Bringing natural killer cells to the clinic

NK cells are innate lymphoid cells that differentiate into the bone marrow and, once reaching maturation, circulate in the peripheral blood, patrol the body, and may lead to tissue-resident natural killer (NK) cells. NK cells are cytotoxic effectors that can recognize and lyse stressed cells including tumor cells. NK cells are tightly regulated by a repertoire of inhibitory and activating receptors, enabling them to kill tumor cells while sparing normal cells. In particular, tumor cells may present a decrease in HLA class I molecule expression, which identifies them as preferential targets for NK cells. Indeed, NK cells express HLA class I–specific inhibitory receptors, such as killer cell Ig-like receptors (KIRs) and NKG2A, that block their effector function against HLA class I⁺ cells (Chiossone et al., 2018). The triggering of NK cell effector functions also requires the recognition of ligands expressed on tumor cell membrane by activating NK cell receptors including NKp46, NKp30, NKp44, NKG2D, and DNAM-1 (Bottino et al., 2006). Most mature NK cells also express CD16a (FcγRIIIA), a low-affinity receptor for the Fc region of IgG responsible for antibody-dependent cell-mediated cytotoxicity (ADCC). NK cells may therefore contribute to therapies with anti-tumor antigen IgG1 mAbs.

Some important features of NK cell biology encourage their manipulation in cancer treatment: (i) they are able to kill tumor cells that lack HLA class I molecules and would be neglected by T cells, (ii) they do not need prior antigen-specific sensitization for harboring effector functions, (iii) they do not cause graft-versus-host disease (GvHD), and (iv) upon target recognition, NK cells can secrete pro-inflammatory cytokines, such as IFN-γ and TNF-α, which have direct antitumor effects, but also chemokines, including CCL2, CCL3, CCL4, CCL5, XCL1, and IL-8, and growth factors, such as FLT3 ligand and GM-CSF, which contribute to the onset, the orientation, and the maintenance of the adaptive immune response (Chiossone et al., 2018; Vivier et al., 2011). Harnessing NK cells in cancer patients presents the dual advantage of inducing the killing of tumor cells, but also of participating in a multicellular immune response against tumor cells.

NK cell anti-tumoral function has been demonstrated in vitro using human cells and in vivo in the mouse against tumor cells of all histiotypes. Correlations have been observed between patient clinical outcome and NK cell infiltration at the tumor bed or cytotoxicity of peripheral NK cells (Table 1). In the past two decades various approaches have been developed to exploit the capacity of NK cells to control tumor growth (Fig. 1). Clinical studies demonstrated the safety of NK cell infusions for immunotherapy with encouraging results from preclinical or clinical studies in hematopoietic malignancies (Myers and Miller, 2021; Laskowski et al., 2022).

NK cell infusions were the first NK cell therapy approach developed in cancer patients. The first clinical trials were based on cytokine-activated autologous NK cells, but their clinical efficacy was not satisfactory. HLA-haploidentical hematopoietic stem cell transplantation (HSCT) revealed the antitumor effect of allogeneic NK cells, based on mismatched between donor KIRs and patient HLA class I molecules (Ruggeri et al., 2002). Subsequently, allogeneic NK cells from HLA-related or unrelated healthy donors have been administered after HSCT for therapy of hematological malignancies with positive results in the control of relapse and GvHD. In a non-transplantation setting, infusion of HLA-haploidentical NK cells and IL-2 led to tumor regression in AML (acute myeloid leukemia) patients (Miller et al., 2005). Afterward, clinical trials in other hematopoietic diseases and solid tumors in combination with chemotherapy showed that allogenic NK cell infusions were safe and, in some cases, also effective (Myers and Miller, 2021; Laskowski, 2022).

Impressed by the success of chimeric antigen receptor (CAR)-T cell therapy in acute lymphoblastic leukemia and non-Hodgkin’s lymphoma, several groups concentrated their efforts in the engineering of CAR-NK cells. CAR-NK cells have been developed from various sources, such as immortalized cell lines (NK-92), peripheral blood mononuclear cells, cord blood, hematopoietic stem and progenitor cells, and induced pluripotent stem cells (iPSCs). Recently, HLA-mismatched anti-CD19 CAR-NK cells showed remarkable clinical efficacy in a phase 1/2 trial in non-Hodgkin’s lymphoma and chronic lymphocytic leukemia (Liu et al., 2020). iPSC-derived anti-CD19 CAR-NK cells also exhibited encouraging results in phase 1/2 clinical trial in relapsed/refractory B cell lymphoma (Bachanova et al., 2021).

Combined cytokine activation with IL-12, IL-15, and IL-18 is used to induce the in vitro expansion of NK cells with potent cytotoxic properties, referred to as cytokine-induced memory-like (CIML) NK cells. After infusion in preclinical models, CIML NK showed superior responses and higher persistence compared with conventional NK cells. Remarkable responses have been reported in AML patients receiving haploidentical cell transplant and CIML NK cells derived from the same donor (Berrien-Elliott et al., 2022). These efforts are being completed by the development of biologicals, such as immune checkpoint inhibitors and NK stimulators, the clinical evaluation of which is being tested at present in various hematological malignancies and solid tumors.

NKG2A is an inhibitory cell surface receptor shared by NK and T cells. Its ligand, HLA-E, is a non-classical MHC class I molecule highly expressed in many solid tumors and hematologic malignancies (André et al., 2018). Monalizumab, a humanized mAb specific for NKG2A, showed prolonged progression-free survival (PFS) in combination with durvalumab compared to durvalumab alone in patients with unresectable and chemo-radio-resistant NSCLC (non-small cell lung cancer; Herbst et al., 2022). Following these results, monalizumab is currently being tested in phase 3 in combination with durvalumab in NSCLC.

Another promising inhibitory receptor to block in cancer therapy is LAG3 (lymphocyte-activation gene 3). It is expressed by tumor-infiltrating CD8⁺ T cells, NK cells, B cells, and plasmacytoid dendritic cells. Several clinical trials are evaluating various LAG-3 targeting molecules in solid tumors, including blocking mAbs, soluble LAG-3–Ig fusion proteins and anti–LAG-3 bispecific drugs coupling LAG-3 to PD-1, PD-L1, or CTLA4 blocking.

TIGIT (T cell Ig and ITIM domain) is a receptor that is expressed on NK cells and on effector, memory, and regulatory T cells, where it functions as a co-inhibitory receptor in synergy with PD-1 and Tim-3. Anti-TIGIT blocking mAbs are being tested in various solid tumors in combination with PD-1, PD-L1, or PD-L2 blocking mAbs. Recently, two phase 3 trials revealed no clinical efficacy of tiragolumab combined to atezolizumab alone or in combo with chemotherapy in NSCLC. Anti-TIGIT/anti–PD-1 bispecific antibodies have also been developed and are now in phase 2 clinical trials in solid tumors.

The NK cell stimulatory cytokine IL-2 was approved for the treatment of several malignancies almost 20 yr ago. Yet, its administration revealed limited efficacy due to its short half-life and severe toxicity, and several modifications have been engineered to address these limitations, including polyethylene glycol conjugation, fusion to tumor-targeting antibodies, and alteration of receptor-binding affinity. IL-15 entered into cancer therapy thanks to its capacity to activate NK and CD8⁺ T cells while sparing regulatory T cells. It displayed a safer profile compared to IL-2 but limited in vivo efficacy. Recently, an IL-2 prodrug masked by the IL2 receptor α chain linked to a tumor-associated protease substrate has been proposed as an innovative strategy to efficiently target tumor-infiltrating lymphocytes, minimizing systemic toxicity (Hsu et al., 2021).

Bi-specific, tri-specific, or tetra-specific molecules targeting one or more activating NK cell receptors and/or cytokine receptors represent a new class of therapeutic molecules, referred to as NK cell engagers, that are being developed to favor tumor cell recognition and enhance NK cell activation (Demaria et al., 2021). Some of them are being tested in phase 1/2 clinical trials, such as GTB-3550, a tri-specific killer engager targeting CD16 and CD33 and bearing an IL-15 domain that promotes NK cell proliferation; DF1001, a tri-specific antibody targeting HER2, CD16, and NKG2D; or AFM13 and AFM24, which are trispecific NK cell engagers targeting CD16a on NK and myeloid cells and the tumor-associated antigens CD30 or epidermal growth factor receptor, respectively. We developed a series of NKp46 targeting NK cell engagers that provided encouraging results in preclinical models (Gauthier et al., 2019). IPH6101/SAR′579, an IgG1 antibody co-engaging NKp46, CD16, and the tumor antigen CD123 is now in phase 1/2 in acute leukemias and myelodysplastic syndromes. IPH64/SAR′514 is an engineered IgG1 for enhanced ADCC, targeting CD16, NKp46, and the B cell maturation antigen. Finally, we are also developing tetra-specific molecules engaging CD16, NKp46, a tumor antigen, and the IL-2 receptor β and γ chains by an IL-2 variant, leading to activation and proliferation of NK cells, as well as control of invasive and subcutaneous tumors in mice.

The future of cancer immunotherapy lies in combined therapeutic approaches. Along this line, several interesting possibilities include the combination of NK cell infusions with immune checkpoint inhibitors or with NK cell engagers. It is also important to take into consideration the factors present in the tumor microenvironment, which limit the anti-tumor function of NK cells (Fig. 1). Although the targeting of these inhibitory pathways has shown disappointing clinical results so far, the combination of some of these drug candidates with NK cell therapies will be quite interesting to test. Finally, it remains to assess NK cell therapies in solid tumors, where mechanisms leading to recruitment and activation need to be more clearly understood to reach clinical benefits in patients.

The E. Vivier laboratory at Centre d'immunologie de Marseille-Luminy and Assistance-Publique des Hôpitaux de Marseille are supported by funding from the European Research Council under the European Union’s Horizon 2020 research and innovation program (TILC, grant agreement No. 694502 and MInfla-TILC, grant agreement No. 875102—MInfla-Tilc), the Agence Nationale de la Recherche including the PIONEER Project (ANR-17-RHUS-0007), MSDAvenir, Innate Pharma, and institutional grants awarded to the Centre d'immunologie de Marseille-Luminy (Institut National de la Santé et de la Recherche Médicale, Centre national de la recherche scientifique, and Aix-Marseille University) and Marseille Immunopole.

Disclosures: L. Chiossone and E. Vivier are employees of Innate Pharma. No other disclosures were reported.