Endothelial Dysfunction and Risk of Acute Graft-Versus-Host Disease after Allogeneic Stem Cell Transplantation: END-GAME Prospective Study Final Results

Background. Endothelial dysfunction has been increasingly correlated to the very early phases of acute graft-versus-host-disease (aGVHD) with endothelial cells acting both as starting elements of the inflammatory process and as a target of alloreactive donor T cells (“endothelial GVHD”).

Aim and Methods. This observational prospective single-center study investigates the role of endothelial dysfunction at the time of allogeneic hematopoietic stem cell transplantation (HSCT) in predicting the incidence of aGVHD. Inclusion criteria were: adult patients (≥18 years old) undergoing their first HSCT for any hematological disease, from any donor and any stem cell source. The primary endpoint was cumulative incidence of aGVHD at day 100 post-HSCT. Levels of circulating endothelial cells (CECs), circulating endothelial progenitors (CEPs) and some soluble biomarkers (plasminogen activator inhibitor-1 [PAI-1], vascular endothelial growth factor [VEGF-A], angiopoietin-2 [ANG-2], soluble VCAM [sVCAM-1]) were assessed on peripheral blood samples collected on the day of HSCT (T0).

CECs and CEPs were identified by flow cytometry according to Lanuti et al. 2018 and Mancuso et al. 2020 protocol, respectively. CECs >30 cells/ml were considered elevated. Since a normal reference range for CEPs has not yet been established, they were divided into tertiles. Soluble biomarkers were analyzed using the automated microfluidic analyzer ELLA (BioTechne, Minneapolis, USA), according to the manufacturer's protocol and results expressed by pg/ml. Statistical analysis was performed in accordance with the EBMT recommendations. As this is a proof-of-concept study, p-values <0.2 were considered significant.

Results. A total of 51 patients were recruited from August 2021 to July 2023. The clinical and transplant characteristics of the patients were as follows: median age at HSCT was 54 years (interquartile range [IQR], 39-64); 29 patients (57%) were female; 31 patients (61%) had acute leukemia; 14 patients (27%) had a high/very-high disease-risk index (DRI) and 16 patients (31%) had a high hematopoietic cell transplantation-specific comorbidity index (HCT-CI). About transplant features, the donors were matched related donors for 4 patients (8%), mismatched related donors for 12 patients (23%), matched unrelated donors for 19 patients (37%) and mismatched unrelated donors for 16 patients (31%); stem cell sources were peripheral blood stem cells for 47 patients (92%) and bone marrow stem cells for 4 patients (8%); 3 transplants (6%) involved female donors to male recipients. Myeloablative conditioning was used in 44 cases (86%). GVHD prophylaxis consisted of post-transplant cyclophosphamide (PTCy) in 24 patients (47%), anti-thymocyte globulin (ATG)-based prophylaxis in 25 patients (49%) and methotrexate plus calcineurin inhibitors in 2 cases (4%).

After a median follow-up of 23 months (IQR, 18-28), a total of 29 (57%) patients developed acute GVHD (14/51, 27% grade 2 or higher), for a 100-day cumulative incidence of aGVHD of 47% (95% confidence interval [95%CI], 33-60%). Two-year cumulative incidence of chronic GVHD, non-relapse mortality (NRM), cumulative incidence of relapse (CIR) and overall survival (OS) were: 27% (95%CI, 15-40%), 15% (95%CI, 6-26%), 10% (95%CI, 4-20%) and 78% (95%CI, 63-87%).

Median CECs at the time of HSCT (T0) were 46 cells/ml (IQR, 28-80 cells/ml). In univariable analysis, CECs >30 cells/ml (n=36, 71%) were found to be associated with higher risk of aGVHD of any grade (SHR=3.0, 95%CI 1.0-9.0, p=0.053). Conversely, CEPs did not associate with cumulative incidence of aGVHD (p=0.631).

Among soluble endothelial biomarkers at T0, VEGF-A (low/normal/high 94%/6%/0%), ANGP2 (low/normal/high 62%/20%/18%), sVCAM-1 (low/normal/high 34%/48%/18%) showed no association with cumulative incidence of aGVHD at univariable analysis (all p>0.2). PAI-1 was not evaluable, as all the patients presented values below the range of reference.

In multivariable analysis, increased CECs at T0 were confirmed as independent risk factor for aGVHD (SHR 2.5, 95%CI 0.8-8.2, p=0.120).

Conclusion. Overall, our findings suggest that early endothelial damage may contribute to the risk of aGVHD. Identifying endothelial stress at the time of transplantation could therefore guide the selection and adjustment of GVHD prophylaxis. Larger study cohorts are warranted to confirm our preliminary results.