Molecular Approaches to Transplant Monitoring; Is the Horizon Here? Free

Transplantation is a remarkable treatment that gives a “second chance” of life to patients with end-stage organ failure. Unfortunately, this exposes the transplanted organ or allograft to the recipient’s immunity, predisposing the allograft to acute rejection, a dreadful complication, and a major risk factor for allograft loss. To preserve the allograft’s health, providers carefully monitor their patients to maintain immunosuppression adequacy, toward balancing the risk of rejection and infection. Monitoring approaches vary substantially among providers and transplant programs. Generally, monitoring encompasses biopsy of the allograft for histopathology and light microscopy to assess rejection, measurement of the allograft function (creatinine for kidney, pulmonary function test for lung, liver function test for liver, echocardiography for heart), measurement of immune activity with donor-specific antibody (DSA) testing, and assessment of immunosuppression adequacy by measuring immunosuppression drug blood concentrations.

Assessment of allograft function or immunosuppression drug concentrations uses noninvasive testing approaches. Thus, these tests are generally implemented as surveillance, wherein patients undergo testing at predetermined posttransplant intervals independent of symptoms or clinical suspicion of allograft dysfunction. When patients present with signs for allograft dysfunction, additional clinical indication testing is performed.

Biopsy, on the other hand, is an invasive procedure associated with procedure-related complications. Further, biopsy samples are analyzed by routine histopathology and light microscopy, which have limited sensitivity and precision to adequately detect and phenotype rejection and nonrejection complications. As a result, practice varies; some centers perform surveillance plus clinically indicated biopsies. Other centers only perform clinically indicated biopsies. In lung transplantation, for example, 70% of centers in the USA perform surveillance plus clinically indicated biopsies while 30% of centers perform only clinically indicated biopsies. The true benefit of these different monitoring approaches remains undefined.

Fortunately, in the last few decades, novel molecular approaches have been introduced; many show promising results with benefits that address the limitations of biopsy plus conventional histopathology. In this Q&A, we focus on 2 prototype tools that recently received approval for patient use by the Centers for Medicare and Medicaid Services of the United States: plasma-based donor-derived cell-free DNA (dd-cfDNA) and “molecular biopsy,” the latter to indicate gene expression profiling of biopsy tissue samples using microarray and other platforms. Five experts discuss the potential use of these novel tools to monitor solid organ transplant patients.

From your perspective, should dd-cfDNA be utilized for surveillance or for the clinically indicated testing?

Michael Oellerich: dd-cfDNA is a minimally invasive quantitative biomarker for the detection of graft injury. As organ transplants are also genome transplants, dd-cfDNA opens up the possibility to monitor allograft health. In case of graft cell death, nucleosomes are released into the bloodstream as cfDNA. The clinical validity of dd-cfDNA to detect or exclude rejection and other graft injuries has been documented in more than 50 studies. Increases of dd-cfDNA levels were found in transplant recipients several days or even up to 3 months before clinical manifestation of acute rejection. There has been discussion that increased dd-cfDNA levels may be a trigger of inflammation. Early detection of in particular subclinical antibody-mediated rejection enables adapted therapeutic interventions and may improve outcomes. This needs to be further investigated in controlled studies with different dd-cfDNA tests. In kidney transplantation, dd-cfDNA may be useful to early detect antibody-mediated rejection in DSA positive patients. An adequate diagnostic performance of dd-cfDNA is suggested by currently published studies for detection of acute rejection [ROC area under curve (AUC) = 0.81]. Clinical sensitivity and specificity are around 80% and 76%, respectively. The high negative predictive value of 90% suggests that dd-cfDNA is useful to exclude rejection. Therefore, dd-cfDNA could be helpful to avoid unnecessary biopsies triggered by increased plasma creatinine. After successful rejection treatment, dd-cfDNA rapidly declined in liver, heart, or kidney transplant recipients. There are several limitations of dd-cfDNA. Increases of dd-cfDNA are not rejection specific. Other sources of graft injury associated with increase of dd-cfDNA include for example, pyelonephritis, acute tubular necrosis, and BK-virus nephropathy in kidney transplant recipients. In general, patients with advanced interstitial fibrosis tubular atrophy showed only a relatively small increase of dd-cfDNA and a discrimination of values from patients with normal histology was not possible. During stages of active disease progression, dd-cfDNA levels might be increased in interstitial fibrosis tubular atrophy. The detection of T cell-mediated rejection seems to be influenced by the employed test procedure. False-negative results may be due to the use of relatively long amplicons (100–130 bp) in the assay used. dd-cfDNA test results should be interpreted in context of all available relevant clinical data and diagnostic findings to achieve individualized transplant patient therapy with the potential to reduce graft loss and to save costs. Based on currently available evidence regarding clinical validity, dd-cfDNA seems to be useful for surveillance of transplant recipients and for clinically indicated testing. Further optimization of dd-cfDNA testing for effective clinical use is an ongoing challenge.

Philip Halloran: Surveillance: no. Clinically indicated: cautious yes, if a probabilistic algorithm to guide the use of the information to avoid biopsy can be defined based on continuous numbers, not cutoffs. The utility of commercially available dd-cfDNA tests in surveillance is limited by high cost. Repeated use of an expensive surveillance test in stable patients is probably not justified without indications. dd-cfDNA has not been shown to actually reduce biopsies and to create savings, and it may be producing unnecessary biopsies.

We need to know what to do with the many ambiguous values for dd-cfDNA in the commercial assays because they have the potential to trigger unnecessary biopsies and investigations. We need good algorithms that state the probability of rejection over the whole range of dd-cfDNA values.

Michael Keller: Plasma dd-cfDNA is a noninvasive quantitative molecular biomarker that increases in the setting of allograft injury. Recent evidence suggests that the rise in dd-cfDNA may precede clinical manifestations of allograft injury by several months. Several observational cohort studies performed in kidney, heart, and lung transplant recipients indicate that dd-cfDNA displays acceptable performance characteristics for the detection of acute rejection—both acute cellular rejection and antibody-mediated rejection. In particular, dd-cfDNA demonstrates a high negative predictive value for acute rejection, providing the potential to effectively “rule out” the majority of acute rejection events. These characteristics make plasma dd-cfDNA appealing as a method of surveillance monitoring for underlying allograft injury. Despite a paucity of high-quality evidence supporting its use, most heart and lung transplant recipients undergo routine surveillance biopsy as standard of care to screen for acute rejection. As a screening tool, these biopsies provide procedural risks, are costly, and are often inconvenient. In addition, analysis of histopathology has high interobserver variability in the pathologic grading of biopsy samples. Dd-cfDNA provides a safe, accurate, and minimally invasive method of screening patients who may benefit most from proceeding to biopsy. A high-quality, randomized control trial comparing a dd-cfDNA method of surveillance monitoring in comparison to traditional surveillance biopsy in heart and lung transplant recipients would be especially helpful to further validate its use. In addition, further validation of specific threshold values to indicate the presence of acute rejection would be beneficial. Based on the available evidence, it is reasonable to utilize dd-cfDNA for surveillance monitoring in select patient populations such as heart and lung transplant recipients at high risk for procedural complications from biopsy or during periods of time when the performance of surveillance biopsy is limited, such as during periods of the COVID-19 pandemic.

Iwijn De Vlaminck: One promising feature of dd-cfDNA is that it is an early marker of graft injury. It’s therefore reasonable to integrate dd-cfDNA in surveillance testing and not just clinically indicated testing. There are practical limitations that need be overcome, however, including the current high cost of the assay.

Should dd-cfDNA analytic approach and reporting (absolute copies/mL vs %dd-cfDNA) be considered to improve its clinical validity?

Michael Oellerich: Both approaches, fractional and absolute determination of dd-cfDNA in kidney transplant recipients, have strengths and weaknesses. Fractional determination has the advantage that it is less sensitive to preanalytical variables (e.g., DNA extraction efficiency). Measurements can be easily compared between different studies. Fractional determination is insensitive to changes in the rate of degradation of cfDNA in blood circulation. It has, however, the disadvantage that it is affected by changes in recipient cfDNA (e.g., by infection or exercise). Leukocytosis or leukopenia can alter dd-cfDNA fraction as recipient cfDNA accounts for the major part of the denominator in fractional quantification. This can result in false-negative or false-positive results. During long-term surveillance dd-cfDNA fraction increases due to a decrease of total cfDNA so that the threshold would have to be adapted. The decline of total cfDNA with time after kidney transplantation is presumably due to a decrease in apoptosis rate for white blood cells as immunosuppressant drug doses are tapered off. Absolute quantification (cp/mL) has the advantage that it is not affected by changes in recipient cfDNA by other factors and that there is no influence on the threshold due to total cfDNA decline over time. In clinical practice, the combination of fractional and absolute determination seems to provide the most comprehensive diagnostic information in kidney transplant recipients.

Philip Halloran: Absolutely. The use of ratios is never desirable when both the numerator and the denominator can change.

Michael Keller: Plasma dd-cfDNA is traditionally reported as a percentage, representing the fraction of donor to donor + recipient cfDNA. While this method may help account for differences between the relative sizes of recipients and donor allografts, there are inherent limitations to this approach. Due to its presence in the denominator, levels of recipient cfDNA in the plasma will influence %dd-cfDNA. Recipient cfDNA may rise in a number of conditions irrespective of the status of the allograft including in the setting of sepsis, multiorgan failure, extreme exercise, and leukocytosis. This may make the interpretation of %dd-cfDNA difficult under conditions that alter the levels of recipient cfDNA. It also raises questions surrounding the validity of specific threshold values of %dd-cfDNA in the detection and diagnosis of various types of allograft injury. Established threshold values for %dd-cfDNA may carry more weight if validated in stable controls and performed on otherwise stable outpatient transplant recipients. However, it may be difficult to interpret %dd-cfDNA in the setting of systemic illness or nonallograft related organ failure, such as in hospitalized patients. Measurements of the absolute amount of plasma dd-cfDNA are not influenced by recipient cfDNA. However, differences in the sizes of allografts—particularly in lung transplant where total tissue mass may differ between individuals to a considerable degree—may impact the ability to interpret prespecified threshold values for the detection of allograft injury. Observational cohort studies in kidney transplant patients have demonstrated similar performance characteristics between absolute and %dd-cfDNA for the detection of acute rejection. There are strengths and weaknesses of both methods, however, developing a diagnostic algorithm using a combination of the 2 methods may be the optimal approach.

Iwijn De Vlaminck: It’s clear that a readout of the relative proportion of donor DNA can be confounded by changes in recipient DNA that are not related to the health of the transplant organ. The absolute burden of donor DNA in blood does not have that issue, and may therefore provide a more robust readout of transplant health, but this will need to be confirmed in clinical studies. Measurements of the absolute amount of donor DNA may prove more difficult to standardize and may be sensitive to changes in the rate of clearance of cfDNA from the blood circulation. Perhaps models that integrate both the absolute and relative burden of cfDNA will prove to be optimal.

Should the “molecular biopsy” approach be utilized for surveillance or for the clinically indicated testing?

Philip Halloran: Surveillance: no for kidney, surveillance biopsies are not justified for kidney; yes for heart because histology kappa values are poor. Clinically indicated: cautious yes. The clinician needs the quantitative assessments that only molecular assessment can provide, although we need a probabilistic algorithm to guide the use of the information to avoid biopsy based on continuous numbers not cutoffs.

Iwijn De Vlaminck: The molecular biopsy is a great technology that enables detailed profiling of different pathways of transplant rejection. It represents an important advance in transplant medicine, but because it relies on invasive tissue biopsies it has a limited role to play as a surveillance tool, and is more relevant for clinically indicated testing. An ideal test would achieve the resolution in cellular pathology of the molecular biopsy but from blood rather than from tissue.

To improve clinical validity and adoption, should dd-cfDNA or “molecular biopsy” be performed at a centralized location or at local transplant centers?

Angela Wu: Most of our work has been focused on monitoring the interactions between microbes and their hosts in a clinical setting, whether these microbes are pathogenic or not. We have found that in addition to monitoring immunosuppression, the metagenomic composition of “molecular biopsy” samples can also be highly informative as to the health status of the patient. When used in combination with dd-cfDNA, host-associated biomarkers, and other clinical measures, this metagenomic information can enhance the accuracy of diagnosis and outcome prediction.

Based on our investigations, “molecular biopsy” results can be strongly affected by different practices at different centers. In particular, if considering additionally the microbial component, the environmental microbes can be hugely different at different sites, affecting analysis outcomes and downstream decision-making. To ensure robustness of the sample handling and results, it is preferable to perform sample processing at a central location. Additional molecular based methods for preserving sample integrity and reducing loss/degradation of cfDNA are also being developed and streamlined to improve the logistical challenges associated with sending samples to central facilities.

Michael Oellerich: For routine implementation of dd-cfDNA testing there are a number of challenges. The laboratory requires specific instrumentation, such as a ddPCR or a next generation sequencing instrument as well as specialized laboratory professionals. Standard clinical chemistry assay validation is necessary for these tests and the development of external quality management programs is recommended. Furthermore, clinical practice recommendations have to be developed. Centralized reference laboratories may have the advantage of lower testing costs and high quality. Individual laboratories in transplant centers have the advantage of shorter turnaround times of dd-cfDNA test results for hospitalized patients if the test would have to be used for actual decision-making. For out-patients, services from centralized reference laboratories would allow for a broader adoption. A cost-effective monitoring frequency should be established in prospective clinical outcome trials. Using reasonable priced technologies, the cost for dd-cfDNA monitoring would be relatively low. By reducing graft damage, it is to be expected that costs for a return to dialysis or retransplantation could be substantially reduced using serial dd-cfDNA testing and other relevant approaches like the molecular biopsy.

Philip Halloran: Always central, if possible, at present. The performance of molecular assays as kits has been poorly reproducible in local laboratories, as shown by the measurement of DSA using Luminex. Machine variation, inter- and intralaboratory variation, and local decisions (e.g., shortcuts) introduce unacceptable variations. This is particularly true when the results use centrally derived machine learning algorithms for interpreting local results on a different machine. Molecular assays should generally be central until the reliability of local machines and practices are established and monitored for variation on an ongoing basis.

Michael Keller: At the moment, assays detecting and quantifying plasma dd-cfDNA should ideally be performed in centralized locations to maintain assay quality, reduce interlaboratory variations in results, and leverage expertise. The adoption and familiarization of the sophisticated technology required to run the assays may limit its implementation in many local transplant centers. Differences in analytical methods and assay reproducibility also become a concern when implemented at local centers. While performing the assay at a centralized location may increase turnaround time, this may not present an issue in several clinical circumstances such as surveillance monitoring or assessing response to treatment.

Iwijn De Vlaminck: The molecular biopsy is a great technology that enables detailed profiling of different pathways of transplant rejection. It represents an important advance in transplant medicine, but because it relies on invasive tissue biopsies it has a limited role to play as a surveillance tool and is more relevant for clinically indicated testing. An ideal test would achieve the resolution in cellular pathology of the molecular biopsy but from blood rather than from tissue.

In regard to monitoring immunosuppression adequacy, is there any role for the dd-cfDNA, “molecular biopsy,” or another novel molecular diagnostic?

Angela Wu: As mentioned earlier, our work has focused on the metagenomic aspect of cfDNA, which includes both donor and recipient-derived microbes (including bacterial, fungal, and viral species). Monitoring of microbial and viral abundance in cfDNA has been demonstrated as a potentially feasible way to evaluate immunosuppression adequacy and also the need for prophylactic antivirals. There are 2 major directions that can be further investigated in this context: (a) precise detection and monitoring of specific microbes of concern, and (b) broad surveillance of microbial diversity and richness as a proxy for the general immunosuppressive condition.

For the first, it is already known that certain viruses are particularly problematic, and could be transmitted from donor to recipient via the transplanted organ, e.g., CMV, HBV/HCV (particular problematic in some Asian regions). In this case, broad genome- or transcriptome-wide detection with sequencing-based methods may be unnecessary, since targeted detection of cfDNA or RNA originating from these viruses afford greater ease of adoption and lower cost. However, targeted PCR assays suffer from other limitations, such as difficulties in the case of low viral load, or viral mutations leading to false negatives. The feasibility of an intermediate approach between total cfDNA sequencing and targeted PCR can be further explored, e.g., targeted sequencing or multiplexed/panel-based PCR. Targeted sequencing that uses semispecific probes to capture potential viral molecules followed by NGS to determine specific species-level abundance could be a way to ensure both specificity and sensitivity. In addition, current commercial PCR assays target replicating viruses, which have relatively complete and long viral DNA; if cfDNA can be incorporated into the detection workflow by also isolating short DNA fragments, this could also improve detection sensitivity.

Michael Oellerich: dd-cfDNA may be helpful to assess minimal necessary exposure with immunosuppressive drugs and to provide actionable information for optimization of immunosuppressive therapy. It can guide tapering in patients to prevent immune activation. Reduction of immunosuppressant dosages is often necessary to treat viral infections and dd-cfDNA may be helpful to detect under-immunosuppression and improve immunosuppressant management. Under-immunosuppression favors DSA development that is a risk factor for antibody-mediated rejection and late graft loss in kidney transplantation. Low tacrolimus concentrations were associated in individual patients with increased dd-cfDNA values indicating immune activation. dd-cfDNA is superior to immunosuppressive drug monitoring that indicates toxicity but is a poor predictor of graft damage due to under-immunosuppression. Also, conventional biochemical markers have limitations. In kidney transplant recipients, for example, a significant degree of graft damage can already be present by the time a rise in plasma creatinine is evident. Increased plasma creatinine is not specific for allograft injury. Due to a lack of validation studies alternative strategies for monitoring immune system activity or expression of rejection-related genes are not widely used. Serial biopsies are clinically unpractical and associated with a 1% major complication rate. Surveillance with dd-cfDNA seems to be particularly useful in patients on anticoagulation who are at increased risk of complications from biopsies. dd-cfDNA is more practical than protocol biopsies to detect under-immunosuppression resulting in immune activation. This is especially important in kidney transplant patients with nonadherence, high epitope mismatch burden, and high immune competence. dd-cf DNA is a valuable diagnostic tool for monitoring transplant patients undergoing immunosuppression minimization.

Philip Halloran: Probably not, except that they detect rejection. Monitoring endogenous viruses would be better. Certain scenarios suggest recent under-immunosuppression and nonadherence, and these can be identified in indication biopsies by central molecular testing. dd-cfDNA identifies probable active rejection, so in a sense that this a reflection of immunosuppression.

Michael Keller: Immunosuppressive agents such as the calcineurin inhibitors tacrolimus and cyclosporine are generally dosed to achieve target concentrations in solid organ transplant recipients. Due to the complex interplay of various metabolic factors and drug interactions, the doses required to achieve these concentrations may vary considerably between individuals. Certain target trough levels may provide adequate immunosuppression for some patients yet other patients may develop episodes of acute rejection at the same trough target. Additionally, variability and fluctuations in target trough concentrations frequently occur. In a sense, dd-cfDNA’s ability to accurately detect episodes of acute rejection serves as a marker of immunosuppression adequacy; however, one of the goals of immunosuppression is to prevent the onset of acute rejection in the first place. As a quantitative biomarker of underlying allograft injury, dd-cfDNA may allow for the personalized titration of immunosuppressive drugs to achieve a precise level of immunosuppression that is individualized to the recipient and prevent the onset of acute rejection. While promising in theory, further studies are needed to clarify and validate its role for this use.

Iwijn De Vlaminck: There’s no doubt that quantitative readouts of immunosuppression are needed in transplant medicine. dd-cfDNA, however, provides only an indirect indication of immunosuppression. Molecular diagnostics that interrogate the immune cell repertoire, transcriptional signatures of immune cells or the burden of endogenous viruses in blood may provide a more direct readout of immunosuppression adequacy in the future.

Is the horizon here? What molecular approaches should be combined to improve monitoring/detection/or phenotyping of the rejection?

Angela Wu: Emerging studies of dd-cfDNA that focus not only on the content of the DNA sequence but also additional features such as methylation signatures, fragmentation profiles, or microbial composition are all contributing additional information that improves the utility of cfDNA as a proxy for organ health, disease state, and host immune condition. In my opinion, the combination of these approaches will ultimately allow us to produce assay for monitoring and detection, as well as prognostication. There is no question that the inclusion of these types of assays and assessments are informative and can improve decision-making, but ultimately the quantity of information obtained needs to be weighed with the clinical need and costs to be applicable in a real-life setting. Additional technologies that can produce this information from cfDNA at lowered cost and with ease of operation will be needed to enable mass adoption.

Michael Oellerich: Serial dd-cfDNA determinations will be helpful to achieve personalized immunosuppression in transplant recipients with the potential to reduce premature graft loss. Additional research is needed to explore combinations of dd-cfDNA with other biomarkers such as chemokines (CXCL9, CXCL10), DSA, and microRNAs as well as a mRNA tissue-based gene signature (MMDx) to further improve diagnostic precision. Prospective randomized clinical utility studies are necessary to show that dd-cfDNA guided changes in treatment will improve outcomes.

Philip Halloran: The emerging picture for the commercial assays is that low values indicate very low probability of active rejection, and high values are usually due to active antibody-mediated rejection or T cell-mediated rejection (but are sometimes unexplained). Intermediate values are common and greater resolution on how such values should guide actions is needed. Overall, we need more probabilistic information about how a result is “actionable” given the time posttransplant: (a) in a patient with no indications; and (b) in a patient with indications for biopsy and strong suspicion of rejection.

Michael Keller: dd-cfDNA is an intriguing tool with the potential to provide a variety of useful clinical applications including the detection of allograft injury, an assessment of response to treatment of acute rejection, monitoring of the adequacy of immunosuppression, and prognostic value. Further validation of its clinical utility and incorporation into routine clinical care may usher the field of solid organ transplant into an era of precision medicine and personalized care yet realized. The combined application with other molecular approaches demonstrates further promise. Micro-array-based systems that analyze expression patterns of messenger RNA, such as molecular microscope diagnostic systems (MMDx) may establish molecular phenotypes that aid in the detection and discrimination of various types of allograft pathology. They may also provide further clarity into circumstances in which elevations in dd-cfDNA are not accompanied by histopathological evidence of allograft injury. While the horizon may be approaching, further research is necessary to establish several aspects pertaining to its proper interpretation and clinical applications.

Iwijn De Vlaminck: dd-cfDNA is fundamentally a marker of graft tissue injury. In this simplicity lies a major strength, but also a major limitation; donor DNA does not provide deeper insight into the cause of the graft injury. Going forward, molecular approaches can be envisioned that provide more detailed information about the cell and tissue types that are injured, and this in turn may enable distinguishing between different causes of graft injury, including infection and rejection-related injury, and different pathways of transplant rejection. Recent advances in assays that interrogate cfRNA and cell-type specific epigenetic marks within cfDNA may offer a path forward.

Author Contributions

All authors confirmed they have contributed to the intellectual content of this paper and have met the following 4 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; (c) final approval of the published article; and (d) agreement to be accountable for all aspects of the article thus ensuring that questions related to the accuracy or integrity of any part of the article are appropriately investigated and resolved.

Authors’ Disclosures or Potential Conflicts of Interest

Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest:

Employment or Leadership

None declared.

Consultant or Advisory Role

P.F. Halloran, Natera Inc.; M. Oellerich, Chronix Biomedical GmbH, Liquid Biopsy Center; I. De Vlaminck, Karius, Kanvas Biosciences.

Stock Ownership

P.F. Halloran, Transcriptome Sciences, Inc.; I. De Vlaminck, Karius, Kanvas Biosciences.

Honoraria

P.F. Halloran, One Lambda, Inc., Natera Inc.; M. Oellerich, Astellas.

Research Funding

A.R. Wu, Hong Kong University of Science and Technology (HKUST); I. De Vlaminck, NIH DP2AI138242, R01AI146165, R01AI151059; S. Agbor-Enoh, the Division of Intramural Research of NHLBI through the Lasker Clinical Research Fellowship Program, the NIH Distinguished Fellowship Program, a grant from the Cystic Fibrosis Foundation (AGBORE20QI0), and the Intramural Anti-COVID-19 Award (Award ID 733235).

Expert Testimony

None declared.

Patents

None declared.

Other Remuneration

P.F. Halloran, One Lambda, Inc., Natera Inc.; M. Oellerich, IATDMCT; I. De Vlaminck, Viracor Eurofins, Kanvas Biosciences, Karius.

Acknowledgment

Kelly Byrne at the NIH Clinical Center, Department of Critical Care Medicine for editorial review.