NK cell–based immunotherapies have been gaining traction in the clinic for treatment of cancer. IL-15 is currently being used in number of clinical trials to improve NK cell expansion and function. The objective of this study is to evaluate the effect of repetitive IL-15 exposure on NK cells. An in vitro model in which human NK cells are continuously (on on on) or intermittently (on off on) treated with IL-15 was used to explore this question. After treatment, cells were evaluated for proliferation, survival, cell cycle gene expression, function, and metabolic processes. Our data indicate that continuous treatment of NK cells with IL-15 resulted in decreased viability and a cell cycle arrest gene expression pattern. This was associated with diminished signaling, decreased function both in vitro and in vivo, and reduced tumor control. NK cells continuously treated with IL-15 also displayed a reduced mitochondrial respiration profile when compared with NK cells treated intermittently with IL-15. This profile was characterized by a decrease in the spare respiratory capacity that was dependent on fatty acid oxidation (FAO). Limiting the strength of IL-15 signaling via utilization of an mTOR inhibitor rescued NK cell functionality in the group continuously treated with IL-15. The findings presented here show that human NK cells continuously treated with IL-15 undergo a process consistent with exhaustion that is accompanied by a reduction in FAO. These findings should inform IL-15–dosing strategies in NK cell cancer immunotherapeutic settings.

We improved the bispecific antibody platform that primarily engages natural killer (NK) cells to kill cancer cells through antibody-dependent cellular cytotoxicity (ADCC) by adding IL-15 as a crosslinker that expands and self-sustains the effector NK cell population. The overall goal was to target B7-H3, an established marker predominantly expressed on cancer cells and minimally expressed on normal cells, and prove that it could target cancer cells in vitro and inhibit tumor growth in vivo. The tri-specific killer engager (TriKE^TM) was assembled by DNA shuffling and ligation using DNA encoding a camelid anti-CD16 antibody fragment, a wild-type IL-15 moiety, and an anti-B7-H3 scFv (clone 376.96). The expressed and purified cam1615B7H3 protein was tested for in vitro NK cell activity against a variety of tumors and in vivo against a tagged human MA-148 ovarian cancer cell line grafted in NSG mice. cam1615B7H3 showed specific NK cell expansion, high killing activity across a range of B7-H3+ carcinomas, and the ability to mediate growth inhibition of aggressive ovarian cancer in vivo. cam1615B7H3 TriKE improves NK cell function, expansion, targeted cytotoxicity against various types of B7-H3-positive human cancer cell lines, and delivers an anti-cancer effect in vivo in a solid tumor setting.

Natural killer (NK) cells are potent immune modulators that can quickly lyse tumor cells and elicit inflammatory responses. These characteristics make them ideal candidates for immunotherapy. However, unlike T cells, NK cells do not possess clonotypic receptors capable of specific antigen recognition and cannot expand via activating receptor signals alone. To enable NK cells with these capabilities, we created and have previously described a tri-specific killer engager (TriKE) platform capable of inducing antigen specificity and cytokine-mediated NK-cell expansion. TriKE molecules have three arms: (i) a single-chain variable fragment (scFv) against the activating receptor CD16 on NK cells to trigger NK-cell activation, (ii) an scFv against a tumor-associated antigen (CD33 here) to induce specific tumor target recognition, and (iii) an IL15 moiety to trigger NK-cell expansion and priming. Here, we demonstrate that by modifying the anti-CD16 scFv with a humanized single-domain antibody against CD16, we improved TriKE functionality. A CD33-targeting second-generation TriKE induced stronger and more specific NK-cell proliferation without T-cell stimulation, enhanced in vitro NK-cell activation and killing of CD33-expressing targets, and improved tumor control in preclinical mouse models. Given these improved functional characteristics, we propose rapid translation of second-generation TriKEs into the clinic.

Chronic lymphocytic leukemia (CLL) is characterized by chronic clonal expansion of mature CD19-expressing B lymphocytes and global dysfunction of immune effectors, including natural killer (NK) cells. CLL remains incurable, and novel approaches to refractory CLL are needed. Our group has previously described trispecific killer engager (TriKE) molecules that redirect NK cell function against tumor cells. TriKE reagents simultaneously bind an activating receptor on NK cells, CD16, and a tumor antigen while also providing an NK cell expansion signal via an interleukin-15 moiety. Here we developed the novel CD19-targeting 161519 TriKE. We demonstrate that 161519 TriKE induced killing of a CD19-expressing Burkitt’s lymphoma cell line and examined the impact on primary CLL targets using healthy donor and patient NK cells. 161519 TriKE induced potent healthy donor NK cell activation, proliferation, and directed killing. Furthermore, 161519 TriKE rescued the inflammatory function of NK cells obtained from CLL patient peripheral blood samples. Finally, we show that 161519 TriKE induced better directed killing of CLL in vitro when compared with rituximab. In conclusion, 161519 TriKE drives a potent activating and proliferative signal on NK cells, resulting in enhanced NK cell expansion and CLL target killing. Our findings indicate the potential immunotherapeutic value of 161519 TriKE in CLL. Key Points 161519 TriKE induces NK cell–mediated killing of primary CD19⁺ CLL tumor cells while also inducing NK cell proliferation. 161519 TriKE enhances interferon γ production on CLL patient NK cells. Visual Abstract

NK cell exhaustion is caused by chronic exposure to activating stimuli during viral infection, tumorigenesis, and prolonged cytokine treatment. Evidence suggests that exhaustion may play a role in disease progression. However, relative to T cell exhaustion, the mechanisms underlying NK cell exhaustion and methods of reversing it are poorly understood. Here, we describe a potentially novel in vitro model of exhaustion that uses plate-bound agonists of the NK cell activating receptors NKp46 and NKG2D to induce canonical exhaustion phenotypes. In this model, prolonged activation resulted in downregulation of activating receptors, upregulation of checkpoint markers, decreased cytokine production and cytotoxicity in vitro, weakened glycolytic capacity, and decreased persistence, function, and tumor control in vivo. Furthermore, we discovered a beneficial effect of NK cell inhibitory receptor signaling during exhaustion. By simultaneously engaging the inhibitory receptor NKG2A during activation in our model, cytokine production and cytotoxicity defects were mitigated, suggesting that balancing positive and negative signals integrated by effector NK cells can be beneficial for antitumor immunity. Together, these data uncover some of the mechanisms underlying NK cell exhaustion in humans and establish our in vitro model as a valuable tool for studying the processes regulating exhaustion.