Mature T Cells in the Graft Improve NK Cells Reconstitution after Haploidentical Hematopoietic Stem Cell Transplantation.

We previously demonstrated that natural killer (NK) cells generated after haploidentical stem-cell transplantation (SCT) are blocked at an immature state characterized by phenotypic features and impaired functioning, a blockage that may affect transplantation outcome (Nguyen et al. Blood 2005). Hypothesizing that the absence of mature donor T cells in the graft may affect NK cell differentiation and function, we examined NK cells from 21 patients who received haploidentical SCT from relatives for advanced malignant hematopoietic disease and underwent either partial (pTCD) (CD3+ in the graft >1x10⁵/Kg, mean: 6.9x10⁵/Kg; n=11) or extensive (e-TCD) (CD3+ in the graft<1x10⁵/Kg; mean: 0.35x10⁵/Kg; n=10) T cell depletion and compared them with NK cells from their healthy donors. As previously described, compared with donor cells, recipient NK cells expressed lower levels of inhibitory KIR (in particular KIR2DL1 and KIR2DL2) and higher levels of CD94/NKG2A receptors after transplant (mean expression of CD94/NKG2A on recipient NK cell at 3 months post-transplant: 93.4%±7.2% versus 49.6%±10.9% on donor NK cells, p<0.0001), but these levels did not differ significantly between the pTCD and eTCD groups. However, the frequency of the immunoregulatory CD3⁻CD56^bright NK subset was sharply lower in the pTCD than eTCD groups after transplantation (25.0%±9.6% versus 53.3%±18.0 at 3 months; p<0.001). The level of NKp30 receptors on NK cells was also higher after pTCD than eTCD transplantation (70.3%±7.1% versus 58.0%±6.5%, p=0.013) and that of pTCD patients resembled the donor NK repertoire. NK cytotoxicity against primary haplomismatched AML blasts was significantly more pronounced after pTCD than eTCD transplants (29.0%±8.9% specific lysis versus 6.7%±4.1% at a ratio Effector/Target (E/T):20/1, p=0.002), although still lower than in donor NK cells (mean specific lysis of donor NK cells from both groups against AML blasts: 43.5%±13% at a ratio E/T: 20/1). This more mature phenotypic and functional profile of NK cells after pTCD transplant was clinically associated with a lower rate of relapse and superior survival (1/11 relapse, 3/11 patients alive in complete remission at 11, 10 and 3 years) than in eTCD group (8/10 relapse; no patient alive at 1 year). These results support a model in which mature donor T cells in the graft may play a key role, in vivo, in NK cell differentiation by improving NK cell maturation and cytolytic function against leukemic blasts. They point to the dilemma of haploidentical hematopoietic SCT in leukemic patients: on the one hand, extensive T-cell depletion is associated with a risk of fatal leukemia relapse due to the loss of the GvL effect T cells, which can not be replaced by immature NK cells; on the other hand, partial T-cell depletion might increase the risk of GvHD but also improves the GvL effect mediated by NK cells. New treatments infusing mature haploidentical NK cells in leukemic patients should be used to test the efficiency of NK alloreactivity.