Haploidentical Hematopoietic Cell Transplant with Post-Transplant Cyclophosphamide Expedites Donor-Derived Immune Reconstitution in a Pediatric Severe Aplastic Anemia Cohort

Severe aplastic anemia (SAA) is a life-threatening, idiopathic process involving the immune-mediated destruction of hematopoietic stem cells. Treatment involves immunosuppressive therapy (IST) and/or allogeneic hematopoietic cell transplantation (HCT). The latter has recently been optimized with use of a post-transplant cyclophosphamide (PTCy) based graft-versus-host disease (GvHD) platform, permitting successful HCT with only half-HLA-matched haploidentical donors. In pediatrics, this platform results in >90% overall survival with low rates of GvHD and graft failure. However, reports on donor chimerism and immune reconstitution are limited. We hypothesized that a PTCy platform would expedite CD3 donor engraftment and CD4 T cell recovery.

In a phase III clinical trial, we treated three cohorts of pediatric (age </=25 years) patients with SAA requiring alloHCT: MUD, MSD, and HAPLO. The MUD cohort included n=15 recipients of 7-8/8 HLA-matched unrelated donor grafts (n=5 7/8, n=10 8/8) with median age 14 years (range 2-17 years), following cyclophosphamide (Cy) 50 mg/kg, fludarabine (Flu) 200 mg/m ², rabbit anti-thymocyte globulin (rATG) 9 mg/kg total dose rabbit and low dose total body irradiation (TBI) at 200 cGy. The MSD cohort included n=9 recipients of 8/8 HLA-matched sibling donor grafts with median age 15 years (range 12-24 years), following Cy 200 mg/kg and 9 mg/kg rATG. The HAPLO cohort included n=4 recipients of 4/8 HLA-matched related donor grafts with median age 8.5 years (range 2-21 years) following Cy 29 mg/kg, Flu 150 mg/m ², 9 mg/kg rATG, and low dose TBI at 200 cGy. GvHD prophylaxis consisted of 30 days of mycophenolate mofetil, and 100 days of a calcineurin inhibitor (followed by an 8-10 week taper), with the HAPLO cohort additionally receiving post-transplant Cy (PTCy) 50 mg/kg/day on days +3 and +4. Peripheral blood lymphocyte counts and peripheral blood and bone marrow donor chimerism were evaluated on days +100, +180 and +365.

Recipient characteristics were largely equivalent between cohorts, with 43% receiving IST prior to HCT. Clonality was overrepresented in the HAPLO cohort (2 of 4, 1 with monosomy 7 myelodysplastic syndrome and 1 with a large PNH clone), with only 2 of 15 in the MUD and none in the MSD cohorts. Active controlled infections at the time of HCT were additionally overrepresented in the HAPLO cohort, with disseminated pseudomonas in 1. The MUD cohort had 1 with Rhizopus invasive sinusitis (none in the MSD cohort). Achievement of neutrophil engraftment was statistically significantly delayed in the HAPLO cohort at a median of 17 days compared to 12 for MSD and 14 for MUD, p=0.004), however there were no episodes of primary or secondary graft failure. With only a single late death from intracranial hemorrhage in the MUD cohort, OS was equivalent at 2 years (100% in MSD and HAPLO cohorts, 93.3% in MUD). GvHD rates were low with only 1 (MUD) experiencing grade II acute GvHD, and 1 (MUD) experiencing mild chronic GvHD.

Immune reconstitution varied by cohort with statistically significantly greater CD3 peripheral blood donor chimerism (p<0.001) as well as CD4 (p=0.04) and CD8 (p=0.03) T cell counts over time for the HAPLO recipients (CD3 chimerism in Fig 1, CD4 count in Fig 2; 2-way ANOVA fixed effects model). Considering a CD4 >200 cell/uL as clinically meaningful, we assessed the proportion of each cohort achieving this milestone at each timepoint. While each cohort had an increase in proportion achieving CD4>200 over time, the HAPLO cohort reached 50% with CD4 >200 at day +100, compared to 11% MUD and 0% MSD (p=0.09). At day +180, the proportion with CD4 >200 were 36% MUD, 25% MSD and 75% HAPLO (1 yr: 90% MUD, 100% MSD and HAPLO).

Donor availability remains a challenge to successful transplant for patients with SAA, and haploidentical matches overcome donor limited accessibility. PTCy as GvHD prophylaxis is associated with low incidences of graft failure and severe GvHD. While PTCy might be anticipated to inhibit donor T cells and immune reconstitution, our data, albeit limited by small cohorts, demonstrates superior and more efficient CD3 donor engraftment and CD4 T cell count recovery following HAPLO HCT with PTCy compared to non-PTCy based MUD and MSD cohorts.