PD-1, a member of the B7-CD28 receptor family, is a transmembrane protein and widely expressed in B cells, T cells, natural killer (NK) cells, and myeloid cells (7). As the ligand of PD-1, programmed cell death ligand 1 (PD-L1) can be expressed in dendritic cells (DCs), macrophages, T cells, NK cells (8, 9), and tumor cells (10). Generally, when PD-L1 binds to PD-1 in the presence of the T cell receptor (TCR) signaling complex, PD-1 delivers a co-inhibitory signal, leading to the termination of TCR signaling and inhibition of T cell proliferation (11). PD-1 often uses mono-tyrosine signaling motifs which present in its cytoplasmic tail, such as immunoreceptor tyrosine-based inhibitory motif (ITIM) and immunoreceptor tyrosine-based switch motif (ITSM) (12), to end the CD28/TCR signal by PD-1 phosphorylation and the recruitment of SHP-2 and SHP-1 (13–15). In the tumor immune context, antigen-presenting cells (APCs) and tumor cells highly express PD-L1, and they can interact with PD-1-overexpressed T cells, leading to T-cell anergy or exhaustion (16, 17). Programmed cell death ligand 2 (PD-L2) is the second ligand for the PD-1 molecule, which is expressed predominantly by DCs, macrophages, B cells, and cancer cell populations, depending on microenvironmental stimulation (18, 19). Similar to PD-L1, PD-L2 plays a crucial role in evading antitumor immunity. The engagement of PD-1 and PD-L2 can lead to the downregulation of T cell responses, which inhibits TCR-mediated proliferation and cytokine production by CD4+ T cells by blocking cell cycle progression (18). Although PD-1/PD-L2 blockade must be considered for optimal immunotherapy in antitumor immunity (20), since most of the research results are focused on the PD-1/PD-L1 pathway, we mainly discuss the PD-1/PD-L1 axis in this article.

As surface molecules, the activity of PD-1 and PD-L1 can be easily inhibited by blocking antibodies. Anti-PD-1 therapy is one of the most successful immune checkpoint blockade therapies that have been approved to treat a wide variety of cancer types ( Table 1 ). PD-1 inhibitors competitively bind to PD-1 and block PD-1/PD-L1 interactions, which subsequently regulate negative signals on the T cell surface to enhance the functions of effector T cells and promote the proliferation of T cells (54). Nivolumab and pembrolizumab are the primary clinically approved PD-1 inhibitors. They are humanized IgG4 antibodies targeting PD-1 with high affinity (55). To ensure that they elicit their inhibitory effects of PD-1/PD-L1 interactions primarily by direct occupancy and steric blockade of the PD-L1-binding site of PD-1 (56), they minimize the function of effector cells engaging other antibodies.

Pembrolizumab was initially approved for refractory unresectable melanoma in 2014 (57), known as the first PD-1-targeted therapy to gain Food and Drug Administration (FDA) approval. Not long, in 2015, it becomes the first immune checkpoint inhibitor to be approved as a first-line treatment, also in melanoma therapy (21). Pembrolizumab is thus approved to treat a wide variety of cancer types. To date, pembrolizumab therapy has been licensed in many cancers (27, 30, 58, 59) and was often conducted primarily in patients with PD-L1-positive disease (31, 34). In general, a higher level of PD-L1 expression is associated with a more effective clinic outcome of pembrolizumab. However, in some cancer types, such as non-small cell lung cancer (NSCLC) (60), classical Hodgkin’s lymphoma (cHL) (25), and urothelial carcinoma (UC) (61), PD-L1 expression did not explicitly correlate with response to pembrolizumab.

Nivolumab also displays a good response and favorable safety profile, particularly in melanoma and NSCLC. Nivolumab was approved by FDA following its showing a clear advantage in response over chemotherapy in refractory unresectable melanoma (62). Soon after, the FDA approved nivolumab for the treatment of NSCLC after progression on a platinum-based chemotherapy regimen (41, 63). Also, nivolumab has been demonstrated durable effects in other cancers (47, 64, 65), and it appears that combination therapy may further improve them (50, 66). Nevertheless, research has demonstrated a low response rate in some hematological tumors, such as follicular lymphoma (FL) (67) and diffuse large B cell lymphoma (DLBCL) (68). It may appear to correlate positively with 9p24.1 translocation and increased PD-L1 expression (69). In addition to nivolumab and pembrolizumab, cemiplimab is also approved by FDA for the treatment of advanced cutaneous squamous cell carcinoma (70) and first-line NSCLC (53). Up till now, more than 1,500 clinical trials involving PD-1 inhibition are currently supported by the National Cancer Institute (NCI).

NK cells can spontaneously kill cells and thus are presumed to be key innate immune effectors in cancer immunosurveillance; it belongs to the family of innate lymphoid cells (ILCs) (98). IFN-γ produced by NK cells during early-phase immune responses can directly kill tumor cells and promote the differentiation of naive CD4+ T cells toward Th1 cells to facilitate cell-mediated immunity (99). Thus, NK cells are critical components both in humoral immunity and in cellular immunity.

As an inhibitory receptor, PD-1 can express on NK cells (100, 101) and prevent the activation of NK cell function when engaging with its ligand which is expressed on the surface of target tumor cells or APC (102). PD-1⁺ NK cells may be inhibited in killing tumor cells instead of being anergic in PD-L1⁺ tumors, which means that PD-1 is an important checkpoint for NK activation and PD-1 blockade might elicit an antitumor NK cell response (102). In high PD-L1 expression head and neck cancer (HNC) patients, the study observed that PD-1 blockade increased cetuximab-mediated NK cell activation and cytotoxicity (103). Besides, tumors might drive the development of PD-L1-expressing NK cells that acquire immunoregulatory functions; such cell population can directly inhibit CD8+ T cell proliferation in a PD-L1-dependent manner (104). These results show the importance of the PD-1/PD-L1 axis in inhibiting NK cell responses in vivo, and future research is needed to determine the specific mechanism of the PD-1 pathway in the antitumor response of NK cells.

DCs, known as specialized APC, transport tumor antigens to draining lymph nodes and cross-present antigens via MHC I and II to activate cytotoxic T lymphocytes (105). DC maturation is necessary to T cell proliferation and differentiation; the final antitumor immunity is also associated with co-stimulatory molecules and cytokines which are expressed as the mature markers on DCs, such as CD80/CD86 and IL-12 (106).

DCs are necessary for anti-PD-1 efficacy. Anti-PD-1-activated T cells secrete IFN-γ, which in turn primes a transcriptomic shift in DC phenotype; DCs produce IL-12 upon sensing IFN-γ to stimulate effector T cell responses (107–109). The activation of the non-canonical nuclear factor kappa-light-chain enhancer of the activated B cell (NF-κB) pathway is also required for checkpoint efficacy, for it can enrich IL-12-producing DCs (107). Additionally, evidence of direct regulation is still emerging. PD-1 expression has recently been identified on DCs in the specific tumor context (110, 111). The result of an ovarian study demonstrated that PD-1 expressed on the tumor-associated DC can suppress NF-κB activation and the release of immune regulatory cytokines and restrict the upregulation of co-stimulatory molecules (111), which mediate immune suppression. PD-1 inhibition seems to increase the co-stimulatory molecule expression of DCs (112). In addition, the specific ablation of PD-1 on intratumoral DCs resulted in enhanced priming of tumor-specific CD8+ T cells to secrete IL-2 and IFN-γ (110). While DCs are the major antigen-presenting cells for cross-presenting tumor antigens to T cells and promoting antitumor response, PD-L1 expression on DCs can be upregulated by inflammatory cytokines, especially IFNs. Such upregulation is likely to prevent the overexpansion of tumor-infiltrating lymphocytes and eventually dampen the antitumor responses (113, 114). These results might provide additional insights into the role PD-1/PD-L1 plays on DCs to facilitate antitumor response and the mechanisms of immune checkpoint blockade therapy efficacy.

Immunosuppressive cells, unlike immune cells, have a positive effect on antitumoral immunity. There are some immunoinhibitory cells that display negative cross talking in TME, including cancer-associated fibroblasts (CAFs), regulatory T cells (Tregs), myeloid-derived suppressor cells (MDSCs), and M2 TAMs (mentioned above). Tregs repress the proliferation of both CD8+ and CD4+T cells through releasing transforming growth factor β (TGF-β) (125). CAFs promote the rate of glycolytic metabolism and further constitute a glucose-deficient TME. CTLs tend to decrease their number when encountered with such conditions (126). It is not yet known whether these have a role in promoting the efficacy of PD-1, but studies have shown that they are crucial in immune resistance, which will be discussed in detail in a subsequent paragraph.

Some studies have shown that carcinomas with robust T cell immunosurveillance can evade recognition through diverse genetic and immune-related mechanisms, including loss of tumor neoantigens and defect in antigen presentation.

Cancer is a genetic disease that can be induced by multiple genetic alterations, which are commonly caused by abnormalities of several key oncogenic pathways (149), like the phosphatase and tensin homolog (PTEN) signal pathway and mitogen-activated protein kinase (MAPK) signal pathway. Here, we mainly describe the two most common pathways, which have been proven to be closely related to PD-1 inhibitor resistance.

Research found that loss of PTEN in tumor cells in clinical patients of melanoma correlates with decreased T-cell infiltration, expansion, and inferior outcomes with PD-1 inhibitor therapy (150). PTEN loss-of-function mutations in tumors were significantly increased in non-responders who were treated with anti-PD-1 antibodies (151). Additionally, one of the most common pathways activated by loss of expression of the tumor suppressor PTEN is the phosphatidylinositol 3-kinase (PI3K) pathway, which plays a critical role in cancer by regulating several critical cellular processes. Thus, the PI3Kβ inhibitor, which is thought to regulate AKT activity in tumors with PTEN loss, has been applied to PTEN-deficient melanoma mouse models and demonstrated to enhance the efficacy of both PD-1 and CTLA-4 inhibitors (150). Accordingly, the regime that anti-PD-1 combined with PI3K-AKT pathway inhibitors may benefit cancer patients in the future.

The RAF/MEK/ERK pathway which is the classic routine in the MAPK pathway is also critical for human cancer; the pathway can be primed by activated RAS interacting with RAF kinase (152–154). Furthermore, RAS, RAF, and MEK are also frequently amplified or mutated in various cancers, accompanied by the activated MEK-ERK signaling pathway (155). KRAS, the component of RAS, is one of the most frequently mutated oncogenes in human cancers and participates in the mechanism of PD-1 inhibitor resistance (156). Similarly, BRAF, another mutated oncogene, has the vast majority in number harboring an activating point mutation (V600E) (157). This oncogenic mutation leads to constitutive activation of the MAPK signaling pathway and increased oncogenic potential through a variety of mechanisms, including reduced apoptosis, increased invasiveness, and increased metastatic behavior (158). Recent in vitro data suggest that BRAF V600E could also contribute to immune escape (157, 159). Based on these, selective inhibition of BRAF has been shown to induce an activated CD8+ T cell infiltrate, as well as increase melanoma MHC expression and melanoma antigen presentation early during treatment both in preclinical models and in human melanoma tissue samples (159–161). The study also suggested that combined BRAF and MEK inhibition with PD-1 blockade immunotherapy in BRAF-mutant melanoma can increase the frequency of long-lasting antitumor responses (162). Thus, the inhibition of the RAF/MEK/ERK signaling pathway may be a promising therapeutic strategy for cancer dysregulated in this pathway.

IFN-γ, effector cytokines of T cells, can directly exert an effective antitumor immune response by recognizing the corresponding receptors on tumor cells or indirectly promote the cross-activation of CD8+ T cells by upregulating antigen-presenting machinery to attack tumor cells (163). Classically, IFN-γ inhibits the proliferation of tumor cells and promotes their apoptosis, as it can activate signal transducer and activator of transcription 1 (STAT1) through using the Janus kinase (JAK) signal transducer and activator of the transcription pathway (127). Recent studies have implicated that defects in such pathways involved in IFN-receptor signaling and antigen presentation are associated with primary and acquired resistance to PD-1 blockades, such as inactivating mutations in JAK1 and JAK2 (143, 164). It may result in PD-L1 not being able to be reactively expressed and failing to attract T cell infiltration due to lack of chemokine production which is controlled by the IFN-γ pathway downstream of JAK1/2 (165). Considering that preexisting T cells in the tumor are a requisite for response to anti-PD-1 therapy (166), the absence of reactive PD-L1 expression may implicate a poor response to PD-1 blockade therapy, because of the impairment of tumor-infiltrating T cells (164).

IFN-β, belonging to type I IFN that is associated with innate immune responses (167), was proved to be suppressed by lysine-specific histone demethylase 1 (LSD1) (168). Ablation of LSD1 in cancer cells increases repetitive element expression; this leads to dsRNA stress and activation of type 1 IFN, which promotes antitumor T cell immunity and sensitizes refractory tumors to PD-1 blockade in a melanoma mouse model (168). The remarkable ability of LSD1 inhibition to convert a tumor resistant to PD-1 blockade to a tumor responsive to PD-1 blockade provides a means to increase the efficacy of anti-PD-1 cancer therapy and potentially turn “cold” tumors “hot” (169). It may suggest LSD1 inhibition combined with PD-1 blockade as a novel cancer treatment strategy. In addition, long-term IFN-β transcription can also promote the occurrence of resistance to anti-PD-1 therapy by inducing intratumoral augment of Tregs and myeloid cells, which cause T cell depletion and immunosuppression (170). Thus, IFNs display the consequence of resulting in T cell depletion and immunosuppression, although they can also promote the effect of tumor-specific CD8+T cells.

As noted, research of immune checkpoint blockade therapy was concentrated on reversing tumor-specific T cell dysfunction. CD8+T cells play an essential role in the scope of T cell-directed immunotherapy. Thus, the exhaustion of CD8+T cells induced by several factors can also be a crucial reason for PD-1 blockade resistance (143).

Epigenomic modifications might underlie CD8+T cell exhaustion. These long-lasting, exhaustion-associated epigenetic programs limit the rejuvenation of antigen-specific CD8 T cells during PD-1 blockade therapy. A study displayed that initial DNA-methylation programs could restrict T-cell expansion and clonal diversity during PD-1 blockade treatment (171). The administration of DNA-demethylating agents before ICB therapy reversed these programs and enhanced the reinvigoration of antitumor CD8 T cells. Moreover, the latest clinic trials concerning epigenetic therapies also suggest that histone deacetylase inhibitors may synergize with PD-1 blockade to overcome resistance (172, 173). What they found highlights epigenetic programs among exhausted T-cells as a potential mechanism to explain PD-1 blockade therapeutic failures. Besides, research found that co-stimulatory molecules like CD28 can also suppress the function of effector T cells and reduce the response to anti-PD-1 therapy by blocking the CD28-B7 co-stimulatory pathway (13). In addition to the regulation of epigenetic change and co-stimulatory pathway over CD8+T cells, other immune-suppressive cells also have more or less indirect effects on it, impacting drug resistance of anti-PD-1 therapy.

MDSCs are defined as immature myeloid cells, which can be induced to expand by tumor progression and play an immunosuppressive role in multiple cancers (174, 175). The recruitment of immunosuppressive MDSCs has shown complex protumorigenic outcomes following anti-PD-1 therapy (176). One mechanism of this recruitment may be driven by anti-PD-1-activated T cells, which partially trigger a tumor-intrinsic NLRP3 inflammasome signaling cascade (176, 177). This signaling cascade constitutes an adaptive resistance pathway, the genetic and pharmacological inhibition of which can enhance the efficacy of anti-PD-1 immunotherapy by inhibiting the tumor infiltration of MDSCs (176). On the other hand, checkpoint-activated CD8+ T cells can induce the differentiation and survival of protumorigenic TAMs and MDSCs by stimulating tumor production of CSF1 by secreting more TNF-α (178). These prompt us to hypothesize that neutralizing MDSCs and preserving T cell function may elicit robust immunotherapy responses by the combined actions of ICB agents together with targeted agents (179). Paradoxically, in HNSCC, it demonstrates reduced granulocytic MDSC infiltration post-PD-1 blockade (180). Thus, it is still unclear whether this model involves different mechanisms of MDSC recruitment or whether blockade of PD-1 inhibits MDSC proliferation directly.

TAM is another type of myeloid cells. It can impact the response to immunotherapy by activating triggering receptors expressed on myeloid cells 2 (TREM2) (181). TREM2 deficiency was associated with the transformation of macrophage subsets and an increase of intratumoral CD8⁺ T cells, some of which expressed PD-1. The observation found that tumor macrophage infiltrates enhanced T-cell-mediated control of tumor growth after the anti-TREM2 therapy; the anti-TREM2 mAb to tumor-bearing mice blunted tumor growth and strongly enhanced the efficacy of anti-PD-1 immunotherapy (181). Efforts are currently ongoing to complement checkpoint blockade with treatment targeting myeloid cells (115), including depletion of myeloid cells from tumors, blocking their pro-tumoral functions, or restoring their immunostimulatory properties (182, 183). These results may be applied as a theoretical basis to clinical trials.

Tregs can inhibit TCR-mediated activation and proliferation of CD4+/CD8+T cells to promote tumor immune evasion. Simultaneously, EZH2 has a critical role in maintaining the identity and function of Tregs; it has been proved that Ezh2 deficiency in Tregs stimulates antitumor immunity with enhanced T cell infiltration and elevated effector function (147). Mechanistically, Ezh2 functioned in regulating the stability of Foxp3 protein which is specifically expressed by Tregs. Based on these, the synergistic impact of the combination of EZH2 inhibition and anti-PD-1 has been found in an anti-PD-1-resistant model of HNSCC. It is explained that EZH2 inhibition can enhance tumor cell Class I MHC expression in vivo including in highly resistant models (147). Thus, it is promising that try to improve the efficacy of anti-PD-1 therapy by combining it with Ezh2 inhibitors.

CAFs are activated fibroblast cells during cancer development, contributing to the establishment of an immunosuppressive TME (184). Despite T cells being recovered from the capability against tumor cells following anti-PD-1 therapy, CAFs can act as a formidable barrier to T cells by secreting-related factors, resulting in T cell exclusion from tumor nests (185). TGF-β, a factor released by CAFs, promotes T-cell exclusion and blocks Th1 effector phenotype acquisition, which eventually results in resistance to PD-1 blockade therapy (186, 187). Inhibition of TGF-β unleashed a potent, enduring cytotoxic T-cell against tumor cells to prevent refractory. In mice with progressive liver metastatic disease, blockade of TGF-β signaling improves the susceptibility to anti-PD-1 therapy and suggests that TGF-β inhibition could prevent, but not reverse, CAF differentiation (186). NOX 4 is a specific downstream target of TGF-β. Inhibition of NOX 4 can “normalize” CAF to a quiescent phenotype and promote intratumoral CD8+ T-cell infiltration, overcoming the exclusion effect (185). These trials show that the regulation of CAFs through repressing the related downstream pathway or factors may have a synergistic effect on the anti-PD-1 therapy.

As mentioned above, one of the major obstacles that remain to be overcome is the restriction of T cells’ function in the immunosuppressive microenvironment formed by Tregs, MDSCs, and TAMs. The adoptive cell therapy (ACT) with chimeric antigen receptor (CAR)-redirected T cells is an attractive anticancer strategy. The breakthrough with CAR-T cell therapy was achieved, targeting B-cell hematologic tumors (188–191), while there is less efficacy in solid tumors. Research shows that TGF-β can be produced in most human tumors and markedly inhibits tumor antigen-specific cellular immunity. CAR-T lymphocytes have generated the resistance to TGF-β suppression, which expresses dominant-negative TGF-β receptors, to counteract these immunomodulatory activities (192). Such a result demonstrates their superior antitumor activity in animal models. Thus, combining engineered CAR-T cells with PD-1 antagonists makes a great deal of sense. There are promising results in both the pre-clinic model and case report (193, 194), presenting a large opportunity for the field of cellular engineering and immune checkpoint therapy.

Accordingly, the abovementioned studies indicate that the resistance to PD-1 inhibitors is directly related to the dysfunction of T cells caused by its epigenetic change, while other immune-related cells can also indirectly result in immune evasion via impacting the antitumor immunity progression of T cells.

The angiogenic tumor vasculature plays a vital role in regulating the response to cancer immunotherapy. Vascular abnormalities restrict T cell trafficking into the intratumor via upregulating vascular endothelial growth factor (VEGF) and gene-related to proangiogenic (195). Study has suggested that the VEGF signal induces the expression of the factor-related apoptosis antigen ligand (FasL)-mediated cell death on vascular endothelial cells, which in turn poses a formidable physical barrier to vascular material exchange (195). Additionally, the tumor neovasculature also decreases immature DCs and expands Treg cells and MDSC populations (195, 196). The modulation of tumor vasculature includes anti-angiogenesis and vascular normalization, which can induce the depletion of Tregs and regulatory B cells, enhancement of M1 TAMs, and activation of T cells, to reduce immunosuppression. The modulation can make favorable conditions for the infiltration of CD8+ cells and allow the effectiveness of immune checkpoint blockade (197). Immune checkpoint inhibitors have also shown promise in combination with anti-angiogenic in solid tumors (198), such as NSCLC and colorectal cancer (199). Thus, anti-angiogenesis and immunotherapy are documented to work synergistically together, showing promise for the resistance of PD-1 inhibitors.

Immune cells undergo complex shifts in metabolic states; immunosuppressive metabolites in TME can inhibit antitumor immunity by inhibiting immune cell infiltration (200–203).

Aerobic glycolysis is indispensable to CD8+T effector cells. It can be restricted by tumor cells that outcompete T cells for glucose uptake (81). In pretreatment of melanoma tumors, hypoxia-associated genes are highly expressed in the tumors that are subsequently resistant to PD-1 blockade compared with those from responding tumors (204). A high concentration of lactic acid can also blunt aerobic glycolysis of CD8+T cells and correlate with primary resistance on PD-1 blockade (205). A database analysis of patients with melanoma revealed strong negative associations between tumor lactate dehydrogenase expression and markers of CTL activation (201). Separately, indoleamine 2,3-dioxygenase (IDO), generated by tumors and immune cells, can enhance Treg and MDSC production and activity and inhibit the effect on T-cell immunity (206). IDO is the initial and rate-limiting enzyme in the degradation of tryptophan through the kynurenine pathway. A report found a significantly higher kynurenine/tryptophan ratio in NSCLC patients with early progression on nivolumab, suggesting that IDO might contribute to primary resistance to anti-PD-1 monoclonal antibodies (207). Despite that, the following clinical studies have shown that the efficacy of the IDO1 selective inhibitor plus PD-1 inhibitor is not as good as that of PD-1 blockade treatment alone (208). The combination therapy of IDO inhibitors and PD-1 antibodies may become a study direction for overcoming immunotherapy resistance. In addition, adenosine also is an immunosuppressive molecule that can suppress effector T cells and NK cells and increase Treg numbers (209, 210). Accordingly, metabolic disorders can encumber proper T cell activation and effector functions, which is a potential mechanism of resistance to PD-1 blockade. It is believed that the combined strategy based on this can bring gratifying results.

The gastrointestinal microbiome has been demonstrated to play an essential role in regulating the immune response function during cancer therapy (211–214). There is a group of active microorganisms that live in symbiosis with the host in the human intestinal tract and may cause tumor resistance to anti-PD-1 when it gets disordered (215, 216). Concordantly, a result has displayed that the responders to PD-1 blockade had a differential composition of gut bacteria (217). It has shown an “unfavorable” gut microbiome with low diversity and high relative abundance. Such a population may impair systemic and antitumor immune responses mediated by the limited intratumoral T cells, myeloid infiltration, and weakened antigen presentation capacity (211). Enhanced responses of anti-PD-1 therapy have been observed in mice that accepted fecal microbiome transplantation of the responder to PD-1 blockade. On the other hand, the efficacy of anti-PD-1 in mice receiving a non-responder could be restored by administration of specific genera enriched in responding patients in these mice. In addition, these specific genera were associated with increased intratumoral immune infiltrates mediated by the recruitment of CD4⁺T cells into the tumor bed and increased ratio of CD4⁺T cells to Tregs in response to PD-1 blockade (217). Besides, fecal microbiota transplant also overcomes resistance to anti-PD-1 therapy in melanoma patients (218). This suggests that regulating the gut microbiota may potentially enhance antitumor immune responses as well as response to immune checkpoint blockade.

During checkpoint blockade with anti-PD-1 inhibitors, other inhibitory checkpoints might become coordinately upregulated and in turn lead to therapeutic failure (219). T-cell immunoglobulin mucin 3 (TIM-3), a member of the TIM family of immunomodulatory proteins, has been identified as a critical regulator of CTL exhaustion with co-expression of PD-1 (220). Such co-expression means that the most dysfunctional subgroup of T cells does not produce IL-2 and IFN-γ and eventually causes adaptive resistance. The mechanism has demonstrated that the increased Tim-3-mediated escape of exhausted TIL from PD-1 inhibition was mediated by PI3K/Akt complex downstream of TCR signaling in HNSCC (219). In vitro, the anti-Tim-3-blocking antibody reverses resistance to anti-PD-1 in PBMC from lung cancer patients (221). On the other hand, significant antitumor activity was observed after sequential addition of anti-Tim-3 mAb to overcome adaptive resistance to anti-PD-1 mAb in a murine HNSCC model (219). Thus, combination therapy targeting TIM-3 and PD-1 signaling pathways might be effective against the resistance of mono-immunotherapy.

Lymphocyte activation gene 3 (LAG-3) can selectively be expressed on activated T cells, NK cells, DCs and may get compensatory upregulation. The regulatory function of LAG-3 on T cells is similar to that of PD-1, which delivers suppressive signaling to hinder antitumor response (222). LAG-3 also competes for binding to MHC class II, which leads to decreased efficacy of MHC class II-mediated antigen presentation (223). The upregulation of LAG-3 in tumors of melanoma and lung cancer patients with acquired resistance to anti-PD-1 therapy has been demonstrated (223). There appeared to be a synergistic benefit of anti-LAG-3/anti-PD-1 combinatorial immunotherapy compared with anti-PD-1 monotherapy. In addition, a higher proportion of effector T cells were observed in mice treated with anti-LAG-3/anti-PD-1 than in PD-1 monotherapy groups. These suggest that anti-LAG-3/anti-PD-1 combinatorial immunotherapy may act synergistically (224). The roles of other checkpoints are still unconfirmed in anti-PD-1 resistance, such as TIGIT. Thus, a more particular knowledge of these new immune checkpoints may provide a rationale for designing combination treatments in the future.