The role of NK cells in regulating tumorimmunity: current state, challenges and future strategies

Abstract

Natural killer (NK) cells are lymphocytes of the innate immune system. Their multifaceted cytotoxic mechanisms, coupled with their capacity to modulate immunity through cytokine secretion, underscore their pivotal role in orchestrating the immune microenvironment within tumors. This comprehensive review aims to elucidate the biological properties of NK cells, shedding light on their intricate involvement in tumor immunity. Furthermore, this review provides a comprehensive summary of the crosstalk between NK cell and other cells, and therapeutic strategies that target both T cells and NK cells, thereby offering potential avenues for therapeutic interventions. Finally, we have summarized the current status, advantages, and limitations of CAR-NK cell therapy, providing a reference for future CAR-NK cell engineering designs.

Introduction

Natural killer (NK) cells, which are characterized as large granular lymphocytes, play a critical role in the human immune system by swiftly recognizing and eliminating abnormal, virally infected, and tumor cells [1]. NK cells possess the remarkable ability to rapidly target and attack cells through cytotoxic mechanisms without prior sensitization, setting them apart from T and B cells [2]. Notably, NK cells have a relatively short lifespan in mammals, with a turnover rate of approximately two weeks among effector cells [2]. These cells are terminally differentiated and exhibit restricted transcriptional activity [3].

The progression of cancer is dependent on the evasion of immune system. It has been demonstrated that the TME fosters immunosuppression, hindering the presence of cytotoxic lymphocytes, including NK cells [4]. Unlike T cell activation, NK cell stimulation is governed by the interaction between their receptors and target cells, operating independently of antigen processing or presentation [5] (Fig. 1). Moreover, NK cell therapy typically causes fewer side effects. Compared to T cell therapy, NK cell therapy is less likely to induce severe side effects such as cytokine release syndrome. This unique characteristic has prompted intense investigation in the field of tumor research, shifting the focus from T lymphocytes to NK cells.

Fig. 1.

Interactions of Ligands and Receptors between NK Cells and Tumor Cells. On the left side of the figure are NK cells, and on the right side are tumor cells. Various ligands are present on the NK cells, and there are also various ligands on the tumor cells. The dashed line in the middle indicates that these ligands can bind to each other. Although we have labeled NKp46 and NKp44 in the figure, the killing dependent on NKp46 and/or NKp44 is not applicable in all experimental systems using the aforementioned candidate ligands. Therefore, the information regarding NKp46 and NKp44 here is for reference only

Since the discovery of NK cells in the 1970 s, there has been a growing body of research on these cells [6–11]. Preclinical investigations have demonstrated that NK cells can regulate tumor growth and inhibit metastatic dissemination [12]. The activity of NK cells can be enhanced by the administration of IL-2 in interferon stimulation deficiency [13]. In the 1980 s, clinical trials were conducted to explore the potential of NK cell-based immunotherapy through the adoptive transfer of autologous lymphokine-activated killer cells into patients with advanced cancers, resulting in significant tumor regression [14, 15]. The introduction of the inaugural chimeric antigen receptor (CAR), CD4 zeta, to NK cells in 1995 generated considerable excitement [16]. In 2002, the strength of the anti-leukemia effect of allogeneic NK cells in patients undergoing hematopoietic stem cell (HSC) transplantation for acute myeloid leukemia (AML) was found to directly influence their prognosis, leading to more favorable outcomes [17]. Subsequently, allogeneic NK cells were activated and administered to patients with AML, displaying significant anti-tumor responses [18]. In 2018, a human clinical trial was conducted to assess the safety of CAR NK-92 cells in treating refractory and relapsed AML. The results demonstrated that no significant adverse effects were observed even at doses up to 5×10⁹ (5 billion) cells per patient [19]. Recently, the feasibility of generating CAR NK cells from iPSCs expressing CAR for targeted antitumor immunotherapy against pre-existing lymphocytes, marked a significant milestone in this field [20]. In 2020, HLA-mismatched NK cells derived from cord blood and combined with CD19-CAR exhibited significant efficacy in treating relapsed or refractory CD19-positive lymphomas and leukemias, while maintaining low toxicity levels [21]. These studies suggest that CAR NK cells hold substantial potential for antitumor immunotherapy and offer innovative concepts for advancing immune cell adoptive therapy. Nevertheless, the efficacy of NK cell adoptive therapy in treating solid tumors has not met expectations, as evidenced by the results of studies [22–24]. The development of NK cell adoptive therapy is summarized in Fig. 2.

Fig. 2.

Key findings in research on NK cell in cancer; Number of publications on NK cell in cancer. The left side of the image presents key findings regarding NK cells in tumors, arranged in chronological order from top to bottom. The right side of the image shows the publication volume of NK cells in the tumor field over the years

The objective of this review is to present a comprehensive overview of the biological functions of NK cells, including their inherent heterogeneity. Furthermore, we delve into an exploration and summarization of the interactions between NK cells and other immune cell populations within the TME. Finally, we summarize the mechanisms that promote NK cell homing and the mechanisms and therapeutic strategies for simultaneously targeting both NK cells and T cells.

The roles of NK cells

Despite the significant successes in CAR-T cell therapy [25–30], its clinical applications have been limited to some patients due to stringent resource requirements and emerging clinical side effects (cytokine-release syndrome and neurotoxicities). This underscores the need for novel strategies. Refining the innate immune system to enhance the anti-tumor immune response presents an alluring avenue. Among innate lymphoid cells (ILC1s), NK cells, the predominant members, were discovered more than four decades ago for their unique ability to lyse tumor cells without prior sensitization [31, 32]. The multifaceted functions of NK cells have sparked inquiries into the existence of distinct Subsets with varying Functional profiles. In 1986, the conventional subset of human peripheral blood NK cells was first identified immunophenotypically based on differential expression levels of CD56 (encoded by NCAM) and CD16a (encoded by FCGR3A) [33, 34]. CD56^bright CD16^low/- NK cells, the immature subset of NK cells, maintain immunomodulatory and cytokine-producing functions, with limited cytotoxicity responses. These capabilities are achieved through their ability to extravasate from the peripheral blood into tissues and lymph nodes, along with enhanced expression of L-selectin (CD62L) and CC-chemokine receptor 7 (CCR7) [33–36]. CD56^dim CD16⁺ NK cells exhibit potent cytotoxicity against susceptible target cells and can mediate ADCC due to their expression of granzymes (GZMA, GZMB) and perforin (PRF1), and can also produce cytokines, growth factors and chemokines [33, 35–37]. However, in mice, the maturation stage of NK cells is determined by the expression levels of CD27 and CD11b. As the functionality of NK cells increases, their cell surface markers transition from CD27⁺CD11b^- to CD27⁺CD11b⁺ and ultimately to CD27^-CD11b⁺ [38–40]. However, it is important to note that both CD27 and CD11b antigens, in both humans and mice, lack specificity in identifying NK cell subpopulations. The quest to further characterize and understand the distinct NK cell subsets remains a significant focus of ongoing research.

The rapid advancement of single-cell sequencing has provided insights into the diversity of NK cells [41–49] (Table 1). The expression level of CD57 changes under continuous antigenic or inflammatory stimulation, leading to the differentiation of CD56^dimCD16⁺ NK cells into distinct cell subsets [50, 51]. Human peripheral blood NK cells can be classified into three subsets based on surface markers: CD56^brightCD16^-CD57^-, CD56^dimCD16⁺CD57^- and CD56^dimCD16⁺CD57⁺ [42]. CD56^dimCD16⁺CD57^- NK cells, a transitional phase of transcription, were generally consistent with the baseline trajectory of CD56^dimCD16⁺CD57⁺ NK cells. However, some CD56^dimCD16⁺CD57^- NK cells deviated from this main path due to abnormally high expression of IL-2R and NKG2A [42]. In a previous study, CD56^brightCD16^- NK cells were identified as an intermediary stage in the progression towards CD56^dimCD16⁺ NK cells [52]. Further investigations revealed that the majority of interactions between CD56^brightCD16^-CD57^- NK cells and CD56^dimCD16⁺CD57⁺ NK cells were facilitated by CD56^dimCD16⁺CD57^- NK cells [42]. Subsequently, different research teams have observed variability in the distribution of NK cell subsets based on their respective tissue locations [43]. CD56^brightCD16^- NK cells are the predominant resident populations in tonsils, lymph nodes, and the gastrointestinal tract, whereas CD56^dimCD16⁺ NK cells constitute the majority of cellular constituents in blood, bone marrow, spleen, and lung [43]. The degree of CD57 expression in CD56^dimCD16⁺ NK cells further provides information on the cellular subgroups. Specifically, CD56^dimCD16⁺CD57⁺ NK cells are predominantly located in blood, bone marrow, spleen, and lung, while CD56^dimCD16⁺CD57^- NK cells are mainly found in tonsils, lymph nodes, and the gastrointestinal tract [43]. Another investigation revealed the existence of two subpopulations, NK1 and NK2, which resemble CD56^dim and CD56^bright, respectively. The NK1 subgroup exhibits a multitude of highly expressed genes consistent with those found in CD56^dim, while the NK2 subgroup significantly overexpresses CD62L, a gene encoding homing receptor in humans [44]. Furthermore, these subgroups exhibit characteristics specific to the tissues in which they reside. A more refined subdivision of tissue-resident cells may yield novel insights into tissue identification. The ongoing advancements in single-cell analysis techniques are expected to unravel the intricate diversity and functions of NK cell subsets, shedding further light on their roles in immunotherapy and disease treatment. Indeed, the diversity of NK cells changes throughout an individual’s lifespan [41]. The distribution of specific key nodes in terminally differentiated NK cells, such as CD56^dimCD16⁺CD57⁺ NK cells, may also be influenced by age [43]. As individuals age, the composition and functional characteristics of NK cell subsets may undergo alterations, which could potentially impact their immune response and effectiveness in immunotherapy.

Table 1.

NK cell phenotypes in cancer

A conserved functional differentiation exists within the subpopulation of NK cells, which also exhibits tissue-specific cytotoxic potential [53]. Granzyme B, one of the primary cytolytic agents [54], was found to be highly expressed in CD56^dimCD16⁺ NK cells in lung tissue compared to other tissues. In contrast, no significant difference was observed in CD56^brightCD16^- NK cells [43]. This indicates that the cytotoxic potential of NK cells may vary based on their tissue localization.(Table 2)

Table 2.

Immunotherapeutic approaches to mobilize antitumour responses by NK cells

Tumor immune microenvironment: interactions between NK cells and other cells

The interactions between NK cells and other cells within the TME are crucial determinants of the immune response against cancer. NK cells are insufficient in number in various TME [55–58]. However, their function is influenced by the other cells, including cancer-associated fibroblasts (CAFs) [59], tumor-associated macrophages (TAMs) [60], dendritic cells (DCs) [61], and various immune cells [62, 63] (Fig. 3). These interactions are highly dynamic and can have both stimulatory and inhibitory effects on NK cell activity. For example, certain tumor-resident cells, such as CAFs and TAMs, may secrete factors that suppress NK cell function and hinder their ability to recognize and kill cancer cells [64, 65]. In contrast, DCs and certain activated immune cells, can promote NK cell activation and enhance their cytotoxicity against tumor cells [66]. Subsequently, we summarized the relevant clinical studies on NK cell therapy for tumors.(Table 3).

Fig. 3.

Interactions between NK cells and tumor-resident cells within the TME The figure shows the interactions of various immune cells in the tumor immune microenvironment. Among them, NK cells can directly bind to tumor cells through ligands to exert anti-cancer effects; they can also interact with other immune cells through cytokines to exert their functions

Table 3.

Clinical trails of tumor immunotherapeutic by using NK cells in recent years

Interactions between NK cells and cancer-associated fibroblasts

CAFs are found in abundance within solid tumors, exerting their influence on tumor progression and therapeutic resistance through the inhibition of ligand-mediated activity in NK cells [67, 68]. This inhibition is achieved through the production of TGFB1 [69–71], the secretion of PGE2, and the expression of IDO1(leading to the downregulation of NKG2D and NKp30 receptors) [72–74], or the reduction of PVR expression [75], thereby impeding the functionality of NK cells [76]. Furthermore, CAFs can facilitate the shedding of MICA and MICB from the surface of cancer cells, effectively suppressing NK cell activation [77, 78]. Another study reported that melanoma-associated fibroblasts (MAFs) contribute to a decrease in tumor cell susceptibility to NK cell-mediated killing. This effect is attributed to the secretion of matrix metalloproteinases (MMPs) by MAFs [77]. While CAFs may represent a promising target for enhancing NK cell activity within the solid TME, the lack of specificity and reliable markers for CAFs poses a significant obstacle to this approach.

Interactions between NK cells and platelets

The interaction between tumor cells and platelets is crucial for facilitating the successful dissemination of blood metastases. Platelets play a protective role by shielding tumor cells from both shear stress and NK cell attacks. Moreover, they contribute to the recruitment of bone marrow cells through the secretion of chemokines. These collective functions demonstrate the significance of platelets in supporting the process of blood metastases [79].

A study conducted in 1999 revealed that platelets impede the lysis of tumor cells by NK cells in mice [80]. In 2005, another study found that Galphaq is a G protein essential for platelet activation [81]. Interestingly, in mice with lifelong genetic defects in NK cells, a reduction in lung tumor cell survival was observed, but this decrease did not coincide with fibrinogen depletion [81–83]. These findings suggest that platelets and fibrinogen jointly protect tumor cells in blood vessels from being cleared by NK cells. MHC class I derived from platelets bestows a pseudonormal phenotype upon cancer cells, which effectively undermines the antitumor reactivity of NK cells [84]. The downregulation of MHC class I, a critical molecule necessary for initiating and activating CTLs, represents a commonly observed evasion mechanism in EBV-associated Nasopharyngeal Carcinoma (NPC) [85]. Subsequently, a positive correlation has been established between specific NK cell immune markers within tumors and a favorable prognosis in NPC. This finding strongly suggests the involvement of NK cells in the TME in effectively controlling the progression of EBV-associated gastric carcinoma [86]. EBV adopts a strategy to evade NK cell-mediated immune surveillance in both NPC and EBV-associated stomach carcinoma. This evasion mechanism involves involves the upregulation of F3, which promotes platelet activation, by EBV [87].

NKG2D is a receptor type present on the surface of NK cells and several other immune cells, including CD8⁺ T cells and γδ T cells. The shedding of NKG2D ligands by platelets impairs the immune surveillance of tumor cells by NK cells [88]. Platelet-derived RGS18, exerts a protective role for circulating tumor cells (CTCs) against immune surveillance by NK cells through its interaction with the immune checkpoint HLA-E: CD94-NKG2A. The disruption of this suppressive signaling pathway leads to the prevention of tumor metastasis in vivo, achieved by the immune-mediated elimination of CTCs [89].

Interleukin-15 (IL-15) is a crucial cytokine that plays a pivotal role in the immune system, particularly in the regulation and activation of NK cells [90]. The interaction between platelet-derived growth factor (PDGF) and its receptor-beta (PDGFRβ) is known to significantly influence cell growth and motility [91]. The combination of immunotherapy with N-803 (an IL-15 superagonist) and dinutuximab, in conjunction with ex vivo expanded NK cells, markedly enhances in vitro cytotoxicity against pediatric solid tumors expressing GD2 and improves the in vivo survival of xenografted immunodeficient NSG mice. This enhancement was accompanied by a significant reduction in tumor necrosis factor-related apoptosis-inducing ligand (TRAIL), PDGF-BB, and stem cell growth factor beta (SCGF-β) [92]. The PDGF-D-PDGFRβ signaling pathway serves as a mechanism through which IL-15 selectively regulates the survival of human NK cells without modulating their effector functions [93](Table 2)..

TGFβ has been demonstrated to influence platelet function and activation, as well as regulate NK cell activity. Platelets have been shown to release TGFβ, which can modulate NK cell responses [94]. The downregulation of TGFβ through PFTBA@HSA treatment has been shown to have significant effects on tumor cells. Specifically, it has been observed to impaire the epithelial-mesenchymal transition (EMT) process of tumor cells, leading to reduced migration and invasion capabilities. Furthermore, this downregulation enhanced the immune surveillance by NK cells, resulting in the inhibition of tumor metastasis [95].

Interactions between NK cells and regulatory T cells

Tregs are a specialized subset of T cells that play a critical role in maintaining immune tolerance and preventing excessive immune responses [96]. Targeting MARCO or IL-37 receptor (IL-37R) in lung cancer cell lines by using antibodies or CRISPR to knock out IL-37 to achieve TAM re-polarization [97]. Consequently, this repolarization led to the restoration of cytolytic activity and antitumor capacity in NK cells and T cells, while simultaneously downregulating Treg cell activity [97]. The same effect on immune cells is reproduced in the scenario of adenosine binding to the G protein-coupled receptor A2AR and the chemotherapeutic drug doxorubicin (DOX)-loaded phosphorus dendron micelles acting on tumor cells [98, 99]. In the field of engineering and technology, there is a fascinating development involving pathogen-mimicking hollow nanoparticles [100, 101]. These nanoparticles are designed to display mannan, a polysaccharide found in microbial cell walls that activates TH17 cells. Through this mechanism, these nanoparticles effectively restrict the fraction of regulatory T cells and elicit TH17-cell-mediated anti-tumor responses. Consequently, they contribute to increased proportions of CD8⁺ T cells, NK cells, and M1-like macrophages within the immune system [102]. Indeed, cytokines can modulate the activity of Tregs and NK cells. A particularly effective approach in tumor immunotherapy involves the combined use of the hydrophobic chemotherapeutic drug paclitaxel (PTX) and IL-12. This combination acts by activating T lymphocytes and NK cells, stimulating the release of IFN-γ, which selectively inhibits Tregs. Moreover, it induces the differentiation of tumor-associated macrophages into the M1-type, thereby enhancing the tumor microenvironment and reducing immunosuppression [103]. The convergence of tumor immunotherapy with the realm of engineering and technology represents a significant and promising avenue for research [98, 100, 101, 104]. Nevertheless, before these findings can be practically applied, further investigations through clinical studies are imperative.

Clinical studies have revealed a noteworthy observation regarding the TME. Specifically, increased infiltration of NK cells and natural killer T cells (NKT), along with a reduction in Tregs, appears to be associated with a favorable clinical response. Among the studied cell populations, CD16⁺CD56^-CD134⁺NK/NKT cells, and CD3⁺CD4⁺FOXP3⁺CD134⁺ Tregs exhibited the most significant changes in response to treatment. Conversely, CD3⁺CD8⁺granzyme B⁺PD-1⁺CD134⁺ cytotoxic T cells exhibited the least noticeable alterations [105]. The tumor-promoting influence of Tregs has been evident in hematologic disorders as well. This phenomenon has been demonstrated in models of multiple myeloma, where the absence of Tregs in vivo effectively halted the dissemination of the disease. Notably, the transient depletion of Tregs in mice with multiple myeloma led to a robust immune response orchestrated by CD8 T cells and NK cells. This response resulted in a state of complete and enduring remission [106].

The STAT3 pathway and TGFB1 secretion mechanisms, when engaged, result in the reduction of NK cell activating receptors, such as NKG2D, while elevating co-inhibitory receptors, namely PD-1 and IL-1R8. This process enables Treg cells to inhibit NK cell-mediated tumor immunotherapy [107–109]. Furthermore, studies have demonstrated that exposure to IL-2, IL-7, or IL-12 in conjunction with neutralization of the IL-1R8 ligand IL-37 enables NK cells to counteract the immunosuppressive effects of Treg cells [107, 110]. Despite the challenges associated with targeting the downregulation of Treg cells, there is considerable promise in this area.

Interactions between NK cells and cytotoxic T lymphocytes

Cytotoxic T Lymphocytes (CTLs), also known as CD8⁺ T cells, are key effector cells in the immune system, primarily responsible for eliminating infected cells, tumor cells, and other abnormal cells [111]. T cells and NK cells play complementary roles in tumor immunity, with NK cells mediating the clearance of CD8⁺ T cell-resistant tumors under the action of STING agonists [112]. Therefore, the Dual attack by both T cells and NK cells provides an opportunity to deepen the impact of immunotherapy. Additionally, the Toll-like receptor 5 agonist entolimod inhibits metastasis and induces immunity by stimulating the NK-DC-CD8⁺ T cell axis [113]. Further studies have shown that NK cells play a crucial role in recruiting DC cells to tumors [114, 115], thereby enhancing the induction of CD8⁺ T cell responses [116–118], while IL-2 secreted by T cells activates NK cells [119]. However, in a mouse model of chronic HBV infection, NK cells limit the therapeutic vaccine-induced CD8⁺ T cell immunity in a PD-L1 dependent manner [120]. Subsequently, scholars discovered that in the submandibular gland, NK cells inhibit the activation of cytotoxic CD8⁺ T cells through the PD-1-PD-L1 [121]. Whether these results can be applied to tumors remains unknown. If NK cells in the TME have dual roles towards CD8⁺ T cells, then research on the different NK cell subsets with bidirectional effects will become an emerging hotspot.

Interactions between NK cells and dendritic cells

In the TME, NK cells play a vital role by exerting cytotoxic effects on tumor cells and facilitating the recruitment of dendritic cells (DCs) to infiltrate into the tumor site. This collaboration between NK cells and DCs leads to the amplification of the T-cell tumor immune response, creating a more robust and coordinated anti-tumor immune reaction [114, 115].

The conventional DC population can be divided into cDC1 and cDC2 [122]. The cDC1 plays a crucial role in promoting antitumor immunity by presenting cell-associated antigens derived from apoptotic tumor cells to CD8 and CD4 T cells [116, 117]. cDC1 cells are distinguished by their selective expression of CLEC9A and XCR1 receptors, which bind XCL1 and XCL2 chemokines [123, 124]. Upon uptake of apoptotic tumor cells, cDC1 cells translocate tumor cell-associated antigens to the tumor-draining lymph nodes, where they subsequently transfer the antigens to other dendritic cells [116, 125]. Furthermore, cDC1 cells can recruit and activate tumor-specific CD8 T cells, thereby contributing significantly to antitumor immune responses [125]. Numerous studies have demonstrated that the presence of cDC1 is essential for the generation of protective antitumor immunity. This has been corroborated through the observation that Batf3 knockout mice, which lack cDC1 cells, are unable to mount effective antitumor immune responses [117, 126–128].

cDC1 has the potential to advance cancer immunotherapies and enhance existing treatments [114]. NK cells play a pivotal role in producing chemoattractants, such as CCL5 and XCL1/2, which facilitate cDC1 accumulation during the early stages of tumor formation and bolster tumor immune control [114]. However, this axis can be disrupted by tumor-derived PGE2, resulting in compromised NK cell function and downregulation of XCR1 and CCR5 chemokine receptors on cDC1 [114, 129, 130]. Notably, the mechanisms governing cDC1 accumulation in human tumors closely resemble those observed in mice, further reinforced by a positive correlation between cDC1 accumulation and improved patient outcomes across various cancer types [114]. At this juncture, it would appear that prioritizing the XCL1/XCL2-XCR1 axis over the CCL5 axis for guiding cDC1 infiltration into the tumor is a more favorable approach. This strategy serves to mitigate the recruitment of pro-tumorigenic immune cells, including macrophages and regulatory T cells, associated with CCL5 signaling. cDC1 in human melanoma correlates with intratumoral NK cell levels, and both innate immune cell types correlate with responsiveness to anti-PD-1 immunotherapy. In human melanoma, the abundance of cDC1 is positively correlated with the expression of the gene encoding the cytokine FLT3LG within the tumor. FLT3LG is primarily produced by lymphocytes, with a notable contribution from NK cells [115](Fig. 4).

Fig. 4.

NK Cells Facilitate Recruitment of Dendritic Cells into Tumor Microenvironment. NK cells establish cellular microenvironments alongside cDC1 within neoplastic lesions. Cross-presenting cDC1 exhibit the XCR1 chemokine receptor, facilitating their attraction towards NK cells that secrete XCL1 and XCL2. Additionally, NK cells release FLT3L, which triggers the differentiation of dendritic cells. The recruitment of cDC1 by NK cells plays a crucial role in initiating the activation of CD8 and CD4 T cells in response to cell-associated antigens. cDC1 efficiently capture apoptotic fragments derived from tumor cells undergoing cell death and transport these antigenic materials to the lymph nodes that drain the tumor site. Within these lymph nodes, cDC1 effectively stimulate the activation of CD8 and CD4 T cells. Consequently, NK cells actively contribute to the initiation of T cell-mediated immune responses against tumors

Interactions between NK cells and neutrophils

The role of neutrophils in solid tumors is a topic of ongoing debate [131]. Keisuke et al. found that neutrophils acquired a pro-carcinogenic phenotype in NK cell-depleted mice [132]. Interestingly, in the mouse model, neutrophils inhibit tumour metastasis when NK cells are absent and activate tumour metastasis in the presence of NK cells, suggesting that the abundance of host NK cells governs the dual role of neutrophils [133]. Subsequent studies have identified opposing functions of neutrophils in regulating metastasis in mice with normal NK cell function versus those lacking NK cells [133]. This regulation is mediated through reactive oxygen species (ROS) [133]. The dual identity of neutrophils is attributed to their heterogeneity and functional plasticity in different environments; they can change dynamically to adapt to different environments.

Neutrophil can secrete various factors, including neutrophil extracellular traps (NETs). Multiple studies have identified an association between the ROS signaling pathway and NETs [134–137]. These NETs protect tumor cells from the cytotoxic effects exerted by CD8⁺ T cells and NK cells [138]. This result was subsequently translated in engineering techniques for the treatment of hepatocellular carcinoma (HCC). Enhancement of NK cell therapy by combining a tumour acid neutraliser and NET lyase in a hydrogel to prevent postoperative recurrence of HCC [139]. Similar to NETs in their release pathway, extracellular vesicles (originating from neutrophils) have been shown to bolster the production of anti-inflammatory cytokines within NK cells [140]. Furthermore, neutrophils also orchestrate the responsiveness of NK cells by secreting cathepsin G [141] and ROS [142], thereby modulating NKp46. Neutrophils additionally mitigate NK cell infiltration by downregulating the expression of NKp46 and NKG2D through the suppression of CCR1 [143]. In summary, neutrophils suppress NK cell activity to facilitate the evasion of tumor cells from NK cell cytotoxicity [144].

Interactions between NK cells and macrophages

A significant number of human tumor lesions lack NK cells and instead exhibit a high infiltration of monocytic origin macrophages (mo-macs) [145–147]. These mo-macs are believed to possess immunosuppressive properties [148, 149]. However, targeting mo-macs has proven to be challenging, primarily due to a limited understanding of the heterogeneity in macrophage states [150]. TREM2 macrophages diminish NK cell activity through the modulation of IL-18/IL-18bp decoy interactions and the production of IL-15 [65]. MARCO⁺ macrophages were found to enhance Treg cell proliferation and IL10 production while reducing CD8⁺ T cell activity. To restore the immune response in lung cancer, targeting MARCO or the IL-37 receptor (IL-37R) in lung cancer cell lines showed promising results. These interventions repolarized TAM, resulting in restored cytolytic activity and antitumor capacity of NK and T cells, and concurrently downregulated Treg cell activity [97]. Intratumoral SIRPα-deficient macrophages trigger the activation of cytotoxic T cells targeting tumor antigens upon exposure to radiotherapy. This activation subsequently engenders a tumor-killing microenvironment characterized by a robust infiltration of tumor-specific cytotoxic T cells, NK cells, and inflammatory neutrophils [151]. Therefore, the targeting of macrophages represents a potential avenue for therapeutic intervention, although further investigation is required in clinical studies.

Macrophage phagocytosis plays a crucial role in the elimination of tumors [152], while the activation of inflammasomes induces adaptive immunity against tumors [153]. Antibody-dependent cellular phagocytosis (ADCP) can lead to the overexpression of PD-L1 and IDO in immunosuppressed macrophages, which is mediated through inflammasome activation. This mechanism inhibits both ADCC in NK cells and the immune response in T cells, thus dampening the anti-tumor effects of cancer cell phagocytosis and inflammasome activation mediated by macrophages during antibody therapy [154].

Development of immunotherapies for cancer that involve both T cells and NK cells

Tumor-specific CD8⁺ T cells fulfill their crucial role by detecting and eliminating cancer cells through the recognition of tumor cell-derived peptides presented by MHC-I proteins via the TCR [155]. Upon activation, these T cells secrete IFN-γ, which sensitizes tumor cells to T-cell attack by instigating the up-regulation of numerous genes involved in the MHC-I antigen presentation pathway. This, in turn, increases the density of TCR ligands on the cell surface [156]. CD8 T cells exhibit remarkable sensitivity in recognizing target cells. Even the recognition of a minimal number of peptide-MHC ligands, often just one or a few, is sufficient to trigger T-cell activation [157, 158]. Nonetheless, the potency of this recognition system hinges entirely upon the continued expression of MHC-I by tumor cell proteins. Tumor cells that lose MHC-I expression, whether through mutation or epigenetic mechanisms, effectively render themselves invisible to CD8 T cells [159, 160]. In cases where cancer cells evade CD8 T cell recognition, NK cells come into play. Notably, the deletion of MHC-I expression renders tumor cells more susceptible to NK cell-mediated cytotoxicity, as several MHC-I proteins serve as ligands for inhibitory receptors on NK cells [161]. As previously stated, NK cells play a pivotal role in recruiting DC cells to the TME, thereby initiating T cell responses [114, 115]. These lymphocyte populations employ disparate recognition strategies when identifying cancer cells, making this combined approach highly effective (Table 4).

Table 4.

Inhibitory receptors shared by T cells and NK cells

Concurrent receptor targeting in T and NK cells for immunotherapeutic strategies

Numerous studies have consistently identified a significant degree of overlap within the receptor-ligand systems of both NK cells and CD8 T cells [162–165]. Early studies identified CD161 (encoded by the KLRB1 gene) as a receptor capable of inhibiting NK cell function [166]. Subsequent research has highlighted the significance of CD161 as an inhibitory receptor on tumor-infiltrating T cells [162]. CLEC2D serves as a ligand for the C-type lectin receptor CD161, exhibiting high expression levels in various tumors, particularly in hematological malignancies [162, 167] (Fig. 5A). Elevated expression levels of KLRB1 were observed in tumor-infiltrating T cells across diverse cancer types, including glioblastoma, hepatocellular carcinoma, colorectal cancer, and melanoma [162, 163]. Within glioblastoma, the KLRB1 gene exhibits heightened expression in clonally expanded CD8 T cells when compared to their non-expanded counterparts [162]. This implies an upregulation of KLRB1 expression concomitant with the proliferation of CD8 T cells upon antigen recognition. Therefore, we posit that exploring the role of CD161-CLEC2D in solid tumors holds promise as a fruitful avenue for further investigation.

Fig. 5.

A: CD161-CLEC2D pathway inhibits T cells B: Precision therapy targeting the NKG2D-MICA/B pathwayThe NKG2D receptor specifically binds to stress-induced ligands MICA/B present on tumor cells. However, tumors have developed mechanisms to evade immune attacks mediated by NKG2D. One such evasion strategy involves the proteolytic shedding of MICA/B from the tumor cell surface. This shedding process is orchestrated by the coordinated action of the ERp5 disulfide isomerase and the proteases ADAM10 and ADAM17. A monoclonal antibody targeting the α3 domain of MICA/B has been developed. By inhibiting the proteolytic shedding of MICA/B, this antibody effectively promotes an NK cell-mediated immune response against metastatic tumors. C: Microbiota triggers monocyte reprogramming of the TME The symbiotic gut microbiota plays a beneficial role by generating specific molecules that activate the STING-IFN-I axis in monocytes. As a result, a subset of these monocytes undergoes differentiation into macrophages, which exhibit anti-tumor effects. Simultaneously, another fraction of the microbiota-derived molecules stimulates the activation of NK cells, leading to the secretion of XCL1 and CCL5. These chemokines, in turn, promote the activation of DCs. The activated DCs establish a binding interaction with NK cells through the IL15-IL15R pathway, thereby enhancing NK cell function and overall anti-tumor immune responses. D: Prospective Avenues of Investigation In the future, research efforts will be directed towards investigating the precise functions and interactions of DC cells, NK cells, and T cells within the TME. This will involve studying how these immune cell populations communicate and collaborate to mount effective anti-tumor immune responses. Additionally, there will be a focus on exploring the distinct subsets of NK cells and their unique characteristics and roles within the TME. Understanding the intricacies of these immune cell populations and their interplay will provide valuable insights for the development of novel therapeutic strategies against tumors

The NKG2A/CD94 receptor is ubiquitously expressed in both circulating and tumor-infiltrating NK cells, as well as in CD8 T cells that specifically bind to HLA-E, a nonclassical MHC-Ib molecule [164]. Consequently, a reduction in surface HLA-E levels can result in the transcriptional downregulation of MHC-I genes, rendering these cells susceptible to NK cell-mediated cytotoxicity [168, 169]. While only some CD8 T cells express NKG2A/CD94, another subset of tumor-infiltrating CD8 T cells, characterized by co-expression of PD-1 and other inhibitory receptors, can upregulate the expression of NKG2A/CD94 [170]. In addition to elevated HLA-E expression on antigen-presenting cells within tumors, the inhibitory CD94/NKG2A ligand on TILs is also elevated in larger tumors. This phenomenon may elucidate why previous DC vaccines targeting tumor DCs presenting tumor neoantigens on MHC-Ia did not yield significant improvements in murine cancer models and cancer patient outcomes [171–173]. The immune checkpoint interaction between HLA-E and CD94-NKG2A mediates the escape of circulating tumor cells from NK cell surveillance [89, 169]. Additionally, the interaction between HLA-E and CD94-NKG2A inhibits the signaling of activated MHC-like receptors, leading to impaired initiation of antitumor responses and insufficient T cell activation [174]. Mouse tumor models targeting CD94/NKG2A have been extensively studied. In a mouse model of head and neck squamous cell carcinoma (HNSCC), the use of NKG2A-blocking monoclonal antibodies augmented the efficacy of peptide-based vaccines [175]. In an interim analysis of monalizumab (a humanized NKG2A-blocking antibody) used in conjunction with cetuximab for the treatment of HNSCC, the combination therapy yielded partial remission in 31% of patients according to response evaluation criteria in solid tumors (RECIST) [176]. Additionally, combining monalizumab with durvalumab, a PD-1 monoclonal antibody, improved remission rates, and extended progression-free survival compared to durvalumab monotherapy [177]. These findings warrant Further investigation in phase 3 clinical trials.

The CD226 receptor serves as a crucial costimulatory receptor expressed by both NK cells and CD8 T cells [165, 178] and is closely linked to enhanced tumor growth during antitumor immune responses. It has several ligands, including CD155 and PVRL2 (CD112). Notably, the activating function of the CD226 receptor is counteracted by three inhibitory receptors: TIGIT, CD96, and PVRIG. Interestingly, both TIGIT and CD96 exhibit a higher affinity for the shared CD155 ligand when compared to CD226, making them superior to CD226 in terms of ligand binding efficiency [179]. TIGIT exhibits high expression levels on tumor-infiltrating CD8 T cells in various solid tumor types and lymphomas, serving as a crucial inhibitory receptor [180]. Additionally, it plays a pivotal role as an inhibitory receptor on NK cells, contributing to the dysfunction of NK cells within TME [181]. Notably, in non-small cell lung cancer (NSCLC) patients undergoing PD-L1 monotherapy, elevated CD226 mRNA levels were correlated with extended overall survival. This observation helps elucidate why the combination therapy of TIGIT and anti-PD-L1 exhibits a higher remission rate in NSCLC patients compared to anti-PD-L1 monotherapy, as demonstrated in clinical trials [182]. Currently, this approach is under evaluation in a Phase III clinical trial for NSCLC patients. Furthermore, extensive research has unveiled distinct mechanisms between PD-1 and TIGIT in inhibiting CD226. PD-1 achieves this by dephosphorylating Tyr-322 within the cytoplasmic domain of CD226, whereas TIGIT surpasses CD226 in terms of its binding affinity to CD155 [183]. CD96, along with the PVRIG receptor (with the ligand PVRL2), has received comparatively less attention in research h, despite being expressed in both T and NK cells [184, 185]. In mouse models, the deactivation of the CD96 receptor on CD8 T cells or NK cells has been shown to enhance anti-tumor immunity [184, 186]. However, its specific role in human cells remains largely unexplored. Currently, a phase I clinical trial involving a blocking monoclonal antibody targeting PVRL2 is in progress [179].

The NKG2D receptor, characterized by ligands such as MICA/B and ULBP1-6 proteins in humans, Rae1α-ε, H60a-c, and Mult1 in mice, is prominently expressed on various immune cell subsets in humans, including NK cells, CD8 T cells, and intrinsic T cells, such as NKT cells and γδ T cells [187, 188]. This receptor is instrumental in inducing perforin-dependent cytolysis within NK cells, γδ T cells, and NKT cells, mediated through the junction protein DAP10 [189]. Furthermore, it provides essential co-stimulatory signals to CD8 T cells. This receptor-ligand system assumes significant importance due to the upregulation of ligands in response to DNA damage and cGAS-STING signaling, a phenomenon seldom observed in healthy cells [190, 191] (Fig. 5B). This family of ligands is induced by cellular stress and transformation. Essentially, the upregulation of NKG2D ligands serves as a signal for the elimination of stressed cells by cytotoxic lymphocytes [192]. Shed MICA, in particular, has been strongly associated with the progression of numerous human cancers but is conspicuously absent in the serum of healthy individuals [193]. The proteolytic shedding of MICA/B represents an immune evasion strategy, significantly reducing the density of stimulatory NKG2D ligands on the surface of tumor cells and triggering the internalization of the NKG2D receptor on cytotoxic lymphocytes [194, 195]. Mechanistically, the disulfide isomerase ERp5 plays a pivotal role in initiating shedding by disrupting the structural disulfide bond in the MICA/B α3 structural domain, rendering it susceptible to hydrolysis by ADAM-10/17 and MMP14 proteins [194] (Fig. 5C). Administration of a MICA/B monoclonal antibody, which targets the α3 structural domain and inhibits hydrolytic shedding, substantially enhances the surface density of NKG2D ligands and leads to their destruction by human NK cells [195]. In a mouse metastasis model, this monoclonal antibody significantly augments the therapeutic efficacy by exploiting its high-affinity binding to Fc receptors, including the CD16 receptor on NK cells, thereby activating NK cells via NKG2D and CD16 receptors [195]. Importantly, this monoclonal antibody demonstrates activity against metastatic tumors that exhibit resistance to cytotoxic T cells due to deficient MHC-I expression [196]. In addition to antibodies, cancer vaccines targeting the MICA/B α3 structural domain involved in protein hydrolytic shedding have been actively investigated [197]. These vaccines induce distinct populations of T and NK cells within mouse tumors, resulting in an increased density of stimulatory MICA/B proteins on the tumor cell surface. Surgical removal of highly metastatic primary tumors is followed by a substantial reduction in late metastatic growth following immunization with this vaccine [197]. Remarkably, this vaccine remains effective against CD8 T-cell tumors, even in cases of mutational inactivation of genes in the MHC-I presentation pathway (B2m) or IFN-γ receptor signaling (Ifngr1). Furthermore, tumor-binding MICA/B antibodies enhance the cross-presentation of tumor antigens by DCs to CD8 T cells and bolster NK cell-mediated killing of tumor cells through the activation of the CD16 Fc receptor on NK cells [197]. The therapeutic approach involving NK cells necessitates not only the activation of NK cells through the receptor-ligand system but also relies on a robust CD4 T-cell response, effectively recruiting NK cells into the TME.

Targeting cytokines and small molecules-sparrow’s nest

TGF-β, being a prominent immunosuppressive cytokine, not only robustly suppresses T-cell proliferation and effector functions but also hinders NK-cell proliferation and cytotoxicity by inhibiting the mTOR pathway or multiple activation receptors on NK cells’ surface, such as the NKG2D receptor [198–202]. Nonetheless, targeting the TGF-β pathway in tumor immunity presents challenges due to its pleiotropic nature. TGF-β exists in substantial quantities as an inactive precursor within the extracellular matrix, and its activation from this latent complex depends on integrin αv receptors through a force-dependent mechanism. Among these integrins, αvβ6 and αvβ8 exhibit the highest affinity for the latent complex, efficiently releasing active TGF-β [203, 204]. However, integrin β6 is expressed at low levels in healthy epithelial cells but markedly elevated in epithelial carcinoma cells. This correlates with reduced tumor aggressiveness and enhanced survival. Consequently, the targeting of the integrin αvβ6 receptor to inhibit TGF-β has gained prominence as a strategy for treating epithelial cancer. The use of an integrin αvβ6/8 mAb significantly amplified T cell-mediated cytotoxicity against human and murine triple-negative breast cancer cells in vitro. Furthermore, it sensitized these cells to PD-1 inhibition by inducing a substantial influx of CD8 T cells in two murine models of aggressive triple-negative breast cancer [205, 206]. Another study found that immature NK cells promote the progression of triple-negative breast cancer [207]. Among the various TGF-β isoforms, TGFB1 exhibits the highest mRNA expression levels across various human tumor types. Combining TGFB1 targeting with PD-1 mAb yielded a significant survival advantage in a mouse tumor model refractory to ICB, affirming the therapeutic potential of TGFB1 targeting [208].

NK cell therapy for hematologic tumors - significant efficacy

Recent studies have found that NK cell functional deficiencies are associated with the occurrence and progression of various hematologic malignancies [209–214]. In the progression of multiple myeloma (MM), there is a reduction in NK cell numbers and impaired cytotoxicity [209, 210]. In T-cell lymphomas, defects in NK cell differentiation are linked to MYC-induced tumorigenesis [211]. Additionally, NK cell functional deficiencies have also been reported in T-cell and B-cell acute lymphoblastic leukemia (T-ALL) [212], chronic myeloid leukemia (CML) [213], chronic lymphocytic leukemia [214], acute myeloid leukemia (AML) [215], and classical Hodgkin lymphoma [216].

In preclinical studies, it was found that FT555, as a GRPC5D CAR-NK cell, has a longer persistence when used in combination with anti-CD38 mAb (Daratumumab) compared to the use of CAR-NK cells alone in a MM mouse model [217]. Another study also indicated that BCMA/GRPC5D dual CAR-NK cells exhibit higher antitumor efficacy in the MM model [218]. The CD33 domain plays a critical role in CAR-NK therapy. Blocking NKG2A significantly enhances the antileukemia efficacy of CD33 CAR NK cells in cytotoxicity assays and AML mouse models [219]. By using a trispecific killing engager (TriKE) that combines IL-15 and the CD16 Fc receptor with the CD33 binding domain, in conjunction with α3 MICA/B CAR-NK cells, further control of leukemia progression can be achieved [220]. However, using α3 MICA/B CAR-NK cells alone in vitro can lead to disease progression [220]. Additionally, CD33/FLT3 CAR-NK cells demonstrated good antileukemia efficacy in animal models, killing over 90% of leukemia cells, and the addition of mucin-1 inhibitory CAR in in vitro experiments can protect approximately 42% of healthy primary human HSCs and HPCs from cytotoxicity [221]. In preclinical studies, CD123 CAR-NK cells exhibited lower toxicity compared to CD123 CAR-T cells, yet their efficacy in AML mouse models was comparable [222].

In a Phase I clinical trial of CD33 CAR-NK cell therapy derived from human umbilical cord blood for patients with relapsed or refractory AML [223], results indicated that among the 10 evaluated patients, only 1 experienced grade 2 CRS, with no higher-grade CRS reported. No cases of immune effector cell-associated neurotoxicity syndrome (ICANS) were observed at any grade. All patients who experienced grade 3-4 bone marrow Suppression recovered within one month. Regarding anti-leukemia efficacy, 60% (6/10) of the patients achieved complete remission (CR) 28 days after CAR-NK cell infusion [223]. Another Phase I trial of BCMA CAR-NK cell therapy derived from induced pluripotent stem cells (iPSCs) for MM showed that among the 9 patients receiving CAR-NK cell infusion (3 of whom received daratumumab as a combination therapy), no CRS or ICANS were observed. One patient receiving 300 million BCMA CAR-NK cells as monotherapy achieved a very good partial response (VGPR). Two patients achieved remission after receiving 100 million BCMA CAR-NK cells in conjunction with daratumumab [224]. Although there was some efficacy, challenges remain due to amplification conditions, transduction efficiency, and anti-tumor efficacy. Future optimization of CAR-NK cell therapy will require genetic modification methods, preconditioning regimens, cell dosing, combination immunotherapies, or hematopoietic stem cell transplantation to facilitate the clinical translation of CAR-NK cells.

NK cell therapy into solid tumors -- a new generation of anticancer weapon

Cancer immunomodulation entails the orchestration of immune cells to sculpt the immunogenicity of burgeoning tumors, thus thwarting the incipience of cancer. Schreiber et al. pioneered an immunomodulatory paradigm that underscores the proactive role of NK cells in combating early metastatic dissemination through hematogenous spread, rather than during the more mature stages of tumor clonality [225]. Consequently, despite their sparse presence within solid tumors, the distinctive anti-tumor effects of NK cells, inclusive of their MHC-unrestricted cytotoxicity, cytokine production, and immunological memory, position them as pivotal entities within the innate and adaptive immune response apparatus. NK cell therapy finds widespread application across various malignancies, including but not limited to pulmonary carcinoma [226–230], acute myeloid leukemia (AML) [231–233], ovarian neoplasms [234, 235], melanoma [236, 237], breast cancer [238, 239], gastric carcinoma [240], and gastrointestinal stromal tumors [241, 242] exhibiting commendable therapeutic efficacy across these investigative endeavors.

NK cell therapy efficacy in solid tumor patients correlates closely with intratumoral NK cell abundance [243]. The preceding discussion elucidated immune dysregulation targeting NK cell activity (tumor-resident cells), hence obviating further elaboration. Here, we primarily encapsulate mechanisms about impaired NK cell trafficking. NK cell trafficking and homing are intricately linked with type I IFN response [244]. Post type I IFN response, secretion of CXCL9, CXCL10, and CXCL11 ensues, fostering binding to NK cell surface receptor CXCR3, thereby inducing heightened infiltration of NK cells [245–247]. Consequently, augmenting type I IFN response may lead to increased recruitment of NK cells within solid tumors. Furthermore, therapy involving type I IFN or IFNG can promote the secretion of CXCR3 ligands. Additionally, augmented intratumoral NK cell recruitment correlates with various chemokines. Abundant secretion of CXCL10 and CCL5 fosters continual accumulation of intratumoral NK cells, enhancing CAR-NK therapeutic efficacy [248]. Overexpression of CXCR4 [249], CCR7 phagocytic activity [250–252], or interaction between CX3CR1 and CX3CL1 [253] can facilitates NK cell chemotaxis into solid tumors. However, the complete functionality of these factors remains elusive. For instance, overexpression of CX3CL1 correlates with adverse prognosis in cancer patients [254]. Thus, strategies to elevate levels of these factors necessitate careful consideration. For example, while increased secretion of CXCL16 and CXCL8 leads to infiltration of NK cells into tumors [255–257], CXCL8 correlates with immune evasion, immunotherapy resistance, and tumor progression [258, 259].

Discussion, concluding marks and Key questions

Development status, advantages, and limitations of CAR-NK cell therapy

Development status of CAR-NK cell therapy

Currently, there are no approved drugs for NK cell therapy. However, the number of clinical studies in this field has surged, with most being in Phase I or Phase II and still ongoing, thus not fully reporting results yet [260]. Early outcomes, however, suggest that CAR-NK cell therapy is safe. CAR-NK therapy targets a variety of antigens and malignancies, but a significant number of CAR-NK trials focus on CD19 for hematologic cancers, a target that has seen success with CAR-T therapies [21, 261].

Advantages of CAR-NK cell therapy

(1) High safety, reduced cytokine release syndrome and neurotoxicity

CAR-T cell activation often leads to the release of inflammatory cytokines, including IL-1, IL-2, IL-6, TNF-α, MCP1, IL-8, IL-10, and IL-15 [262–267]. Among these cytokines, particularly pro-inflammatory TNF-α, IL-1, and IL-6, are responsible for causing CRS (Cytokine Release Syndrome) and neurotoxicity [268]. However, CAR-NK cells release different types of cytokines (IFN-γ and GM-CSF) [269]. In fact, patients injected with CAR-NK cells did not experience CRS or neurotoxicity, and no increase in inflammatory cytokines (such as IL-1β, IL-6, and IL-10) was detected above baseline levels [18, 270–272]. This difference in cytokine release profile is considered one of the reasons for the enhanced safety of NK cells, reduced toxicity, and decreased occurrence of cytokine storms. It is important to note that these results are based on a limited number of clinical studies with CAR-NK cells [19, 261, 273], involving a small number of patients, and there has been no direct comparison with CAR-T therapy. Additionally, recent reports have observed T-cell malignancies in a small number of patients receiving CAR-T therapy [274]. Although the incidence is low (22 cases among more than 27,000 treated patients reported to date), such events have been observed with five CAR-T cell therapy products (out of a total of six approved by the FDA). In three of these cases, the cancer cells expressed the CAR. This prompted the FDA to add a class-wide black box warning to these therapies [274]. This indicates that, in terms of these severe adverse events, CAR-NK cells may be a safer alternative to CAR-T cells. Although CAR-NK cells have better safety profiles (lower risk of CRS), their durability of efficacy is insufficient (NK cells Survive in the body for only about 2 weeks) (Table 5). Most clinical studies are in phase I/II, and the data scale is limited. Therefore, CAR-NK therapy still has a long way to go.

Table 5.

CAR-NK vs. CAR-T therapies(safety, efficacy, cost)

(2) Multiple tumor recognition mechanisms

CAR-NK cells can recognize and eliminate cancer cells through both antigen-dependent and antigen-independent mechanisms, which may provide significant advantages in the treatment of solid tumors compared to CAR-T cells [275]. The inherent antigen-independent mechanisms of NK cells could be beneficial for tumors with fine subclonal heterogeneity [276]. Additionally, NK cells can mediate ADCC through CD16, which binds to the Fc domain of IgG bound to cancer cells. This further enhances their potential efficacy against solid tumors [277].

Limitations of CAR-NK cell therapy

Similar to CAR-T cells, CAR-NK cells have poor tumor infiltration [278, 279] and harmful interactions with the TME [280]. Additionally, NK cells exhibit resistance to viral transduction [281]. Recent clinical trials have indicated that the median Transduction efficiency of retroviral vectors producing CAR-NK cells is 72.4% (Ranging from 22.7% to 91.1%) [261]. This efficiency can impact the statistical power of controlled Phase II or III clinical trials. Therefore, strategies must be further explored to address the antiviral responses triggered by lentiviruses in NK cells, which can induce apoptosis [282]. Pre-treating primary human NK cells with combinations of cytokines such as IL-2, IL-12, and IL-21 [282–284], or using the TBK1 inhibitor BX759 to interfere with downstream signaling of pattern recognition receptors [282, 285], can increase lentiviral Transduction efficiency to 25−40% [282, 285]. However, the process of optimizing the design of CAR-NK cells still requires further refinement. Moreover, NK cells have a short half-life [286] and poor persistence [287]. This necessitates multiple infusions, which not only increases toxicity risk but also doubles the financial burden on patients. Although several early trials have suggested that multiple injections of CAR-NK cells do not result in severe adverse events, the sample size has been limited [19, 273], and further trials are needed to evaluate safety. Reprogramming CAR-NK cells to endow them with memory or memory-like characteristics is a feasible and necessary approach. This research direction holds promise for addressing the limitations related to the persistence of CAR-NK cell therapy.

Summary and future directions

This article reviews the characteristics of NK cells, their interactions with other immune cells within the TME, and their shared receptors with T cells. Finally, it summarizes the advantages and limitations of CAR-NK cell therapy, from NK cells to CAR-NK cells. The limited infiltration of NK cells into tumor sites, their short half-life, poor persistence in vivo, and low transduction efficiency due to viral resistance are all areas that warrant research attention. To address the issue of limited infiltration into tumor sites, strategies such as local infusion [288], integrating chemokine receptors into the CAR-T domain [289], or degrading the extracellular matrix (ECM) to enhance penetration [289] could be explored. For the immunosuppressive microenvironment, solutions could include converting inhibitory signals into immune stimulatory signals [290], combining with immune checkpoint inhibitors (ICIs), and blocking inhibitory signals [291]. Future investigations will delve deeper into the roles of immune cell cohorts, particularly dendritic cells, NK cells, and T cells, within the specific milieu of the TME, spanning both preclinical and clinical studies (Fig. 5D). To overcome the short half-life and poor persistence in vivo, approaches such as using fully humanized scFv [292], engineering active domains into the CAR structure [293], employing memory-like cells or stem cells [294, 295], and cytokine stimulation [296] could be considered. The low transduction efficiency due to viral resistance may be addressed through cytokine stimulation [282, 283] or inhibitor interference with downstream signaling [282, 285]. The use of CAR-NK therapy is still in its early stages, but it has demonstrated strong antitumor activity in preclinical and early clinical trials across various cancer types. Effective non-viral engineering strategies for NK cells need to be developed to improve the manufacturing process of CAR-NK therapy, supporting the increasing demand for advanced therapies. So far, an ideal non-viral delivery system that can efficiently deliver CAR transgenes into NK cells without compromising cell viability or receptor status has not been developed; considering the importance of receptor status for NK cell function. Further research into new delivery systems and clinical trials is necessary to fully realize the therapeutic potential of CAR-NK cells and translate these findings into routine clinical practice.

Author contributions

WXZ: Conceptualization, Writing – original draft, Writing – review & editing. CF: Software, Writing – original draft. YQZ: Investigation, Methodology. WCL: Data curation. JBN: Investigation. DS: Data curation. ZXY: Conceptualization, Funding acquisition, Project administration. WCZ: Conceptualization, Funding acquisition, Writing – review & editing.All authors reviewed the manuscript.

Funding

This work was financially supported by the following funding: (1) National Natural Science Foundation of China [grant number 82260555]; (2) Medical Innovation and Development Project of Lanzhou University [grant number lzuyxcx-2022-177]; (3) Major Science and Technology Projects of Gansu Province [grant number 22ZD6FA021-4]; (4) Science and Technology Program of Gansu Province [grant number 23JRRA0996].

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.