is crucial to prevent further complications. SubAR is linked to worse outcomes, including higher risks of subsequent clinical acute rejection, donor-specific antibody (DSA) formation, and graft fibrosis.

Here is where non-invasive biomarkers come into the picture as molecular events precede clinical and histological changes. Biomarkers in blood, plasma, or urine could provide a more accurate, reliable, and patient-friendly approach to monitoring graft health. This allows for earlier interventions, potentially improving transplant outcomes. Biomarkers can also play a role in identifying medication nonadherence, particularly in high-risk young adult patients.

Biomarkers can be predictive, diagnostic, prognostic, or used to monitor treatment efficacy. Regardless of their type, they need to be non-invasive, reproducible, reliable, and cost-effective. It is important also to understand their context of use as this will help how they can be used and benefited from.

This review will discuss the most important non-invasive biomarkers and their clinical utility in the evolving field of transplantation, aiming to protect the invaluable gift of life.

“Invasive vs Non-Invasive Biomarkers in Kidney Transplant.” Image generated by Saed Shawar using DALL-E 3 program. Accessed July 1, 2024.

Blood biomarkers:

1. Donor-derived cell-free DNA:

Donor-derived cell-free DNA (dd-cfDNA) quantifies the amount of intracellular DNA released from an injured allograft into the recipient’s circulation, and its levels typically increase during rejection or other forms of allograft injury. It has emerged as a promising diagnostic and prognostic biomarker for alloimmune events. It can be measured and reported as a fraction of the total cell-free DNA in the blood or as an absolute value. Currently, three commercially available assays in the United States can analyze dd-cfDNA: AlloSure (CareDx), Prospera (Natera), and TRAC (Viracore Eurofins).

Existing research highlights the value of dd-cfDNA in transplant rejection:

• A pioneering study, The Donor-Derived Cell-Free DNA in Blood for Diagnosing Acute Rejection in Kidney Transplant Recipients (the DART trial), demonstrated that dd-cfDNA is a superior biomarker to serum creatinine for diagnosing acute rejection (AR). A cut-off value of 1% effectively distinguished between any rejection, T-cell Mediated Rejection (TCMR) or Antibody Mediated Rejection (ABMR), and no rejection, with a high negative predictive value (NPV 84%); this test was particularly useful for identifying ABMR (PPV 44%; NPV 96%). However, a limitation of this study is its moderate sensitivity (59%) for detecting rejection and inability to differentiate between a specific type of TCMR like 1A and no rejection.
• Encouragingly, the Assessing Donor-derived cell-free DNA Monitoring Insights of kidney Allografts with Longitudinal surveillance (ADMIRAL), which monitored over 1,000 adult kidney transplant recipients for three years, established that a lower dd-cfDNA cut-off (>0.5%) could predict the development of de novo donor-specific antibodies. These antibodies can significantly increase the risk of future graft dysfunction. Furthermore, persistently elevated dd-cfDNA levels predicted over a 25% decline in the estimated glomerular filtration rate over three years. However, like the DART trial, dd-cfDNA could not differentiate between borderline TCMR and no rejection.
• The Trifecta trial compared gene expression in kidney biopsies with dd-cfDNA levels. This study found the strongest correlation between dd-cfDNA and active ABMR, followed by mixed rejection and active TCMR. It also found that dd-cfDNA was increased in both donor-specific antibodies positive and donor-specific antibodies negative molecular ABMR/mixed biopsies and was a better predictor of ABMR/Mixed than DSA, though the best predictions use both.
• A recent encouraging multinational study comprising 2882 kidney transplant recipients found that dd-cfDNA was strongly correlated with AMBR, TCMR and mixed rejection (p-value < 0.0001 for all). In fact, dd-cfDNA was highly associated with the presence, activity, and severity of all types of kidney allograft rejection. This study also showed that dd-cfDNA enhanced rejection detection independent of standard monitoring, showing promise in predicting subAR and improving diagnostic accuracy when integrated into care models.

As a result of these promising findings many US transplant centers have adopted dd-cfDNA assays to monitor the occurrence of graft rejection and injury in kidney transplant recipients.  Use of this method continued to grow, with Medicare reimbursement coming in 2017 and several commercial payers agreeing to coverage.

In summary, dd-cfDNA offers several advantages over traditional methods for monitoring kidney transplants. It surpasses serum creatinine in detecting early signs of kidney injury and declining kidney function. It has been associated with transplant rejection, particularly ABMR. It might also predict long-term graft survival by identifying patients at risk of developing new DSA.

It is important to note that factors other than rejection can influence dd-cfDNA levels. For example, infections like the BK polyomavirus (BKPyV) virus can cause false-positive results. This supports the idea that dd-cfDNA correlates with unspecific parenchymal injury and not primarily with alloimmune mediated injury. Therefore, we need to be aware of other factors that interfere with its elevation and the best context to use it.

Finally, research into urine dd-cfDNA is ongoing. Studies have shown a correlation between dd-cfDNA levels in urine and protein levels, suggesting its potential as a marker for allograft injury. Also, Elevated urine dd-cfDNA could be used in discriminating BKPyV in recipients infected with BK virus.

2. Gene Expression profile (GEP):

GEP analyzes the messenger RNA (mRNA) patterns and expression in circulating leukocytes, reflecting the recipient’s immune response to the allograft. Specific gene panels have been identified that can act as a “signature” for graft health, subclinical acute rejection (subAR), or acute rejection (AR). Various assays have been developed and tested for these purposes:

• TruGraf (Eurofins Transplant Genomics): This microarray-based GEP included 57 classifier genes and was evaluated in the CTOT-08 study. TruGraf associates with a normal protocol kidney biopsy (Transplant eXcellent [TX]) or the absence of a standard biopsy (not TX) in stable allograft patients. This cohort (n = 382) showed a specificity of 87%, NPV of 88%, and AUC of 0.84 for subAR. The study aimed to determine if the kidney allograft is “immune quiescent,” reducing the need for unnecessary surveillance biopsies. This level of performance held up in two external validation cohorts, and the GEP independently correlated with the development of de novo DSA and clinical allograft outcomes.
• AlloMap Kidney (CareDx): This next-generation sequencing (NGS) assay profiles gene expression and assesses immune quiescence using a five-gene classifier. It is a quantitative score, and validation demonstrated an excellent ability to rule out rejection (NPV 95.3%). However, its PPV was lower (23.3%), suggesting its utility in confirming immune quiescence rather than definitively diagnosing rejection.
• Tutivia (Veloxis): This is a blood test that analyzes 17 specific genes using NGS. It combines this genetic data with a proprietary artificial intelligence (AI) algorithm to classify kidney transplant patients into low—or high-risk AR. Using a preset threshold, the gene signature yielded a PPV of 60% and an NPV of 79%, with higher accuracy in for-cause biopsies than in surveillance biopsies.
• Kidney Solid Organ Response Test (kSORT) (Immucor Inc): This real-time PCR method analyzes a 17-gene set associated with rejection. While initial studies reported promising results, including high PPV and NPV for AR, follow-up studies yielded conflicting data. A recent study by Lee and colleagues showed some promise, with PPV ranging from 54% to 58%, depending on the sort of score threshold used. These conflicting results highlight the importance of validating GEP assays.

Urine Biomarkers

Urine is the net result of all functioning nephrons. Golden fluid may be the ideal biomarker and a reliable proxy for allograft health. It offers a convenient and noninvasive means of collection, making it an ideal source for biomarkers.

1. Chemokines CXCL9 and CXCL10:

Both are members of the CXC chemokine family, induced by interferon-gamma (IFN-γ) and bind to the CXCR3 receptor on activated T-cells and natural killers. These chemokines are promising biomarkers for monitoring immune responses and predicting rejection in kidney transplantation.

Elevated levels of urinary CXCL9 and CXCL10 have been associated with graft inflammation and acute rejection, as demonstrated in the CTOT-01 study. The study showed that urine CXCL9 and CXCL10 protein levels had a robust diagnostic performance for Banff >= IA TCMR, with a sensitivity of 74%–85% and PPV of 68%–71%. The study also showed that CXCL9 levels at 6 months identified a low-risk subgroup for subsequent rejection and decline in allograft function.

The study by Van Loon et al. showed that, when integrated into a clinical model with eGFR, donor-specific antibodies, and polyoma viremia, urinary CXCL9 and CXCL10 effectively predict graft rejection. This integration led to a reduction of 59 protocol biopsies per 100 patients when the risk for rejection was predicted to be below 10%, with AUC of 81.3%, PPV of 22.2%, and NPV of 96.3%.

Several studies showed that persisting low urine CXCL10 levels reflected a low risk for SubAR.   The CTOT-09 study showed that urine CXCL9 protein levels began to increase weeks before the development of acute rejection in subjects undergoing the Tacrolimus withdrawal group. On the other hand, other factors like UTI or BKPyV infection can also trigger chemokine elevation. Therefore, interpreting these results requires careful consideration of a patient’s overall clinical picture and ruling out infection.

2. Urine gene expression profiles

Urinary cell mRNA profiling represents a groundbreaking advancement in the noninvasive monitoring of kidney allograft status.

CTOT-04 study was a landmark multicenter trial that found and validated the diagnostic and prognostic utility of a three-gene urinary cell mRNA signature (CD3ε mRNA, IP-10 mRNA, and 18S rRNA). This signature distinguished acute rejection due to TCMR vs non-rejection (NR) states, achieving an AUC of 0.74. and ABMR vs NR with AUC 0.78.  It even demonstrated that molecular monitoring could detect rejection before clinical symptoms and histological changes, allowing for earlier and more targeted interventions.

The mRNA levels of regulatory T-cell markers, such as Forkhead box P3 (FOXP3), and T-cell costimulatory proteins, like OX40, have been studied for their predictive value in reversing acute rejection. Studies have shown a correlation between increased urinary FOXP3 mRNA, OX40 levels, and successful reversal of acute rejection.

Final word and how we should look at these biomarkers:

Biomarkers are revolutionizing rejection diagnosis, prognosis, and immunosuppression management. As we advance into personalized medicine, the goal is to move from a one-size-fits-all general protocol-driven approach to individualized decisions based on a patient’s unique biological fingerprints. Ideally, we would love a single biomarker that is perfectly accurate to guide our decision. However, we must understand that each biomarker has its strengths and limitations. The beauty lies in combining them strategically to create a comprehensive approach to allograft health and enhance the accuracy of rejection diagnosis and prognosis.

Based on this strategy, a combination of biomarkers has shown better diagnostic performance rather than one alone:

1. The Q-score, which is consistent with six urinary biomarkers (Cell-Free DNA (cfDNA), Methylated cfDNA (m-cfDNA), Clusterin, CXCL10, Creatinine, and Total Protein), provides a quantitative assessment and demonstrates good NPV. A score >32 accurately discriminated both ABMR and TCMR from non-rejection. Performance included 91–100% sensitivity, 92–96% specificity, and AUC 0.96–1.0.
2. GEP paired with dd-cfDNA test (OmniGraf). individually had NPV of 82%–84%, PPV of 47%–56%, AUC of 0.72–0.75. When GEP and dd-cfDNA were negative, NPV increased to 88%; more impressively, when GEP and dd-cfDNA were both positive, PPV increased to 81%. The two tests agreed (positively or negatively) in 70% of patients. This synergy highlights the power of combining different biomarker platforms.
3. Multimodal approach: Friedewald et al. presented a research abstract in ATC 2024, showing that combining urine CXCL9 and CXCL10 with blood GEP and dd-cfDNA would further improve diagnostic performance for rejection, with NPV 87% and PPV 63%, AUC 0.853.

Biomarker research, a rapidly evolving field, holds the key to the future of personalized medicine in kidney transplantation. The goal is to create a comprehensive biomarker panel that offers a complete picture of the golden gift of health. This will enable clinicians to tailor immunosuppression with precision, leading to improved graft survival and a better quality of life for patients.

– Post prepared by Muhammad Azam and Saed Shawar @kidneyshawar 

To view Van Loon et al [OPEN ACCESS], please visit AJKD.org.

Title: Automated Urinary Chemokine Assays for Noninvasive Detection of Kidney Transplant Rejection: A Prospective Cohort Study
Authors: Elisabet Van Loon, Claire Tinel, Henriette de Loor, Xavier Bossuyt, Jasper Callemeyn, Maarten Coemans, Katrien De Vusser, Virginia Sauvaget, Juliette Olivre, Priyanka Koshy, Dirk Kuypers, Ben Sprangers, Amaryllis H. Van Craenenbroeck, Thibaut Vaulet, Dany Anglicheau, and Maarten Naesens
DOI: 10.1053/j.ajkd.2023.07.022