Outcomes of haploidentical transplantation with umbilical cord blood co-infusion in aplastic anemia: Clinical and single-cell insights

Abstract Aplastic anemia (AA) is a life-threatening disorder characterized by pancytopenia and bone marrow failure. Hematopoietic stem cell transplantation (HSCT) from matched sibling or unrelated donors remains the standard curative approach, but suitable donors are not always available. Haploidentical HSCT (haplo-HSCT) provides an alternative but carries increased risks of graft failure and graft-versus-host disease (GVHD). Umbilical cord blood (UCB), with its immunologic naivety, may reduce GVHD incidence, though its limited cell dose may delay engraftment. Combining haplo-HSCT with UCB infusion has been proposed to harness the benefits of both sources. However, clinical evidence in AA is limited. We retrospectively analyzed outcomes of 62 patients with severe AA (SAA) or transfusion-dependent non-severe AA (TD-NSAA) who underwent haplo-HSCT between May 2018 and September 2024. Thirty-one patients received a single unit of unrelated UCB infused 4–6 hours before haplo-HSCT (UCB group); 31 underwent haplo-HSCT alone (non-UCB group). Outcomes included engraftment, graft failure, GVHD, and overall survival (OS). Baseline characteristics were generally balanced. FAC conditioning was more common in the non-UCB group (90.3% vs. 53.3%), while TBF was more frequently used in the UCB group (33.3%). In the UCB group, the median total nucleated and CD34⁺ cell doses from cord blood grafts were 14.92×10⁸ and 5.67×10⁶, respectively. Single-cell RNA sequencing (scRNA-seq, BD Rhapsody) was performed on bone marrow samples from 6 patients (3 per group) at pre-HSCT and engraftment. Data were analyzed using STARsolo and Seurat (Wilcoxon test, log2FC >0.25, p<0.05, min.pct >0.1). One patient in the UCB group died before engraftment. Neutrophil graft failure occurred in 3 patients (10%) in the UCB group and 4 (12.9%) in the non-UCB group. Median neutrophil engraftment was 13 days (range: 9–22) in the UCB group and 12 days (range: 8–21) in the non-UCB group (p=0.10). Platelet graft failure was observed in 3 (10.0%) vs. 4 (12.9%) patients, respectively. Median platelet engraftment was 15 days (range: 9–24) vs. 13 days (range: 6–70) (p=0.64). The cumulative incidence of grade II–IV acute GVHD was 10.35% in the UCB group and 12.9% in the non-UCB group (p=0.76), and 2-year chronic GVHD incidence was 20.0% vs. 18.9% (p=0.99). The 2-year OS rates were 81.02% and 80.65%, respectively (p=0.94). ScRNA-seq revealed that post-HSCT, hematopoietic stem/progenitor cells (HSPCs), neutrophils, and macrophages increased, while monocytes and T cells (CD4⁺, CD8⁺) decreased. T cell reclustering showed expansion of effector CD4⁺/CD8⁺ T cells, particularly in cord blood recipients. These patients also demonstrated increased naïve and central memory CD8⁺ T cells. Expression of cytotoxic genes (PRF1, GZMA) was upregulated and resting-associated genes (LEF1, CCR7) were downregulated in T/NK cells, especially in the UCB group. Notably, CXCR4 expression—crucial for stem cell homing and bone marrow adhesion—was elevated in UCB recipients, while GVHD-related gene TNFRSF1A was reduced. Haplo-HSCT combined with a single unit of unrelated UCB is a feasible and effective treatment for patients with SAA and TD-NSAA, with engraftment rates, GVHD incidence, and survival outcomes comparable to haplo-HSCT alone. scRNA-seq analyses suggest that UCB may enhance immune reconstitution by promoting cytotoxic T/NK phenotypes and improving stem cell homing via CXCR4 upregulation, while potentially mitigating GVHD through TNFRSF1A downregulation. These findings provide mechanistic insights supporting the use of UCB in haploidentical transplantation for AA. Further prospective studies are warranted to confirm these results.