Co-Transfusion of Donor NK Cells in Haploidentical Stem Cell Transplantation

39 pedatric patients with acute leukemias (ALL (n=19), AML (n=14) and MDS (n=6)) received T and B cell depleted grafts from full haplotype mismatched related donors. Depletion of the G-CSF stimulated leukapheresis products was carried out with CD3/CD19 coated magnetic microbeads and the CliniMACS device and resulted in a median number of 15.9×10⁶ CD34 (2.5–41) stem cells, 147×10⁶ CD56 NK-cells (9–552) and 413×10⁶ CD14 monocytes (101–1100) per kg body weight. Median numbers of residual T and B cells were 56 000 (10 000–192 000) and 26 000 (2000–149 000) respectively. A reduced intensity regimen (melphalan (140mg/m²), thiotepa (10mg/kg), fludarabine (160mg/m²), OKT3 (0.1mg/kg)) was given in most patients. Co-transfused, HLA mismatched NK cells were traced in peripheral blood of 26 patients starting on day +1 with flow cytometry and appropriate HLA antibodies. Mean numbers of donor derived CD56+ cells/μl were: 3 (day 1), 22 (d 3), 17 (d 7), 75 (d 10), 197 (d 14). Theoretically, the mean absolute number of 4.8×10⁶ co-transfused NK cells should have resulted in a mean number of 2000 cells/μl in peripheral blood of the patients. Comparison of this expected amount with the mean number of NK cells measured within the first week postransplant (25/μl, n=17 data points) showed, that only 1.2% of the cells remained in circulating blood. Thus, the majority of donor NKs did not circulate and probably homed to other compartments (bone marrow, lymph nodes). The number of NK cells cotransfused at day 0 partially influenced the speed of NK cell recovery: patients, who received > 100×10⁶ donor NK cells/kg had significantly higher amounts of circulating cells at day 14 than patients, who received <100×10⁶ donor NKs (240 vs. 140/μl, p<0.05). No significant difference was observed after d 14. Recovery of T cells was not influenced. Graft rejection occurred in 13%. This rate was similar to that of a historical control group (15% in patients who received CD34 positive selected grafts and standard conditioning regimens), although our study patients mainly received an intensity reduced regimen. We conclude, that co-transfused cells facilitated hematopoietic engraftment. Our approach resulted in low TRM (10% at d 365) and in a low relapse rate (20% at 2 years) in patients with microscopical remission (<5% blasts), but was insufficient in patients with active disease (80% relapse rate). We therefore investigated options to increase NK cell activity. Cytotoxicity against K562 cells and thymidine-uptake after PHA stimulation were measured prior and post depletion in 30 procedures. Median specific lysis at E:T ratio = 20:1 was 15% prior and 23% post depletion. Thus, NK activity was not hampered by the procedure. Specific lysis was significantly enhanced by pre-incubation with 1000 U/ml Interleukin (IL) 2 (44%, median) or 2ng/ml IL 12 (40%, median) or 1ng/ml IL 15 ( 53%, median) in vitro. In contrast, thymidine-uptake was reduced from 170 000 to 3000 counts due to profound T-cell-depletion. NK activity was weak against patient derived cryopreserved leukemic blasts without stimulation, but could be significantly increased by cytokine incubation in vitro. Therefore, a pilot study with infusions of IL 15 stimulated NK cells in vivo was started. Up to now, 6 patients received a total of 8 infusions with 12×10⁶ - 150×10⁶ ex vivo stimulated NK cells per kg bw without any side effects. Conclusions: co-transfusion of donor NK cells in haploidentical transplantation is feasible. Only a small portion of cells remained in circulating blood and homing to other organs is likely. NK activity could be increased by cytokines; the use of ex vivo IL 15 stimulated NK cells is currently evaluated. Clinical results suggest antileukemic and graft facilitating effects of donor NK cells.