Fewer Circulating Natural Killer Cells 28 Days after Double Cord Blood Transplantation (dUCBT) Predicts Inferior Survival and IL-15 Response

Previously we demonstrated that following dUCBT, increased absolute lymphocyte (ALC) counts early are associated with improved disease free survival (DFS) (Burke, BBMT, 2011). Considering that a significant proportion of the lymphocytes at this time point are NK cells, we hypothesized that higher NK cell counts would be associated with improved transplant outcomes. We further hypothesized that patients with higher NK numbers would have more mature NK cells with increased NK function (cytotoxicity and cytokine production). To test this hypothesis we used a cohort of dUCBTpatients (n=111, separate from Burke, BBMT, 2011) and examined the number of NK cells (CD3-CD56+) in the peripheral blood after dUCBT. Patients were stratified into low (<50 NK cells/mm3), medium (50-120 NK cells/mm3), and high (>120 NK cells/mm3) groups based on absolute NK counts at D+28. The 3 patient groups did not vary based on age, gender, conditioning intensity, degree of HLA mismatch, UCB cell dose or CMVserostatus/reactivation. As shown in Figure 1, in multiple variate regression analysis, patients with low NK cell numbers experienced significantly lower DFS (HR=1.96, 95% CI: 1.02-3.77, p=0.04) (Fig 1A). There was a trend toward higher non-relapse mortality (NRM) in the low group (41% vs. 26% vs. 18% for low, medium, and high NK groups; p=0.08, Fig 1B), but no difference in relapse or aGVHD (Fig 1C and D). We used a multicolor FACS panel that includes a lineage cocktail (CD14, 19 and 3), CD56, CD117, NKG2A, KIR cocktail and CD57, which allowed us to classify the circulating D+28 NK cells into stage III, IV, V and stage VI NK cells. Interestingly, between the three groups, the proportions of NK cells in the various developmental stages did not differ significantly. We next tested the D+28 NK functionality upon a 4 hr.coculture with K562 cells, by staining NK cells for intracellular cytokines (IFN-g, TNF-a) and degranulation (CD107a). Although we tested a considerable number of patients (n=69), there was no difference in any of the three measures of NK functionality between the three patient groups, but all 3 measures were significantly lower than healthy controls. Considering that IL-15 is a key cytokine that drives NK cell maturation proliferation, and survival we tested the 3dUCBT groups for differences in serum IL-15 concentrations, but found no differences (Fig 2A). Given that the groups differed in cell numbers, we used Ki67 staining to assess whether the D+28 NK cells were differentially in cell cycle and undergoing proliferation. While the percentage of Ki67+ NK cells was significantly higher than controls, the 3 patient groups did not significantly differ from one another. We next tested the cells for the response to IL-15 stimulation. To do this, D+28cryoperserved PBMCs were thawed and rested overnight in media without cytokines. The next morning, PBMCs were stimulated with 0.2 ng of IL-15 for 15 min and the NK cells were assessed for the phosphorylation of STAT5 by FACS. Patients with low numbers of NK cells at D+28 had a defect in IL-15 signaling as demonstrated by a lower percentage of NK cells with p-STAT5, compared to patients with high NK cells (Fig 2A, 26% vs 37%, p=0.04). Many signals, including IL-15 drive the expression of key transcription factors includingTbet andEomes, which control NK cell development and functionality. To investigate the expression of these we used multicolor FACS. For T-bet, there was no difference in the mean fluorescent intensity (MFI) or percent expression between patients with either low or high NK cells at D+28 or healthy controls (Fig 2C). In contrast, patients with high numbers of NK cells at D+28 showed significantly moreEomes (by MFI) than patients with low NK numbers (Figure 2D, 675 vs 552, p=0.025). Therewas also a significantly higher proportion of NK cells expressingEomes in patients with high NK numbers compared to patients with low NK numbers or healthy controls (Figure 2E, 72.5% vs 61.3% vs 42.8%, p < 0.01). Thus, patients with low NK numbers at D+28 afterdUCBT have impaired NKEomes expression. Considering that STAT5 regulates T-bet andEomes, these studies uncover an association between IL-15 andEomes that leads to a reduction in NK cell numbers afterdUCBT which is associated with reduced DFS, likely due to increased NRM. Exogenous, supra-physiological IL-15 early afterdUCBT may overcome this defect and improve transplant outcomes. Figure 1 Figure 1. Figure 2 Figure 2. Disclosures Miller: Oxis Biotech Scientific Advisory Board: Membership on an entity's Board of Directors or advisory committees.