CD45RA-Depleted Haploidentical Transplantation Combined with NK Cell Addback Results in Promising Long-Term Outcomes in Pediatric Patients with High-Risk Hematologic Malignancies

Novel graft engineering strategies that allow selective T-cell depletion have made haploidentical donor (haplo) hematopoietic cell transplantation (HCT) feasible for patients lacking a matched donor. We hypothesized that selective CD45RA+ cell depletion can limit graft-versus-host disease (GVHD) through removal of naive T cells but allow for rapid immune reconstitution through adoptive transfer of memory T cells, minimizing risk of graft failure, infections and relapse. We performed a prospective clinical trial utilizing CD45RA+ depleted haplo HCT followed by donor NK cell addback to maximize the graft-versus-leukemia effect in children with high-risk hematological malignancies (NCT01807611), and present here an interim analysis.

Between 2013 to 2019, 72 patients (42 males, 30 females) were enrolled. The median age was 8.1 years (range 0.6-20.8). Twenty-nine patients had acute lymphoblastic leukemia (ALL), 39 had myelodysplastic syndrome/acute myeloid leukemia (MDS/AML), 3 had biphenotypic leukemia and 1 had Non-Hodgkin's Lymphoma. At time of HCT, 25 patients were in CR1, 24 in CR2, 4 in CR3/> and 19 had active disease. Of the 53 patients in CR, 22 had evidence of minimal residual/detectable disease (MRD) at HCT. Donors used were mothers (n=30), fathers (n=35), or sibling/other (n=6) with majority (n=62) being KIR-ligand mismatched to the recipient. All patients received serotherapy-free, reduced intensity conditioning with fludarabine, melphalan, cyclophosphamide and Total Lymphoid Irradiation. GCSF mobilized peripheral blood grafts were infused on i) Day 0 CD34 + selected graft (median CD34+: 9.85 x10 ⁶cells/kg, range 1.96-44.64) and ii) Day +1 CD45 RA+ depleted graft (median: CD34+: 5.82 x10 ⁶cells/kg, range 0.58-39.43); providing a large number of CD45RO+ T-cells (median CD3+ : 60.1 x10 ⁶cells/kg, range 16.08-528.43) after depletion of CD45RA+ cells (CD3+CD45RA+ median 0, range 0-0.2 x10 ⁶cells/kg). NK cells (median: 11.7 x10 ⁶cells/kg; range: 1.65-99.2), isolated from a non-mobilized apheresis by CD3 depletion/CD56 selection, were infused on day +6. GVHD prophylaxis included MMF (n=61) and/or Sirolimus (n= 8) until day +42 post-HCT.

The median time for neutrophil and platelet engraftment was 11 (range 9-13) and 17 days (range 10-100) respectively. One patient had primary graft failure and was successfully re-transplanted. Immune reconstitution was rapid and robust and recapitulated CD45RA depleted graft content with mean CD3, CD4 and CD8 T-cell count at day +30 of 1080, 640 and 230 cells/ul respectively. TCR Vbeta spectratyping revealed a broad repertoire by day +100 prior to the emergence of naïve T cells. The incidence of Cytomegalovirus and Adenovirus viremia was 38% (28/72) and 13% (10/72) respectively. The cumulative incidence of aGVHD was 36.1% (25.1-47.2%), aGVHD grade III-IV was 29.2% (19.1-39.9%) and chronic GVHD was 20.8% (CI: 12.3-30.9%). With a median follow-up of 3.1 years (range 0.04-8.0 years) the 3-year leukemia free survival (LFS) was 88% (CI: 76.1-100%) for patients in CR1, 70.8% (CI: 54.8-91.6%) for patients in CR2, 50% (CI: 18.8-100%) for patients in CR3/> and 21.1% (CI: 8.81-50.3) for patients with active disease (p = <0.0001). Amongst patients in CR, those with no detectible disease had 3-year LFS of 84.4% (CI: 72.7-97.9%) as compared to 68.2% (CI: 51.3-90.7%) for patients with detectible MRD at time of HCT (p=0.087). Three-year LFS was comparable for ALL and AML at 62.1% (CI 46.7-82.5%) and 58.5% (CI 44.8-76.4%) respectively (p=0.95). For the entire cohort, the 3-year OS and LFS was 68.9% and 62.2% (CI 58.9-80.6% and 51.9-74.6%), respectively. Despite the inclusion of patients with active disease and detectible MRD the cumulative incidence of relapse for the entire cohort was 26.5% (CI:16.8-37.1%), comparable to other studies with patients transplanted in MRD-negative CR. The cumulative incidence of non-relapse mortality was 11.5%, (CI, 5.3-20.4%), with 5/10 deaths occurring in the 19 patients with active disease.

In conclusion, selective CD45RA+ depletion allowed for adoptive transfer of abundant memory T cells that along with NK cell addback was associated with rapid, functional immune reconstitution, resulting in low rates of graft failure, viral reactivation and relapse. Despite a higher incidence of aGVHD compared to other T-cell depleted grafts, our approach translated into promising long-term outcomes in this high-risk patient population.