A Prospective Study of Tocilizumab for the Prevention of Graft Failure and Graft-Versus-Host Disease in Haplo-Cord Transplantation, As a Potential Alternative to Anti-Thymocyte Globulin

Introduction Haplo-Cord transplantation (HCT), the co-infusion of an umbilical cord blood (UCB) graft with CD34-selected cells from a haplo-identical adult donor, provides an alternate source of stem cells for patients who lack a suitable HLA-matched donor. Traditionally, anti-thymocyte globulin (ATG) is administered prior to HCT, to reduce the incidence of graft failure, graft-versus-host disease (GVHD) and graft versus graft reactions. However, excessive ATG-mediation of cord lymphocytes may compromise immune reconstitution (IR) leading to increased infections, and relapse due to impaired graft versus tumor effects. In this Phase II study (NCT04395222), we assess the safety of reducing ATG in our haplo-cord platform while using tocilizumab as an alternative immunomodulatory agent. Methods Single arm, Bayesian study design. Conditioning was with fludarabine, melphalan and total body irradiation, followed by tocilizumab 8 mg/kg on day-1. Accrual was planned to four successive cohorts (C), each with up to 10 subjects. In the first cohort (C1), patients received the standard 3 doses of ATG 1.5 mg/kg (total dose 4.5 mg/kg). The primary endpoint was successful haplo-derived neutrophil engraftment by day 21. In the absence of safety signals and ≥6 out of 10 patients achieving outcome, one dose of ATG was dropped in successive cohorts until eliminated. Results Between 10/2020-9/2022, 21 patients were enrolled (Table). 8/10 patients in C1 and 6/10 pts in C2 (ATG 3 mg/kg) met the primary endpoint. In C3 (ATG 1.5 mg/kg), the first patient did not meet outcome, prompting study halt due to safety concerns. All patients engrafted. The median time to ANC and platelet engraftment was 16d (14-17d) & 18d (16-25d) in C1 and 26d (18-33d) & 30d (19-46d) for C2. The patient in C3 had ANC and platelet engraftment on days 34 and 67. Of patients not meeting the primary endpoint; 2/2 patients in C1 had haplo-derived CD33 chimerism detected, however only 1/4 in C2 did. The remaining 3 along with the single patient from C3 had only cord engraftment. The IR, measured by median CD4 count at day 100, was better with reduced ATG - C2 vs C1 (144 vs 47 cells/ul (p=0.002). The 1y CI of Relapse was 50% in C1 vs 20% in C2 (p=0.15), with NRM 20% for C1 vs 10% (p=0.65) for C2. The 1y OS for C1 was 50% (95% CI 47-100%) and 90% (95% CI 73-100%) for C2 (p=0.028). Conclusion Decreasing the cumulative dose of ATG in our haplo-cord platform positively impacted CD4 IR post-transplant. However, lower ATG doses increased the risk of early rejection of the haploidentical graft, leading to myeloid bridge failure and prolonged myelosuppression. Tocilizumab did not provide additional immune-modulation to prevent haplo-rejection and also delayed engraftment by a median of 2-3d compared to our standard ATG regimen. This study was halted for futility. Future studies to optimize early engraftment without affecting IR are needed.