NK cells (natural killer cells) are important effector cells of the immune system responsible for anti-tumor and anti-viral resistance, derived from bone marrow hematopoietic stem cells and matured through a multi-step differentiation process (31, 32). Human NK cells are classified into two major subpopulations based on the expression of CD16 and CD56. The majority (85%-95%) of NK cells in peripheral blood are of the CD56-CD16+ subpopulation, displaying a developmentally mature phenotype and possessing a high degree of cytotoxicity (33, 34). Unlike T cells, which depend on antigen presentation, NK cells are capable of lysing without the need for antigen presentation, and are able to pass through lysates containing granzyme B and perforin. Cell particles containing granzyme B and perforin cross the immune synapse to mediate sequential killing of infected or malignant cells (35).

In contrast, the CD56+CD16- subpopulation has fewer NK cells and exhibits an immature phenotype. Although the latter have low cytotoxicity when not activated, they are capable of producing large amounts of cytokines and exerting potent immunomodulatory effects when stimulated by pro-inflammatory cytokines (e.g., IL-15) (36, 37). The two subpopulations complement each other functionally, with the former being primarily responsible for the direct killing effect, and the latter playing an important role in immune response via cytokine secretion.

Upon maturation, NK cells have a potent killing capacity, although this cytotoxicity is not associated with major histocompatibility complex (MHC) expression because NK cells do not express somatically rearranged antigen receptors (38, 39). Instead, NK cells regulate NK cell activity through a balance of their activating (e.g., NKG2D, NKp46) and inhibitory (e.g., KIR family) receptors, thereby killing or causing tolerance in target cells (40, 41).

CD16 (low affinity IgG Fc region receptor III, FcγRIII), the most potent activating receptor expressed by NK cells and the only receptor that can activate NK cells on its own, can assist antibody-mediated immune responses through the antibody-dependent cell-mediated cytotoxicity (ADCC) pathway (42–44). The Fc portion of the antibody was engineered to increase its affinity for CD16a and enhance the ADCC effect. For example, replacing four amino acids in the Fc region to become a GASDALIE mutant significantly enhanced the affinity of Fc for CD16a, while the affinity for CD32b remained almost unchanged (45).

NKG2D protein is an important activating cell surface receptor protein, which is mainly expressed on cytotoxic immune cells, such as NK cells, CD8+ T cells, etc (46, 47). The first known NKG2D receptor is MICA with MICB, which is expressed in a wide variety of tumors (liver cancer, breast cancer, lung cancer) and organ transplant recipient MIC is expressed in tissue cells (48–51). MICA exhibits a very low tumor mutation burden, suggesting that its expression is not significantly affected by DNA editing mechanisms, so NKG2DL overexpression may be a potent strategy for anti-tumor progression (52, 53).

NK cells are distributed in high concentrations in the liver, accounting for 50% of hepatic innate immune cells, making the liver one of the major sites of NK cell residency (31, 54, 55). This feature arises from the fact that embryonic hepatic hematopoietic stem cells are divided into two fractions, one of which continues to remain in the adult liver, generating the characteristic tissue-resident NK cells (LrNK) (56).

Compared to peripheral blood, NK cells in the liver differ in both effector molecule expression and cellular activity. The CD56+CD16- subpopulation is predominant in the liver (33). LrNK cells have a suppressive function in the immune tolerance microenvironment of the liver, particularly in inhibiting the antiviral response of T cells through the PD-1/PD-L1 pathway. For example, Zhou et al. found that exogenous transfusion of LrNK cells to normal or LrNK cell-deficient mice suppressed antiviral T-cell responses in the liver and was dependent on the PD-1-PD-L1 axis. In contrast, NK (cNK) cells circulating in the peripheral blood promoted T-cell responses (57).

TIM-3 is highly expressed in the depleted state of NK cells and acts as a co-inhibitory receptor involved in the modulation of type I immune responses. The immunomodulatory mechanism of TIM-3 is dependent on its binding to several ligands, such as Galectin-9, phosphatidylserine (PtdSer), HMGB1 and CEACAM-1 (81). PtdSer acts as an “eat-me” signal that promotes the clearance of apoptotic cells by binding to TIM-3. HMGB1, a damage-associated molecular pattern (DAMP), regulates the innate immune response by suppressing the inflammatory response when it binds to TIM-3. Binding of CEACAM-1 is thought to be closely related to inhibitory signaling by TIM-3.

These ligands, including CEACAM-1, Galectin-9, PtdSer, and HMGB1, bind to different regions of TIM-3, respectively, triggering intracellular inhibitory signals (82). For example, upon binding of TIM-3 to Galectin-9, the Y256 and Y263 sites in its cytoplasmic domain are phosphorylated, leading to dissociation of the articulator BAT3 from TIM-3, which in turn inhibits TCR signaling and reduces NK cell immune response, especially reducing the secretion of key cytokines such as IFN-γ (82). Phosphorylation of the Y256 and Y263 sites is not only a critical step in TIM-3-regulated signaling, but also promotes the activation of other inhibitory signals by preventing the binding of BAT3 to TIM-3.

The co-inhibitory receptor, TIGIT, blocks the direct interaction of NK cells with tumor cells by binding to the ligands CD155 and CD112, which are highly expressed in antigen-presenting cells (APCs) and tumor cells, diminishing their killing ability and further inhibiting cytokine secretion by NK cells, such as TNF-α and IFN-γ (83, 84). In addition, TIGIT interferes with tumor recognition by NK cells through competitive inhibition of CD226. Although TIGIT shares the same ligand as CD226, it binds to CD155 and CD112 with higher affinity, thereby inhibiting CD226-mediated activation signaling. This competitive mechanism further exacerbates the depleted state of NK cells (84).

LAG-3 is a structurally similar inhibitory receptor to CD4 that inhibits NK cell activation mainly through binding to MHC class II molecules (85). LAG-3 is highly expressed not only in T cells but also upregulated on NK cells, and this upregulation leads to suppression of both innate and adaptive immune functions in tumor patients.

LAG-3 hinders effective recognition and clearance of tumor cells by NK cells through interaction with FGL1 (fibrinogen-like protein), which is highly expressed in hepatocytes and tumor cells. In addition, LAG-3 regulates downstream molecules (e.g., SHP-1 and SHP-2) through inhibitory signaling motifs (e.g., FXXL motifs and KIEELE motifs) in its intracellular structural domains, further blocking the activation signaling pathway of NK cells by dephosphorylating activating signaling molecules (26, 86).

Co-upregulation of these inhibitory receptors is particularly evident in response to chronic antigenic stimulation, especially in NKG2C+ NK cells, where the expression of LAG-3 and PD-1 rises progressively over time (87).

Immune checkpoint inhibitors (ICIs), especially PD-1/PD-L1 inhibitors, have demonstrated significant clinical efficacy in the treatment of a variety of solid tumors, such as lung, breast, advanced hepatocellular, and pancreatic cancers (6, 88–92). These inhibitors work by blocking the binding of PD-1 to its ligand PD-L1, restoring the anti-tumor activity of NK cells and T cells, and increasing their cytotoxicity and secretion of immune factors such as IFN-γ and TNF-α. In different types of tumors, including HCC, lung cancer, and melanoma, NK cells have shown variable responses to PD-1/PD-L1 inhibitors, influenced by the tumor microenvironment and the extent of PD-1 expression on NK cells. It has been found that not all of the PD-1 in NK cells is derived from endogenous expression, and that NK cells also acquire PD-1 and other inhibitory substances from the membrane of the tumor cells through SLAM receptor-mediated trogocytosis. NK cells can also acquire inhibitory molecules such as PD-1 from the membrane of tumor cells through SLAM receptor-mediated trogocytosis (93). This process results in the suppression of the anti-tumor function of NK cells, which can be reversed by PD-1 inhibitors.

Although PD-1/PD-L1 inhibitors show good single-agent efficacy in some tumors, single-agent efficacy is typically lower in metastatic tumors of the hepatobiliary system, and patients experience initial resistance or subsequent decreased drug sensitivity. Therefore, investigators are exploring further enhancement of the anti-tumor effects of NK cells through combination therapies (6, 94–96). The combined blockade of PD-1/PD-L1 inhibitors with other inhibitory pathways has shown promising potential in reversing immune exhaustion during hepatocellular carcinoma treatment. TIM-3, is highly expressed in the TME of HCC, especially on tumor-infiltrating NK cells (e.g., cNK and LrNK cells). Studies have shown that TIM-3, through binding to its ligand phosphatidylserine (PtdSer), induces inhibition of downstream signaling pathways such as PI3K/mTOR, which in turn leads to dysregulation of NK cells and tumor evasion of the immune (6). Through gene ablation, antibody blockade, or lentiviral-mediated TIM-3 disruption experiments, the researchers succeeded in restoring NK cells’ cytokine secretion (e.g., IFN-γ, TNF-α) and cytotoxicity, significantly inhibiting HCC growth (88).

IL-15 and IL-2, two of the most widely studied cytokines, are able to enhance anti-tumor efficacy by stimulating NK cell proliferation and enhancing their effector functions. IL-15 significantly increases NK cell cytotoxicity by activating the downstream JAK/STAT signaling pathway through binding to IL-15Rα (97, 98). IL-2 is also capable of enhancing NK cell activation through the CD25 receptor, but is less used due to its tendency to induce proliferation of Tregs. Currently, researchers are developing modified versions of IL-2, such as mutant IL-2 with selective activation of NK cells, to avoid the side effects of Tregs (99).