Abstract 1000

Natural killer (NK) cells exhibit in vitro cytotoxicity against many tumor cell types and have an important role in controlling tumor growth, as depletion of NK cells from tumor-bearing mice hastens tumor growth and impairs survival. These data, in combination with results from clinical trials of haploidentical killer immunoglobulin-like receptor (KIR)-ligand mismatched bone marrow transplantation, led to interest in the use of adoptive NK immunotherapy for the treatment of malignancy. Recent clinical results have shown that allogeneic NK cells can be safely administered after chemotherapy and/or irradiation but have also demonstrated limited persistence of the infused NK cells without clear evidence of efficacy. We traced the fate of adoptively infused NK cells in order to delineate the barriers to successful NK immunotherapy using several NK-sensitive murine tumor models.

Mice bearing established lymphoma or leukemia received an intravenous infusion of 0.5–1.0×106 (approximately 2.5–5×107/kg body weight) NK cells from allogeneic or syngeneic donors after or concurrently with total body irradiation and bone marrow rescue. Using luciferase +ve NK cells, we first showed that in animals bearing subcutaneous tumors, NK cells homed to lymphoid organs in the first week, followed by progressive localization to and accumulation within the tumor site (Figure 1). In contrast, in non-tumor bearing animals NK cells homed to lymph nodes, spleen and liver with maximal proliferation at the end of the second week. These observations indicated that NK cells fail to eradicate the tumor despite prior demonstration of in vitro sensitivity, successful homing and local accumulation. As expected from these findings, survival was not prolonged by the NK cell infusion. Reisolation of donor NK cells within 18 hours of transfer showed enhanced cytotoxicity (14% vs 2%, p = 0.004) and IFNγ production (48% vs 22%, p = 0.04) compared with naive resting NK cells. In contrast, NK cells isolated from tumor-bearing mice at later time-points beginning d+5 showed loss of IFNγ production (48% early vs 3% late, p = 0.01), decreased expression of the activating receptor NKG2D, and impaired cytotoxicity in chromium release assays. These observations did not relate to over-stimulation through NKG2D, as NK cells from NKG2D−/− animals were also susceptible to acquired dysfunction (although their baseline cytotoxicity was lower than WT NK toward A20 lymphoma).
Fig. 1

Adoptively transferred luciferase-transgenic NK cells accumulate within the tumor over time

Fig. 1

Adoptively transferred luciferase-transgenic NK cells accumulate within the tumor over time

Close modal

Eomesodermin and T-bet are transcription factors with important roles in effector functions of CD8+ T cells and NK cells. In T cells T-bet downregulation has been shown to correlate with exhaustion (Kao et al, Nat Imm 2011). Flow cytometry of reisolated NK cells revealed downregulation of Eomesodermin (naive splenic control, d+1 reisolated and d+10 reisolated cells showing 85%, 96%, and 29% expression, respectively) and T-bet (naive splenic control, d+1 reisolated and d+10 reisolated cells showing 82%, 99%, and 59% expression, respectively), correlating with loss of IFNγ production.

The phenotype described herein was most dramatic within the tumor and within mice carrying high tumor burdens, but was also present in NK cells reisolated from non-tumor bearing animals that received NK cells, suggesting that homeostatic proliferation after transfer of mature NK cells could also contribute to exhaustion. CFSElo proliferated NK cells showed the most dramatic loss of effector function (chromium release = 42% in unproliferated vs 18% in proliferated cells, p = 0.03) and transcription factor expression (Eomesodermin positive 83% in unproliferated cells vs 18% in proliferated cells, p = 0.002).

Collectively, our results suggest that the success of NK cell immunotherapy is limited by an acquired dysfunction that occurs within days after homeostatic proliferation and target encounter and that may be related to the downregulation of transcription factors required for NK effector function. These findings illuminate a previously unappreciated phenomenon and explain why short-term in vitro killing assays have limited utility in predicting the in vivo behaviour of transferred NK cells. Hence, these findings suggest that transferred NK cells become dysfunctional in vivo and that novel approaches may be required in order to circumvent the described dysfunction phenotype.

Disclosures:

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution