Search

The current paradigm defines acute rejection (AR), a composite of acute cellular rejection or antibody-mediated rejection, as distinct acute events. Yet, affected patients often progress to chronic lung allograft dysfunction despite initiation of targeted therapy at diagnosis of the acute process. We hypothesize that AR is not an acute event, but rather, an exacerbation of an on-going early post-transplant disease. Herein, we leverage donor-derived cell-free DNA, a sensitive marker, to define post-transplant allograft injury pattern associated with AR. 141 prospectively enrolled lung transplant patients (NCT02423070) were grouped as acute rejection ("AR") or no rejection ("NR") as determined by adjudication. Allograft injury was measured by ddcfDNA via shotgun sequencing in 1557 prospectively collected plasma samples; medians and interquartile ranges (IQR) are reported Over the median 19.6 months follow-up, 87 episodes of AR were detected in 51 patients (34.7%) at a median 7.6 (IQR =2.4 - 12.7) months post-transplant. %ddcfDNA was high after transplant surgery and then decayed. "AR" showed slower decay. Their baseline ddcfDNA levels were higher than for "NR" well before AR diagnosis (FigA), starting from Day 7 post-transplant and reached 4X higher than for "NR" by 3 months post-transplant (0.88%, IQR=0.31% - 1.93% vs. 0.23%, IQR=0.13% - 1.59%). At AR diagnosis, ddcfDNA levels rose (1.95%, IQR=1.14% - 5.04%) and were ∼2 fold higher than the within subject high baseline levels (FigB), and ∼6X higher than for time-matched "NR". ddcfDNA reduced after AR treatment but remained higher than for time-matched "NR" (0.52%, IQR=0.22% - 1.59% vs. 0.25%, IQR=0.08% - 0.57%, p< 0.01). Early after transplant, "AR" patients show higher levels of baseline allograft injury that persist and then rise with diagnosis of AR. If ongoing studies confirm these findings and identify mechanisms that predict elevated baseline injury and AR phenotype, treatment paradigms could be revised to improve allograft outcomes. [ABSTRACT FROM AUTHOR]

Non-invasive monitoring of heart allograft health is an important clinical goal and circulating microRNAs (miRs) have been associated with acute rejection in limited studies. The Genomic Research Alliance for Transplantation (GRAfT) is a multicenter prospective study. Patients with no history of rejection after transplant were selected as controls. Cases were patients with acute cellular rejection (ACR) or antibody-mediated rejection (AMR). Small RNA sequencing was performed on plasma samples collected at the time of an endomyocardial biopsy. Differential miR expression analysis and LASSO regression was used to develop ACR and AMR miR panels while adjusting for clinical covariates. The panels were validated in an independent sample from GRAfT and Stanford University. Receiver operating characteristic curves were generated and area under the curve (AUC) statistics are reported. Distinct ACR and AMR clinical scores were developed to translate gene expression data for clinical use. The GRAfT cohort had a median age of 52 years, with 35% females and 45% Blacks. We included 85 controls and 34 patients with rejection (20 ACR, 14 AMR). In ACR and AMR, using differential gene expression and regression analysis, we identified 11 and 15 miRs that accurately discriminated ACR and AMR (p<0.05). Independent validation of the miR panels within GRAfT led to excellent diagnostic accuracy for ACR AUC 0.92 (95% CI: 0.86-0.98) and AMR AUC 0.82 (95% CI: 0.74-0.90). The externally validated ACR AUC was 0.91 (95% CI: 0.82-0.99). Distinct ACR and AMR miR clinical scores were developed ranging from 0-100; using a threshold of 65 identified ACR with a sensitivity of 86%, specificity of 78% and negative predictive value of 98%, for AMR the respective performance was 82%, 84% and 97% (Figure). The unique clinical ACR and AMR scores usher in an era whereby genomic biomarkers are used not only for rejection surveillance but permit accurate, non-invasive diagnosis of ACR and AMR. [ABSTRACT FROM AUTHOR]

African American patients (AA) have poorer outcomes than White patients (W) after heart and kidney transplant. Little is known about differential outcomes in lung transplantation (LTx). This study compares allograft injury and chronic allograft dysfunction (CLAD) in AA and W LTx. The 469 LTx enrolled in two multicenter prospective cohort studies were categorized by self-reported race. Outcomes: Time from transplant to CLAD-free survival adjudicated by a committee using ISHLT standard definitions. Measure: Serial post-transplant plasma samples (n=2152) were assayed for %ddcfDNA by shotgun sequencing. Analyses : Baseline characteristics, post-transplant %ddcfDNA trends, tacrolimus trough levels and CLAD-free survival were compared between AA and W. LTx recipients were 79% W and 14 % AA. Donor-recipient race mismatch was more commonly observed for AA than for W (75.0% vs. 27.3%, p< 0.01). Other baseline characteristics were similar in AA and W. During follow-up (median 19.8 months, IQR = 10.7 - 36.9), tacrolimus trough levels were similar in AA and W (9.2±0.1 vs. 9.0±0.1, p = 0.19). However, AA showed 1.5X higher rejection rates (HR = 1.6, CI = 1.0 - 2.6), and higher %ddcfDNA levels particularly beyond 6 months when maintenance immunosuppression doses were lowest (Figure A). Higher %ddcfDNA trends correlated inversely with CLAD-free survival; AA developed CLAD and/or died 18.8 (mean) months before W (Figure B). Similarly, acute rejection was inversely with CLAD-free survival. AA LTx show higher allograft injury and poorer CLAD-free survival than W LTx despite equivalent tacrolimus blood levels. Studies to understand the mechanisms of allograft injury and poor outcomes in AA are warranted across all solid organ transplants. [ABSTRACT FROM AUTHOR]

Purpose Plasma donor-derived cell-free DNA (ddcfDNA) is a biomarker of organ transplant rejection, but the use of urine ddcfDNA as a clinical biomarker has not been thoroughly investigated. We hypothesize ddcfDNA is present in urine with levels that are predictive of plasma ddcfDNA. To test this, we assessed the physical characteristics of urine ddcfDNA and determine its quantitative relationship to plasma ddcfDNA in heart (HT) and lung transplant(LT) recipients. Methods We evaluated %ddcfDNA in concurrently collected urine and plasma samples from 40 patients (16HT and 24LT). Pre-transplant donor and recipient blood was collected for DNA extraction and genotyping. After transplantation, we collected serial urine and plasma samples concurrently for cfDNA extraction, library construction, and shotgun sequencing. Sequence reads were separated as recipient-derived or donor-derived using single nucleotide polymorphisms from the pre-transplant genotype data. The %ddcfDNA was computed as percentage of donor reads to donor plus recipient reads. We compared fragment length, nucleotide composition, and amount of %ddcfDNA in urine and plasma cfDNA. Results 51 urine cfDNA samples were sequenced (mean depth of 30 million reads). Urine and plasma cfDNA were both Adenine/Thymine rich compared to genomic DNA (60% vs. 40%), however urine cfDNA was shorter than plasma cfDNA (peak length 80 vs. 167 base pairs). Immediately after transplant surgery, urine ddcfDNA levels were highest (x̅= 1.013%) and then decayed logarithmically to a baseline level of 0.176%. Plasma ddcfDNA also showed a post-transplant logarithmic decay pattern. Concurrently measured urine and plasma %ddcfDNA values showed a predictable relationship with saturation kinetics (r2=0.449) (Figure). Conclusion Urine %ddcfDNA levels were predictive of plasma %ddcfDNA levels suggesting that like plasma %ddcfDNA, urine %ddcfDNA is a potential biomarker of allograft rejection. Future studies should validate these findings and assess the clinical utility of urine %ddcfDNA. [ABSTRACT FROM AUTHOR]

Summary of Objectives Genomic biomarkers are used to non-invasively screen cardiac transplant recipients for allograft rejection. Gene expression profiling has no ability to diagnose antibody-mediated rejection (AMR) and has a poor positive predictive value for T-cell mediated rejection (TCMR). An emerging class of genomic biomarkers are circulating, cell-free microRNAs (cf-miR). MicroRNAs modulate gene expression and have been implicated in rejection, they are found in the circulation, and early studies, using targeted genomic approaches suggest that cf-miRs may be a helpful biomarker to detect rejection. We hypothesize that cf-miRs can non-invasively diagnose rejection and distinguish AMR from TCMR. Methods The Genomic Research Alliance for Transplantation (GRAfT) is a prospective, multicenter study enrolling transplant recipients since 2014. All patients with AMR or TCMR were included and a group of controls without rejection were selected based on age, sex and race. Each biopsy was reviewed by two blinded, independent, expert cardiac pathologists. Plasma aliquots will undergo small molecular RNA isolation and the circulating cf-miR transcriptome will be sequenced using a next-generation platform. We expect to generate ∼6 million, single-end, 50bp reads per sample. Sequence data will be annotated using the FASTX_Toolkit and DESeq2 will be used to identify differentially expressed cf-miRs. Machine learning algorithms will be used to refine the cf-miR panel and maximize the C-statistic. Endpoints The study includes 35 patients with allograft rejection and 70 controls (Table). For the TCMR cohort there were 27 episodes of grade ≥ 2R rejection. For AMR there are 11 episodes of ≥ pAMR2 and 16 episodes of pAMR1 h or i. In total ∼300 plasma samples will be sequenced. Our primary objective is to develop a cf-miR panel that diagnoses allograft rejections, distinguishes rejection subtypes and allows for a blood-based, biochemical assessment of rejection severity. [ABSTRACT FROM AUTHOR]

Higher rates of acute rejection in African American transplant recipients are reported across multiple organ types. Mechanisms underlying poorer outcomes vs. other races are poorly defined. Because race is a social construct, most studies use self-identified race to categorize patients. Reports of genetic unique to individuals of African ancestry (AA) led us to hypothesize that ancestry markers might enhance the precision of race categorization in studies of transplant injury and rejection. Using the Genomic Research for Transplantation (GRAfT) cohort, we performed whole-genome genotyping assay and measured the donor-derived cell-free DNA (dd-cfDNA), a biomarker of graft injury. We genotyped 171 heart recipients and estimated the genetic ancestry proportions by GRAF-pop software referencing 1000Genome data. Applying the filters for African genetic ancestry proportion (AGAP), patients were categorized as AA if AGAP was > 85%, and not of AA if AGAP < 15%. Serial post-transplant plasma samples were analyzed for dd-cfDNA by Illumina HiSeq3500. We consolidated the highest dd-cfDNA level along the time series to represent each individual; data are presented as median + the standard error. To assess potential relationship between AGAP and dd-cfDNA, we selected patients at two extremes: 31 AAs (>85% AGAP) and 97 non-AAs (< 15% AGAP). Standard statistical analyses assessed interaction of self-reported race or AGAP with dd-cfDNA. Among the 171 recipients, 69 self-reported as AAs (40%), while the AGAP reported 72 AAs (42 %). The consistence between self-reported and AGAP was 97.7% (167/171). AAs at the estimated extremes (> 85% AGAP) had significantly higher dd-cfDNA level than non-AAs (< 15% AGAP): 3.64 + 0.73 vs. 1.79 + 0.41, p=0.03). Self-reported race showed a non-significant trend towards higher dd-cfDNA levels in AAs compared to non-AAs: 2.91 + 0.48 vs. 1.79 + 0.40, p =0.074). This study documents a high consistence between self-reported race and genetic ancestry markers. It also confirms our prior reports of increased graft injury reflected as higher levels of dd-cfDNA in AAs at the extremes of genetic ancestry proportion. Supporting our hypothesis, ancestry markers might augment self-reported race data, warranting further studies on the interaction genetic heterogeneity with social determinants of health. [ABSTRACT FROM AUTHOR]

Antibody-Mediated Rejection (AMR) that occurs latter in the lung transplant (LTx) course show greater risk of chronic lung allograft dysfunction (CLAD) than early AMR. This suggest distinct mechanisms. This study leverages a large cohort and sensitive genomic tools to profile early and late AMR. 435 LTx patients enrolled in two cohort studies were included. Endpoint and measures: CLAD-free survival is defined as being alive and CLAD-free. A committee adjudicated for both AMR and CLAD. Allograft injury at AMR was assessed as FEV1 decline and as plasma donor-derived cell-free DNA-%ddcfDNA, a validated biomarker. Then, bronchoalveolar lavage cells obtained at AMR (n=11) and no rejection control time points were subject to bisulfite sequencing; only AMR that met ISHLT probable AMR criteria was analyzed to reduce heterogeneity. Analyses : AMR was grouped as late (detected beyond 1 year post-transplant) or early AMR to compare donor-specific antibody (DSA) profile, allograft injury, and CLAD-free survival. Sequence reads for AMR and within subject control were compared to identify differentially methylated genes in early and late AMR. Early AMR (n=75) was more common than late AMR (n=47) over the 20 (IQR = 10 - 37) months follow-up. Both showed similar DSA profiles. However, late AMR showed higher FEV1 decline (0.8 L vs. 0.4L, p < 0.01), %ddcfDNA (5.2% vs. 2.1%, p < 0.01), and hazard of dying or developing CLAD than early AMR (Figure A). Epigenetic analysis identified distinct gene sets in early (n=1381) or late (n=994) AMR. There were only 264 (10%) overlapping genes. However, these genes were predominantly hypermethylated in early AMR and hypomethylated in late AMR. Pathway enrichment analysis also show distinct mechanisms; complement pathways in early AMR and NK-cells pathways in late AMR (Figure B) Early and late AMR show distinct clinical and molecular profiles, suggesting the need for different treatment strategies. Validation studies are needed and may guide treatment decisions. [ABSTRACT FROM AUTHOR]

Donor-specific antibodies (DSA) is a major risk factor for chronic lung allograft dysfunction (CLAD) after lung transplantation (LTx). Unfortunately, DSA management (Rx) varies significantly. To date, there is no accepted standard on when to initiate treatment. This study examines whether the timing of DSA treatment alters outcomes. 435 LTx subjects enrolled in two cohort studies were included. They underwent surveillance and clinically-indicated HLA DSA testing. DSA+patients were treated with a Rituximab or Bortezomib-based regimen either before they showed any clinical or histological signs of AMR (Early Rx) or when they developed these signs of AMR (Late Rx); choice was based on provider's practice. Outcomes: Time from transplant to CLAD-free survival, defined as being alive and CLAD-free. CLAD was adjudicated by a committee. Measure: Allograft injury at time of Rx was quantified via plasma donor-derived cell-free DNA (ddcfDNA), a validated biomarker. Analyses: Baseline characteristics, allograft injury, DSA clearance and CLAD-free survival were compared between treatment groups and controls (DSA- patients). Median follow-up was 10 (IQR = 11 - 37) months. 58% of patients developed HLA DSA, which was predominantly Class II than class I (85% vs. 15%). DSA+patients were treated either early (n = 49) or Late (n = 97). The remaining 32% of DSA+patients lacked relevant clinical data or were not treated. Baseline characteristics and DSA specificities were similar between treatment groups. However, Late AMR showed higher median %ddcfDNA (4% vs 1%, p < 0.01), lower percentage of DSA clearance (40% vs 63%, p = 0.02), and a 3-fold hazard of dying or developing CLAD than Early Rx (Figure). Early Rx extended CLAD-free survival by 34 months compared to Late Rx, and showed similar CLAD-free survival as DSA-patients (p = 0.67). This study supports Early Rx of DSA over Late Rx since Early Rx shows improved CLAD-free survival. Multivariate analyses to identify other potential covariates are being planned. [ABSTRACT FROM AUTHOR]

Purpose AMR show strikingly similar molecular mechanisms (chronic inflammation, fibrogenesis, and others processes) to conditions that accelerate lung aging. Clinically, both AMR and age acceleration manifest as progressive loss of lung function leading us to hypothesize that AMR is associated with age acceleration. Age acceleration triggers signature methylation changes at cytosine-guanosine (CpGs) motifs of "aging" genes. Thus, to test this hypothesis, we used unbiased bisulfite sequencing. Methods Six lung transplant patients with AMR (positive allograft dysfunction, DSA and histopathology and/or C4d) were identified. Two bronchoalveolar (BAL) samples were collected from each patient, one pre-AMR control (no rejection, no infection) and one at AMR diagnosis. BAL cells were used for DNA isolation, bisulfite treatment, and sequencing. Sequences were analyzed via a custom-built computation workflow (bismark, bsseq, bumphunter) that assigns CpGs to DNA regions, normalizes data for cell composition, performs paired analysis to identify differentially methylated regions (DMRs), and maps the DMRs to genes for pathway enrichment analysis. Thresholds for CpG coverage (6X) and DMR mean difference (10%) were arbitrarily selected. P-values were corrected for multiple analyses. Results We identified 900K DMRs between AMR and control. Average DMR length was 238 base pairs with median 4.7 CpGs per DMR (range 3-64). 38.8K DMRs showed 10% or higher methylation difference between AMR and control, with 67% of the DMRs mapped to genes within 1 Killobase. "Aging" genes were overrepresented (q=2.880 × 10−13): genes that promote lung age acceleration (n=370) were predominantly hypo-methylated in AMR, while genes that inhibit lung age acceleration (n=454) were predominantly hyper-methylated. Antibody-mediated pathways such as macrophage activation, Fc-gamma-mediated cell-death, and cytotoxic T-cell activation were also overrepresented (q=0.0007 - 0.0040). Conclusion Both antibody-mediated pathways and "aging" genes were differentially methylated in AMR suggesting epigenetic regulation of these processes and a strong association between AMR and age acceleration. Future studies are needed to validate these findings and assess whether age acceleration contribute to the progressive loss of lung function often observed with AMR. [ABSTRACT FROM AUTHOR]

Prior studies have demonstrated significant variability amongst pathologists in the histopathological interpretation of the endomyocardial biopsy (EMB) for acute cellular rejection (ACR) and antibody-mediated rejection (AMR). Percent donor-derived cell-free DNA (%ddcfDNA) has been shown to be a sensitive and specific marker of both ACR and AMR. We hypothesized that combining %ddcfDNA to histopathology can enhance interrater reliability (IRR) in EMB interpretation. The Genomic Research Alliance for Transplantation (GRAfT) is a multicenter, prospective cohort study that enrolled patients from 2015 to 2020. In this substudy, the center pathologist read was compared to two blinded core cardiac pathologists. ACR was graded according to the ISHLT criteria as 0, 1R, 2R, and 3R while AMR was graded as 0, 1H+, 1I+, 2, and 3. %ddcfDNA was determined by shotgun sequencing of donor and recipient DNA. Weighted Cohen's kappa was used to analyze IRR between the center and core reads. The analysis included 94 patients (median age 55 years (IQR: 45-62), 30% Female sex, 41% Black race) with a total of 428 EMBs with paired %ddcfDNA measures. The kappa coefficient comparing the center to core pathology read showed moderate agreement for ACR 0.47 (0.29, 0.66) and fair agreement for AMR 0.30 (0.08, 0.53). Using our previously validated threshold of %ddcfDNA ≥0.25, we reclassified 13 contested EMBs in ACR improving the kappa coefficient to 0.78 (0.64, 0.92, p<0.0001). Similarly, for AMR, we reclassified 14 contested EMBs improving IRR to 0.83 (0.70, 0.96, p<0.0001, Figure). Similar to prior studies, we found limited IRR in the histopathologic interpretation of EMB. Using %ddcfDNA we reclassified contested samples and significantly enhanced IRR in EMB interpretation for ACR and AMR. These results suggest that combining %ddcfDNA with EMB may be a more practical "gold standard" for determining acute rejection in heart transplantation. [ABSTRACT FROM AUTHOR]

Proteomic analysis has identified circulating biomarkers of primary graft dysfunction after heart transplant, but their role in the biological pathways underlying acute cellular rejection (ACR) or antibody-mediated rejection (AMR) is unknown. Adult heart transplant patients from the Genomic Research Alliance for Transplantation (GRAfT, a prospective, multicenter, post-transplant study from 2015 to 2019) had the Olink (Uppsala, Sweden) proximity extension assay (PEA) performed on blood samples collected at time of endomyocardial biopsy. Patients were categorized as having ACR (grade ≥2R), AMR (grade ≥1) or no rejection. Biomarkers of inflammation and cardiovascular disease were assessed, and the log2 fold change was compared between patients. Student's T test was utilized with p<0.025 considered significant. REACTOME pathway analysis was used to evaluate protein expression. The analysis included 113 GRAfT patients (35% Female, 52% Black race, mean age 50.6±11.8 yrs); 84 patients (74%) had no rejection, 17 (15%) had ACR, and 12 (11%) had AMR. With the PEA, 152 unique proteins were analyzed. In both ACR and AMR, CD5, CXCL9, CXCL10, and IL-6 were upregulated (Figure) compared to controls, suggesting common pathways of cytokine-receptor interaction and chemokine (specifically, IL-10) regulation. Protein dysregulation was more pronounced in AMR with 10 inflammatory and 11 cardiovascular proteins having >1 log2 fold change, compared to 1 inflammatory protein in ACR. Patients with AMR and not ACR had upregulation of known cardiovascular biomarkers: myoglobin (MB), ST2, and NT-proBNP, suggesting more myocardial injury, fibrosis and wall stress. Proteomic analysis reveals unique biomarkers and pathways upregulated in AMR and ACR. Cardiac biomarkers were more pronounced in AMR, suggesting more severe allograft injury. These protein markers may provide additional insights into the pathophysiologic processes that underly acute rejection. [ABSTRACT FROM AUTHOR]

Lung transplant patients are at risk of infection due to immunosuppression state. Current practice uses tacrolimus level to monitor immunosuppression, which lacks precision. Invasive surveillance bronchoscopies are used to monitor for pathogens. Given that plasma anellovirus abundance correlates with degree of immunosuppression, we sought to assess the association of plasma anellovirus-specific cell-free DNA (cfDNA) abundance and tacrolimus levels with pathogen detection. A total of 139 prospectively enrolled lung transplant patients provided 1197 serial plasma samples. Patient clinical timepoints were grouped as pathogen positive or negative based on microbiological recovery of respiratory pathogens. CfDNA was isolated from plasma for shotgun sequencing. Sequence reads were mapped to identify viral cfDNA. Anellovirus abundance was calculated as the fraction of anellovirus cfDNA reads to total viral cfDNA reads. Anellovirus abundance and tacrolimus trough levels were compared between groups using Wilcoxon rank-sum test. Median post-transplant follow-up was 365 days. Prevalence of pathogen recovery was 28.9% and median time of detection was 184 days post-transplant (Interquartile range [IQR] = 107 - 311). Median anellovirus abundance was higher in pathogen positive (n= 65, 81.1% [IQR = 2.5 - 94.8%]) than pathogen negative timepoints (n=1132, 4.2% [IQR = 0.0 - 85.7%], p=0.0005) (Fig. A). Tacrolimus levels were not significantly different (9.9 [IQR = 7.7 - 11.7] ng/ml vs. 9.3 [IQR = 6.5 - 12.1] ng/ml, p=0.49) between the two groups (Fig. B-C). Plasma anellovirus abundance is significantly associated with isolation of pathogen in lung transplant recipients. [ABSTRACT FROM AUTHOR]

The majority of lung transplant centers in the United States perform surveillance bronchoscopy with transbronchial biopsies (TBBx) to monitor for acute rejection (AR) and infection after lung transplantation. However, no high-quality studies have assessed the benefit of surveillance TBBx, an invasive, costly procedure with a risk of complications. Donor-Derived Cell-free DNA (dd-cfDNA) is a non-invasive molecular biomarker of allograft injury that increases in the setting of AR and infection. Prior observational studies support the utility of dd-cfDNA in detecting AR and infection for surveillance purposes, however, no studies have directly compared a dd-cfDNA-based surveillance strategy to surveillance TBBx. We describe here the study design of a randomized controlled trial, Lung Allograft Monitoring with Blood dd-cfDNA Assessments (LAMBA 001) to test the hypothesis that a dd-cfDNA-based surveillance strategy is non-inferior to surveillance TBBx during the first-year post-transplant. LAMBDA 001 is a multicenter, parallel group, open label randomized controlled non-inferiority trial that aims to randomize 400 adult lung transplant recipients via stratified block randomization to surveillance with dd-cfDNA (the ProsperaTM test) vs. surveillance TBBx, performed according to the participating institution's standard of care surveillance schedule. Patients will be followed for 1-year post-transplant. The primary endpoint of the study is Hospital-Free Days (HFD) at 1 year, defined as days alive and spent outside of an acute-care hospital, long term care facility or emergency department. Secondary endpoints include forced expiratory volume in 1 second at one year, incidence of acute rejection, infection, mortality at one year, and health-related quality of life. Assuming a mean (standard deviation) HFD of 311 (37.6) days starting at 4 weeks post-transplant, a non-inferiority margin of 2% (6 HFD) and a one-sided alpha of 2.5%, a sample size of 400 patients provides 80% power to detect non-inferiority. Analysis will be run on both an intention-to-treat and per-protocol population. The LAMBDA 001 trial will provide high quality evidence comparing the effectiveness of dd-cfDNA-based surveillance vs. surveillance TBBx for lung transplant patients in the first-year post-transplant. [ABSTRACT FROM AUTHOR]

Pleuroparenchymal fibroelastosis (PPFE) is a rare entity of interstitial lung disease previously reported in both lung and hematopoietic stem cell transplant (HCT) recipients. The association between PPFE and chronic Graft-versus-Host Disease (cGVHD) has not been established. Patients on cGVHD protocols (NCT00092235; NCT02669251) between 10/01/2004 and 01/31/2020 were included. An independent committee of two transplant hematologists, one cardiothoracic radiologist, and three transplant pulmonologists reviewed patients' demographic and HCT information, pulmonary function tests (PFT), and thoracic CT scans. The diagnosis of PPFE was adapted from the 2013 ATS/ERS Interstitial Lung Disease Consensus Statement (Table 1). PPFE was diagnosed in 0.8% (3/374) of patients (Table 2). All were female with body mass index (BMI) <18.5kg/m2. Patients 2 and 3 had classical Hodgkin lymphoma (cHL) with intrathoracic involvement, received bleomycin (patient 2 and 3) and mediastinal radiation (patient 2) prior to HCT. All had severe cGVHD, requiring 1-5 lines of systemic immunosuppressive therapy. Before the development of PPFE, Patients 2 and 3 had viral pneumonia, while patient 1 had no documented pulmonary infection. Lung biopsies in all patients showed non-specific inflammation and fibrosis; of note, all were performed before the definition of PPFE as a clinical entity in 2013, and did not include the pleura or subpleural parenchyma. At data cut-off (April 2021), all patients were alive. Median time from HCT to PPFE diagnosis was 5.5 years (range, 2-6.5), and median survival post-PPFE diagnosis was 7.5 years (range, 5.5-12). PPFE is a rare, delayed complication of HCT which can be seen with concurrent cGVHD. The pathogenesis of PPFE is unclear but may be associated with the immune-mediated lung injury in cGVHD, infection, preexisting malignancy and lung-toxic treatment. Heightened clinical awareness for PPFE among the cGVHD population warrants further study. [ABSTRACT FROM AUTHOR]

Plasma donor-derived cell-free DNA (ddcfDNA) is a sensitive biomarker for acute rejection (AR). Percent ddcfDNA, expressed as percent of donor DNA to donor plus recipient cfDNA, is reliant on stable recipient cfDNA. However, lung transplant patients are predisposed to kidney and other end-organ injury, which changes recipient cfDNA. In this study, we assessed the test performance of absolute amount of ddcfDNA measured in copies per mL of plasma and %ddcfDNA for detecting lung transplant rejection. Prospectively enrolled patients (n=56) were adjudicated for antibody-mediated rejection (AMR) and acute cellular rejection (ACR) using ISHLT consensus guidelines. Measurement : Serially collected plasma samples (n=729) were assayed for %ddcfDNA by shotgun sequencing and for absolute ddcfDNA by digital droplet PCR using primers that target donor recipient single nucleotide polymorphisms. Analysis: We assessed the area under the receiver operator characteristics curve (AUROC) of %ddcfDNA and absolute ddcfDNA to detect AR, a composite endpoint of AMR and ACR. Over the median 19.6 months follow-up, there were 60 episodes of AR (AMR = 31, ACR = 29). ddcfDNA levels, measured as absolute or as %, were highest after transplant surgery and decayed to baseline levels of 74.7 (IQR = 45.7-116.9) Copies/mL or 0.34% (IQR = 0.21%-0.82%), respectively. Levels increased with AR compared to controls for %ddcfDNA (1.17% vs. 0.34%, p = <0. 0001) and absolute ddcfDNA (232 versus 74 copies/mL, p = 0. 0014). %ddcfDNA showed an AUROC of 0.73 to detect AR, higher than the AUROC of 0.66 for absolute ddcfDNA (Fig.1A). In select patients with increased recipient cfDNA (Fig.1B), absolute ddcfDNA was better to detect AR than %ddcfDNA. %ddcfDNA performed better than absolute ddcfDNA quantitation to detect AR. However, absolute ddcfDNA detects AR better when concurrent recipient tissue injury occurs. A combination of these strategies may provide the optimal method to diagnose acute rejection. [ABSTRACT FROM AUTHOR]

Respiratory viral infection (RVI) is common in lung transplant recipients (LTRs), and a risk for subsequent chronic lung allograft dysfunction (CLAD). Approaches to identify such high risk for post RVI - CLAD are lacking. Our prior research showed viruses thought to be high risk are associated with elevated donor-derived cell-free DNA (%ddcfDNA), a sensitive marker of allograft injury. We hypothesize that high risk RVI are associated with greater allograft injury. We assessed if %ddcfDNA identifies events in which RVI leads to decline in lung function. This study included 39 LTRs enrolled in the prospective Genomic Research Alliance for Transplantation with RVI from January 2016 - March 2019 and 1 year of follow up. Serially collected plasma samples were assayed for %ddcfDNA by shotgun sequencing. We defined significant allograft injury as %ddcfDNA values within 7 days of RVI ≥ 1 % (cases) or < 1% (controls). We examined type of RVI, concurrent histopathology, and lung function decline defined as percent change from personal best. We identified 59 RVI events at a median of 274 days post-transplant (IQR: 142, 425) including 38 controls (%ddcfDNA <1%) and 21 cases (%ddcfDNA ≥ 1%) (medians 0.47% and 3.30%, p < 0.001*). Demographics, personal best FEV 1 , and histopathology were not different between cases and controls. At RVI diagnosis, cases showed significantly greater %FEV 1 decline than controls (medians -13.38% and -1.83%, p = 0.006*). Cases also had significantly greater %FEV 1 decline at 90 (medians -7.97% and 0.91%, p = 0.04*) and 365 (medians -20.05% and 1.09%, p = 0.04*) days post RVI. In contrast, no difference in %FEV 1 decline was seen at any timepoint when RVIs were grouped by abnormal (n=17) vs normal (n=27) histopathology. We found that in subjects experiencing RVI, %ddcfDNA may discriminate between those who will recover their lung function and those who will experience sustained decline, a predictive utility not seen with histopathology or pulmonary function testing. [ABSTRACT FROM AUTHOR]

Restrictive Allograft Syndrome (RAS) and Bronchiolitis Obliterans Syndrome (BOS) are two forms of chronic lung allograft dysfunction (CLAD), an irreversible loss of allograft function and a major cause of death in lung transplant recipients (LTR). Clinical features and trends in allograft injury associated with RAS and BOS remain incompletely understood. This study examines the clinical characteristics and trends of allograft injury, as assessed by donor-derived cell free DNA (ddcfDNA), in RAS and BOS. 369 LTR enrolled in two cohorts were examined for the development of BOS or RAS, as judged by adjudication committees using standardized definitions for CLAD. Demographic and clinical variables were collected prospectively for each patient. Serial plasma samples were collected for ddcfDNA quantification via shotgun sequencing. During the median 38 months follow-up (IQR 24.3-49.7), 116 patients (31.4%) developed CLAD, including 83 (71.5%) with BOS and 33 (28.4%) with RAS. There were no significant differences in demographic and baseline clinical factors between the two groups, including sex, race, BMI, pre-transplant diagnosis, or lung allocation score. RAS patients had a higher percentage of positive donor-specific antibodies (80.6%) compared to BOS (60.5%) (p=0.03). RAS patients had higher levels of ddcfDNA than BOS before CLAD diagnosis (p=0.02), with a rise in ddcfDNA levels preceding RAS diagnosis by 3 months (Fig 1a). RAS patients had significantly lower baseline forced expiratory volume in 1 second (FEV1) (p=0.007) and forced vital capacity (FVC) (p=0.001) six months prior to CLAD diagnosis (Fig 1b). RAS patients showed more prevalent DSA and significantly lower baseline lung function compared to BOS. RAS patients also showed higher allograft injury measured via ddcfDNA than BOS patients, with a ddcfDNA rise preceding RAS diagnosis by 3 months. This study demonstrates differences in clinical variables and molecular injury between RAS and BOS. [ABSTRACT FROM AUTHOR]

Prior studies have demonstrated that percentage donor-derived cell-free DNA (%dd-cfDNA) in lung transplant patients may serve as a noninvasive marker of allograft injury and aid in the detection of acute allograft rejection, infection and chronic lung allograft dysfunction (CLAD). Clinical interpretation of %dd-cfDNA values require an understanding of its normal biological variation in stable lung transplant patients in order to identify abnormal results that may indicate the presence of allograft dysfunction. This study establishes the biological variation and reference change values (RCV) of %dd-cfDNA in lung transplant patients using an analytically validated assay with an established analytic coefficient of variation (CV A) of 6.8%. Levels of plasma %dd-cfDNA (Allosure®) were analyzed from a cohort of patients at 4 lung transplant centers using %ddcfDNA as a method to monitor for allograft rejection in place of surveillance transbronchial biopsy. Patients with stable allograft function and ≥ 3 %dd-cfDNA samples were included. Stable patients were defined as having no evidence of infection, rejection, symptoms or documented decline in forced expiratory volume in 1 second (FEV1) > 10% from baseline. The AlloSure assay, a next-generation sequencing-based approach, was used to measure %dd-cfDNA in the plasma. Intraindividual coefficient of variation (CVI), interindividual CV (CVG), the index of individuality (II) and the RCV were calculated. 35 patients with a combined 124 %dd-cfDNA samples were included in the final analysis. 57% were men, the mean age was 59 years and 77% had bilateral lung transplants. The median %dd-cfDNA was 0.31% (interquartile range 0.18% - 0.68%), the 97.5th percentile and 95th percentile were 1.3% and 1.0% respectively. In 30 patients with an average of 3.7 tests, the CV I was 25%, the CV G was 19%, the II was 1.3 and the RCV was 72%. In stable lung transplant patients, fluctuations in %dd-cfDNA levels of up 72% or levels less than 1% are within normal biological variation. With further validation, these thresholds may be incorporated into surveillance monitoring algorithms to identify potentially abnormal results indicating allograft dysfunction. [ABSTRACT FROM AUTHOR]

African American (AA) patients are at risk for increased rates of rejection after heart transplantation (HT).We compared cell-free DNA (cf-DNA) levels after HT by recipient race. This was a retrospective analysis of 96 HT recipients from the Genomic Research Alliance for Transplantation (GRAFT) Registry, of which 63 patients had cf-DNA values. Cf-DNA values were compared by race withan exponential decay model and Kaplan-Meier (KM) analysis was performed to compare time-to-first rejection. Compared to non-AA patients, AA recipients had a similar prevalence of diabetes and hypertension,proportion of males, and donor characteristics. AA recipients had higher cf-DNA values compared to non-AA recipients for the first five days following transplant (8.3% vs. 3.2% p=0.001 Table 1/figure 1). The stable state cf-DNA values decayed rapidly for AA patients and equalized to non-AA patients over the first 7 days (0.46% vs 0.45%, p=0.8 Table 1). Cellular rejection did not differ by race (HR [CI]=1.4 [0.62,3.2], p=0.4). However AA were at higher risk of antibody mediated rejection (HR [CI]=3.8 [1.3,10.9], p=0.01). African American patients had increased cf-DNA values in the first week following heart transplant. This may be a marker of early injury contributing to increased rates of allograft rejection in AA patients. Further analysisadjusting for confounding variables and determining predictors of clinical outcomes will be included at the time of presentation once follow-up of the GRAFT registry is complete. [ABSTRACT FROM AUTHOR]

Maintenance of immunosuppression in lung transplant recipients (LTx) is integral for preventing allograft rejection, with tacrolimus as a mainstay drug. Prior research from our group showed significantly poorer outcomes for Non-European Ancestry patients (Non-EA) as compared to those of European Ancestry (EA) and that donor-derived cell-free DNA (ddcfDNA) predicts molecular allograft injury in LTx. This study examines differences in immunosuppression between EA and Non-EA LTx in the context of disparate outcomes. 453 LTx within two multicenter cohorts were analyzed based on self-identified race. Serial post-transplant tacrolimus blood levels and dosages were collected. Tacrolimus trough levels were dichotomized as "therapeutic" or "non-therapeutic" based on goal trough level ranges and were analyzed over 6-month periods up to 18 months post-transplant. Serial plasma samples were assayed for %ddcfDNA by shotgun sequencing and log-transformed to adjust for skewness. The study cohort was 80.1% EA and 19.9% Non-EA. The odds and mean number of therapeutic trough levels were similar between EA and Non-EA within the 0-6 (OR=0.98 95% CI 0.71 - 1.35, p=0.91), 6-12 (OR=2.25 95% CI 0.93 - 5.45, p=0.07), and 12-18-month (OR= 1.41 95% CI 0.73 - 2.71, p=0.31) intervals. However, non-EA patients required twofold higher median tacrolimus dosages to reach therapeutic trough levels as compared to EA patients (Fig. 1a). Furthermore, non-EA patients had significantly higher ddcfDNA levels as compared to non-EA (Fig. 1b). Despite disparities in post-transplant outcomes, non-EA achieve similar therapeutic levels of tacrolimus as compared to EA but require significantly higher doses to do so. At a molecular level, non-EA showed significantly higher levels of molecular allograft injury as assessed by ddcfDNA despite equivalent immunosuppression trough levels. Further research is needed to investigate racial outcome disparities and use of ddcfDNA to augment immunosuppressive monitoring in minority LTx. [ABSTRACT FROM AUTHOR]

Select bacterial and mold infections are known risk factors for chronic allograft dysfunction (CLAD) in a subset of lung transplant recipients (LTR). Secondary bacterial (SBI) and fungal infections (SFI) have been described among people with severe influenza infection and COVID-19 requiring mechanical ventilation, but the incidence and clinical outcomes of LTR who develop secondary infections following respiratory viral infections (RVI) are not well described. We conducted a retrospective study of LTR who were diagnosed with either influenza or COVID-19 from January 2011-May 2021. Infection definitions and CLAD stage were defined according the International Society of Heart and Lung Transplantation guidelines. Fifty-seven LTR with influenza and 33 with COVID-19 were identified. Eleven (19%) of the LTR with influenza developed SBI and seven (21%) with COVID-19 developed SBI (p=0.83). Among patients who developed SBI, Pseudomonas aeruginosa was the most common isolated pathogen (44%). Five (9%) and 7 (21%) LTR developed SFI within 90 days post Influenza and COVID-19 infection, respectively (p = 0.09). Among patients with fungal infection, Aspergillus species were the most common pathogen (75%). At 180 days post RVI, all-cause mortality was higher in LTR with SBI (17%) compared to those without (3%), p=0.02. Mortality was similarly higher in LTR with SFI (33%), compared to those without (1.3%), p<0.0001. At 180 days post RVI, LTR with SBI had progression of CLAD stage 43% vs 11%, p=0.004. Based on univariable logistic regression, LTR who had augmented corticosteroids at time of RVI, and lower respiratory tract infection (LRTI) had higher risk of SFI (odds ratio (OR) of 6.57 (1.8, 23.9), p=0.002 and 12.5 (2.53, 61.7), p=0.004, respectively). There were trends of increased OR of SBI among LTR with LRTI and LTR requiring ICU admission due to RVI with the ORs of 2.78 (0.98, 8.01), p=0.06 and 3.77 (0.90, 15.84), p=0.07, respectively. Secondary bacterial and fungal infections are associated with increased all-cause mortality in LTR with influenza and COVID-19, and in the case of SBI, also associated with CLAD stage progression. LTR with LRTI and augmented steroids may be at increased risk of secondary fungal infections. Strategies to mitigate and improve diagnosis may warrant further examination. [ABSTRACT FROM AUTHOR]

Donor specific antibodies (DSA) develop in a significant fraction of heart transplant recipients (HTR). Enhanced characterization of high-risk DSA may permit early treatment and prevent downstream antibody-mediated rejection (AMR) and allograft failure. We hypothesize that high risk DSA is associated with allograft injury. Here, we compared allograft injury between high and low risk DSA using donor-derived cell-free DNA (ddcfDNA), a sensitive marker shown to detect AMR ∼3 months earlier than biopsy. Using the prospective, multicenter study, Genomic Research Alliance for Transplantation (GRAfT), 81 HTR who underwent surveillance and clinically indicated endomyocardial biopsy, and DSA testing were analyzed. Measurement: Serially collected plasma samples (n=860) were assessed for ddcfDNA by shotgun sequencing. Analysis: DSA were categorized as "high risk" if they were associated with AMR or were otherwise considered "low risk". DSA was further categorized by specificity as class II vs. class I. We compared ddcfDNA between groups with generalized mixed linear models; excluding ddcfDNA before Day 14 post-transplant to reduce transplant surgical contribution. 47% HTR (n=38) developed DSA at a median 24 days (IQR = 6 - 141 days) post-transplant; and include 28 de novo and 11 preformed DSA. 15.8% (n=6) were categorized as high risk DSA and 84.2% (n=32) were low risk DSA. High risk DSA showed 4X higher ddcfDNA levels compared to low risk DSA (mean [95% CI] 0.30[0.22% - 0.39%] vs. 0.07[0.06%-0.10%], p <0.01, Figure A). As compared to patients without DSA, low risk DSA showed similar ddcfDNA (p=0.52, Figure A). We also found higher ddcfDNA in class II than in class I DSA (Figure B). High risk DSA show higher allograft injury as measured by ddcfDNA. Similarly, de novo and class II DSA are associated with higher allograft injury than class I or preformed DSA. We provide a novel concept that allows for use of ddcfDNA to stratify high-risk DSA in heart transplantation. [ABSTRACT FROM AUTHOR]

Pulmonary chronic graft-versus-host disease (PcGvHD) is a devastating complication of allogeneic hematopoietic stem cell transplant. The 2014 NIH cGvHD Consensus Criteria only describe the obstructive phenotype of PcGvHD. On the other hand, the 2019 ISHLT Chronic Lung Allograft Dysfunction (CLAD) criteria define obstructive, restrictive and mixed phenotypes of CLAD after lung transplantation. In this study, we adapted the 2019 ISHLT criteria to define phenotypes of PcGvHD and compare its performance to the NIH criteria. 350 consecutive patients enrolled on a cGVHD natural history protocol (NCT00092235) were included. We adapted ISHLT CLAD criteria by using populational predictive PFT, instead of within subject best PFT. An independent committee of one transplant hematologist, two transplant pulmonologist, and one cardiothoracic radiologist reviewed clinical data to adjudicate subjects for PcGvHD based on the CLAD-PcGvHD or NIH Criteria (Table 1). Over a median 112.5 (range, 8-392) months of post-transplant follow-up, half of the patients met the CLAD-PcGvHD criteria (n=166), including 12 (3.4%) obstructive, 67 (19.1%) restrictive, 47 (13.4%) mixed, and 40 (11.4%) undefined. Less than half of them (n= 78) met NIH PcGvHD Criteria (NIH+); the rest (n=88) did not (NIH-) mainly due to restriction without obstruction on PFT. The NIH+ and NIH- groups were similar in age, Karnofsky performance status, race, HCT indication and regimen, and cGvHD severity. The NIH- group had more male recipients, skin and joints/fascia cGvHD (64.8% vs 44.9%, 84.1% vs 65.4%, 77.3% vs 57.7%, respectively, all p<.05) than the NIH+ group. Overall survival post-HCT were similar between the NIH+ and NIH- groups, but significantly poorer than the no CLAD-PcGvHD subjects (figure 1). The CLAD-PcGvHD criteria identified high risk patients with PcGvHD that were missed by the NIH criteria. The proposed criteria could become a valuable clinical tool in studying lung disease in cGvHD. [ABSTRACT FROM AUTHOR]

Plasma donor-derived cell-free DNA (ddcfDNA) is a sensitive biomarker for acute rejection (AR). However, its performance remains unknown for single (SL) vs double lung (DL) transplant recipients, where differences in lung mass may affect interpretation. In this study, we aim to distinguish the performance characteristics of ddcfDNA to detect AR in SL vs DL transplant patients. Lung transplant recipients (n=105) enrolled in two prospective cohort studies were adjudicated for acute cellular rejection (ACR) and antibody mediated rejection (AMR). Measurement: 337 serially-collected plasma samples were assayed for ddcfDNA by shotgun sequencing. Analysis: Area-under-the-receiver-operating-characteristic-curves (AUROC) of ddcfDNA and performance characteristics for the detection of AR were calculated separately for both DL and SL transplant recipients; cfDNA values were not doubled for SL transplant as in prior studies. The AUROC for ddcfDNA to detect AR was 0.86 and 0.89 in DL and SL transplants, respectively. The optimal ddcfDNA threshold of 0.54% for SL transplant showed a 92% sensitivity and 80% specificity for detecting AR. For DL transplant, the optimal threshold of 1.1% showed 78% sensitivity and 83% specificity for detecting AR. When ddcfDNA was doubled for single lung transplant, the 1.1% threshold showed 92% sensitivity and 80% specificity for detecting rejection. The AUROC for detection of AMR was 0.91 for both DL and SL transplants. However, the AUROC for detection of ACR was 0.73 for DL and 0.87 for SL transplants. Plasma ddcfDNA showed similar accuracy to detect AR in both SL and DL transplant recipients, however, the accuracy of ddcfDNA for the detection of ACR appears higher in SL vs DL transplant recipients. Correcting ddcfDNA levels in SL transplant patients by doubling their value appears to be a reasonable method of preserving the performance characteristics of the test for the detection of acute rejection. [ABSTRACT FROM AUTHOR]

Acute rejection (AR) causes significant allograft injury. However, its systemic effect on recipient tissues remains poorly defined. Donor-derived cell-free DNA (ddcfDNA) is a sensitive marker of allograft rejection, but does not capture any concurrent recipient tissue injury. Herein, we leverage cfDNA epigenetics as marker of tissue injury to profile recipient tissue injury associated with AR. cfDNA was extracted in 92 plasma samples from 33 heart transplant recipients (HTR) (19 with AR and 14 stable controls) from the GRAfT study (NCT 02423070), and 19 healthy controls (HC). Measurements: Total cell-free nuclear (ncfDNA) and mitochondrial DNA (mtcfDNA) were assayed via digital-droplet PCR. To quantify recipient cfDNA, we first measured ddcfDNA fraction by shotgun sequencing; recipient cfDNA is the difference of total ncfDNA and ddcfDNA. Next, we performed whole-genome bisulfite sequencing on cfDNA and used tissue-specific DNA methylation signatures to deconvolute recipient tissue sources of cfDNA. Median [interquartile range (IQR)] of cfDNA percent or copies (cp)/mL are reported. ddcfDNA levels increased 6 fold with AR compared to stable HTR (0.54 [0.32-1.21] % vs. 0.09 [0.03 - 0.12] %, p <0.01). ncfDNA, which is >99% recipient-derived, was 5X higher in stable HTR (7997 [3974-9740] cp/mL) than HC (1224 [921-1985] cp/mL, p <0.01). With AR, ncfDNA doubled (14949 [8460-24018] cp/mL) compared with stable HTR (p <0.01). mtcfDNA tripled with AR than stable HTR, and showed excellent discriminatory performance (AUC = 0.92, 95%CI [0.84-0.99]). Genome-wide methylation analysis showed significantly higher cfDNA from hematopoietic cells, vascular endothelium, hepatocyte and kidney with AR than Stable HTR, Fig.1. Genome-wide cfDNA methylome analysis shows increased recipient tissue injury with AR. Ongoing studies intend to validate this finding and test if this approach can distinguish antibody-mediated from acute cellular rejection phenotypes. [ABSTRACT FROM AUTHOR]

Clinical infection or colonization of the lung, specifically with Pseudomonas aeruginosa (PsA), is associated with increased incidence of lung allograft injury in lung transplant recipients (LTRs) and with an increase in BAL (bronchoalveolar lavage) pro-inflammatory cytokines. However, the effect of bacterial colonization on sustained innate immune activation and the mechanisms involved are not well understood. We sought to determine the mechanisms underlying persistent pro-inflammatory alveolar macrophage response to PsA stimulation. We stimulated THP-1 derived macrophages and BAL-derived alveolar macrophages from LTRs with bacteria representative of infectious pathogens and upper respiratory commensals for up to 96 hours. Using a high dimensional flow cytometry panel, we evaluated macrophage responses including expression of costimulatory molecules (CD80, CD86, CD40); co-inhibitor molecules (B7-H1), pro-inflammatory cytokines (TNF-α, IL-6, IL-8, IL-1β), and anti-inflammatory mediators (IL-10, IL-1RA, and TGF-β). We observed an upregulation of pro-inflammatory cytokines (TNF-α, IL-6, IL-8, IL-1β) following stimulation by PsA compared to stimulation with other pathogens (Staphylococcus aureus) and oral-commensal bacteria (Prevotella melaninogenica, Streptococcus pneumoniae) in THP-1 macrophages. IL-1 β elevations were sustained for 96 hours after PsA exposure. Alveolar macrophages from LTRs undergoing surveillance bronchoscopy demonstrated elevated levels of IL-6 and showed a similar IL-1β responses to that observed in THP-1 macrophages when exposed to PsA. PsA induces a sustained increase in IL-1β production in alveolar macrophages from LTRs. Given the prior described role of IL-1β in lung allograft injury and CLAD, further investigations are ongoing to determine clinical correlates of allograft injury with PsA and IL-1β. [ABSTRACT FROM AUTHOR]

Lung transplant patients with AMR often fail treatment. Defining the mechanisms involved may identify better drug targets, as well as biomarkers that can be used to tailor therapies and prevent downstream chronic lung allograft dysfunction (CLAD). Here, we perform whole-genome DNA methylome analysis to define the mechanisms associated to AMR non-responders. The case-control design included 26 patients with AMR and 21 controls, matched for race, sex and age. DNA was extracted from BAL cells for whole-genome bisulfite sequencing; controls samples were post-transplant time-matched to AMR samples. AMR patients were adjudicated as Non-responders if they developed CLAD within 2 years of diagnosis, otherwise, AMR patients were grouped as Responders. Bisulfite sequence reads were analyzed with an in-house computational workflow to map BAL cell-type composition, and molecular pathway differences between groups. AMR (14 Non-responders, 12 Responders) were diagnosed at a median 9.6 months post-transplant. We identified different BAL cell-type compositions; monocyte predominance for Responders vs. neutrophilic predominance for non-responders (p< 0.01). The different cell composition was present before AMR diagnosis and persistent after treatment. Cell-composition was similar for Responders and Controls (Fig A). We also identified pathway differences; Responders showed classic complement activation pathways, while Non-responders showed NK-cells and other antibody-mediated cytotoxic pathways (Fig B). We identified different BAL cell composition and mechanisms that correlate with response to AMR treatment. If validated, these features are poised to identify novel drug targets and may serve as biomarkers to tailor AMR treatment. [ABSTRACT FROM AUTHOR]

African American (AA) patients are at risk for increased rates of rejection after heart transplantation (HT). We compared cell-free DNA (cf-DNA) levels after HT by recipient race. This was a retrospective analysis of 96 HT recipients from the Genomic Research Alliance for Transplantation (GRAFT) Registry, of which 63 patients had cf-DNA values. Cf-DNA values were compared by race with an exponential decay model and Kaplan-Meier (KM) analysis was performed to compare time-to-first rejection. Compared to non-AA patients, AA recipients had a similar prevalence of diabetes and hypertension, proportion of males, and donor characteristics. AA recipients had higher cf-DNA values compared to non-AA recipients for the first five days following transplant (8.3% vs. 3.2% p=0.001 Table 1/figure 1). The stable state cf-DNA values decayed rapidly for AA patients and equalized to non-AA patients over the first 7 days (0.46% vs 0.45%, p=0.8 Table 1). Cellular rejection did not differ by race (HR [CI]=1.4 [0.62,3.2], p=0.4). However AA were at higher risk of antibody mediated rejection (HR [CI]=3.8 [1.3,10.9], p=0.01). African American patients had increased cf-DNA values in the first week following heart transplant. This may be a marker of early injury contributing to increased rates of allograft rejection in AA patients. Further analysis adjusting for confounding variables and determining predictors of clinical outcomes will be included at the time of presentation once follow-up of the GRAFT registry is complete. [ABSTRACT FROM AUTHOR]

Clinical outcomes remain extremely poor for children with cystic fibrosis undergoing lung transplantation. Since 1961, the Cystic Fibrosis Foundation has maintained a network of accredited Cystic Fibrosis Care Centers (CFCC) designated to deliver specialized, multidisciplinary care for people with cystic fibrosis (CF). While this integrated clinical approach is thought to be beneficial in patients undergoing transplantation, it remains unknown whether lung transplantation at a CFCC is associated with improved long-term outcomes among children with CF. We used the Scientific Registry of Transplant Recipients to identify all children (age <18 years) with CF who underwent first-time lung transplantation between 2005-2018. Clinical, demographic, and donor characteristics were compared between groups. We used multivariable Cox proportional hazards modeling to compare graft survival (defined as freedom from death, graft failure, or retransplantation) in children treated at CFCCs and non-CFCC centers. We included 336 children with CF who underwent lung transplantation during the study period. Of these patients, 229 (68%) were transplanted at CFCCs. During a median follow-up time of 2.5 person-years, 35 (10%) children experienced graft failure or underwent retransplantation, and 7 (2%) children were lost to follow-up. In our multivariable Cox proportional hazards model, there was no significant difference in graft survival (HR 1.13, 95% CI 0.82-1.56) or overall survival (HR 1.10, 95% CI 0.77-1.57). In children with CF, lung transplantation at a designated CFCC was not associated with a significant improvement in overall or graft survival. While treatment at a CFCC may not be associated with improved survival after pediatric lung transplantation, specialized care at a CFCC may result in improvements in other aspects of care which could not be assessed in this analysis. [ABSTRACT FROM AUTHOR]

Extracorporeal Membrane Oxygenation (ECMO) is a lifesaving procedure in lung transplant recipients with severe primary graft dysfunction (PGD). There is limited information regarding the impact of post-op ECMO on lung allograft function. Circulating donor-derived cell-free DNA (dd-cf-DNA) has emerged as a potential biomarker of lung allograft injury. The goal of our study is to compare the dynamic dd-cf-DNA changes in patients that required ECMO support with changes in lung transplant patients that did not require ECMO. Peri-transplant (pre-transplant and days 1, 3, 7 and 10, 14, 30, 60, 90 post-transplantation) plasma samples from lung transplant recipients (LTRs)of the Genomic Research Alliance for Transplantation (GRAfT) were analyzed for dd-cf-dDNA by next generation sequencing. After matching for gender, type of transplant (single versus double), and underlying diagnosis, the dd-cf-DNA levels and 3 month post-transplant lung function was compared among patients who required (n=9) and did not require (n=45) ECMO post-operatively. The dd-cf-DNA levels were similar between patients with PGD that required ECMO compared to non ECMO patients. The dd-cf-DNA levels for both ECMO patients and their matched controls were decreased by day 10 (Figure 1). The figure illustrates dd-cf-DNA trends for 2 matched cases (ECMO) and 10 controls (not ECMO), showing high initial post-transplant dd-cf-DNA levels followed by an exponential decay. There was no difference between lung function (FEV1) at 3 months between the ECMO patients and their matched controls (p=0.27)-Figure 2. This report suggests that despite a higher risk for postoperative mortality, patients with severe primary graft failure requiring ECMO had similar cfDNA dynamics and lung function at 3-months post-transplant. [ABSTRACT FROM AUTHOR]

This study assessed respiratory pathogens isolated in the first year post lung transplant in a recent multicenter cohort. Prior studies may not reflect current practices in antimicrobial prophylaxis and molecular diagnostics. We examined 499 consecutive bronchoalveolar lavage (BAL) samples for microbial isolates in 82 lung transplant recipients (LTR) enrolled in the prospective Genomic Research Alliance for Transplantation from July 2015 to 2017. Isolates were examined for species, timing: early (0-1 months post-transplant (MPT), intermediate (1-6 MPT), and late (6-12 MPT), and associated lung function decline. 389 microbes were isolated in 499 BAL samples, representing 348 unique episodes, over a median of 196 days. Bacterial isolates were more common in the early and intermediate periods, and fungal and viral pathogens more common in the late period. In non-Cystic Fibrosis (CF) populations, the frequencies of most common bacterial isolates S. aureus, P. aeruginosa , and enteric Gram Negative Bacteria were 15.6%, 12.6%, 23.0%, compared to 28.0%, 30.0%, and 18.0% respectively, in CF LTRs. The most common viral pathogens were rhinovirus (60.6%), and enterovirus (12.7%). Non-invasive molds (Penicillium etc) comprised 73.3% of fungal isolates, with potentially invasive molds (Aspergillus, Mucor sp.) isolated in 12.9% of cases. Notably, there were only two isolates of Aspergillus fumigatus (2.0%). 48/389 isolates (12.3%) were associated with > 10% reduction in FEV1 at 90 days post isolation of which rhinovirus was the most frequent isolate (11/48). In a contemporary cohort, bacterial pathogens predominate in the early post-transplant period, while fungal and viral pathogens are more common later. Pseudomonas and Aspergillus sp. comprised a smaller proportion of total than previously described, and do not appear to associate with early lung function decline. Further studies are underway to determine the conditions in which select infections impact early lung function. [ABSTRACT FROM AUTHOR]

Purpose De novo donor-specific antibodies (dn DSA) are a major risk factor for antibody-mediated rejection in lung transplant patients (LTs). Little is known about specific circumstances that predispose LTs to develop dn DSA. Clinicians use arbitrary intervals to monitor for dn DSA. Since inflammation and injury drive alloimmune responses, we hypothesize that events like infection or acute cellular rejection (ACR) may predispose LTs for dnDSA production. To test this, we assessed i) inflammatory events preceding dn DSA development, and ii) whether these inflammatory events trigger allograft injury using a sensitive biomarker (donor-derived cell-free DNA-%ddcfDNA). Methods 311 LTs from three prospective cohort studies with a median follow-up of 3.4 years were analyzed. LTs underwent DSA testing before transplantation, after transplantation at fixed intervals and at time of allograft dysfunction. Positive DSA (mean fluorescent intensity ≥ 1000 units) were adjudicated based on the post-transplant day of first detection, as preformed (<28 days) or dn DSA (≥28 days). Inflammatory events were defined as any one of the following preceding dn DSA detection by 0-90 days: histopathologic or clinically-diagnosed ACR, new respiratory bacteria, viruses or fungi on microbiological testing. Allograft injury was assessed using %ddcfDNA in serially collected plasma samples via shotgun sequencing (n=135). Results 42.3% of LTs showed positive DSA with 16.9% preformed (n=54) and 24.8% dn DSA (n=79). Median time from transplantation to first dn DSA detection was 210 days. Class II dnDSA (83.5%) were more common than class I (7.6%) or mixed class I and II (8.9%). We identified an inflammatory event preceding 72.2% of dn DSA: infectious causes (39.2%, n=31), ACR (22.8%, n=18), other histopathology findings (10.2%). Respiratory bacteria (45.0%) and viruses (41.9%) accounted for the most infectious organisms identified. Median time-lag from inflammatory event to first DSA detection was 28.5 days (range = 0- 88 days). These inflammatory events were accompanied by a rise in %ddcfDNA from 0.32% (baseline) to 1.14% (median). Conclusion Primarily, respiratory viruses, bacteria, or ACR preceded dn DSA development, suggesting significant interrelation among allograft inflammation, injury, and DSA development. Incorporating these findings in monitoring protocol may enable timelier detection of dn DSA. [ABSTRACT FROM AUTHOR]

Patients who fail to respond to initial pulmonary antibody-mediated rejection (AMR) treatment may benefit from second generation therapies. Unfortunately, current approaches identify non-responders late after irreversible allograft loss and chronic lung allograft dysfunction (CLAD) has developed. We hypothesize that responders and non-responders show distinct features. In this study, we compared responders and non-responders to identify distinctive features that can be used to recognize the non-responders early before CLAD develops. A committee reviewed clinical data of a prospective cohort of 435 patients to adjudicate for AMR and CLAD. Patients were assigned as responders if they were alive and free from CLAD 1 year after AMR diagnosis, otherwise, AMR patients were categorized as non-responders. Donor-specific antibody (DSA) specificity, strength, lung histopathology, and magnitude of FEV1 decline at the time of AMR diagnosis, 30 days, and 90 days after diagnosis were compared between the groups. Then, we compared molecular biomarker of allograft injury: donor-derived cell-free DNA (%ddcfDNA) - focusing on baseline %ddcfDNA (average of the two lowest values from transplantation) and %ddcfDNA levels at AMR diagnosis. 49 (11.2%) subjects developed AMR including definite (N=1, 2%), probable (N=8, 16%), and possible (N=40, 82%). 73% of AMR patients developed CLAD within 1 year of AMR diagnosis (non-responders) and 27% were responders. Responders and non-responders show important differences at AMR diagnosis, including time from transplantation to AMR (median 107 IQR = 35-224 versus 302, IQR 130-555 days; p=0.002) and FEV1 decline at diagnosis from baseline (12.7% vs 29.0%, p=0.023). Time from transplantation to DSA development, DSA specificity and DSA clearance by 90-day of AMR diagnosis were similar (p>0.05) between the groups. %ddcfDNA rise at AMR diagnosis was higher than for no AMR controls (p < 0.001). The elevated %ddcfDNA at diagnosis (2.06 vs 1.87; p=0.622), as well as baseline %ddcfDNA values (0.597 vs 0.725; p=0.877) were similar between the groups. This study showed that responders and non-responders to AMR treatment have differences in the magnitude of FEV1 decline, time of AMR development, and %ddcfDNA. If validated in another cohort, these features can be used to identify and treat non-responders early to prevent CLAD, thus early death. [ABSTRACT FROM AUTHOR]

Tacrolimus is a primary immunosuppressive drug used in lung transplantation (LTx). Unfortunately, significant variability in drug level monitoring and dosing makes it challenging to determine immunosuppression adequacy. Furthermore, the relationship between drug levels and allograft injury remain poorly defined. This study examines the relationship between tacrolimus blood and donor-derived cell-free DNA (ddcfDNA), a biomarker of allograft injury. 96 LTx from a multicenter cohort were included. Periodic tacrolimus blood levels (n= 4,204) were obtained. Serial plasma samples (n=2,122) were assayed for %ddcfDNA by shotgun sequencing. %ddcfDNA was log-transformed to adjust for skewness. Tacrolimus trough levels were dichotomized to "therapeutic" and "non-therapeutic" based on clinically desirable trough ranges. Tacrolimus trough and %ddcfDNA trends were analyzed. Separate analysis was performed in patients with and without ACR (> Grade 1 per ISHLT criteria). Over a 24-month follow up period, %ddcfDNA showed an inverse relationship to tacrolimus tough levels (Figure 1a); patients showed higher %ddcfDNA with non-therapeutic tacrolimus trough levels than with therapeutic levels (p<0.0001) (Figure 1b). 27% of patients experienced at least one episode of ACR. During the study follow-up, their %ddcfDNA levels were higher (p=0.049) and their percentage of non-therapeutic trough levels was higher (p=0.006) compared to patients with no ACR. However, even when tacrolimus trough levels were therapeutic, patients with ACR showed higher %ddcfDNA levels compared to patients without ACR (p=0.020). %ddcfDNA levels vary inversely with tacrolimus blood trough levels in LTx. Even in patients with therapeutic trough levels, those with ACR demonstrate higher %ddcfDNA levels. Further studies should examine whether sensitive biomarkers such as ddcfDNA can be used to augment immunosuppression monitoring in LTx rather than trough levels alone. [ABSTRACT FROM AUTHOR]

The incidence of acute cellular rejection (ACR) in lung transplant recipients (LTRs) remains about 33% in the first-year post-transplant, the highest for any organ transplant. Although acute rejection frequently responds to treatment, at the present time there is no reliable biomarker to non-invasively monitor response to therapy after ACR treatment. We hypothesized %ddcfDNA levels following acute rejection treatment can predict response to therapy. We retrospectively collected clinical and histopathology data on lung transplant recipients treated for ACR who underwent %ddcfDNA testing as part of clinical care between March 2020 and August 2020. We identified a total of 14 LTRs that were treated for ACR and had a %ddcfDNA value available. One patient had only one %ddcfDNA and was excluded from the analysis. 11/13 patients also had available %ddcfDNA values after treatment. 10/11 of these patients demonstrated a decline in %ddcfDNA level following treatment of rejection(figure 1) The mean reduction in %ddcfDNA after treatment was 55% (Log 10 Ratio -0.35; 95% CI, 9% to 78%; p = 0.03). One patient had a significant increase in follow up %ddcfDNA level attributed to persistent ACR and antibody mediated rejection(AMR) requiring Anti-Thymocyte Globuline(ATG) treatment. Subsequent values for this patient also demonstrate that %ddcfDNA levels decrease following aggressive treatment of rejection. One patient(patient 6) had concurrent infection and rejection. %ddcfDNA may to be a useful biomarker in evaluating allograft post rejection treatment. Persistent high levels are suggestive of ongoing ACR or AMR [ABSTRACT FROM AUTHOR]

The need to reduce in-person clinic visits during the COVID-19 pandemic necessitated a significant reduction in traditional lung transplant surveillance with bronchoscopy and clinic-based pulmonary function testing (PFT). We hypothesized that a screening strategy that utilized cell free DNA (CareDx: Allosure), a non-invasive marker of molecular allograft injury, to determine the need for for-cause diagnostics would yield equivalent lung function stability as traditional monitoring using surveillance bronchoscopy and clinic performed PFTs in lung transplant recipients. We compared the change in FEV1, the prevalence of donor-specific HLA antibodies (DSA), and frequency of hospitalization over parallel 6-month timespans in 2 cohorts of LTRs assessed during the 1st 3 years of their transplant. The 1st cohort (N=65) was monitored from 3/2019 to 9/2019 by in-clinic PFT, surveillance bronchoscopy, DSA, and the 2nd cohort (N=73) was monitored from 3/2020-9/2020 with telemedicine, home spirometry, surveillance dd-cfDNA, and DSA testing every 1-3 months. In cohort 2, patients with >10% decline in home FEV1, and/or dd-cfDNA>1% underwent for-cause bronchoscopy and/or further diagnostic evaluation. Surveillance PFTs were conducted once between 7/20 and 9/20 in cohort 2 to assess the stability of lung function. For both cohorts, a minimum interval of 150 days and a maximum interval of 200 days was used for analysis of spirometric change. There were significantly fewer bronchoscopies/patient (pt) in cohort 2 vs cohort 1 (0.91 vs 1.8, p <0.01). Despite reduced bronchoscopic surveillance, the median % change in FEV1 over 6 months did not differ between cohort 1 vs cohort 2 (1.04% vs 1.01 % p =0.51). Further, there was a comparable prevalence of any level HLA-DSA in cohort 1 vs 2 (32% vs 31% p=0.26). The total incidence of hospitalizations was similar in cohort 1 vs 2 (0.98 events/pt. vs. 1.03 events/pt. P=0.25). The use of non-invasive cell-free DNA testing with home spirometry screening appears to yield comparable medium-term allograft outcomes, as compared to conventional clinic-based PFTs and surveillance bronchoscopy. Further multi-center studies are warranted to determine if this approach can reduce invasive procedures while maintaining equivalent outcomes in clinically meaningful parameters. [ABSTRACT FROM AUTHOR]

Prior observational data suggest that donor-derived cell-free DNA (dd-cfDNA) increases in lung transplant acute rejection and infection. The performance of dd-cfDNA in routine clinical care remains undefined. In response to the COVID-19 pandemic, to mitigate the risk of exposing patients to infection, four centers used dd-cfDNA for surveillance instead of surveillance bronchoscopy, providing a unique opportunity to assess the performance of dd-cfDNA in routine clinical care. As part of routine care during the COVID-19 pandemic, four lung transplant centers implemented a home-based surveillance program using plasma dd-cfDNA (Allosure®) in preference to surveillance bronchoscopy. Based on prior data, dd-cfDNA > 1% triggered further work-up including bronchoscopy. dd-cfDNA testing was also performed in response to a decline in forced expiratory volume in 1 second (FEV1), symptoms or treatment follow up. Data was retrospectively analyzed from 4/1/2020 - 9/1/2020 to assess the performance of dd-cfDNA in diagnosing a composite of ACR, AMR and/or infection. 169 patients underwent 380 dd-cfDNA measurements over the study period. The mean age was 58.5 years, 54% of patients were male and 82% bilateral lung transplants. 99 (58%) patients were <1 year post-transplant. 327 of 380 dd-cfDNA values were drawn for surveillance reasons. 31 patients had a surveillance level > 1%. Of these, 19/31 (61%) had evidence of ACR, AMR or infection. 115 patients had surveillance levels that remained < 1% over the study period with 109/115 (95%) displaying no clinical evidence of ACR, AMR, infection or decline in FEV1 or symptoms. The remaining 23 patients had levels drawn for clinical indications (non-surveillance). 45 surveillance bronchoscopies were performed with concomitant dd-cfDNA (23 triggered by dd-cfDNA > 1%). For diagnosis of ACR, AMR or infection in these patients, dd-cfDNA > 1% yielded a sensitivity of 84%, specificity of 77%, positive predictive value of 73% and negative predictive value of 87%. In this study, dd-cfDNA identified ACR, AMR and/or infection in asymptomatic lung transplant patients that may not have been identified by clinically indicated biopsy alone. Low levels of dd-cfDNA may also be useful in ruling out AMR, ACR and/or infection, supporting its use as a potential non-invasive marker for surveillance monitoring. [ABSTRACT FROM AUTHOR]

Patients with COVID-19 show variable clinical course; transplant patients often show worse outcomes. The effect of COVID-19 on the allograft and the sources of tissue injury that contribute to such poor outcomes are poorly defined. This study leverages cell-free DNA (cfDNA) to measure allograft injury as donor-derived cfDNA (ddcfDNA) and injury from different tissue types using tissue-specific DNA methylomic signatures. 14 consecutive COVID-19 transplant patients (8 Kidney, 3 Lung, 1 Heart, 1 Liver, and one multi-organ transplant patients) and 30 healthy controls were included. Plasma nuclear cfDNA (ncfDNA) and mitochondrial cfDNA (mtcfDNA) level were measured via digital droplet PCR, and ddcfDNA using AlloSure (CareDx). cfDNA whole-genome bisulfite sequencing was performed to identify cfDNA tissues of origin leveraging tissue specific DNA methylomes and deconvolution algorithm. 75% of the COVID-19 transplant patients showed high ddcfDNA level compared to published quiescent values, including all lung, 50% of the kidney, liver and multi-organ transplant patients (8.5, 4.4, 30 and 16-X fold change, respectively). Total ncfDNA and mtcfDNA were 15X and 310X higher in COVID-19 transplant patients compared to controls, respectively; < 0.0001.The predominant tissues contributing to cfDNA were hematopoietic cells (80%) (Figure). More importantly, COVID-19 transplant patients showed 10 to 100 fold higher tissue specific cfDNA derived from monocyte, neutrophil, erythroblast, vascular endothelium, adipocyte, hepatocyte, kidney, heart and lung compared to controls. Analysis comparing cfDNA in transplant and non-transplant COVID-19 patients is on-going. The allograft undergoes significant injury following COVID-19. Further, cfDNA from multiple tissue types is significantly higher in COVID-19 transplant patients. Future studies in a larger cohorts of transplant and non-transplant patients are needed to elucidate why transplant patients show worse COVID-19 outcomes. [ABSTRACT FROM AUTHOR]