Introduction Targeted molecular profiling combined with reproducible formalin-fixed paraffin-embedded (FFPE) EMB-based technology has the potential to support cardiac rejection diagnosis. Objective We aimed to develop and validate targeted gene expression diagnostic models of cardiac rejection and show their association with ISHLT pathological grades. Method We built a multicenter cohort of 591 FFPE-EMBs collected from four international centers between 2011 and 2021. Biopsies were graded according to the ISHLT working formulations, including 188 AMR cases, 289 ACR cases, and 114 non-rejection cases, randomly split in a derivation (n=475) and a validation cohort (n=116). Tissue gene expression was analyzed on FFPE-EMB using the Banff Human Organ Transplant gene set. Molecular classifiers for AMR and ACR were built using a supervised model. Association between molecular scores and pathology severity of rejection were analyzed in both derivation and validation set. Results The derivation included a total of 151 AMR (pAMR1H+: n=46, pAMR1I+: n=36, pAMR2-3: n=69), 232 ACR (ACR 1R: n=143, ACR 2-3R: n=89) and 92 non-rejection cases. The validation set included a total of 37 AMR (pAMR1H+: n=12, pAMR1I+: n=15, pAMR2-3: n=10), 57 ACR (ACR 1R: n=31, ACR 2-3R: n=26) and 22 non-rejection cases. Median AMR scores in AMR, ACR and non-rejection cases were 0.655 (IQR=0.316), 0.217 (IQR=0.254) and 0.140 (IQR=0.209), respectively. Median ACR scores in ACR, AMR and non-rejection cases were 0.678 (IQR=0.347), 0.263 (IQR=0.364) and 0.302 (IQR=270), respectively. AMR and ACR molecular scores were strongly associated with the pathology assessment of severity of rejection according to AMR and ACR international working formulations, respectively (derivation set: ACR: P for trend=1.017E-46, AMR: P for trend=2.911E-52; validation set: ACR: P for trend=1.827E-13, AMR: P for trend=5.174E-10, Fig. 1). Molecular AMR score was not associated with ACR severity; neither was the ACR molecular score with AMR severity. Conclusion Tissue-based molecular diagnostic system developed closely aligned with histological grading of cardiac allograft rejection, enhancing diagnostic precision, and offering a reliable companion tool for routine practice.

Purpose To analyze the differences between risk factors for acute cellular (ACR) and antibody-mediated rejections (AMR) in a large cohort of highly phenotyped heart-transplant recipients. Methods We performed a multicenter, retrospective, observational study (Pitié-Salpêtrière, HEGP, 2004-2016, n=1250 patients). We included all patients with at least one endomyocardial biopsy (EMB): n=1053 patients, 13677 EMB<1-year post-transplant. All EMB performed before 2012 were retrospectively reviewed and graded according to ISHLT guidelines. ACR were defined as ≥ 1R1B (threshold to start acute rejection treatments at our centers) and AMR as pAMR≥1. Pre-formed DSA were evaluated using Luminex Single Antigen Assay (retrospective assessment for patients transplanted before 2009; threshold: MFI ≥ 1000). Risk factors for rejection were identified using Cox proportional hazard model. Results All patients received induction therapy (ATG = 92%). Combined transplantations were performed in 53 patients and 324 were transplanted with pre-formed DSA (31.1%). ACR were diagnosed in 458 patients (43.5%, 896 EMB) and AMR in 82 patients (7.8%, 138 EMB, figure). Risk factors for ACR included recipient's age (HR=0.990 per 1-year increment, p=0.005), type of transplantation (combined heart-kidney and heart-liver compared to isolated heart: HR = 0.243 and 0.161 respectively, p < 0.001), pre-formed DSA (HR=1.250, p=0.03), HLA A-B-DR mismatches (HR=1.130 per 1-mismatch increment, p=0.02) and the type of induction (basiliximab compared to ATG: HR=1.887, p<0.001). Risk factors for AMR included recipient's age (HR=0.976 per 1-year increment, p=0.003), pre-formed DSA (HR=1.971, p=0.02) and ischemic time (HR=1.04 per 10-minute increment, p=0.02). Conclusion We found 6 factors independently associated with the risk of ACR or AMR. Two factors were common between ACR and AMR (recipient's age, pre-formed DSA). These findings might help clinician to customize post-transplant follow-up.

Purpose To identify distinct profiles of cardiac allograft vasculopathy (CAV) trajectory and their determinants. Methods We performed a multicenter study (Pitié & HEGP, Paris). We included all patients transplanted between 2004 and 2011 who had at least 2 coronary angiograms during follow-up (n = 436). We performed an comprehensive evaluation of potential risk factors that might be associated with CAV including in-depth characterization of donor-specific antibodies (DSA) and retrospective endomyocardial biopsies reviewing. A semi parametric mixture model was used to identify distinct trajectories of CAV progression from year 1 to year 10 post-transplantation among the 1407 coronary angiograms (3.4±1.4 angiograms per patient). Multivariable logistic regression model was used to determine risk factors associated with each trajectory. Results Median follow-up post-transplant was 7 years. Four latent classes of CAV trajectories were identified: #1-non-progressors (n=240, 55%), #2- late progressors (n=35, 8%), #3- early progressors (n=99, 22.7%) and #4- severe progressors (n = 62, 14.3%, figure). Taking latent class #1 as reference, 5 parameters were associated with CAV progression: donor age (HR = 1.06 and 1.08 per 1-yr increment for latent classes #3 and #4, respectively, p < 0.001), donor tobacco consumption (HR = 2.49, 2.02 and 2.23 for classes #2, #3 and #4, respectively, p < 0.001), donor gender (male donors: HR = 2.3 and 3.94 for classes #3 and #4, respectively, p < 0.001), 1-year post-transplant dyslipidemia (HR = 2.21 for class #3, p = 0.02) and pre-formed DSA (HR = 3.85 for class #3). Conclusion In a large and highly phenotyped cohort of heart transplant recipients, we identified 4 distinct prototypes of CAV trajectories. Donor-derived cardiovascular risk factors, 1-year post-transplant cardiovascular profile (dyslipidemia) and immunological parameters (pre-formed DSA) were independent risk factors for CAV severity and progression.

Purpose To analyze the incidence and the risk factors for biopsy-proven rejections during the first year following heart transplantation in a large cohort of highly phenotyped heart-transplant recipients. Methods We performed a multicenter, retrospective, observational study (Pitie-Salpetriere, HEGP, 2004-2016, n=1250 patients). We included all patients with at least one endomyocardial biopsy (EMB): n=1053 patients, 13677 EMB Results All patients received induction therapy, mostly with ATG (92%). Cyclosporine was the most prescribed calcineurin inhibitor (92%). Combined transplantations were performed in 53 patients (heart-kidney: n=30; heart-liver: n=23). Almost one-third of patients were transplanted with pre-formed DSA (n=324; 31.1%). Biopsy-proven rejections were diagnosed in 489 patients (46.4%) representing 1,009 EMB (7.3%, mostly low-grade cellular rejections: 1R1B or 1R2: n=724). Risk factors for rejection included recipient's age (HR=0.988 per 1-year increment, p=0.002), type of transplantation (combined heart-kidney and heart-liver compared to isolated heart: HR = 0.336 and 0.204 respectively, p Conclusion Recipient age, type of transplantation (combined vs non-combined), pre-formed DSA, number of HLA mismatches and the type of induction were significantly associated with the risk of allograft rejection during the first-year post heart transplantation. These findings might help clinician to customize immunosuppression regimen and post-transplant follow-up.

Gene expression analysis of allograft tissue have been increasingly recognized as an interesting tool for the diagnosis of rejection as a companion to histopathology. This study evaluates in heart transplantation the diagnostic performance of Reverse Transcriptase-Multiplex Ligase-dependent Probe Amplification (RT-MLPA), a new technique of targeted gene expression analysis, suitable for formalin-fixed paraffin embedded (FFPE) tissue.

Background: Diagnosis criteria of acute humoral rejection in small bowel transplantation (SBT) are not clearly defined, although the presence of preformed or de novo DSA has been reported to be deleterious for SBT survival. Methods: We retrospectively studied all intestinal biopsies obtained in the first year of transplantation from our cohort of SBT patients (n=23) between May, 2009 and November, 2014. We systematically looked for C4d staining, semi-quantitatively assessed according to the Banff 2007 classification, and for signs of cellular rejection (apoptosis), vascular lesions (capillaritis, thrombosis, hemorrage congestion), oedema of the chorion, lamina propria inflammation and mucous ulcerations. Identification of anti-HLA DSAs and their ability to fix C1q was performed by Luminex Single Antigen. Results: We assessed 345 biopsies (17±6 biopsies per patient) from these 23 patients. Among these patients, 3 did not develop DSAs. From the remaining 20 DSA+ patients, 7 (35%) had a C1q-binding DSA. 78 biopsies (22.6%) were C4d+ (grade≥2). Multivariate logistic regression analysis revealed that 3 histological parameters were independently and significatively associated with the presence of C4d on the biopsy: capillaritis (OR 2.883, p=0.003), mucosal ulceration (OR 3.429, p=0.003) and presence of apoptosis (OR 1.957, p=0.036). Patients with less than 15% of C4d+ biopsies had a better graft survival (86% at 2 years) compared to patients with 15 to 30% of C4d+ biopsies (53% at 2 years) and to patients with more than 30% of C4d+ biopsies (18% at 2 years, log-rank test p=0.0015). Conclusion: The presence of C4d positive biopsies (grade≥2) in the first year of SBT is significantly associated with capillaritis, ulceration and apoptosis, as well as a worse graft survival at 2 years. These histological signs are relevant for acute humoral rejection diagnosis in SBT.

Antibody-mediated rejection (AMR) is an important mechanism of graft loss in cardiac transplantation. Endothelial cells (EC), the main target of antibodies, are activated in AMR and in turns recruit blood mononuclear cells in the microcirculation. The aim of the study was to characterize the EC changes in a model of AMR in which the blood monocytes are not engaged to better analyze the complex pathophysiology of cardiac AMR.