In the past few decades, there has been a rapid development in tumor immunotherapy, which has been recognized as a key strategy to control the progression of malignant tumors (1). Among these immunotherapies, immune checkpoint inhibitors (ICIs) (2), chimeric antigen receptor T cells (3) and bispecific antibodies (4) are the most promising immunotherapeutic strategies. ICI therapy is one of the most well-studied immunotherapies (2,5). Inhibitors block the transmission of inhibitory signals, stimulate the activation of cytotoxic T lymphocytes (CTL) and induce the antitumor effects of T lymphocytes. Two ICIs have been investigated, including cytotoxic T lymphocyte antigen 4 (CTLA-4) (6) and programmed cell death 1/cell death 1 ligand, (PD-1/PD-L1) (7). However, the response rate of anti-PD-1/PD-L1 monoclonal antibodies (mAbs) and anti-CTLA-4mAb in patients with cancer overall is poor (8,9). Most patients show primary or acquired resistance to these ICIs (10,11). As a member of the ICI family, LAG-3 is an inhibitory molecule with multiple biological effects on the function of T cells (12). LAG-3 is highly expressed in various types of tumor infiltrating lymphocytes (TILs) and participates in the immune escape mechanism of tumors (12). Therefore, LAG-3 could be used as an indicator of tumor prognosis and a target of tumor therapy. The present review primarily describes the research progress of LAG-3 in immunotherapy.

LAG-3 is a member of the immunoglobulin superfamily, which was identified in 1990 (13). LAG-3, also known as CD223, has a molecular weight of 70kDa and is located on human chromosome 12 (12p13) (14). The LAG-3 gene is located in the same region as the CD4 gene, and there is a certain homology between the two (15); however, they share<20% homology at the protein level (13). Similar to CD4, LAG-3 binds to major histocompatibility complex-II (MHC-II) on antigen presenting cells (APCs), but with a much stronger affinity (13). The LAG-3 gene contains 8 exons, and the corresponding cDNA encodes a membrane protein, with 498 amino acids (16). LAG-3 is composed of an extracellular region, a transmembrane region and an intracellular region (17). The membrane protein contains four extracellular immunoglobulin superfamily domains: One V region and three C2 regions (16,17). The V region is special, as there is an extra ring in the middle of the domain and also contains an abnormal in-chain disulfide bridge (18). The cytoplasmic region of LAG-3 is composed of three parts: A serine phosphorylation site, a ‘KIEELE’ motif and a glutamate-proline dipeptide repeat sequence (EP sequence), among which the ‘KIEELE’ motif is a highly conserved sequence that has not been found in other proteins and is involved in intracellular signal transduction of the LAG-3 molecule (17,19) (Fig. 1). It was initially hypothesized that the EP sequence was the key basis for the downstream signal of LAG-3. However, the result of mutation analysis found that this was not the case. The ‘KIEELE’ motif in the LAG-3 molecule is conserved (17), and molecules that lack this domain are unable to exert an inhibitory effect on T cells, suggesting that this motif may be associated with the downstream inhibitory signal of LAG-3 (17). However, the mechanism of downstream signal transduction is still unclear. Under the action of the membrane-penetrating metalloproteinases disintegrin and metalloproteinase domain-containing protein (ADAM) 10 and ADAM17, the LAG-3 molecule on the cell membrane splits into two parts at the junction peptide: Soluble LAG-3(sLAG-3) and a transmembrane-cytoplasm molecule (20). sLAG-3 has been found to induce maturation of dendritic cells and target tumor cells (21). Previous studies have discovered five primary ligands of LAG-3. In a quantitative cell adhesion experiment, COS-7 cells transfected with LAG-3 and DaudiB lymphocytes labeled with ⁵¹Chromium were co-cultured together, and it was found, for the first time, that a garland was formed between the two cells, which was specifically dependent on LAG-3 and MHC-II molecules (14,22). Anti-LAG-3 or HLA-DR Ab have been found to destroy the formation of this garland, confirming that the ligand of LAG-3 is the MHC II protein and combines with a conservative extension ring in the LAG-3 D1 domain (14,22). LAG-3, once combined with MHC-II, transmits inhibitory signals via the cytoplasmic domain; however, the signal transduction mechanism is still unclear (17). Galectin-3 (Gal-3), a 31-kDa galactose-binding lectin that regulates T cell activation and immunoprecipitation, was found to interact with LAG-3 and inhibits the secretion of interferon-γ by CD8⁺T cells in vitro, indicating that Gal-3 is also an LAG-3 ligand (23). Gal-3 can be expressed on the surface of different cancers, such as lung cancer (24), prostate cancer (25), colorectal cancer (26), breast cancer (27) and so on. Thus it exerting its regulatory function on CD8⁺T cells via multiple mechanisms (28). LSECtin has also been proposed to be a LAG-3 ligand. LSECtin binds to the four glycosylated sites on LAG-3 and is a member of the DC-SIGN family of molecules (28). LSECtin is expressed in the liver and in melanoma tumor cells, suggesting a mechanism by which LAG-3 can regulate CD8⁺T cells and natural killer (NK) cell function in these environments (29). α-synuclein (α-syn) fibrils is a protein aggregation, which exists in the substantia nigra of the brain in patients with paralysis tremor and is a member of the Syn fibrils family (28–30). α-syn was found to bind to LAG-3, which leads to the intercellular delivery of pathological α-syn fibrils, while blocking their combination with a LAG-3 Ab could significantly reduce the toxicity and the intercellular delivery of pathological α-syn fibrils, suggesting that α-syn fibrils were also a LAG-3 ligand (31). Fibrinogen-like protein 1(FGL1) was recently identified as the primary ligand of LAG-3, and a genome-scale reporter array identified that the FGL1 protein was tightly bound to the LAG-3 receptor (32). Functional detection of the FGL1 protein revealed that upon binding to the LAG-3 receptor on the surface of T cells, T cell proliferation was inhibited and immune activity was also affected (33). Blocking the interaction between FGL1 and LAG-3 could enhance the antitumor effect of T lymphocytes, which has important significance in the research of tumor immunotherapy (32). LAG-3 is expressed on activated CD4⁺ (34) and CD8⁺T cells (35), regulatory T cells (Tregs) (36), a subpopulation of NK cells (37), B cells (38) and plasmacytoid dendritic cells (39). As an inhibitory molecule expressed on the surface of lymphocytes, LAG-3 is involved in regulating the proliferation and activation of T lymphocytes and effector T lymphocytes, and plays a specific role in maintaining environmental stability in vivo (36). The process of T cell inactivation and death are present both in cancer and chronic infection (40). As a co-inhibitory receptor of PD-1, LAG-3 is highly expressed in chronic virus infection and various tumors (41). The high expression of LAG-3 is also associated with autoimmune diseases, tumors and chronic toxic infectious diseases (18,42–44).

Immunosuppressive receptor molecules play an important role in the maintenance of immune homeostasis. When T lymphocytes are activated to a certain extent, immunosuppressive molecules, such as LAG-3, CTLA-4 and PD-1 are expressed to maintain the immune response in a stable state (45). The molecule LAG-3 blocks the signal transduction pathway of T cell activation; however, the intracellular segment of the LAG-3 molecule produces immunosuppressive signals, which have been found to regulate CD4⁺T cell activity (46). LAG-3 regulates the immune response of T cells in three ways: Firstly, it directly inhibits the proliferation and activation of T cells via negative regulation of T cells. Secondly, it can promote the inhibitory function of Tregs, and the T cell response can then be indirectly inhibited. Thirdly, it can prevent T cell activation by regulating the function of APCs (47). A LAG-3 knockout mouse model showed increased numbers of CD4⁺T and CD8⁺T cells in the compared with wild-type mice; however, there were no significant changes in the ratio of CD4⁺T and CD8⁺T cells (48). Additionally, studies also found that LAG-3mAbs or LAG-3 knockout cells did not significantly alter the apoptosis of T cells. However, the proportion of S phase cells increased significantly, suggesting that the inhibitory effect of LAG-3 on T cell proliferation may be by controlling the cell cycle rather than through apoptosis (49). Therefore, LAG-3 plays a key role in regulating immune homeostasis. Under normal physiological conditions, naive CD8⁺T cells are expressed at low levels (48). LAG-3 expression was significantly increased when tumor antigen cells were stimulated (50). LAG-3 has also been shown to bind to tumor infiltrating lymphocytes (TILs) in multiple solid tumors, and TILs are highly expressed in human solid tumors (51–53). LAG-3 negatively regulates TILs and anti-LAG-3 Ab can enhance antitumor immunity (54). Woo et al (55) found that LAG-3 and PD-1 were expressed in CD4⁺ and CD8⁺TIL in mouse cancer models and can enhance the cellular immune response of CD8⁺T cells by co-inhibiting LAG-3 and PD-1. Surprisingly, despite the high affinity, only a handful of residues located in the D1 loop of LAG-3 are involved in MHC-II and LAG-3 binding, in contrast to the extensive molecular interaction between MHC-II and CD4 (22). In 2006, Casati et al (56) found that sLAG-3 binds to MHC-II and mediates APC activation, thereby activating and promoting the production and proliferation of CD8⁺T lymphocytes. This indicated that sLAG-3 could compete with LAG-3 molecules to bind MHC-II and thus inhibiting LAG-3 from exerting a negative regulatory role on the proliferation and activation of T cells. Blocking with an inhibitor or knockdown of the LAG-3 gene has been found to accelerate the progression of autoimmune diseases in several animal models of autoimmune disease (17). The progression of diabetes was found to accelerate in mice lacking the LAG-3 gene in an obese diabetic mouse model (57). The infiltration of CD4⁺T and CD8⁺T cells around the islet cells in the mice increased significantly, accelerated the destruction of islet cells and all of the mice eventually developed diabetes (57). LAG-3 was also found to be involved in immune homeostasis and autoimmune diseases with other immunosuppressive molecules, such as Tim-3, TIGIT (58).

Typically, LAG-3 is expressed on the surface of activated T and NK cells to maintain immune response homeostasis and prevent the occurrence of immune overreaction or autoimmune diseases (22). However, in the tumor microenvironment (TME) or chronic infectious diseases, the continuous activation of T cells leads to co-expression of immunosuppressive molecules, such as LAG-3, PD-1, T-cell immunoglobulin-and mucin domain-containing (Tim)-3, TIGIT and CD160, resulting in the incapacitant and even apoptosis of T cells, which is the ‘deceptive mechanism’ by which tumors and chronic infectious diseases escape from being killed by the immune system (59). On the surface of activated CD4⁺and CD8⁺T cells, LAG-3 is physically integrated with CD3 (60) and then specifically binds to the CD3-TCR complex after TCR participation acting as a co-receptor for negative signals (61).

The immune system plays an important role in removing malignant cells (62). ICs are inhibitory signals that exist in the immune system which continuously regulates the immune response intensity of peripheral tissues in the normal healthy body and maintains the tolerance of autoantigens to prevent damage to tissues (63). However, immune or tumor cells in the TME highly express inhibitory IC co-stimulatory molecules, leading to the depletion and disability of antitumor T cells, which play an important role in tumor immune escape (63). LAG-3 expression is increased in TILs and in numerous types of cancer (64). In NSCLC, LAG-3 expression has been observed in TILs and studies have shown that its expression is associated with poor prognosis (65). LAG-3 negatively regulates TILs and anti-LAG-3 Ab was found to enhance antitumor immune function (54). LAG-3 and PD-1 are co-expressed on CD4⁺ and CD8⁺TILs, and when LAG-3 and PD-1 are jointly inhibited, the immune response of tumor CD8⁺T cells is enhanced. Blocking the LAG-3 receptor with an inhibitor can be used in combination with antitumor vaccines to improve the activation of tumor-specific CD8⁺T cells. Notably, this effect does not involve CD4⁺T cells, but rather by LAG-3 directly regulating CD8⁺T cells (48). The co-expression of LAG-3 and PD-1 in inactivated and dying T cells has been observed in human ovarian cancer tissues (66). In a mouse model of ovarian cancer, blocking of LAG-3 and PD-1 was found to promote tumor antigen specific CD8⁺T cells to produce cytokines, and joint blocking of LAG-3 and PD-1 in the treatment of ovarian cancer was more effective compared with immune inhibitors and blocking of a single molecule (66). LAG-3 expression in CD4⁺ and CD25⁻T cells in non-small cell lung cancer was found to be upregulated and significantly associated with poor prognosis (67,68). LAG-3 expression on the surface of CD8⁺T cells in TIL was significantly increased in patients with hepatocellular carcinoma compared with on peripheral blood lymphocytes (64). The mRNA expression levels of LAG-3 molecules was increased on the CD8⁺T cell surface of TILs and led to the functional defects of CD8⁺T cells (69). LAG-3 is also highly expressed in Treg cells. LAG-3⁺FoxP3⁺Treg cells have been shown to be highly expressed in the tumor tissues of peripheral blood mononuclear cells (including lymphocyte and dendritic cells), TIL and in patients with melanoma and with colorectal cancer (70). These LAG-3⁺Treg cells show an activation phenotype, producing interleukin (IL)-10 and transforming growth factor (TGF)-β1, and LAG-3⁺CD49b⁺cells are associated with poor prognosis in colorectal cancer (70). LAG-3 expression on Tregs induced IL-10 and TGF-β1 production (71). This indicated that LAG-3⁺Treg cells have an immunosuppressive effect and are associated with tumor immune escape (72). A recent study found that co-culture of DLBCL cell lines with primed T cells in the presence of anti-LAG-3 and anti-TIM-3Ab induced potent dose-dependent increases in in vitro cell death using AcellaTox and IL-2 ELISA assays, suggesting potent antitumor activity of these compounds (73). Woo et al (55) established B16 melanoma, MC38 colorectal adenoma and Sa1N fibrosarcoma mouse models and found that most tumors in groups with double-blocking of LAG-3/PD-1 disappeared. The results of flow cytometry analysis showed that the number of CD4⁺ and CD8⁺T cells in the group of mice with the combined treatment increased significantly. Wierz et al found that (74) in a mouse model of chronic lymphocytic leukemia, double blocking of PD-1 and LAG-3 significantly decreased the number and percentage of chronic lymphocytic leukemia cells in the blood and spleens of mice. Additionally, the serum levels of IL-2, IL-23 and TNF-α in mice with double blocking of PD-1 and LAG-3 was also increased, along with the immune response, proving that the dual target of PD-1 and LAG-3 could control the increase in the yield of chronic leukemia cells in mice. Huang et al (75) randomly assigned ovarian cancer-bearing C57BL/6 mice into different groups and treated them with anti-PD-1, anti-LAG-3 or anti-PD-1/LAG-3 Ab treatments. The results showed that the double blocking of LAG-3 and PD-1 inhibited tumor growth, significantly increased the number of CD4⁺ and CD8⁺TILs and enhanced the antitumor immune response. LAG-3 is often co-expressed with PD-1 on exhausted T cells, and targeting of both using anti-LAG-3 and PD-1Ab was effective at reinvigorating T cells and eliminating mouse tumors (74,76). Accumulating evidence suggested that multiple ICIs could optimize antitumor immunity (63,77,78). However, the mechanism by which PD-1 and LAG-3 pathways operate together to inhibit T-cell functions is not known. The Yale Cancer Center found that the protein FGL1 plays an important role in LAG-3 immunosuppression (32). Inhibition or knockdown of FGL1/LAG-3 promoted the effectiveness of T cells against tumors in the body (32). Blocking this pathway could be used in conjunction with anti-PD-1/PD-L1 therapy to achieve antitumor efficacy (32).

LAG-3 expression in the TME has also been associated with increased tumor mutational burden. For example, cancers with high microsatellite instability (MSI^hi), which includes a subset of patients with colorectal cancer, exhibited higher somatic mutations and higher levels of immunogenic neoantigens. The TME of these MSI^hi tumors are characterized by increased expression of multiple IRs including LAG-3, PD-1, PD-L1, CTLA4 and indoleamine 2,3-dioxygenase (IDO) compared with in microsatellite stable tumors (79). These data may suggest an association between tumor mutational burden, antitumor immune response and the co-expression of LAG-3 and other IRs. In fact, the presence of LAG-3 and other co-expressed IRs in the TME may explain why MSI^hi tumors are not naturally eliminated, despite a hostile immune microenvironment (79).

ICIs have innovated the treatment of cancer, and the combination of LAG-3 and PD-1 drugs has been effective (54). The corresponding clinical trials of LAG-3 and its effects on tumors have been performed in the USA, Australia, etc. (54,80). Until 2019, one LAG-3 fusion protein and 11 LAG-3 inhibitors were in clinical trials or are being used as anticancer therapies (Table I). A phase I clinical trial using LAG-3 as an immunotarget for cancer is under way, including a single center trial for hematological malignancies (clinical trial no. NCT02061761). In addition, LAG-3 specific Abs have been used in the combination of anti-PD-L1 and anti-LAG-3 therapy for solid tumors (clinical trial no. NCT01968109). However, there are still numerous problems to be solved in the investigation of LAG-3. Firstly, the biological function of LAG-3 binding to ligands is still unclear, and the specific mechanism of its negative regulation of T cell function is also still unknown. Therefore, further investigation is required. Secondly, there might be other potential ligands of LAG-3 that could be used as biological markers for the diagnosis and evaluation of prognosis. Thirdly, whether LAG-3 can play an antitumor role by regulating NK cells or B cells remains to be further investigated. Fourthly, how the molecular mechanism of different ICIs plays a co-role is not clear, and it is still important to identify the optimal IC combination. Lastly, whether modern advanced biological technology can be used to optimize the molecular structure of LAG-3 inhibitors and reduce production costs and to ensure improvements in a clinical setting remains to be determined. Taken together, LAG-3 has the potential to target molecules involved in biological functions and play a therapeutic role. There are numerous unknown links in the specific mechanism, and more depth understanding is required, such as ensuring LAG-3 inhibitors can be applied to clinical patients earlier for more beneficial effects, such as improved survival. The immune checkpoints of LAG-3 play crucial roles in cancer development and may be used in future clinical practice for cancer therapy. Along with the progress of additional and further observations, a broad variety of scientific questions will emerge and require to be addressed by scientists and physicians.