Follistatin and Endoglin: Potential Biomarkers of Endothelial Damage and Non-Relapse Mortality after Myeloablative Allogeneic Hematopoietic Cell Transplantation in Blood and Marrow Transplant Clinical Trials Network (BMT CTN) 0402

Abstract

Introduction:Inflammation and angiogenesis are linked biologic processes influencing the onset and recovery of acute graft-versus-host disease (aGVHD) after allogeneic hematopoietic cell transplantation (HCT). In aGVHD, angiogenic factors (AF) contributing to tissue healing/repair (e.g. epidermal growth factor [EGF]) versus AF associated with tissue damage/inflammation (e.g. follistatin [FS]) exist in a dysregulated balance. Whether these AF are associated with, or predictive of, non-relapse mortality (NRM) in the absence of aGVHD, is not understood. Given that regimen-related tissue injury plays a role in NRM, we hypothesized that elevated levels of inflammation-associated AF early after HCT would be associated with NRM even in patients without aGVHD.Methods:We studied plasma samples from patients enrolled on Blood and Marrow Transplant Clinical Trials Network (BMT CTN) aGVHD prophylaxis study 0402, a randomized trial of tacrolimus/methotrexate vs tacrolimus/sirolimus after matched sibling peripheral blood stem cell transplantation with myeloablative conditioning. Levels of AF (angiopoietin-2 [Ang2], EGF, FS, vascular endothelial growth factor [VEGF]-A and -B, endoglin [sEng], placental growth factor [PlGF], and soluble VEGF receptor [sVEGFR]-1 and -2) were quantified by MILLIPLEX magnetic bead panels.We analyzed factors associated with NRM in univariate analysis and used multivariate analysis to develop composite AF scores to predict NRM. In calculating the composite scores, we extrapolated the individual probabilities of NRM based upon regression models, and sorted the individual probabilities into 3 composite scores based on: 1) estimated probability of NRM <10%, 2) probability of NRM between 10-19%, 3) probability of NRM >20%.Results:221 patients from BMT CTN 0402 had pre-transplant plasma samples and samples at day +28 available for analysis. Univariate analysis showed associations of raw levels of Ang2, FS, sEng, and sVEGFR1 on day +28 with NRM. Of these, only higher levels of log-transformed FS (p=0.01) and of sEng (p=0.04) showed a correlation with NRM; these two AF were selected for the final model. In multivariate analysis, patients with the highest levels of day +28 FS and sEng (score 3, n=32, Figure) had a 4.6-fold higher relative risk (RR) of NRM (95% confidence interval 1.7-12.3, p<0.01, Table). The prognostic value of the composite score was stronger than either individual factor alone (RR log[FS]=2.1, p=0.03, and RR log[sEng]=2.3, p=0.07). Grade III-IV aGVHD and patient age>50 years were also independently associated with 1-year NRM. There was no effect of patient gender, disease status, CMV serostatus, or type of GVHD prophylaxis (tacrolimus/methotrexate versus tacrolimus/sirolimus) on NRM. There was no significant effect of the composite score, grade III-IV aGVHD, or age on the competing risk of relapse. FS and sEng measured before HCT were not associated with 1-year NRM (p=0.91 and p=0.62, respectively). There was only a weak correlation between the day 0 and day 28 FS and sEng (Spearman’s rho 0.36 and 0.38, respectively). Among the 221 cohort patients, there were 25 deaths due to NRM, including 5 deaths occurring prior to day 56. Death was more likely to be from organ toxicity in patients with composite score 3 (5/9, 56%) vs those with a composite score of 1 (1/8, 12%) or 2 (3/8, 37%).Conclusions:A composite score using day 28 plasma levels of FS and sEng predicts the risk of 1-year NRM after HCT, providing further evidence that endothelial dysfunction is a relevant biological problem in HCT. Pre-HCT measurements of FS and sEng were not associated with NRM, suggesting that elevations in these factors may reflect susceptibility to regimen-related toxicity. We have previously shown that elevated levels of FS and sEng at the onset of aGVHD are associated with poor survival. Here, we show that these factors are predictive of NRM independent of aGVHD. Follistatin modulates inflammatory responses following tissue injury via interaction with the transforming growth factor β superfamily and may reflect endothelial cell damage. sEng is upregulated in proliferating endothelial cells, in inflammation and tissue ischemia, and can influence leukocyte adhesion and endothelial transmigration. Measurement of these predictive AF may highlight prospective opportunities to intervene and prevent subsequent complications thereby limiting NRM after allogeneic HCT. [Display omitted] [Display omitted] DisclosuresNo relevant conflicts of interest to declare.