Clinical and Recent Patents Applications of PD-1/PD-L1 Targeting Immunotherapy in Cancer Treatment—Current Progress, Strategy, and Future Perspective

Targeting PD-L1 and PD-1 interactions is a relatively new therapeutic strategy used to treat cancer. Inhibitors of PD-1/PD-L1 include peptides, small molecule chemical compounds, and antibodies. Several approved antibodies targeting PD-1 or PD-L1 have been patented with good curative effect in various cancer types in clinical practices. While the current antibody therapy is facing development bottleneck, some companies have tried to develop PD-L1 companion tests to select patients with better diagnosis potential. Meanwhile, many companies have recently synthesized small molecule inhibitors of PD-1/PD-L1 interactions and focused on searching for novel biomarker to predict the efficacy of anti-PD-1/PD-L1 drugs. This review summarized clinical studies and patent applications related to PD-1/PD-L1 targeted therapy and also discussed progress in inhibitors of PD-1/PD-L1.


INTRODUCTION
Programmed cell death protein 1, also referred to as cluster of differentiation 279 (CD279), is a surface protein that can regulate the immune system by inhibiting T-cell activity. PD-1 is constitutively expressed on activated T-cells, B cells, natural killer (NK) cells, macrophages, and dendritic cells (DCs) (1). Programmed death-ligand 1 (PD-L1), also referred to as B7-H1 or CD274, is constitutively expressed on antigen-presenting cells, lymphoid, endothelial, and epithelial cells (2). Interferon gamma (IFN-γ) and tumor necrosis factor (TNF-α) secreted by activated T-cells can also induce PD-L1 expression on tumor cells and antigen-presenting cells (APCs) (3). Figure 1 shows that naïve T-cells are activated through binding between T cell receptors (TCR) and the peptide-MHC complex presented by (APC); T-cell activation can lead to transient upregulation of PD-1, which is the receptor of PD-L1. Binding between PD-1 and PD-L1 negatively regulates downstream signaling mediated by co-activation of TCR and CD28 (4). When PD-L1 interacts with PD-1, the immunoreceptor tyrosine-based inhibitory motifs (ITIM) and immunoreceptor tyrosine-based switch motifs (ITSM), which are on the intracellular domain of PD-1, can be phosphorylated. The Src homology 2 domain-containing protein tyrosine phosphatase 1 (SHP-1) and Src homology 2 domain-containing protein tyrosine phosphatase 1 (SHP-2) are then recruited and bind to ITIM to further inhibit the signaling downstream of the TCR (5). After inhibiting the TCR-mediated signaling pathway, PD-1 prevents the activation of the pathway mediated by PI3K/Akt or Ras/MEK/Erk. This further inhibits the function of CD8+ T-cells (6). Programmed cell death 1 ligand 2 (also known as PD-L2, B7-DC), which is the second ligand of PD-1, is expressed on tumor cells, APCs, cancer associated fibroblasts, and macrophages (7)(8)(9). PD-L2 plays an inhibitory role on the functioning of T-cells, which is similar to that of PD-L1. Meanwhile, PD-L1 also interacts with the surface protein CD80 (B7-1) expressed on activated T-cells. Interacting with PD-L1, CD80 could induce increased expression of Bim, which contributes to the apoptosis of CD8+ T-cells (10). As a result, the PD-1/PD-L1 signaling pathway promotes tumor cells escaping immune surveillance by inhibiting cell survival and activation of T-cells.
Targeting PD-L1 and PD-1 interactions is a novel therapeutic strategy used for cancer treatment. Antibodies targeting PD-1 or PD-L1 have marked a breakthrough in cancer immunotherapy and have become a hot topic in cancer therapy. Many companies have therefore begun studies on cancer immunotherapy and applied a series of related patents and patent applications in this field. To date, there have been about 5,000 patents published, and the number of patents continues to increase (Figure 2). In this review, we demonstrate the development of PD-1/PD-L1 directed immunotherapy and progress in inhibitors disrupting PD-1/PD-L1 binding. Moreover, patents or patent applications related to PD-1/PD-L1 signaling pathway and its inhibitors will also be discussed in this review, which will provide an update on PD-1/PD-L1 targeted cancer therapy.

HISTORICAL OVERVIEW OF RELEVANT PATENTS OF PD-1/PD-L1 TARGETING CANCER IMMUNOTHERAPY
The PD-1 protein was discovered by Tasuku Honjo in 1992, and he was awarded the Nobel Prize in physiology and medicine in 2018. The patent published in 1995 by Honjo firstly proposed the sequence of PD-1 protein and gene encoded PD-1 (11). Honjo's discovery also showed that PD-1 is a protein that negatively regulates the immune system (12). Later, Gordon Freeman identified B7-4 as one of the ligands to PD-1 (13). Meanwhile, Dr. Lieping Chen and his team independently discovered B7-H1. The sequence of B7-H1 protein and gene encoded B7-H1 was published in 1999 by Dong et al. (14). However, they did not mention the correlation between B7-H1 and PD-1. Based on his own findings of B7-H1, Chen (15). Freeman also mentioned that PD-L1 is one of the members of the CD28/B7 immunoglobulin superfamily that could inhibit the T-cell function through PD-1/PD-L1 interactions (15). Table 1 shows patents and patent applications for the finding of PD-1 and PD-L1 proteins and the development of PD-1/PD-L1 blockade therapy. Honjo's studies suggested that suppression of the PD-1 protein could be effective in cancer treatment (12). Studies have shown that expression of PD-1 and PD-L1 was enhanced in cancer cells and was related to defective immune responses (24). These studies suggested that two immune checkpoint molecules may be important therapeutic targets for cancer and infectious disease treatment. Thus, the blockade of PD-1/PD-L1 interactions using inhibitors may be a novel and effective strategy for immunotherapy. Additionally, a previous study showed that blockade of the PD-1/PD-L1 pathway using PD-L1 antibody could inhibit T-cell apoptosis (25). This study also showed that PD-L1 antibody affected the survival of tumor cells in vivo (25). These results proved that PD-L1 antibodies can enhance Tcell growth to further inhibit tumor growth-this suggests that inhibition of the PD-1/PD-L1 interaction could be a new method of cancer treatment.
Honjo cooperated with Ono Pharmaceutical Co. and Medarex to develop an anti-cancer medication targeting PD-1, named nivolumab. Two studies of nivolumab conducted in Phase III trials showed impressive efficacy for this antibody in advanced melanoma (26,27). The results of a phase III trial showed that the overall survival rate at 1 year was significantly different between the nivolumab group (72.9%) and dacarbazine group (42.1%) of previously untreated patients who had advanced melanoma without a BRAF mutation (26). In addition, nivolumab showed higher response rates and lower toxicity rates than ipilimumab and chemotherapy (27). Following the results of these two clinical trials, the Food and Drug Administration (FDA) approved nivolumab for the treatment of advanced melanoma in 2014. The discovery of the PD-1/PD-L1 signaling pathway attracted researchers' attention on developing antibodies against this pathway. The PD-1 protein has led to breakthroughs in cancer immunotherapies in the past decades. Many companies have filed patents related to antibodies during these past 20 years. Table 2 shows the core patents related to FDA-approved antibodies while Table 3 shows patents related to antibodies.

STRUCTURE ANALYSIS OF ANTIBODIES TARGETING PD-1 AND PD-L1
Several structures and classes of antibodies inhibiting the PD-1/PD-L1 interaction have been published recently. Most of these anti-PD-1 antibodies are fully human immunoglobulin G4 (IgG4) antibodies with the S228P mutation, including nivolumab, pembrolizumab, cemiplimab, dostarlimab, MEDI-0680, and SSI-361. These antibodies have similar binding properties to the natural IgG4, which reduce ADCC function and eliminate CDC function, but they still retain function in binding to FcγRI and FcγRIIb. Spartalizumab is a humanized IgG4κ monoclonal antibody with S228P mutations and K447 deletion (44). Tislelizumab was generated via the introduction of several mutations (including S228P, E233P, F234V, L235A, D265A, and R409K) in IgG4 antibodies (45). AMP-224 is an anti-PD-1 recombinant fusion protein that contains the extracellular domain of PD-L2 and Fc domain of human IgG1 (46). Moreover, the crystal structures of PD-1/Anti-PD-1 antibodies have also been explored. The N-terminal extension, BC-loops, and FG-loops are crucial for binding of nivolumab and PD-1. The VL chain of nivolumab and PD-L1 residues shared an overlapping binding surface on the FG loop (47). The C'D loop of PD-1 mainly contributes to the interaction with pembrolizumab (48). Anti-PD-1 antibodies inhibit the  PD-1/PD-L1 interaction by competing with PD-L1 while binding to PD-1. The epitopes of these antibodies directly occupy the partial binding site of the PD-L1 protein. In addition, the binding of PD-1 and its antibodies induces optimal conformational changes in the PD-1 protein, which blocks PD-1/PD-L1 interactions, because PD-1 also interacts with PD-L1 in distinct conformations. Tislelizumab interacts with an IgV-like domain of PD-1 and is different from pembrolizumab and nivolumab, as shown by its unique binding epitopes, including Gln75, Thr76, Asp77, and Arg86 (45). Although SHR-1210 was reported to have unspecific interactions with some human receptors driving angiogenesis, the optimization of complementary determining region (CDR) domains successfully eliminated off-target binding (49). Meanwhile, the binding properties of SHR-1210 have not been reported. Unlike anti-PD-1 antibodies, three approved anti-PD-L1 antibodies include human IgG1 antibodies. Atezolizumab and durvalumab are antibodies of eliminated FcγR-binding and effector functions while avelumab was designed to retain intact Fc functions (50). BMS-936559 is differentiated from three approved PD-L1 antibodies and is an IgG4 mAb with S228P mutations (50). KN035 is a fusion protein containing a single domain of the humanized anti-PD-L1 antibody and the Fc of an IgG1 (51). CX-072 is a human PD-L1 specific protease-activatable antibody prodrug. CX-072 was designed by linking the masking peptide links to the targeted antibody (52).
Recently, the crystal structures of the PD-L1/avelumab complex revealed that avelumab/atezolizumab/BMS-936559 binds to the IgV domain of PD-L1 through its heavy chain (VH) and light chain (VL). These are dominated by the VH chain (53). A comparison of the PD-L1/antibody and human PD-1/PD-L1 complexes demonstrates that antibodies directly occupy the partial binding site of the PD-1 protein. In contrast, the PD-L1/durvalumab Fab complex demonstrated that the binding sites of the antibody are in the N-terminal region of the PD-L1 protein (53). The KN035/PD-L1 complex showed a different pattern. The paratope of KN035 is limited to only two complementary determining regions (CDRs)-one of which contributes to binding with high-affinity (54). This narrow binding area provides an opportunity for rationally designing peptides or small-molecule inhibitors that imitate the nanobody/PD-L1 interface.

CLINICAL APPLICATION OF PD-1/PD-L1 TARGETING CANCER IMMUNOTHERAPY
There have been more than 2,000 clinical trials of anti-PD-1 antibodies and over 1,000 clinical trials of anti-PD-L1 antibodies (Figure 3). Based on the data from several clinical trials, some of these drugs have been approved by the FDA, the National Medical Products Administration (NMPA), and the European Medicines Agency (EMA) for  Table 4) (50). In addition, camrelizumab and toripalimab were approved by NMPA for marketing.
Clinical trials of most antibodies have just started, and the results require further updating. Table 5 summarizes several clinical trials of anti-PD-1/PD-L1 antibodies that are currently being developed with the latest data. The data from clinical trials revealed that newly developed antibodies also showed a durable response. Table 5 also demonstrates that anti-PD-1/PD-L1 antibodies can cause treatment-related adverse effects (TRAEs) and immune-related adverse effects (IRAEs). In some patients, these AEs led to treatment discontinuation and treatment interruption. The Objective response rate (ORR) is 47% among the 75 patients with metastatic CSCC who received cemiplimab-rwlc. Complete response was achieved in 4% of patients (81). Among patients with relapsed/refractory cHL (NCT02961101 and NCT03250962), the response duration rate at 6 months was 76% in patients treated with camrelizumab monotherapy (n = 19) compared to 100% in those treated with decitabine plus camrelizumab (n = 42) (95). Among the 127 patients with advanced melanoma (NCT03013101), the ORR is 17.3% in overall population after treatment with toripalimab. The disease control rate (DCR) was 57.5% and median progression free survival (PFS) was 3.6 months (96). Based on the clinical results shown above, cemiplimab, camrelizumab, and toripalimab were approved for clinical use.

THE CURRENT OPTIMIZATION OF ANTI-PD-1/PD-L1 TREATMENT STRATEGY
Several clinical trials using antibodies targeting the interaction of PD-1 and PD-L1 for cancer treatment have shown promising abilities in prolonging survival, but not all patients respond to PD-1/PD-L1 inhibitors (97). In addition, clinical results have also shown that anti-PD-1 or anti-PD-L1 treatment caused TRAEs and IRAEs, although anti-PD-1/PD-L1 drugs have shown lower toxicity than standard chemotherapy (98). Most seriously, AEs caused by these antibodies sometimes could lead to treatment discontinuation and treatment interruption (98). Due to the limited success and disadvantages of anti-PD-1/PD-L1 antibodies, effective strategies are needed to improve the efficacy of PD-1/PD-L1 targeted immunotherapy. Detecting PD-L1 expression in tumor cells and tumor infiltrated T-cells would be useful for targeting patients with a big likelihood of responding to PD-1/PD-L1 treatment. Meanwhile, it is also crucial to search for potential biomarkers that could selectively reflect the efficacy and feasibility of anti-PD1/PD-L1 therapy. Furthermore, small molecule inhibitors targeting PD-1 and PD-L1 are emerging as their potential advantages are realized vs. monoclonal antibodies.

The Application of PD-L1 Immunohistochemistry (IHC) Assays
Some clinical trials have shown that more than half of patients had no response to anti-PD-1 drugs, and some responders even experience tumor relapse within 2 years after treatment of anti-PD-1 drugs (26,99). Studies suggest that clinical efficacy of PD-1/PD-L1 targeted immunotherapies may be predicted by PD-L1 expression on tumor cells and tumor-infiltrating immune cells (100). Developing PD-L1 IHC test compounds have begun to attract scientists' attention during the past 5 years. Several companies have designed commercially available PD-L1 IHC tests, including 22C3, 28-8, SP263, SP142, E1L3N, and 73-10 assays. Merck developed a PD-L1 IHC test using 22C3 antibody and also applied for a patent (US9709568B2), which claimed the use of the 22C3 antibody for diagnostic purposes (101). In addition, BMS designed a different companion assay for PD-L1 expression using its 28-8 antibody and had a patent application (WO2013173223A1) that described a method of detecting PD-L1 expression using the clone 28-8 antibody (102). The SP142 assay was developed by Ventana and was described in patent application WO2015181343A2 (103). These PD-L1 IHC assays are currently being tested in clinical trials, and some of them have been approved by the FDA as companion diagnostics for PD-1/PD-L1 targeted immunotherapies. Table 6 shows that PD-L1 expression was first reported to be associated with higher response rates to pembrolizumab/atezolizumab and was approved by the FDA to guide the selection of patients for anti-PD-1/PD-L1 treatment. For example, the DAKO 22C3 IHC assay is approved for use as a companion diagnostic with pembrolizumab immunotherapy in NSCLC, gastric cancer, cervical cancer, HNSCC, and ESCC (68,(104)(105)(106)(107)(108). In addition, the Ventana PD-L1 (SP142) assay has also been approved as a companion diagnostics test for atezolizumab in UC and TNBC (109,110). IHC 28-8 and SP263 (nivolumab and durvalumab, respectively) are complementary diagnostics and have not been approved by the FDA. Recent studies (e.g., shown in the meta-analysis) have also confirmed Frontiers in Immunology | www.frontiersin.org that efficacy of PD-1/PD-L1 inhibitors was more sensitive in PD-L1 positive patients compared to negative groups (111). Each PD-L1 IHC assay, performed in different IHC staining platforms, is independently developed for a specific anti-PD-1 or anti-PD-L1 drug. As is shown in Table 7, differences between six commonly used PD-L1 IHC assay were shown by detection system, staining platform, and antibody epitope. Thus, each assay potentially displayed different staining sensitivities. Different PD-L1 IHC assays and different PD-L1 tumor expression cut-off points are used in clinical trials, which raises concerns about whether the tests can be used interchangeably. The Blueprint PD-L1 IHC Assay Comparison Project was founded to enable a better understanding of the similarities and differences between these four PD-L1 IHC systems. This project is an industrial-academic partnership seeking to harmonize IHC PD-L1 testing. The result from phase I of the Blueprint project showed that the 22C3, 28-8, and SP263 assays displayed comparable sensitivity and the SP142 assay showed significantly less sensitivity (112). The phase I of the Blueprint project detect PD-L1 expression on TCs using four PD-L1 IHC assays performed in different staining platforms, and the result of staining was evaluated independently by three pathologists (112). Phase 2 of the Blueprint project compares 73-10 assay with four other PD-L1 IHC assay (including 22C3, 28-8, SP263). The results from phase 2 showed highly comparable sensitivity between 22C3, 28-8, and SP263 assays, less sensitivity with SP142 assay, and higher sensitivity with 73-10 assay when detecting PD-L1 expression on TC (113). The high concordance was observed between scorings by glass slide and scorings by digital image (113). Most importantly, a recent study has investigated the cause of distinct immunohistochemical staining generated by SP142 assay. The results suggested that discordances are more likely caused by differences of staining platform rather than antibody epitope (114).

The Current Potential Biomarkers Used to Evaluate the Feasibility of Anti-PD-1/PD-L1 Therapy
PD-1/PD-L1 inhibitors represent a breakthrough in cancer therapy. However, the response rates of PD-1/PD-L1 inhibitors in patients is, overall, unsatisfactory and results in limited applications in clinical practice. Therefore, searching for biomarkers predicting the efficacy of PD-1/PD-L1 inhibitors is crucial for patient selection. There are several biomarkers associated with the response to anti-PD-1/anti-PD-L1 therapy ( Table 8) including PD-L1 expression, lactate dehydrogenase (LDH), mismatch-repair (MMR) deficiency, gene alteration, tumor mutational burden, etc. A clinical study conducted by Diem showed that patients with an elevated baseline LDH showed a significantly shorter OS (P = 0.0292) and lower response rate compared with patients with normal LDH at baseline and during treatment. This suggests that LDH could predict early response or progression in advanced melanoma patients with anti-PD-1 therapy (115,116). In addition, patients who achieved clinical benefit after treatment of anti-PD-1 therapy were detected with a higher percentage of Bim + PD-1 + CD8 + T-cells in the peripheral blood (117). The levels of Bim in PD-1 + CD11a hi CD8 + T-cells (also indicated tumor reactive T cell) could be a predictive factor of clinical benefit in patients with metastatic melanoma treated with anti-PD-1 therapy (117). High pretreatment lymphocyte count (LC) and relative eosinophil count (REC) were associated with improved overall survival of melanoma patients with pembrolizumab treatment (118). Patients with T-cells expressing SRY-Box 2 (SOX-2) experienced disease regression following the treatment of nivolumab, suggesting that SOX-2 is associated with a clinical response upon immunotherapy with anti-PD-1 monoclonal antibodies (119). A retrospective study showed that the median  PFS of patients with a neutrophil-lymphocyte ratio (NLR) of ≥3 was shorter than that in patients with a NLR of < 3 (2.0 vs. 5.3 months, p = 0.00515) at 4 weeks after treatment (120).
The clinical data suggested that the NLR ratio might be an indicator of a poor prognosis in patients with advanced NSCLC receiving nivolumab (120). Patients with a 1.5-fold increase in circulating soluble PD-L1 (sPD-L1) concentrations were more likely to achieve partial responses to anti-PD-1 antibodies after 5 months upon anti-PD-1 therapy. This shows the predictive effect of sPD-L1 on clinical response to anti-PD-1 therapy (121). Among 36 EGFR-mutated metastatic NSCLC patients, compared with patients detecting decreased levels of sPD-1, patients with an increased or stable sPD-1 level achieved longer PFS (p = 0.004) and OS (p = 0.002) after two cycles of nivolumab (122). In melanoma, the pre-treatment tumors in responding patients were detected with higher expressions of IFN-γ and IFN-γ-inducible genes, including indoleamine 2,3-dioxygenase 1 (IDO1) and C-X-C motif Chemokine Ligand 9 (CXCL9) (123). These associations were also found in NSCLC or renal cell carcinoma patients (123). In addition, genetic aberrations within tumors were also found to be associated with clinical efficacy in anti-PD-1/PD-L1 therapy. For example, among 155 patients, six patients with MDM2/MDM4 amplification and seven of eight patients with Epidermal Growth Factor Receptor (EGFR) alterations were found to have time-to-treatment failure (TTF) <2 months (124). Meanwhile, hyper-progressors harbored MDM2/4 amplifications or EGFR alterations (124).
A retrospective analysis showed that EGFR-mutant and ALKpositive NSCLC patients receiving anti-PD-1/PD-L1 therapy showed lower ORR (P = 0.053) (125). Immunotherapeutic  Objective response rate (complete or partial response) was significantly correlated to tumor microenvironment-related SNPs concerning CCL2, NOS3, IL1RN, IL12B, CXCR3, and IL6R genes.  (129). Gene variations that occur in at least 1% of the population used to be called polymorphism. Single nucleotide polymorphisms (SNPs) of tumor microenvironmentrelated genes (including CCL2, NOS3, IL1RN, IL12B, CXCR3, and IL6R) were significantly associated with ORR of patients treated with anti-PD-1/PD-L1 therapies (130). And safety of anti-PD-1/PD-L1 targeted therapies was significantly associated with gene SNPs including UNG, IFNW1, CTLA4, PD-L1, and IFNL4 genes (130). Besides that, rs17388568, which maps to a locus of IL2 gene and IL21 gene, was correlated with a higher response to anti-PD-1 targeting therapy (131). CD8, PD-1, and PD-L1 expression in the tumor and at the invasive margin significantly correlated with treatment outcome (P = 0.001) (132). Versus the progression group, the response group had significantly higher numbers of CD8 + , PD-1 + , and PD-L1 + cells (CD8, P = 0.0001; PD-1, P = 0.0002; PD-L1, P = 0.006) (132). Among HNSCC patients treated with pembrolizumab, PD-L2-positive patients showed higher ORR compared with PD-L2-negative patients (133). And longer PFS and OS were observed in PD-L2-positive patients (133). Except for the biomarkers mentioned above, the tumor mutation burden/load (TMB) also served as a predictive or prognostic factor for response to anti-PD-1/PD-L1 immunotherapy. TMB is an estimate of somatic mutations by accessing the data from whole exome sequencing (WES) or sequencing a select panel of genes. Foundation Medicine has developed clinical testing platforms to measure TMB using hybrid capture-based next generation sequencing. FDA has approved FoundationOne CDx to be used as a companion diagnostic for therapy selection. Several studies have shown that TMB is associated with a clinical response to anti-PD-1/PD-L1 treatment in melanoma and NSCLC (Table 9). Recently, a novel blood-based TMB (bTMB) assay was developed for cell-free DNA by researchers from Foundation Medicine. A retrospective analysis using bTMB assay showed that bTMB is correlated with significant PFS benefit (P = 0.013) and TMB (Spearman rank correlation = 0.64) in patients with NSCLC treated with atezolizumab (139). Neoantigens derived from mutated genes are tumor-specific and show significant correlation with the clinical response to anti-PD-1/PD-L1 treatment. A significantly higher candidate neoantigen burden was detected in patients with CB vs. those with NCB and associated with improved PFS (median 14.5 vs. 3.5 months, log-rank P = 0.002) (134). The PFS in patients with a higher non-synonymous burden were higher than those with low non-synonymous burden (median PFS 14.5 vs. 3.7 months, log-rank P = 0.01) (134). These data suggested that higher non-synonymous mutation or candidate neoantigen burden in tumors were associated with improved PFS of anti-PD-1-treated NSCLC patients. A recent study has shown that a minority of somatic mutations in tumors could lead to neoantigens and TMB could be used to estimate tumor neoantigen load (140).

Discovery of Small Molecule Compounds Inhibiting PD-1/PD-L1 Interactions
The limited success and disadvantage of antibodies prompted researchers to search for more effective strategies for PD-1/PD-L1 targeted therapy and improve the efficacy of cancer immunotherapy. Thus, studies on the discovery of low-molecular-weight compounds inhibiting PD-1/PD-L1 interaction have begun to attract scientist's attention. During the past 5 years, many companies, such as Arising International Inc, Chemocentryx Inc, Institute of Materia Medica, Guangzhou Maxinovel Pharmaceuticals Co, Incyte Corporation, Bristol Myers Squibb (BMS), and Aurigene, have discovered a series of small molecule chemical compounds and peptides. Meanwhile, these companies have applied for a series of patents related to inhibitors ( Table 10). Most of these patents presented not only the structure of PD-1/PD-L1 inhibitors, but also the method of compound synthesis and the use of inhibitors as immunomodulators. In addition, the patents showed verified inhibitory effects of these inhibitors. Some of these inhibitors could only block PD-L1/PD-1 interactions. Other inhibitors, such as the peptides discovered by BMS company, could inhibit interactions of PD-L1 with PD-1 or CD80. All inhibitors discovered by Aurigene, including small molecule chemical compounds and peptides, showed an inhibitory effect on the PD-1 signaling pathway.

DISCUSSION AND PERSPECTIVE
Anti-PD-1/PD-L1 antibodies have achieved success in the field of cancer immunotherapy during the past decade and mark a breakthrough in oncology. Eight antibodies blocking PD-1 and PD-L1 interactions have been approved for several indications. Despite the promising results reported in some clinical trials, limited drug efficacy caused by IRAEs has been observed and durable responses have been found in only a limited number of patients. In addition, immune-related adverse events caused by anti-PD-1 drugs have been reported in several clinical trials. Due to the limited successes and disadvantages of anti-PD-1/PD-L1 antibodies, more attention has been given to developing more effective strategies to improve clinical response rates. However, using PD-L1 expression as a biomarker of response is important in identifying patients who could obtain a positive clinical response from PD-1/PD-L1 targeted immunotherapy. The use of a single PD-L1 IHC assay with immunotherapy using a specific anti-PD-1/PD-L1 antibody would be one strategy for improving clinical trial outcomes. However, responses were also seen in patients with negative or low PD-L1 expression. For example, in three trials (CheckMate 017, CheckMate 025, and OAK), favorable long-term outcomes were achieved in PD-L1-negative patients (26,188,189). The CheckMate 227 trial among NSCLC patients with a high tumor mutational burden showed that progression-free survival was significantly longer with first line nivolumab plus ipilimumab than with chemotherapy, regardless of PD-L1 status (190). These studies also suggested that a higher mutation or neoantigen load could potentially result in a higher likelihood of response to PD-1 or PD-L1 inhibitors.
Apart from TMB, there are several other biomarkers including LDH, MMR-deficiency, gene alteration, and IFN-γ related gene. These are useful biomarkers for the response to anti-PD-1/PD-L1 cancer therapy in solid tumors. Some studies have shown dynamic PD-L1 expression in the tumor cells further limits the feasibility of PD-L1 IHC (191). PD-L1 expression could be regulated through extrinsic and intrinsic signaling pathways such as mitogen-activated protein kinase (MAPK) signaling pathway, Janus kinase/signal transducers and activators of transcription (JAK/STAT) signaling pathway, miRNA-related pathway, as well as IFN-γ and TNF-α (192)(193)(194). An understanding of the mechanism of regulation of dynamic PD-L1 expression may be useful for developing novel strategies to improve the efficacy of anti-PD-1/PD-L1 drugs. On the other hand, small molecules are expected to reduce immune-related adverse events and promote higher efficacy. Studies on small molecule PD-1/PD-L1 inhibitors have just begun within the preclinical stage. CA-170 is the first PD-1/PD-L1 inhibitor successfully entering clinical trial, and it is potentially a small molecule PD-1/PD-L1 inhibitor in cancer therapy. Future clinical trial results of CA-170 would be important for developing small molecule inhibitors.

AUTHOR'S NOTE
This review has made a summary about clinical studies and patent application of PD-1/PD-L1 targeted therapies. The paper has also shown the promising result of anti-PD-1/PD-L1 drug in various cancer types and several kinds of strategies improving efficacy of anti-PD-1/PD-L1 drug have been mentioned in the paper, including developing companion PD-L1 test, searching for biomarkers, and discovering small molecule PD-1/PD-L1 inhibitors. The paper has shown the development of anti-PD-1/PD-L1 therapies and provided broad knowledge of PD-1/PD-L1 targeted therapies.

AUTHOR CONTRIBUTIONS
LG, RW, and HK contributed conception and design of the review article. LG organized the database collection. LG and RW wrote the first draft of the manuscript. HK wrote and revised sections of the manuscript. All authors contributed to manuscript revision, read and approved the submitted version.