Differences in Reconstitution of NK-Cell Subsets and Expression of Effector Molecules after HLA-Identical and Haploidentical Allogeneic Stem Cell Transplantation in Patients with AML.

The aim of this investigation was to compare NK-cell reconstitution together with the expression of effector molecules in AML patients after CD34+ enriched haploidentical allogeneic stem cell transplantation (HASCT) (group A, n=9) or myeloablative HLA-identical blood stem cell transplantation (BSCT, group B, n=8) with healthy donors (group C, n=15).

The median percentage of NK-cells was significantly different between normal donors (C: 19 %, range 9–27 %) and patients in group A on day +21 (79 %, 58–95 %; p = 0.001) but not in group B (26 %, 14–59 %). In healthy donors a median of 10 % (4–18 %) of NK-cells were CD56_brightCD16_neg, and a median of 90 % (82–96 %) were CD56_dimCD16_bright, whereas in patients the predominating subset on day +21 was CD56_brightCD16_neg with a significant higher median percentage in group A than in group B (A: 89 %, 77–98 % vs. B: 57 %, 32–80 %; p = 0.01). The mean fluorescence intensity (MFI) of NK-cell receptors NKp30 and NKp44 were similar but of NKp46 was significantly higher on day 21 on CD56_dim in transplanted patients compared to normal controls (group A: 78, group B: 76, group C:44, p=0.01). In contrast, NKG2D expression was significantly lower only in CD56_bright cells of patients on day 21 (group A: 62, group B: 75, group C: 108, p=0.01). In group B HLA DR was higher expressed in both subsets where this was true only for CD56_dim in group A compared to C (group A_dim: 62, groupB_dim: 63, group C_dim:7, p=0.01, group A_bright: 21, group B_bright: 75, group C_bright: 24, p=0.02). On day +70 the expression pattern became similar to that of normal controls.

On day +21 as well as on day +70, group A but not B showed a significant higher expression of TRAIL on CD56_dim cells compared to healthy donors (p=0.001). In contrast to normal donors and group B, significant more FAS-ligand could be detected on CD56_dim than on CD56_bright cells at both time points in group A (p = 0.01). Interestingly, in comparison to normal donors the expression of perforin was significantly lower in group A and B on day +21 (p=0.01) and group A on day +70 (p=0.01), whereas granzyme B was equally expressed in all groups. In line with this observation we documented the absence of perforin expression in 4 patients on day +21 (1x A, 3x B) and 3 patients on day +70 (2x A, 1x B), which was never observed in group C. Two patients in group A received an add-back of unstimulated donor NK-cells and concomitant IL-2 in-vivo. As a result a significant but transient increase in the expression of TRAIL, FAS-ligand and perforin could be detected whereas granzyme remained unaffected. Additionally, we found also a significant and again transient increase of NKG2A, CD69 and HLA DR and a decrease of NKp30, NKp44 (only on CD56_bright), NKp46 and NKG2D. This was associated with a significant better efficacy of NK-cell lysis of K562 cells in-vitro (pre IL-2 vs. IL-2 vs. post IL-2: 28.8% vs. 51.7% vs. 27.7%; p=0.01).

These data provide new insights into the dynamics of NK-cell reconstitution, which differ between haploidentical T-depleted and matched allogeneic T-repleted allogeneic transplants and might be modulated by IL-2 in-vivo. Whether these findings translate into different anti-leukemic and anti-infectious capacities of NK-cell subsets needs to be determined.