In lung transplantation (LTx), while survival after transplantation has increased over time due to improved survival in the early post-transplant period, chronic allograft rejection remains a major cause of morbidity and mortality [1
]. Infections and chronic lung allograft dysfunction (CLAD) are major reasons for this inferior long-term outcome [2
]. Furthermore, the diagnosis of CLAD has proven challenging and evaluation typically includes bronchoscopy, mostly to rule out other causes such as infection. However, the sensitivity of transbronchial biopsy for the diagnosis of chronic rejection is only 30% [3
]. Given the importance of diagnosing graft rejection and dysfunction, many investigators have explored the use of biomarkers in bronchoalveolar lavage (BAL) primarily in acute rejection (AR) [4
] as an adjunct to diagnosis, but there are as yet no universally accepted diagnostic or prognostic BAL biomarkers for acute rejection nor for CLAD diagnosis or biomarkers for prediction of lung transplant survival [8
C-X-C motif chemokine 10 (CXCL10) is increasingly being explored as a biomarker for the early detection and monitoring of renal function in renal transplant patients [9
]. CXCL10 is an interferon gamma-induced, small cytokine, belonging to the C-X-C motif chemokine family. Previous studies have demonstrated that CXCL10 is a potent chemoattractant for various immune cells including CD4 and CD8. Additionally, elevated CXCL10 reflected tubulointerstitial inflammation and peritubular capillaritis [10
]. Donor-derived cell-free DNA (cfDNA) has also been increasingly explored in the transplant space, especially for the detection of acute rejection in renal transplantation [11
]. Measurements of donor-specific cfDNA in the plasma [12
] and urine [14
] (specifically in renal transplantation) have been shown to detect episodes of acute rejection. These methods typically require extensive sequencing, high throughput SNP-testing, or a priori knowledge of specific genetic differences between donor and recipient in order to work.
Cell-free DNA consists of fragments of nucleic acids that circulate in the biofluids of both healthy individuals and patients with a variety of diseases [15
]. Multiple studies have shown that circulating plasma and urine cfDNA is elevated in transplant patients undergoing rejection of allografts [11
]. Recent studies have utilized digital PCR to quantify donor-derived cfDNA from blood and demonstrated that donor-derived cfDNA could be used as an early non-invasive biomarker for acute lung allograft rejection [18
]. In this study we explored the use of cfDNA and CXCL10 derived from BAL specimens as diagnostic biomarkers of CLAD and possibly as prognostic markers of overall survival.
In this study, we used BAL fluid samples from lung transplant patients to evaluate the predictive capacity of the biomarkers CXCL10 and cell-free DNA to detect and distinguish between subphenotypes of chronic lung allograft disorder as well as predict transplant survival. Our results showed that both CXCL10 and cfDNA together could distinguish between stable patients and the BOS and RAS subphenotypes of CLAD. Furthermore, both CXCL10 and cfDNA together could segregate patients into low- and high-survival patient populations, which superseded the utility of traditional biomarkers such as IL-6 and IL-8 in their predictive capacity.
Such findings are important because, of all major solid organ transplants, lung transplantation has the worst median overall survival of approximately 6 years as compared to >10 years for both heart and kidney transplantation [24
]. This is largely due to the development of CLAD in 50% of recipients after 5 years, ultimately leading to significant morbidity and mortality. The mechanisms underlying CLAD development are poorly understood in part because a diagnosis of CLAD describes not a single entity, but a heterogeneous group of phenotypes characterized by varying degrees of airway neutrophilia, fibrosis, histological features, and responsiveness to therapy [27
]. While two major phenotypes of CLAD have been described, BOS and RAS [21
], the differential phenotyping of these two is confounded by a lack of biomarkers able to effectively differentiate the two [8
]. Identification of such biomarkers is essential because they may elucidate mechanistic differences between these phenotypes and lead to the development of individual therapies specific to each condition.
Our group and others have sought to identify such biomarkers in the BAL fluid of LTx patients, as collection of this biofluid is relatively noninvasive and directly reflects the pulmonary milieu. Biomarkers in the BAL fluid such as neutrophil proportions [23
], IL-6 [28
], and IL-8 [23
], have been found to differentiate between CLAD and stable phenotypes, but do not reliably distinguish between BOS and RAS. As such, additional biomarkers are necessary to understand the differences
The current study is the first study to jointly assess CXCL10 and cfDNA in BAL. CXCL10, a pro-inflammatory cytokine that is a CXCR3 ligand, has been previously found to be associated with diffuse alveolar damage in and development of CLAD [30
] and also as a risk factor for CLAD development in the context of organizing pneumonia post-transplantation [31
]. Previously, our group found that CXCL10 was elevated in RAS versus stable patients and was associated with survival [28
]. While donor-derived cfDNA has been evaluated in the peripheral blood of LTx patients for monitoring infection, acute rejection, and CLAD [17
], levels of cfDNA in the BAL fluid of LTx patients has not been explored.
In the BAL of LTX patients, the combination of cfDNA and CXCL10 were found not only to differentiate between BOS and RAS patients, but also associate with graft survival after diagnosis. This suggests that high levels of both may identify patients with a poor prognosis. Our data show that BOS patients had a significantly greater amount of cfDNA in BAL than stable or RAS patients. Additionally, BOS patients had greater neutrophil and IL-8 counts compared to stable patients. Given the role of IL-8 as a neutrophil chemotactic factor, it is not surprising to see its levels elevated in the BOS group, who also had elevated neutrophils counts. The elevated presence of cfDNA in the BAL of BOS patients is likely multifactorial and we hypothesize it is likely from a combination of tissue injury in CLAD and release from granulocytic inflammatory cells, such as that in the form of neutrophil extracellular traps, resulting in both donor-derived and host cfDNA.
Given that CXCL10 was not significant in univariate analysis, its key role in differentiating between the different phenotypes and in predicting survival in the context of cfDNA was surprising. DNA, through activation of the TLR9 receptor, has been shown to be upstream of CXCL10 release by BAL fluid cells [32
]. Given that CXCL10 can bind to CXCR3 on neutrophils and this receptor-ligand interaction has been shown to enhance neutrophil activation and neutrophil-mediated lung injury [33
], it is possible that cfDNA and CXCL10 work cyclically to amplify inflammation and contribute to CLAD development. Mechanistic studies are needed to elucidate the exact roles that each of these biomarkers play in the post-transplantation pulmonary environment. Such studies could elucidate the relationship between these biomarkers and the differences in survival and help to develop targeted therapeutics to prevent CLAD development or progression.
Limitations of this study include the small sample size but, as an exploratory study, our results suggest cfDNA and CXCL10 measurements in BAL as highly synergistic biomarkers could be utilized prognostically in in LTx patients. These markers together appear to provide useful information above and beyond standard BAL biomarkers including IL-6, IL-8, and neutrophils. Larger prospective studies in LTx patients utilizing cfDNA and CXCL10 are needed to further assess our findings here and are currently underway. Future studies with more robust characterization of serial cfDNA and CXCL10 levels in BAL samples from lung transplant could lead to a biomarker to help clinicians anticipate the development of CLAD in lung transplant recipients, and/or monitor its progression and the relative effectiveness of current and future therapeutic interventions.