Lung cancer is the leading cause of cancer-related death worldwide; more than 1.7 million people succumbed to lung cancer in 2018 (1). Based on origin, lung cancer can be divided into small cell lung cancer (SCLC) and non-SCLC (NSCLC), of which NSCLC accounts for 80–85% of the cases (1). Early lung cancer often lacks symptoms, which may lead to delayed diagnosis and treatment. In the late stage, both SCLC and NSCLC can metastasize to other organs; SCLC can metastasize considerably more rapidly, and patients develop metastatic symptoms (bone pain, nervous system changes such as dizziness and seizures, jaundice, enlarged lymph nodes, and/or other conditions such as syndrome of inappropriate antidiuretic hormone and Horner syndrome) faster than those with NSCLC. Due to the large proportion of patients diagnosed with locally advanced or widely metastatic cancer at the time of diagnosis, the 5-year relative survival rate for NSCLC is poor, from 68% in patients with stage IB disease to 0–10% in patients with stage IVA-IVB disease (2). Although SCLC is characterized by rapid responses to chemotherapy and sensitivity to radiotherapy, given the early treatment resistance, the 5-year overall survival (OS) is <10% (3). Traditional treatments for lung cancer include chemotherapy, radiotherapy and surgery. Fortunately, advances in the knowledge of lung cancer and technologies for its detection have promoted marked progress of theories and molecular methods in diagnosing lung cancer and have revolutionized the relevant therapeutics. Researchers have already extensively described the characteristics of the lung cancer genome, and several major pathways sensitive to targeted therapy have been identified (4). Drugs that target these pathways have improved response and survival in patients with metastatic disease (5), some of which have replaced chemotherapy as first-line treatment drugs. Unfortunately, the efficacy in most patients is limited by the emergence of resistance mechanisms, while these interventions are effective initially (6,7). Therefore, investigation of effective strategies to eliminate these resistant tumor cells is urgently needed.

In recent decades, a series of studies have reported the importance of the immune system in malignant disease control, and immunotherapy has gradually attracted the attention of researchers (8,9). Inducing passive or active antitumor responses by the immune system against malignant tumors is an attractive therapeutic strategy. As a critical part of immune surveillance, natural killer (NK) cells exhibit cytotoxic activity against diverse tumor cell types; furthermore, NK cells bridge the innate and adaptive immune responses (10). With the development of methods to regulate NK cell function and enhance tumor sensitivity to NK cell cytotoxicity and the ability to expand NK cells in vitro and manipulate their homing, numerous NK cell-based immunotherapy methods and strategies have been developed (9). In physiological conditions, lung tissue has a considerable amount of NK cells, which may be important antitumor effector cells of lung tissue. Therefore, immunotherapy strategies based on NK cells may confer great clinical benefit to lung cancer treatment. In the present review, the distribution and function of NK cells, the control effect of NK cells on lung cancer, and the effect of the lung cancer tumor microenvironment (TME) on NK cells were briefly introduced and some NK cell-based immunotherapy strategies were described. Given the advances summarized in the present review, an exciting future for NK cell-based cancer immunotherapy is foreseen and the challenges that remain to be tackled are presented. Although enormous steps have been taken in understanding NK cell biology, more work is required to fully explore the anticancer potential of these cells.

A search for scientific papers published between 1975 and 2020 focusing on NK cells, lung cancer and NK cell-based immunotherapy was performed in PubMed. The search terms used were ‘NK cell’, ‘lung’, ‘cancer’, ‘immunotherapy’, ‘tumor microenvironment’, ‘cytokine’, ‘monoclonal antibodies’, ‘adoptive transfer’, ‘CAR’, alone and in combination. A total of 176 scientific papers were selected, 117 of which were original studies.

NK cells are innate lymphocytes that can directly eliminate target cells without prior exposure (11,12) and play a key role in antiviral and antitumor immunity. NK cells, mainly present in the peripheral blood, comprise approximately 15% of all circulating lymphocytes (13), while they are also distributed in multiple tissues including the liver, lung, skin, kidney and bone marrow. Moreover, based on the expression of CD49a (i.e., integrin α1), CD69 and CD103 (i.e., integrin aE) (14–17), NK cells can be subdivided into circulating and tissue-resident NK cells. Tissue-resident NK cells usually display high expression of CD49a, CD103, and CD69 (18). More commonly, researchers subdivide human NK cells into two major subsets with distinct maturation and functional properties according to the expression of CD56 and the antibody binding-Fc receptor CD16 (13). CD56^brightCD16⁻ NK cells (approximately 10% of NK cells in the peripheral blood) are specialized in secreting cytokines and are abundantly located in secondary lymphoid organs (lymph nodes, tonsils, and spleen) (19), most of which exhibit characteristics of tissue-resident lymphocytes and tissue-specific adaptations. Furthermore, they can also reveal cytotoxicity under prolonged stimulation with cytokines such as interleukin (IL)-15, IL-12, and IL-18 (13,20–24). CD56^dimCD16⁺ NK cells (approximately 90–95% of NK cells in the peripheral blood) (12) are potent cytolytic effector cells, which can rapidly secrete pro-inflammatory cytokines such as interferon (IFN)-γ and cytotoxic mediators such as granzyme once activated. Most of them exhibit characteristics of circulating cells, but they can also show a resident phenotype while located in the lymph nodes, mucosa, and other parts.

Activation of NK cells is regulated by stimulatory and inhibitory signals (25,26). The activation signals are mainly provided by NKp46, NKp30, NKp44, natural killer group 2 member D (NKG2D), CD16 and killer cell immunoglobulin-like receptor (KIR)-S (27), which usually recognize self-ligands expressed on infected or transformed tissues [known as ‘recognize non-self’ and ‘stress-induced self’ (28)]. The inhibitory signals are mainly provided by the classic inhibitor, KIR, which usually identifies diseased cells that are lacking ligands such as major histocompatibility complex (MHC) class I molecules [known as ‘missing self’ (29)]. Activated NK cells can exert cytotoxicity via several distinct mechanisms: i) They release cytoplasmic particles containing granzymes and perforin through immune synapses with target cells to induce target cell apoptosis (30); ii) they play a role through the tumor necrosis factor (TNF) family (31). They express a death-inducing factor ligand [factor-associated suicide ligand, (FASL)] after activation and induce FAS expression on malignant cells, which leads to target cell apoptosis (32). Moreover, TNF-α produced by activated NK cells can also induce tumor cell apoptosis (33); iii) NK cells secrete various effector molecules (including multiple cytokines, chemokines and growth factors) interacting with dendritic cells (DCs), macrophages, T cells and endothelial cells to limit tumor angiogenesis and activate adaptive immunity (10,34–36). For example, IFN-γ produced by NK cells increases the expression of MHC class I molecules on transformed cells to promote their recognition by cytotoxic lymphocytes (CTL) (36). Other NK cell-derived factors, including TNF-α, IL-10, granulocyte-macrophage colony stimulating factor (GM-CSF), and chemokine C-C motif ligand (CCL)-5, can also regulate the immune response (10); iv) by modifying the Fc portion of IgG antibodies, NK cells can eliminate target cells through antibody-dependent cell-mediated cytotoxicity (ADCC) (37) (Fig. 1).

NK cells exhibit marked mobility, thereby circulating between organs to promote immune surveillance (38). NK cells can respond to multiple chemokines and be recruited to different tissues or inflammation sites owing to the expression of several chemokine receptors (39). The lung is a critical organ of body-environment interaction and is rich in NK cells (40,41). It is generally considered that NK cells originate and develop in the bone marrow and then migrate to the lung (42). The proportion of NK cells in the lung is similar to, and even slightly higher, than that in the peripheral blood, accounting for 10–20% of the lymphocytes in the lung (43). IL-15 secreted by bronchial epithelial cells and alveolar macrophages may be responsible for the high proportion of NK cells (44,45) because it is the main cytokine supporting NK cell cytotoxicity, homeostasis, and development (46,47). Unlike the liver and secondary lymphoid organs rich in CD56^brightCD16⁻ NK cell subpopulation, most NK cells in the lung exhibit the CD56^dimCD16⁺ phenotype (48), indicating that most are circulating subsets and highly differentiated (49,50). Despite their high differentiation, human NK cells in the lung exhibit a weaker response to target cell stimulation than peripheral blood NK cells (51), which may be attributable to the inhibition of alveolar macrophages (52) and soluble factors of the lower respiratory tract (51). Perhaps because the pulmonary mucosa is continuously exposed to the environment and autoantigens, NK cells with restricted function in physiological conditions may be more conducive to the maintenance of pulmonary homeostasis (49). Although circulating CD56^dimCD16⁺ NK cells are the major subpopulation in the lung (51), CD49a⁺ tissue-resident NK cells (mainly CD56^brightCD16⁻ NK cells) also account for approximately 15% of human NK cells in this organ (53). Studies have revealed that CD56^brightCD49a⁺ NK cells in the lung strongly co-express CD103 and CD69, significantly different from CD56^brightCD16⁻NK cells in the peripheral blood (18,53). In in vitro experiments, CD49a⁺ tissue-resident NK cells exhibited a higher ability to degranulate and produce IFN-γ when in contact with virus-infected autologous macrophages than NK cells in the peripheral blood (50). Collectively, these results indicated that circulating NK cells in the lung have a larger number and highly differentiated phenotype but exhibit depressed function, while tissue-resident NK cells have stronger function. It is necessary to further study the characteristics of circulating and tissue-resident NK cells in the lung to understand their roles in the physiological condition and in the occurrence and progression of lung tumors, which may provide novel directions for the development of therapeutic strategies.

In the 1980s, some studies revealed that cancer incidence was higher among individuals with NK cell dysfunction (54,55). Since then, studies have increasingly confirmed that the antitumor effect of NK cells can act against multiple tumor types (56,57), including head and neck (58), pharyngeal (59), colorectal (60), and lung (38,61) cancers. The direct evidence that NK cells act against lung cancer is supported by Kras-driven spontaneous lung cancer and cancer cell implantation experiments in mice (61,62), both of which revealed that mice lacking NK cells have a greater lung tumor burden. However, the antitumor effect is limited to the early stage of Kras-driven lung cancer in mice due to NK cell dysfunction in the advanced stage (61).

In the past few years, some studies have revealed that NK cells can infiltrate lung cancer and that the number of tumor-infiltrating NK (TINK) cells is significantly associated to postoperative patient survival, indicating that the infiltration of NK cells into tumors may benefit patient prognosis (63–65). Similar phenomena have been observed in patients with breast cancer (66) and renal clear cell carcinoma (67). Interestingly, most TINK cells are of the CD56^brightCD16⁻ NK type (68), and they only exist in the intratumoral fibrous septum and the interface between stromal and surrounding tumor cells, which appear to indirectly be in contact with cancer cells (68,69). Conversely, in renal clear cell carcinoma, NK cells infiltrate the entire tumor tissue (70). Although the mechanism of TINK enrichment remains unclear, homing restriction and an immunosuppressive microenvironment may play an important role (71–73). A previous study revealed that the proportion of CD56^brightCD16⁻ NK cells in tumoral and non-tumoral lung tissues is similar (74), indicating that the enrichment of CD56^brightCD16⁻ NK cells in tumors may be driven by the rejection of CD56^dimCD16⁺ NK cells by the tumor. Conversely, it may be related to the chemokine spectrum of CD56^dimCD16⁺ NK cell subsets. For example, it has already been confirmed that the adhesion signal of heterodimerization of chemokine receptor CCR5 (i.e., MIP-1β receptor), which is only expressed by CD56^bright CD16⁻ NK cells, could force leukocytes to stay in the tissue (75,76). In contrast, the viability of CD56^dimCD16⁺ NK cells may be impaired in the TME. Interestingly, it is generally considered that tumor rejection is mainly due to direct killing by lymphocytes. However, some studies have revealed that IFN-γ and other lymphocyte-derived cytokines such as TNF-α can promote tumor rejection to control tumor progression (77,78), which indicates that tumor rejection is a more complicated event than previously considered and that the cytokine secretion function of tumor-infiltrating CD56^brightCD16⁻ NK cells cannot be ignored in tumor control.

In summary, NK cells can infiltrate and eliminate tumor cells; therefore, targeting NK cells through immunotherapy is an attractive anticancer strategy. Based on the available literature, it can be theorized that the localization of NK cells in tumors (NK cells are most successful in the treatment of hematopoietic malignancies such as leukemia because NK cells are abundant in the peripheral blood) may be an essential factor of NK cell-based immunotherapy. Exploring the role of NK cells in survival and the lung tumor environment may enable the development of methods to improve the ability of NK cells to migrate and infiltrate into tumor tissues, thereby effectively improving the antitumor immunity of the body.

Although it has been determined that NK cells have antitumor effects, malignant tumors continue to develop in the presence of NK cells, which does not mean that NK cells do not contribute to tumor control, but that their antitumor activity may be impaired to some extent (79,80). In this regard, the TME, which is composed of cell components, growth factors, proteases, extracellular matrix, and lymphatic and vascular systems, plays an important role (81). The TME allows tumor cells to obtain cancer markers, establishing a chronic inflammatory environment that maintains tumor growth and induces dysfunction of NK cells in various ways (79,82). An increasing number of studies have revealed that the phenotype and function of NK cells are altered in the tumor microenvironment (68–69,79). A comprehensive understanding of the factors and mechanisms that cause NK cell changes in the TME may help reveal means to restoring their antitumor potential.

Several mechanisms have been revealed to be related to the phenotype and function alterations of NK cells (Fig. 2). Firstly, tumor cells can affect the phenotype of NK cells depending on cell-to-cell contact (83). In lung cancer, an in vitro Transwell experiment revealed that the communication between NK and tumor cells is associated with the downregulation of active receptors including NKp30, NKp80, DNAM-1, and NKG2D on the surface of TINKs (69). Another study revealed that the expression of CD155 on tumor cells is related to the downregulation of DNAM-1 on NK cells in NSCLC (84). Conversely, the inhibitory receptors of NK cells are upregulated in cancer. In humans, the expression of T-cell immunoglobulin and ITIM domain (TIGIT) on NK cells was further upregulated in tumor regions compared with peritumoral regions in colorectal tumors (85).

Secondly, the modification of the NK cell phenotype can be altered due to the high expression of immunosuppressive factors such as transforming growth factor (TGF)-β, indoleamine 2,3-dioxygenase (IDO), IL-4, and prostaglandin E2 (PGE2) (86–89). TGF-β has been revealed to be overexpressed in lung cancer cells (90,91), and its expression level can be a prognostic marker in lung cancer (92). Some mechanisms by which TGF-β inhibits the function of NK cells have been identified in lung cancer: i) TGF-β changes the receptor spectrum of NK cells in patients with lung cancer (68,69,93). TGF-β downregulates the expression of NK activating receptors NKp80, NKp30, and NKG2D (68,69,94), thereby inhibiting the cytolytic activity of NK cells (86,94). Neutralizing TGF-β inhibits the downregulation of NKG2D expression and restores the antitumor response of NK cells (93,95); conversely, TGF-β upregulates inhibitory receptors including NKG2A (96) and programmed cell death-1 (PD-1) (97) in NK cells in tumors (96,98). ii) TGF-β affects the metabolism of NK cells in lung cancer. In Kras-driven lung cancer, high levels of TGF-β in the TME cause aberrant expression of the fructose 1,6-biophosphatase (FBP1) protein in NK cells, thereby inhibiting NK cell glycolysis and reducing cellular activity, eventually leading to NK cell dysfunction (61); iii) TGF-β mediates NK cell polarization toward angiogenesis (99). Furthermore, another study revealed that the expression of PGE2, which inhibited the antitumor activity of NK cells in NSCLC tumor tissue, was significantly increased (100). Notably, immune cells which are a major component of the TME, inhibit NK cell function mainly by secreting immunosuppressive molecules (101–103). Studies have revealed that the number of myeloid-derived suppressor cells (MDSCs) (104) and regulatory T cells (Tregs) (105) in the lung cancer TME is higher than that in normal tissues and peripheral blood adjacent to cancer, both of which can secrete TGF-β (86,93,106,107).

In addition, the physical and chemical conditions of the TME, including hypoxia, low pH and low glucose concentration, can also impair NK cell function (108). Previous studies have confirmed that hypoxia downregulates the expression of NCR and NKG2D on NK cells (109) and damages their cytotoxicity (110). In NSCLC, high HIF-1α levels of tumor negatively impact (111,112) the OS of patients; the associated mechanisms may include adenosine generation and accumulation, lactate accumulation and extracellular acidosis. Both adenosine accumulation and extracellular acidosis can block NK cell activation, proliferation and cytotoxicity (113,114), while lactate accumulation mainly inhibits the cytotoxic activity of NK cells and increases the number of MDSCs that inhibit NK cytotoxicity (115).

Clearly, various components of the TME affect the antitumor functional activity of NK cells in different ways during the progression of lung cancer. Among them, TGF-β is the main inhibitor of NK cell function. Notably, intratumoral NK cells may have a negative effect on other immune cells located in the TME after their own antitumor function decreases. For example, DC maturation was impaired due to the lack of IFN-γ secretion by NK cells and Tregs were profusely recruited through CCL22 secretion induced by NK cells (116,117). An important question is whether the modifications in the intratumoral NK phenotype and function are reversible. If they are, enhancing NK cell function with immuno-stimulatory cytokines such as IL-15 or by neutralization of immunosuppressive factors produced in the environment may improve the efficacy of NK cell-based immunotherapy and further ameliorate the clinical outcome of lung cancer.

Based on the important role of NK cells in tumor control, NK cell immunotherapy has developed rapidly. Several approaches have recently been proposed to boost NK cell antitumor function, to support in vivo persistence and homeostatic proliferation, and to promote homing to the tumor microenvironment (33) (Fig. 3).

In recent years, molecular targeted therapy and immune checkpoint inhibitor therapy have led to marked progress in the treatment of lung cancer. However, a considerable number of patients remain unresponsive to treatment, and the need for new treatment strategies is still urgent. Numerous studies have confirmed the critical role of NK cells in lung cancer control. Immunotherapy targeting NK cells may be an effective strategy for lung cancer treatment. The growing insight into the NK cell potential for lung cancer treatment provides a platform for the development of NK-based immunotherapy. However, numerous obstacles remain to be overcome to derive the full benefit of the NK cell antitumor potential. First, the poor ability of NK cells to reach tumor tissues limits their application in solid tumor therapy, which is a common problem with cellular immunotherapy strategies. As aforementioned, when NK cells are present in tumor tissues, they are preferentially localized in the matrix without coming into contact with the tumor cells. Second, changes in NK cell-activated receptors and ligands in tumors may result in decreased antitumor activity. Finally, the TME remains the main obstacle to the effectiveness of the adoptive transfer of NK cells. Despite these challenges, as more data are gathered on the lung cancer TME, immune regulatory cell populations, cancer-related changes in NK cell biology, function, and transport, NK cell immunotherapy will become increasingly effective. Key components to the success of future trials include the incorporation of modalities that harness NK cell cytotoxicity while promoting in vivo survival, homeostatic proliferation, and trafficking to the tumor and the development of drugs that trigger NK cell tumor killing via ADCC or sensitization of the target and drugs that promote NK cell tumor homing, such as the development of monoclonal antibody-chemokine fusion proteins to promote the infiltration of cytolytic NK cells into tumor tissues. Efforts should be made to solve the problems of clonal expansion and genetic modification of NK cells. Currently, several phase I and II clinical trials for the treatment of targeted NK cells for lung cancer are underway, including chemotherapy combined with NK cell adoptive transfer therapy (NCT03366064, NCT03410368), immune checkpoint inhibitors combined with NK cell adoptive transfer (NCT03958097), surgery combined with NK cell adoptive transfer therapy (NCT02843815), and CAR NK cell therapy (NCT03656705). In the future, combined standard radiotherapy, chemotherapy or radiochemotherapy, targeted therapy, ex vivo stimulation or CAR-NK cells and other targeted NK cell methods may eventually change the treatment mode of lung cancer, providing hope to patients with limited treatment options.