Donor Natural Killer (NK) Cell Immunotherapy Following Non-Myeloablative HLA-Haploidentical Hematopoietic Cell Transplantation (HCT) Improves Relapse and Progression Free Survival (PFS) in Patients with Hematologic Malignancies

A non-myeloablative conditioning regimen with post-transplant cyclophosphamide (CY) was developed to provide potent in vivo T cell depletion for patients (pts) with high-risk hematologic malignancies undergoing human leukocyte antigen (HLA)-haploidentical HCT (Luznik, O'Donnell BBMT 2008). This regimen was well-tolerated but overall survival (OS) and event-free survival at 2 years (yrs) post-HCT was low at 36% and 26%, respectively, due to high relapse rates. One explanation could be that while post-HCT CY promoted low rates of acute graft-versus-host disease (GVHD), it also eliminated early T and NK cell clones important for disease surveillance. Based on that hypothesis, and using these published outcomes as our comparison group, we developed a Phase I/II clinical trial for hematological malignancies built on this regimen (CY 50 mg/kg day +3), with the addition of a single dose of donor NK cells at day +7 after HCT. Thirty-six pts with median age of 46 (range 8-75) yrs having ALL (n=10), AML (n=8), MDS (n=6), HL (n=5), MM (n=3), NHL (n=3), and CLL (n=1) have been treated to date. Pts were high risk due to underlying disease, potential for relapse, and/or risk for transplant-related mortality (TRM). Most pts were heavily pre-treated, with median time from cancer diagnosis to transplant being 2.1 (0.3 – 9.9) yrs, including 7 pts with prior autologous HCT and 6 pts with 1 or more prior allogeneic HCTs; 58% had HCT-CI scores ≥ 3 indicating high risk for TRM. Non-mobilized peripheral blood mononuclear cells were collected from donors on day+6 using apheresis and stored overnight. NK cells were isolated on day+7 using the Miltenyi CliniMACS system (CD3 depletion/CD56 selection) and infused fresh. The Phase I dose-finding study enrolled at 2 NK doses, with extended enrollment at the 2nd dose level for Phase II [2.5 (n=11) or 5 (n=25) x 106/kg +/- 20%, respectively], with 83% of pts meeting NK cell dose parameters. NK products had a median log T cell depletion of 5.4 (4.1-7.1), median NK recovery of 54% (39-68%), and median NK purity of 92% (74-99%). All pts tolerated the NK cell infusion well without adverse reactions. Full donor chimerism (>95% CD3) was seen in 74% of pts by day +28, while 14% and 11% experienced graft rejection or graft failure, respectively. Of the 30 evaluable pts, grades 2-4 and 3-4 acute GVHD occurred in 37% and 10% of pts, respectively (none with grade 4), and of 29 evaluable pts, 24% developed chronic extensive GVHD. Relapse or progression occurred in 36% of pts. With a median follow-up of 1 yr (range, 0.1 – 4 yrs), 10 pts have died from relapse/progression (n=7) or infection (n=3), giving a probability of OS and relapse/PFS at 1 yr of 74% and 69%, and at 2 yrs of 63% and 51%, respectively (Fig 1). These results provide a platform to further augment NK cell alloreactivity in the post-HCT setting to prevent relapse and disease progression.