Donor-derived Cell-free DNA—Ready for Prime Time Yet?

Heart transplantation (HT) is still the only treatment markedly changing the disease course in individuals with end-stage heart failure.1 However, allograft rejection is still a major issue that contributes to mortality, even if clinical results after HT have significantly improved over time. As such, monitoring and diagnosing rejection of cardiac allografts remains challenging: despite being the "gold standard," endomyocardial biopsy (EMB) has shown several limitations because of its invasiveness and interobserver variability.2 To overcome EMB limitations, several noninvasive methods based on molecular biology assays from peripheral blood have been developed.3-5 However, none of these methods have yet achieved enough validation to be fully recommended in clinical practice. In this issue of Transplantation, Halloran et al6 present early results from the Trifecta-Heart cfDNA-MMDx (Comparing the DD-cfDNA Test to MMDx Microarray Test, Central HLA Antibody Test, and Histology) study—an investigator-initiated, prospective trial—to examine the correlations between donor-derived cell-free DNA (dd-cfDNA) and gene expression profiling of the EMB (by Molecular Microscope read). Seventy HT patients were included with 137 biopsies and analysis performed during 18 mo of the study duration. The main focus of the present report was the correlation of plasma dd-cfDNA with gene expression in EMBs and comparing these correlations with those in 604 kidney transplant biopsies from the Trifecta-Kidney cfDNA-MMDx (Comparing the DD-cfDNA Test to MMDx Microarray Test, Central HLA Antibody Test, and Histology) study.7 Here, the authors could show that top transcripts correlating with dd-cfDNA were related to genes increased in rejection, such as interferon gamma-inducible genes (eg, major histocompatibility complex, class II, DM alpha), but also with genes expression of myocardial injury and expressed in macrophages (eg, serpin family A member 1, heme oxygenase 1), and myocytes (eg, the atrial natriuretic peptide A and B genes, natriuretic peptide A and natriuretic peptide B). After a gene enrichment analysis, the top dd-cfDNA-correlated genes reflected inflammation and rejection pathways. Furthermore, it was shown that dd-cfDNA correlations with rejection genes in EMBs were similar to those in kidney transplant biopsies, but in EMBs, there were stronger associations for T cell–mediated rejection (TCMR) genes than in kidneys, whereas antibody-mediated rejection (AMR) genes were more correlated to dd-cfDNA in kidney than in heart biopsies. Moreover, correlations with parenchymal injury-induced genes and macrophage genes were much stronger in hearts. The authors conclude that in this first analysis of the Trifecta-Heart study, dd-cfDNA correlates with molecular rejection and injury and macrophage infiltration. These results are encouraging for the utility of dd-cfDNA in the clinical management of HT recipients. After several studies showing an association between an increase in dd-cfDNA and histology-proven rejection,8 this article shows the correlation between dd-cfDNA and molecular profile of graft tissue. However, we do not believe that, at this early stage, this small and preliminary analysis is adding significant substance to what we already know from histology-based studies to support dd-cfDNA clinical utility. To this extent, the information that dd-cfDNA is higher in EMBs showing rejection is not new; the way the analysis is performed does not allow for assessment of the diagnostic accuracy of dd-cfDNA and the logistics and time needed to have both molecular test results to take a therapeutic decision is not compatible with clinical practice. In contrast, we believe that these data provide significant knowledge into the interpretation of the molecular pathways of rejection and myocyte injury occurring in heart allografts compared with the kidney. The first important finding, to be confirmed with a larger sample size, is that as opposed to studies based on pathology-diagnosed rejection, and in contrast to Trifecta-Kidney data, dd-cfDNA is more correlated with TCMR-related genes than with AMR-related genes in heart allografts. This is also a partial discrepancy with pathology-based studies, where dd-cfDNA was more associated with AMR than with TCMR and suggests that the archetypes of the Molecular Microscope Diagnostic Systems may require a more complex interpretation. A second important finding is the strong association with "injury"-related genes. In particular, dd-cfDNA is associated with macrophage genes, natriuretic peptide genes, and myocyte dedifferentiation. This finding provides additional relevance to the "injury" archetype that was identified in heart allograft,9 because it is associated with a biomarker of graft damage, such as dd-cfDNA. However, the authors provide a too vague hypothesis about the cause of the "injury," and the analysis lacks deeper patient phenotyping, which includes granular clinical and functional information. It needs to be underlined that AMR is a form of vasculitis; in its chronic form, it may be associated with loss of interstitial capillaries, and AMR in the heart can lead to ischemic-type injury.10 In light of the strong association between the injury archetype and dd-cfDNA, witnessing significant myocyte lysis, we believe that an alternative hypothesis to interpret this finding is that at least some of the patients showing the "injury" archetype are subsequent to AMR-mediated damage. The classical AMR archetype, which was originally derived from kidney gene expression studies and included endothelial inflammation genes, might be shadowed by an overexpression of injury-related genes, that is, typical of heart allografts in a context where capillary density is reduced and inflammation is diffused. Indeed, injury archetype and not AMR showed prognostic implications in data from the INTERHEART (Diagnostic and Therapeutic Applications of Microarrays in Heart Transplantation, a Multicenter Study) consortium.9 Moreover, the injury archetype reflects macrophage gene activation, and these cells are highly relevant in the histology of severe AMR. The hypothesis that the backbone of some of the "injury" cases is related to ongoing or previous AMR may reconcile the discrepancy with previous studies showing that dd-cfDNA is more associated with AMR than TCMR, and it would be aligned with the notion that in HT, AMR has a wide clinical and pathological phenotype that may lead to allograft vasculopathy and is associated with adverse prognosis. Taken together, these preliminary data from the Trifecta study provide the rationale to revise the interpretation of the molecular phenotypes of heart allograft rejection and the mechanisms associated with graft damage. The limited sample size represents a significant warning against any firm conclusion, and we look forward to further analysis of this relevant project. Therefore, in our opinion, "Prime Time" for dd-cfDNA as a single biomarker in clinical routine surveillance of HT patients has yet to come.