Generation, secretion and degradation of cancer immunotherapy target PD-L1

Abstract

Cancer immunotherapy is a rapidly developing and effective method for the treatment of a variety of malignancies in recent years. As a significant immune checkpoint, programmed cell death 1 ligand 1 (PD-L1) and its receptor programmed cell death protein 1 (PD-1) play the most significant role in cancer immune escape and cancer immunotherapy. Though PD-L1 have become an important target for drug development and there have been various approved drugs and clinic trials targeting it, and various clinical response rate and adverse reactions prevent many patients from benefiting from it. In recent years, combination trials have become the main direction of PD-1/PD-L1 antibodies development. Here, we summarized PD-L1 biofunctions and key roles in various cancers along with the development of PD-L1 inhibitors. The regulators that are involved in controlling PD-L1 expression including post-translational modification, mRNA level regulation as well as degradation and exosome secretory pathway of PD-L1 were focused. This systematic summary may provide comprehensive understanding of different regulations on PD-L1 as well as a broad prospect for the search of the important regulator of PD-L1. The regulatory factors of PD-L1 can be potential targets for immunotherapy and increase strategies of immunotherapy in combination.

Introduction

In human physiological condition, immune system plays a vital role in immune surveillance, defense and regulation [1]. In the complex immune system, T cell-based cellular immunity can recognize and clear aberrant cells, such as pathogen-infected cells and cancer cells. T cells undergo a series of orderly differentiation processes under the induction of thymosin in the thymus, and gradually mature into a T cell library that recognize various antigens [1]. Due to the effective role of T cell in cancer immunity, the mechanism of T cell action has attracted more and more attention in recent years. T cell is activated through the recognizing of T cell receptor (TCR) peptide-major histocompatibility complexes (MHC) in antigen-presenting cell (APC) or other target cells. Other than this, T cell response also needs a series of co-stimulatory and co-inhibitory receptors and their ligands which are also known as immune checkpoints. The co-stimulatory factors include T-cell-specific surface glycoprotein CD28 (CD28) and its ligand T-lymphocyte activation antigen CD80 (CD80) or T-lymphocyte activation antigen CD86 (CD86), CD27 antigen (CD27) and CD27 ligand (CD27L), tumor necrosis factor receptor superfamily member 9 (CD137) and CD137 ligand (CD137L) et al. The co-inhibitory factors include PD-1 and its ligand PD-L1 or programmed cell death 1 ligand 2 (PD-L2), cytotoxic T lymphocyte-associated antigen-4 (CTLA-4) and CD80/CD86, T cell immunoglobulin and mucin domain-containing protein 3 (TIM3) and galectin-9 (GAL9) and so on. The co-stimulatory factors, as the accelerator of a car, can promote T cell stimulation, proliferation and toxicity; while the co-inhibitory factors as the brake of a car can inhibit this effect. In physiological condition, the brake is more important than the accelerator in T cell immune system [1–3].

Among these co-inhibitory factors, PD-1 and PD-L1 play the most significant role in normal immune regulation and tumor immune escape. Because of the abnormal expression and significant role of PD-L1 in many malignant tumors, more and more evidences proved that PD-L1 could be a therapeutic target in numbers of malignancies, and PD-1/PD-L1 axis targeting drug has brought hope for the treatment of many cancers. Since the approval of PD-1 inhibitor pembrolizumab and nivolumab by Food and Drug Administration (FDA) in 2014, progress about PD-1/PD-L1 inhibitors as a strategy of cancer immunotherapy has grown rapidly [4, 5]. Up to now, more than 4400 clinical trials of PD-1/PD-L1 targeting therapies are in progress, and FDA has approved various PD-1/PD-L1 inhibitors for the treatment of various cancers [6, 7].

Due to the important role of PD-L1 in cancer, it is imperative to study its function and properties. In this review, we focused on the born, banish and death of PD-L1 protein, including its gene, mRNA and protein structure as well as its regulation factors in mRNA level, post-translational modification, degradation pathway, its biofunction and key role in various cancers as well as approved antibody drugs, peptides and small molecule inhibitors targeting PD-1/PD-L1. This review may provide a comprehensive understanding of PD-L1 as well as a broad prospect for the search of the important regulator of PD-L1.

Biological structure and post-translational modification of PD-L1

Gene structure

PD-L1, also known as B7 homolog 1 (B7–H1), programmed cell death 1 ligand 1 (PDCD1 ligand 1) and cluster of differentiation 274 (CD274), is the first identified ligand for the inhibitory receptor PD-1. PD-L1 is encoded by the CD274 gene (HGNC accession number: 17635; Ensembl Gene accession: ENSG00000120217) and located in chromosome 9p24.1 spanning approximately 17.6 kb. CD274 gene generates three transcripts. Among them, gene size of the longest one is 3.6 kb (NM_014143.4) and this transcript comprises seven exons, which encodes for a protein with 290 amino acids (NP_054862.1) (Fig. 1a); size of the shortest one is 3.3 kb (NM_001267706.1) and this transcript encodes an isoform with 176 amino acids (NP_001254635); size of the rest one is 868 bp (NM_001314029.2) and this transcript encodes an isoform with 245 amino acids (NP_001300958.1) [8].

Fig. 1.

Structure of PD-L1. a Scheme of PD-L1 in gene, mRNA and protein level. b Crystal structure of PD-L1 IgV domain (blue) with PD-1 (green), PDB code:4ZQK

mRNA and protein structure

The encoded general PD-L1 protein with long transcript has a mass of 33.275 kDa, consisting of N-terminal amino acid sequence and signal peptide (encoded by exon 1), two annotated immunoglobulin V-like (encoded by exon 2) and C-like (encoded by exon 3) domains, a hydrophobic transmembrane fragment and a cytoplasmic tail (encoded by exons 4–7) (Fig. 1a). The first 18 amino acids contain the signal peptide sequence, which are removed during protein processing. The extracellular domains including IgV and IgC-like domains (19–238 amino acids) are important functional domains and play critical roles in interacting with the receptor PD-1 to mediate the activation threshold of T cell and inhibit T cell effector response (Fig. 1b). In structure, PD-L1 IgV domain interacts with PD-1 via the residues of C’CFG strands by forming hydrophobic interactions and polar interactions (Fig. 1b). Following IgV and IgC-like domains, there is the transmembrane region (239–259 amino acids) that connects the intracellular regions. The function of cytosolic tail (260–290 amino acids) is still unclear in signal transduction (Fig. 1a) [8].

Glycosylation modification of PD-L1

Before PD-L1 is translocated to the cell membrane for exerting its biological functions, there are various post-translational modifications emerging as crucial regulatory mechanisms of PD-L1. Among these post-translational modifications, asparagine (N)-linked glycosylation of PD-L1 is a general expression form. Multiple studies have shown that in the majority of PD-L1 expressed cells, PD-L1 is highly glycosylated in heterogeneous expression pattern [9, 10]. This glycosylation is a biosynthetic secretory pathway in the endoplasmic reticulum and Golgi apparatus. Initially, when synthesized nascent PD-L1 enters the endoplasmic reticulum lumen, oligosaccharyltransferase transfers a 14-sugar core glycan from dolichol to an N residue of an N–X–T/S motif (N is asparagine, X is any amino acid except proline, S is serine, and T is threonine). Then, the core glycan is trimmed and further processed in the endoplasmic reticulum and golgi apparatus. After that, the glycosylated PD-L1 is translocated to the cell membrane to exert its biofunctions. In this process, if the glycosylation on PD-L1 is abnormally trimmed, the protein will be translocated to the cytosol and degraded rapidly by endoplasmic reticulum. Because of this common glycosylation, the weight of the majority of PD-L1 ranges from 45 to 55 kDa on western blotting detection, rather than the non-glycosylated PD-L1 in 33 kDa. Mass spectrometric analysis demonstrated that the highly glycosylated asparagine residues of PD-L1 are its extracellular domain-N35, N192, N200 and N219 [11, 12] (Fig. 2).

Fig. 2.

Glycosylation, phosphorylation, ubiquitination, palmitoylation and acetylation of PD-L1. The dotted line refers to the classification of various modifications on PD-L1, and the solid line refers to the specific process of various modifications of PD-L1

Functionally, PD-L1 protein is stabilized by its glycosylation through preventing its degradation from the 26S proteasome, which enhances its interaction with PD-1 on T cell, leading to immunosuppression (Fig. 2) [12]. Therefore, intervention in the glycosylation process of PD-L1 can reduce the stability of PD-L1 and thus enhance immunity.

Phosphorylation of PD-L1

Beside N-linked glycosylation, PD-L1 can also be phosphorylated in its various serine (S) and threonine (T) sites. Glycogen synthase kinase 3β (GSK3β) can phosphorylate nonglycosylated PD-L1 at T180 and S184 of PD-L1 extracellular domain by binding its post-translational motifs (S/TXXXS/T, in which S is serine, T is threonine, and X is any amino acid). The phosphorylation of PD-L1 by GSK3β results in K48 ubiquitination of PD-L1 and induces PD-L1 degradation by proteasome. Glycosylation of PD-L1 in N192, N200, and N219 can form a steric hindrance thereby preventing the interaction between GSK3β and PD-L1, keeping PD-L1 from phosphorylation and subsequent degradation [12, 13]. Similarly, glycogen synthase kinase 3 alpha (GSK3α) also phosphorylates PD-L1 at S279 and S283 and mediates PD-L1 ubiquitination and degradation by proteasome (Fig. 2) [14]. Except to induce degradation PD-L1 by phosphorylating PD-L1, Janus kinase 1 (JAK1) and NIMA related kinase 2 (NEK2) mediated PD-L1 phosphorylation can stabilize PD-L1 through promoting its glycosylation. In pancreatic cancer, NEK2 phosphorylates PD-L1 at its T194/T210 sites by NEK binding motif (F/LXXS/T), thereby stimulating PD-L1 glycosylation at the N192, N200 and N219 sites and promoting PD-L1 stability [15]. In hepatocellular carcinoma, JAK1 induces phosphorylation of PD-L1 at T112 in the presence of interleukin-6 (IL-6), which recruits the N-glycosyltransferase STT3 oligosaccharyltransferase complex catalytic subunit A (STT3A) to aggrandize PD-L1 glycosylation and maintain its stability [16] (Fig. 2). To sum up, phosphorylation of PD-L1 is closely related to its stability and can also be a potential target for enhancing immunotherapy.

Ubiquitination and deubiquitination of PD-L1

In 2017, Zhang et al. found PD-L1 protein is negatively regulated by cyclin d–cyclin-dependent kinase 4 (CDK4) and the E3 ligase cullin3–speckle-type BTB/POZ protein (SPOP) via proteasome-mediated degradation. They indicated that PD-L1 protein expression is fluctuant during cell cycle progression in multiple human cancer cell lines, peaking in M and early G1 phases, and sharp reduction in late G1 and S phases. In this process, SPOP is phosphorylated by cyclin D-CDK4, and it can be stabilized by recruiting 14–3–3 protein gamma (14–3–3γ), thus preventing its binding to fizzy and cell division cycle 20 related 1 (FZR1), a degrader of SPOP. Then, this stabilized cullin3–SPOP can recognize their downstream substrate PD-L1 to promote its ubiquitination and degradation [17]. In lung adenocarcinoma, downregulation of cyclin-dependent kinase 5 (CDK5) can decrease PD-L1 expression through tripartite motif containing 21 (TRIM21) mediated PD-L1 ubiquitination and degradation, which shows TRIM21 ubiquitin modification function in PD-L1 [18]. As mentioned that GSK3β can interact with non-glycosylated PD-L1 and N192, N200 and N219 of PD-L1 are generally glycosylated in cancer cells, PD-L1 glycosylation antagonizes GSK3β binding. If non-glycosylated PD-L1 is phosphorylated by GSK3β, β-transducin repeats-containing proteins (β-TrCP) can be recruited to form a complex, which catalyzes K48 PD-L1 ubiquitination and degradation. In turn, deletion of the F-box of β-TrCP or mutation of the GSK3β phosphorylation motif abrogated GSK3β mediated PD-L1 ubiquitination, suggesting that degradation of PD-L1 is phosphorylation dependent. These findings also reveal the linkage between glycosylation and ubiquitination of PD-L1, which provides some basis for the regulation of glycosylation on PD-L1 level [19]. Recently, Ariadne RBR E3 ubiquitin protein ligase 1 (ARIH1) was identified as another E3 ubiquitin ligase for PD-L1 ubiquitination. Similarly, PD-L1 phosphorylation at Ser279/283 driven by GSK3α is proved to promote K48-linked ubiquitination of PD-L1 by ARIH1 when epidermal growth factor receptor (EGFR) is abrogated, leading to the destabilization of PD-L1 [14]. Under similar conditions with EGFR blocked, PD-L1 can be degraded through membrane-associated RING-CH 8 (MARCH8) interacting with PD-L1 N-terminal region and inducing its ubiquitination [20]. In addition, E3 ubiquitin ligase STIP1 homology and U-Box containing protein 1 (STUB1) is also revealed to promote the polyubiquitination and destabilization of PD-L1 [21] (Fig. 2).

On the other hand, deubiquitinase COP9 signalosome 5 (CSN5) is further demonstrated to catalyze the proteolytic removal of polyubiquitin from PD-L1 to maintain its stability in TNF-α mediated PD-L1 expression. In details, TNF-α upregulated p65 induces transcriptional activation of CSN5, thereby maintaining PD-L1 stabilization via preventing PD-L1 ubiquitination. CSN5 inhibitor combination with cytotoxic T-lymphocyte associated protein 4 (CTLA4) blockade antibody can destabilize PD-L1 and then enhance the therapeutic efficacy in 4T1 breast cancer, B16 melanoma, and CT26 colon cancer syngeneic mouse models [22]. After that, ubiquitin-specific peptidase 9, X-linked (USP9X) was reported to interact with PD-L1 and induce deubiquitination and stability of PD-L1 in oral squamous cell carcinoma. In addition, USP9X is overexpressed and promotes tumor growth in oral squamous cell carcinoma [23]. Otherwise, ubiquitin specific protease 22 (USP22) can interact with the intracellular terminus of PD-L1 directly to induce its deubiquitination and stabilization, which inhibits tumor immunity and caused poor prognosis in liver cancer [24]. On the other hand, USP22 is indicated to stabilize PD-L1 by deubiquitinating CSN5 to suppress T cell cytotoxicity and induce tumorigenesis in non-small cell lung cancer (NSCLC) [25]. Similarly, ubiquitin thioesterase OTUB1 (OTUB1) and ubiquitin specific peptidase 21 (USP21) can interact with intracellular domain of PD-L1 and remove PD-L1 ubiquitin chains to stabilize PD-L1, which causes low sensitivity of tumor cell to T cell cytotoxicity as well as decreased CD8⁺ T cells infiltration in tumor and these two deubiquitinases may be potentially immunotherapeutic targets for lung cancer treatment [26, 27]. Beyond that, ubiquitin specific peptidase 7 (USP7), another USPs, has been reported to upregulate PD-L1 and suppress tumor immunity in gastric cancer, which indicates that USP7 may be a deubiquitinase of PD-L1 [28] (Fig. 2). Taken together, these results indicated that enhancing PD-L1 ubiquitination or inhibiting its deubiquitination through these proteins can be an important way to suppress the immunosuppressive function of PD-L1.

Palmitoylation of PD-L1

Palmitoylation is another modification, whereby the cysteine residue of protein covalently links a palmitate and is mediated by proteins containing a conserved Asp–His–His–Cys (DHHC) domain [29]. In breast cancer, zinc finger DHHC-type palmitoyltransferase 9 (ZDHHC9) can interact with PD-L1 and catalyze its palmitoylation at C272 site to stabilize PD-L1. Blocking palmitoylation of PD-L1 can reduce PD-L1 expression and increase breast cancer cells sensitivity to T cell killing and suppress tumor growth [30]. Further study found that PD-L1 can also be palmitoylated by zinc finger DHHC-type palmitoyltransferase 3 (ZDHHC3) at C272 site and attenuated PD-L1 degradation through lysosomal pathway. The authors also designed peptide targeting PD-L1 palmitoylation to induce PD-L1 degradation and enhanced cancer immunity [31]. Blocking or inhibiting these palmitoyltransferases also show excellent tumor suppressive effect; therefore, palmitoylation of PD-L1 can also be a potential target for tumor immunotherapy [30, 31] (Fig. 2).

Acetylation of PD-L1

Through endocytosis and nucleocytoplasmic transport pathways, PD-L1 can transfer from plasma membrane to nucleus controlled by K263 acetylation in its intracellular region. In addition, this acetylation is mediated by histone acetyltransferase p300 and histone deacetylase 2 (HDAC2). Unacetylated PD-L1 can be accumulated in the nucleus through huntingtin interacting protein 1 related (HIP1R) and adaptin β2 (AP2B1) mediated endocytosis to enhance the activation of multiple immune-response pathways [32]. This finding revealed that PD-L1 can regulate immune-response pathways through the nucleus location, which provide a new strategy for immunotherapy through targeting PD-L1 translocation (Fig. 2) (Table 3).

Table 3.

Regulators of PD-L1 in several levels

Regulator	PD-L1 expression	Cancer/cell type	Tumor infiltrating immune cells	References
GSK3β	Down	Breast cancer	High tumour-infiltrating lymphocyte	[12]
GSK3α	Down	Non-small cell lung adenocarcinoma cancer Lymphosarcoma	High infiltrated CD8+ cytotoxic T cells	[14]
JAK1	Up	Hepatocellular carcinoma	Low CD8+ T cell infiltration	[16]
NEK2	Up	Pancreatic cancer	Low CD8+ T cell infiltration	[15]
CDK4	Down	Colon cancer Melanoma Breast cancer	High tumour-infiltrating lymphocyte	[17]
SPOP	Down	Colon cancer Melanoma Breast cancer	High tumour-infiltrating lymphocyte	[17]
TRIM21	Down	Lung adenocarcinoma	More CD3+ T, CD4+ T and CD8+ T cell	[18]
β-TrCP	Down	Breast cancer	High tumor-infiltrating lymphocyte	[12]
ARIH1	Down	Non-small cell lung adenocarcinoma cancer Lymphosarcoma	High infiltrated CD8+ cytotoxic T cells	[14]
MARCH8	Down	NSCLC		[20]
STUB1	Down	Melanoma		[21]
CSN5	Up	Breast cancer Melanoma Colon cancer	Low tumor-infiltrated activated CD8 + T cell	[22]
USP9X	Up	Oral squamous cell carcinoma		[23]
USP22	Up	Liver cancer NSCLC	Low tumor-infiltrating lymphocyte	[24, 25]
OTUB1	Up	Breast cancer	Low CD8+ T cells infiltration	[26]
USP21	Up	Lung cancer		[27]
USP7	Up	Gastric cancer		[28]
ZDHHC9	Up	Breast cancer	Low CD8+ T cells infiltration	[30]
ZDHHC3	Up	Colon cancer	Low CD8+ T cells infiltration	[31]
HDAC2	Up	Breast cancer	low CD8 + cytotoxic T-cell infiltration	[32]
IFN-γ	Up	Glioma Renal tubular epithelial cells	Low T cells infiltration	[143, 146]
JAK–STAT	Up	NSCLC Triple-negative Breast Cancer	Low T cells infiltration	[213, 214]
NF-κB	Up	Melanoma Natural killer/T-cell lymphoma Gastric Carcinoma	Low tumour-infiltrating Lymphocyte	[151, 215, 216]
TGF β	Up	Lung cancer	Regulatory T cell expansion	[155]
IL-12	Up/down	Endothelial cells and monocyte-derived macrophages/THP-1 derived macrophages		[158]
MYC	Up	Melanoma NSCLC Leukemia Lymphoma Hepatocellular carcinoma	Low tumour-infiltrating lymphocyte	[164]
HIF-1α	Up	Non-small cell lung cancer Pulmonary adenocarcinoma Breast cancer Prostate cancer	High PD-1+ tumor-infiltrating lymphocytes Resistance to cytotoxic T lymphocytes-mediated lysis	[169] [170] [171]
PI3K–AKT	Up	Colon cancer Melanoma NSCLC	Decreased CD3+ T cells and increased FoxP3+ Tregs	[175] [177] [178]
MEK–ERK	Up	Myeloma Breast cancer Melanoma Lung adenocarcinoma	Decreased CD8+ T cells Reduced tumor-infiltrating lymphocytes	[182] [183] [184] [187]
EGFR	Up	Melanoma Glioblastoma Renal Cancer	Reduced CD8 +  T cell infiltration	[191] [192] [193]
KRAS	Up	NSCLC Lung adenocarcinoma	Induced apoptosis of CD3+ T cells	[186] [194]
CDK5	Up	Medulloblastoma	Decreased T cell infiltration	[195]
miR-513	Down	Cholangiocytes		[196]
miR-155	Down	Dermal lymphatic endothelial cells		[197]
miR-34a	Down	AML NSCLC	Decreased PD-L1+/CD8+ T cell apoptosis Increased CD8 +  T cell infiltration	[198] [199]
miR-142-5p	Down	Pancreatic cancer	Increased CD4 + and CD8 + T lymphocytes,	[200]
miR-25–93-106b	Down	Pancreatic cancer		[201]
miR-217	Down	Laryngeal cancer		[203]
miR-152	Down	Gastric cancer		[204]
miR-200b	Down	Gastric cancer		[204]
miR-17-5p	Down	Melanoma		[205]
miR-15a/miR-16/miR-193a-3p	Down	Malignant pleural mesothelioma		[206]
PKM2	Up	DCs, T cells, and colon carcinoma cells		[209]
Glutamine deprivation	Up	Lung cancer, colon cancer cells		[211]
CMTM4	Up	Melanoma		[21]
CMTM6	Up	Melanoma Breast cancer lung cancer	Decreased cytotoxic T lymphocyte	[33]
HIP1R	Down	Colorectal cancer pancreatic cancer, melanoma breast cancer, lung cancer		[212]

The location and functions of PD-L1

Location and distribution of PD-L1

PD-L1 is predominantly located in plasma membrane and endosomes in cells. Through endosome genesis and maturation process, little of PD-L1 located in recycle endosome membrane can be recycled to the cytomembrane by endosome recovery [33]. Moreover, PD-L1 in endosome can be degraded by lysosomes when endosome are maturated [33, 34]. Besides, PD-L1 in endosome can also be secreted outside the cell by exosomes [33–36] (Fig. 3a).

Fig. 3.

Subcellular level (a), cellular level (b) and tissue (c) distribution of PD-L1

In cells, T cells and B cells only express negligible amounts of PD-L1, and a fraction of monocytes macrophages, mast cells and dendritic cells generally express PD-L1. However, PD-L1 can be upregulated in T and B cells, dendritic cells, keratinocytes and monocytes upon lipopolysaccharide (LPS) and interferon gamma (IFN-γ) stimulation. Not only in normal immune related cells, PD-L1 is expressed in a variety of other cells either, including placenta, vascular endothelium, pancreatic islet cells, muscle, hepatocytes, epithelium and mesenchymal stem cells. Furthermore, various of cancer cells overexpresses PD-L1 to escape from immunologic surveillance, leading to the enhancement of cancer cell growth. Diverse solid tumors including melanoma, renal cell carcinoma, NSCLC, thymoma, ovarian, gastric cancer and colorectal cancer overexpress PD-L1 to form an immunosuppressive microenvironment, thereby escaping T cell killing [37–44]. Consistently, increased PD-L1 expression was revealed in multiple hematologic cancers, such as acute myeloid leukemia, B-cell lymphomas, and multiple myeloma to increase immune escape (Fig. 3b) [34, 38, 45]. In tissue specificity, PD-L1 is highly expressed in the heart, skeletal muscle, placenta and lung, and it is low expressed in thymus, spleen, kidney and liver, and there is almost no expression in brain, colon and small intestine tissues, peripheral blood and other tissues and organs [46] (Fig. 3c). The distribution of PD-L1 in a variety of cells and tissues may be one of the reasons for the adverse drug reactions and different patient response rates of drugs targeting PD-1/PD-L1. Limiting the tissue distribution of these drugs may also be a solution to these problems.

Physiological function

In human immune system, T cell based immune regulation plays a key role in recognizing and destroying aberrant cells, especially the pathogen-infected cells and cancer cells [1, 47]. These immunosurveillance and clearance processes are extremely complex physiological response process, which are controlled by a series of co-stimulatory and co-inhibitory receptors and their ligands [48]. In this process, TCR on T cells recognizes MHC on target cells, and coactivator CD28 on T cells binds to the CD80/CD86 on APC or cancer cells to activate T cells functions. To limit tissue damage and maintain immune tolerance, the co-inhibitory receptors and their ligands also play crucial and necessary roles in T cell mediated immune system [49]. The most significant function of PD-L1 is to act as a co-inhibitory receptor that can bind to its ligand PD-1, thereby inhibiting the activation of T cell and limit T cell effector response [46]. The inhibition of T cell activation can negatively regulate T cell proliferation, survival, cytokine production, and other effect on target cells. Mechanically, interaction between the extracellular domain of PD-L1 and PD-1 induces a conformational change in PD-1, where the cytoplasmic immunoreceptor tyrosine-based inhibitory motif (ITIM) and the immunoreceptor tyrosine-based switch motif (ITSM) of PD-1 is phosphorylated by Src family kinases. These phosphorylated tyrosine motifs can subsequently recruit the tyrosine phosphatases Src homology 2 domain–containing tyrosine phosphatase 2 (SHP-2) and Src homology 2 domain–containing tyrosine phosphatase 1 (SHP-1) to inhibit T cell activating signal [50]. In the early reports, this immunosuppressive function was suggested to decrease the activation of the TCR signal, while recent studies revealed that CD28, other than the TCR, may be a primary medium in this phosphorylation process. In addition to its interaction with PD-1, PD-L1 can interact with CD80 as well, and this interaction can competitively inhibit T cell activation pathway which is mediated CD80 binding to its ligand [51, 52]. However, if PD-L1 and CD80 are cis-expressed, they can also interact with each other and this binding can block both PD-1/PD-L1 and CD80/CTLA-4 interactions thus enhancing T cell activity [53] (Fig. 4a, b).

Fig. 4.

PD-L1 function in T cell-mediated immunity and tumor immunology. a Immune checkpoint in T cell-mediated immunity. b Diagrammatic figure of tumor immunology and PD-L1 antibody action mechanism

Immune function of PD-L1

Besides its indispensable function in immune system, the PD-1/PD-L1 pathway is stood out among all suppressed immune checkpoints because of its overexpression and its great value as a therapeutic target in various cancers. In cancer cells, interaction between PD-L1 and PD-1 conducts pro-survival signal to cancer cell, resulting in the resistance to tumor necrosis factor receptor superfamily member 6 (Fas)- or staurosporine-induced cell apoptosis [54]. Furthermore, tumor-associated PD-L1 facilitates Interleukin-10 (IL-10) production in human peripheral blood T cell and results in immune suppression [46]. In addition, PD-L1 can also induce T cell dysfunction and promote T cell anergy in vitro and in vivo [55]. Otherwise, PD-L1 can protect tumor cells from the cytotoxic effects of type I and type II interferons and cytotoxic T lymphocyte (CTL)-mediated cytolysis even without PD-1 in T cell, indicating that the cytoplasmic domain of PD-L1 may be critically required for this cytotoxic effects, but it remains unclear what intracellular factor participates in such proposed function [54].

Due to the immune inhibitory role of PD-L1 in tumor, amount of PD-L1 in tumor microenvironment may be used to predict the therapeutic response to PD-1/PD-L1 interaction blocker when PD-L1 expression in tumor tissue ranges from 14 to 100%, such as in melanoma, NSCLC, bladder cancer [56–58]. While the conclusions are disputed in other related studies, it has been reported that amount of PD-L1 in immune infiltrating cells rather than tumor cells shows more reliable clinical predictive value in bladder cancer [59]. In NSCLC, PD-L1 negatively expressed patients in tumor microenvironment are still sensitive to PD-L1 immunotherapy, and their results indicate that PD-L1 expression in immune infiltrating cells and tumor cells may both contribute to the response of PD-L1 immunotherapy [60]. Besides, PD-L1 in non-tumor cells plays a dispensable role in B16 melanoma model, while in MC38 colorectal cells, PD-L1 in both tumor and non-tumor contributes to tumor growth. These results indicate that PD-L1 expression in both cells may impact on tumor growth and immunotherapy response, but the concrete results should be dependent on the context [61]. Unexpectedly, many studies revealed that clinical response of some PD-1/PD-L1 inhibitors is not dependent on PD-L1 expression as some PD-L1 positively expressed patients do not obtain efficient response, and a fraction of PD-L1 negatively expressed patients perform acceptable response to PD-1/PD-L1 targeting immunotherapy. These results may arise from loss of cancer cell sensitivity to T cells effector mechanisms, or expression of PD-1/PD-L1 outside of the tumor microenvironment or inducing immune inhibition function through PD-1/PD-L2 pathway [62, 63]. Therefore, further research is still necessary to explain this phenomenon to make the application of PD-1 inhibitor more precise.

Nonimmune function of PD-L1

In addition to immunosuppressive function of PD-L1 in T cells immunity, more and more non-immune functions of PD-L1 have also been reported. In 2008, there has been reported that PD-L1 does not entirely depend on PD-1 to induce resistance against T cell mediated killing [54]. Absence of PD-L1 inhibits this resistance, but not PD-1. In addition to resistance against T-cell destruction, PD-L1 in cancer cells also increases resistance against apoptosis induced by Fas ligation or the protein kinase inhibitor, which reveals a new mechanism of PD-L1 in cancer [54]. PD-L1 overexpressed cancer cells are chemoresistant through the activation of mitogen-activated protein kinase/extracellular-signal regulated kinase (MAPK/ERK) pathway. By analyzing its downstream genes, researchers revealed that PD-L1 may be involved in DNA damage repair [64]. Besides, PD-L1 can be induced by serine/threonine–protein kinase B-raf (BRAF) ^V600E, an oncogenic mutant, and then promotes chemotherapy-induced DNA damage and apoptosis in colon cancers, which is unrelated to PD-L1 immune functions [65].

Moreover, basal expression of PD-L1 in the melanoma cells can promote tumorigenesis as well as metastasis to lung in NOD–SCID IL-2 receptor gamma null (NSG) mice by suppressing autophagy and mammalian target of rapamycin (mTOR) signaling [42, 66]. PD-L1 can be localized in nuclear in circulating tumor cells or doxorubicin-treated breast cancer cells, which is associated with prognostic and combination therapy of cancers [67, 68]. Hou et al. also reported that PD-L1 takes on nuclear transcriptional activity and is involved in the pyroptosis pathway, which is mediated by tumor necrosis factor α (TNFα)-activated caspase-8 as well as hypoxia-activated gasdermin C (GSDMC), resulting in tumor necrosis in hypoxic area [69]. Basal expression of PD-L1 is related with gastric cancer cell migration and PD-L1 ablation can induce cell migration and wound reparation which is unrelated to PD-L1 immune functions [70].

In addition to cancer cells, adipocyte can express PD-L1 and adipocyte-specific PD-L1 suppresses cancer immunity against mammary tumors and melanoma. PD-L1 deficiency in adipocyte also aggravates diet-induced body weight gain, pro-inflammatory cells infiltration into adipose tissue as well as insulin resistance, which is correlated with high body mass index and type 2 diabetes [71]. The nonimmune functions of PD-L1 may reveal more applications of drugs targeting PD-L1 and explain the possible reason of the different response rates of antibody drugs targeting PD-L1 in different patients.

Development of tumor immunotherapy targeting PD-1/PD-L1

Antibody blockers

By reason of the abnormal expression of PD-L1 in many malignant tumors, more and more evidences proved that PD-L1 could be a therapeutic target in numbers of malignancies [72]. Antibody drug targeting PD-1/PD-L1 axis brought hope for the treatment of many cancers (Fig. 4b). Pembrolizumab, a humanized monoclonal IgG4 antibody, was the first PD-1 antibody approved for patients with advanced or unresectable melanoma on September 4, 2014 [73]. Subsequently, because of its significant antitumor effect performed in NCT01866319 trial, pembrolizumab was approved for expanded first-line use for various advanced melanoma regardless of BRAF mutation status on December 18, 2015 [74, 75]. After that, pembrolizumab was also approved for treatment of previously treated advanced or metastatic PD-L1 positive NSCLC [76], following with accelerated approval for recurrent or metastatic head and neck squamous cell carcinoma (HNSCC) [77]. Recently, pembrolizumab was used in the treatment of urothelial carcinoma [78], Hodgkin lymphoma [79], microsatellite instability or mismatch repair deficient cancers [80], and gastric cancer [81] (Table 1).

Table 1.

PD‐1 and PD‐L1 antibody inhibitors (Approved agents)

Name	Target	Type	Cancer	Overall response rate (Dose)	Median OS/PFS	References
Pembrolizumab	PD-1	IgG4 antibody	Melanoma PD-L1-positive NSCLC Metastatic HNSCC Urothelial carcinoma Hodgkin lymphoma Microsatellite instability or mismatch repair deficient cancers Locally Advanced or Metastatic Gastric or Gastroesophageal Junction Adenocarcinoma	24% (2 mg/kg/2 weeks) 26% (10 mg/kg/2 weeks) 41% (10 mg/kg/2 weeks) 16.1% (10 mg/kg/2 weeks) 28.6% (200 mg/3 weeks) 69.0% (200 mg/3 weeks) 39.6% (10 mg/kg/2 weeks or 200 mg/3 weeks) 13.3% (200 mg/3 weeks)	18 months; 12-month OS rate was 58% (2 mg/kg) and 63% (10-mg/kg dose) 10.3 months 6–10 months 7.8 months 9-month OS and PFS rates were 97.5% and 63.4% 5.8 months	[73] [76] [77] [78] [79, 124] [80] [81]
Nivolumab	PD-1	IgG4 antibody	Unresectable or metastatic melanoma Metastatic squamous NSCLC Advanced or metastatic Urothelial carcinoma Advanced RCC HNSCC Hematologic malignancy Metastatic colorectal cancer Advanced hepatocellular carcinoma	31.7% (3 mg/kg/2 weeks) 19% (3 mg/kg/2 weeks) 23.8% (PD-L1 expression ≥ 1%); 16.1% (PD-L1 expression < 1%) (3 mg/kg/2 weeks) 21.5% (3 mg/kg/2 weeks) 13.3% (3 mg/kg/2 weeks) 65% (3 mg/kg/2 weeks) 53% (3 mg/kg/2 weeks) 14.3% (240 mg/2 weeks)	8.9 months 12.2 months 11.3 months (PD-L1 expression ≥ 1%); 5.95 months (PD-L1 expression < 1%) 25.0 months 7.5 months PFS was 14.7 months 5 months 15.6 months	[82] [83] [84] [85] [86] [87, 88] [89] [90, 91]
			Esophageal or gastroesophageal junction cancer	5.6% (200 mg/3 weeks)	Disease-free survival was 22.4 months	[92]
Atezolizumab	PD-L1	IgG1 antibody	NSCLC Metastatic urothelial carcinoma	31% (1200 mg/3 weeks)		[93] [78]
Durvalumab	PD-L1	IgG1 antibody	Advanced or metastatic Urothelial carcinoma	17.8% (10 mg/kg/2 weeks)	18.2 months	[94, 95]
			Advanced bladder cancer		14.4 months	[96]
			Small cell lung cancer	68% (1500 mg/3 weeks plus chemotherapy arm)	13.0 months	[97]
Avelumab	PD-L1	IgG1 antibody	Metastatic urothelial carcinoma Merkel cell carcinoma	16.5% (10 mg/kg/2 weeks) 31.8% (10 mg/kg/2 weeks)	7.0 months 12.6 months	[98] [99]
Tislelizumab	PD-1	Monoclonal IgG4 antibody	Hodgkin’s lymphoma NSCLC Hepatocellular carcinoma	87.1% (200 mg/3 weeks) 12.2% (200 mg/3 weeks) 13.3% (200 mg/3 weeks)	9-month OS rates was 98.6% 11.5 months 13.2 months	[100] [101] [102]
Dostarlimab	PD-1	Monoclonal IgG4 antibody	Mismatch repair deficient recurrent or advanced endometrial cancer	43.5% (500 mg/3 weeks)	PFS was 8.1 months	[103]
Camrelizumab	PD-1	Monoclonal IgG4 antibody	Hodgkin lymphoma Nasopharyngeal cancer NSCLC Hepatocellular carcinoma Oesophageal squamous cell carcinoma	77.3% (200 mg/2 weeks) 34% (200 mg/2 weeks) 21.2% (200 mg/3 weeks) 11% (200 mg/2 weeks) 33.3% (200 mg/2 weeks)	Six-month progression-free survival was 48.2% PFS has not yet reached 6-month overall survival was 76.1% Progression free survival was 3.6 months	[125] [104]
Sintilima	PD-1	Monoclonal IgG4 antibody	Hodgkin lymphoma	67.9% (200 mg/3 weeks)		[126] [104, 105]
Cemiplimab	PD-1	IgG4 antibody	Advanced cutaneous squamous cell carcinoma NSCLC Metastatic basal cell carcinoma	47% (3 mg/kg/2 weeks) 39% (350 mg/3 weeks) 21% (350 mg/3 weeks)	8.1 months PFS was 8.2 months 25.7 months	[106] [107] [108]
Toripalimab	PD-1	IgG 4 monoclonal antibody	Unresectable or metastatic melanoma Nasopharyngeal carcinoma Urothelium carcinoma	17.3% (240 mg/2 weeks) 23.9% (240 mg/2 weeks) 27.2% (240 mg/2 weeks)	22.0 months 15.1 months 14.6 months	[109, 127] [110] [111]
Penpulimab	PD-1	IgG1 monoclonal antibody	Hodgkin lymphoma	89.4% (200 mg/2 weeks)	12-month PFS rate was 72.1%	[112]
Zimberelimab	PD-1	Monoclonal IgG4 antibody	Hodgkin lymphoma	90.6% (240 mg/2 weeks)	12-month OS rate was 99%	[113, 128]

In the same year as the discovery of pembrolizumab, nivolumab, another fully human IgG4 PD-1 immune checkpoint inhibitor antibody, was approved for second-line or later-line treatment of unresectable or metastatic melanoma [82]. Next year, nivolumab was approved for the treatment of metastatic NSCLC as it can significantly prolong the overall survival, response rate and progression free survival of patients in a PD-L1 expression independent manner [83]. Afterwards, because of its meaningful clinical benefit, nivolumab was approved by FDA for locally advanced or metastatic urothelial carcinoma [84]. Meanwhile, nivolumab was also the first PD-1 blocker that was approved for treatment of advanced renal cell carcinoma (RCC) [85]. In addition, nivolumab was approved for treatment of HNSCC [86], hematologic malignancy [87, 88], metastatic colorectal cancer [89], and advanced hepatocellular carcinoma [90, 91]. This year, FDA also approved nivolumab for patients with completely resected esophageal or gastroesophageal junction cancer [92] (Table 1).

Atezolizumab, a PD-L1 immune checkpoint antibody, was approved for treatment of metastatic NSCLC on October 18, 2016 based on the clinical study. It can significantly improve survival of NSCLC patients and this response is correlated with PD-L1 expression in cancer cells and tumor-infiltrating immune cells [93]. In advanced or metastatic urothelial carcinoma, atezolizumab was also approved as a first-line treatment and increased levels of PD-L1 expression on immune cells can increase response to atezolizumab [78]. Whereas, another research showed atezolizumab treatment can cause a clinically various improvement of overall survival and be independent of PD-L1 expression [63] (Table 1).

Durvalumab, a monoclonal antibody of PD-L1, is also an approved drug for treatment of platinum-resistant primary advanced or metastatic urothelial carcinoma in PD-L1-positive patients [94, 95] as well as advanced bladder cancer [96], small cell lung cancer [97]. Another PD-L1 IgG1 antibody, avelumab, have received accelerated approval for refractory metastatic urothelial carcinoma [98] and Merkel cell carcinoma [99]. Last year, tislelizumab, another anti-human PD-1 monoclonal IgG4 antibody, was approved for classical Hodgkin’s lymphoma [100], NSCLC [101] and hepatocellular carcinoma [102]. Dostarlimab was also approved by FDA for treatment of women with mismatch repair deficient recurrent or advanced endometrial cancer in this year [103] (Table 1).

Camrelizumab is a PD-1 inhibitor that received conditional approval in China for the treatment of relapsed or refractory classical Hodgkin lymphoma, nasopharyngeal cancer, NSCLC, hepatocellular carcinoma and oesophageal squamous cell carcinoma [104]. In addition, Sintilima, another PD-1 inhibitor, has received approval in China for the treatment of relapsed or refractory classical Hodgkin lymphoma [104, 105]. Moreover, there are other four PD-1 antibodies have been approved in China, including cemiplimab, toripalimab, penpilimab and zimberelimab. Cemiplimab, a PD-1 blocking antibody, was approved by FDA for advanced cutaneous squamous cell carcinoma in 2018 [106], and it was approved for the first-line treatment of patients with advanced NSCLC with high PD-L1 expression [107] as well as metastatic basal cell carcinoma in this year [108]. Toripalimab, a recombinant PD-1 monoclonal antibody, was approved for the treatment of unresectable or metastatic melanoma [109], nasopharyngeal carcinoma [110] and urothelium carcinoma [111]. Penpulimab and zimberelimab, two PD-1 antibodies, were approved for Hodgkin lymphoma recently [112, 113] (Table 1).

Up to now, there are thousands of clinical trials for anti-PD-1/PD-L1 mAbs. In recent 2 years, number of clinical trials of anti-PD-1/PD-L1 mAbs raised more sharply. By analyzing these clinical trials, scientists revealed the number of new monotherapies is declining in recent 2 years. Combination trials develop quickly and have taken over new clinical development, and China has the highest recruitment rate in both monotherapy and combination. Based on the number of clinical trials, vascular endothelial growth factor (VEGF), CTLA4, poly-ADP–ribose polymerase (PARP), TGFβ/TGFβR, EGFR, histone deacetylase (HDAC), CDKs et al. target therapy are all combined with anti-PD1/PD-L1 for cancer treatment [6]. PD‑L1 antibody treatment combined with anti‑VEGF can bring about a higher percentage of CD8⁺ T cells and higher level of C–X–C Motif Chemokine Receptor 3 (CXCR3) ligands were observed in tumor tissues [114]. Blockade of autocrine VEGF also remarkably down‐regulated PD‐L1 expression on M2 macrophages, which decreased CD4⁺/CD8⁺ T cells in the peripheral blood and Treg cells [115]. This combination can not only suppress tumor growth, but also reprogram immunosuppressive microenvironment into an immunostimulatory microenvironment [116]. Combined PD-1 and CTLA-4 blockade were found an increase in CD4⁺ Teff/Treg cell ratio and CD8⁺ T cell/Treg cell ratio in tumor tissue [117]. Similarly, combined inhibition of TGF-β signaling or CDKs and the anti-PD-L1 caused increased influx of CD8⁺ T cells and decreased CD4⁺ T cells in the tumor microenvironment [118, 119]. PARP inhibition contributes to accumulation of DNA damage, which may complement anti-tumor activity of PD-1/PD-L1 blockade by expanding neoantigen expression and greater immune recognition of the tumor. PARP inhibitors also can up-regulates PD-L1 expression by inactivating glycogen synthase kinase 3 [120, 121]. Combination treatment of EGFR antibody and PD-L1 blockade upregulates T cells and myeloid cells infiltration thereby enhancing tumor immune response [122]. HDAC inhibition also can enhance PD-1 blockade through upregulating PD-L1 or decreasing the anti-inflammatory phenotype of macrophages [123]. These studies and trials showed the advantages of combination therapies and combinations can avoid resistance mechanisms that block anti-PD1/PDL1 efficacy. Therefore, searching and discovery of more valuable targets to enhance immunotherapy effect and increase benefit groups will be important issues in the future immunotherapy.

Peptides and small molecular inhibitors

With surprising clinical activity, there are evidences that antibody drugs also produced severe immune-related adverse events due to its effect in breaking of immune self-tolerance. This drove the production of some peptides and small molecule inhibitors which have shorter pharmacokinetic profile and are easy to avoid severe adverse effects [129]. AUNP-12 (SNTSESFK(SNTSESF)FRVTQLAPKAQIKE), a 29-mer peptide, was first identified as a modulator targeting the PD-1/PD-L1. It is highly active in HEK293 cells (EC₅₀ of 0.72 nM) and PBMC (peripheral blood mononuclear cell) proliferation assay (EC₅₀ of 0.41 nM). In addition, AUNP-12 can inhibit the growth of mouse melanoma cells and mouse breast cancer cells in vivo. More importantly, AUNP-12 can alleviate immunotherapy-related adverse events because of its short half-life [130], which bring hope for the development of peptides to inhibit PD-1/PD-L1. Based on technology of mirror-image phage display and bacterial surface display, DPPA-1 (D-peptide antagonist, NYSKPTDRQYHF) and TPP-1 (SGQYASYHCWCWRDPGRSGGSK), were developed as two peptide antagonists of PD-L1 with an affinity of 0.51 μM and 94.67 nM, respectively. Both of them can effectively disrupt the PD-1/PD-L1interaction and reverse PD-L1 mediated inhibition of T cell activation in vitro. Besides, DPPA-1 can also inhibit tumor growth and prolong mice survival in vivo [131, 132]. In this year, CLP002 (WHRSYYTWNLNT) was discovered to bind to human PD-L1 with K_D of 366 nM and block PD-1/PD-L1 interaction with IC₅₀ of 2.17 μM. In Jurkat T cells and cancer cells co-cultured model, CLP002 can inhibit apoptosis and restore proliferation of Jurkat T cells. Compared with PD-L1 antibodies, CLP002 shows more efficient tumor penetration ability and needs lower dose for antitumor application in vivo [133]. Among these four above peptides, AUNP-12 was designed by analyzing strands and loops of PD-1 from the interface of PD-1/PD-L1 interaction [130], while DPPA-1 and CLP002 were identified and optimized by library display on phage with 12-mer peptides. Meanwhile, DPPA-1 is a proteolysis-resistant D-peptide antagonist [131]. TPP-1 was identified by screening a random bacterial surface display library with format X5CWCWRX5 structure [132]. The methods and structural characteristics of these peptides screening will provide the basis for the discovery of more potent PD-1/PD-L1 inhibitors. In addition to peptides, small molecule inhibitors targeting PD-1/PD-L1 interaction have also obtained promising results. To overcome immunological antagonistic diseases caused by blocking PD-1/PD-L1 binding, the Bristol–Myers Squibb (BMS) company explored small molecule compounds based on homogeneous time-resolved fluorescence (HTRF) PD-1/PD-L1 binding assay and discovered a series of inhibitors against PD-1/PD-L1 interaction [134]. Among them, BMS202 ((N-(2-[2-Methoxy-6-(2-methyl-biphenyl-3-ylmethoxy)-pyridin-3-ylmethyl]-amino-ethyl)-acetamide)), a small molecule inhibitor to inhibit the PD-1/PD-L1 binding with an IC₅₀ of 0.018 μM, can induce IFN-γ production in vitro. BMS202 exhibits anti-tumor effects through inhibiting Treg (regulatory T cells) expansion and increasing the cytotoxicity of tumor-infiltrating CD8⁺ T cells by inhibiting PD-1/PD-L1 binding in melanoma [135]. In addition, BMS-1001 and BMS-1166 can bind to PD-L1 and inhibit PD-1/PD-L1 interaction with an EC₅₀ of 253 nM and IC₅₀ of 1.4 nM [136].

To overcome the deficiency of high molecular weight, molecular polarity and poor pharmacokinetic parameters of these inhibitors, in this year, A9 was designed based on structure of BMS-1058 (IC₅₀ = 0.48 nM) with a novel 2-(2-methyl-[1,10-biphenyl]-3-yl) pyridine structure, and it shows comparative activity (IC₅₀ = 0.93 nM) to inhibit PD-1/PD-L1 interaction [137]. All of them show less toxicity and can enhance the activity of T cells [136, 137] (Table 2). In addition, there are many other inhibitors targeting PD-1/PD-L1 interaction with novel structures that were modified from reported inhibitors, and can activate Jurkat T cells through inhibiting PD-1/PD-L1 pathway. [1, 2, 4] triazolo [4,3-a] pyridines A22, with an IC₅₀ of 92.3 nM can induce IFN-γ release in dose-dependent manner by T cells [138]. In addition, C2-symmetric compounds LH1306 and LH1307 have been shown to inhibit PD-1/PD-L1 interaction with IC₅₀ of 25 and 3 nM, respectively [139]. A20 and A22, two novel 4-phenylindoline derivatives structure compounds based on BMS-37, are another PD-1/PD-L1 interaction inhibitors with IC₅₀ of 17 nM and 12 nM [140]. All these inhibitors share 2-methyl-3-biphenyl methanol scaffold on which new structures modification can be carried out.

Table 2.

PD‐1 and PD‐L1 peptides and small molecule inhibitors

Name	Target	Sequence/structure	Activity	References
AUNP-12	PD-1/PD-L1	SNTSESFK(SNTSESF)FRVTQLAPKAQIKE	EC50 = 0.72 nM	[130]
DPPA-1	PD-L1	NYSKPTDRQYHF	KD = 0.51 μM	[131]
TPP-1	PD-L1	SGQYASYHCWCWRDPGRSGGSK	KD = 94.67 nM	[132]
CLP002	PD-1/PD-L1	WHRSYYTWNLNT	IC50 = 2.17 μM	[133]
BMS202	PD-L1		IC50 = 0.018 μM	[135]
BMS-1001	PD-L1		EC50 = 253 nM	[136]
BMS-1166	PD-L1		IC50 = 1.4 nM	[136]
A9	PD-1/PD-L1		IC50 = 0.93 nM	[137]
A22	PD-1/PD-L1		IC50 of 92.3 nM	[138]
LH1306	PD-1/PD-L1		IC50 = 25 nM	[139]
LH1307	PD-1/PD-L1		IC50 = 3.0 nM	[139]
A20	PD-1/PD-L1		IC50 = 17 nM	[140]
A22	PD-1/PD-L1		IC50 = 12 nM	[140]
CA-170	PD-1/ VISTA	Unknown		[141]

Based on AUNP-12, CA-170 was discovered as an orally bioavailable small molecule antagonist of PD-1 and VISTA (V-domain Ig suppressor of T-cell activation) and currently underwent phase I clinical trial (NCT02812875). It can inhibit the growth of mouse syngeneic tumors by enhancing peripheral T cell and tumor infiltrating CD8⁺ T cell activation in melanoma and colon carcinoma. In human beings, CA-170 shows no dose limiting toxicity with T_1/2 of 4–9.5 h. It also increases CD8⁺ and CD4⁺ T cells activation to eliminate the suppression of T cell by PD-L [141]. While, an article has been reported to question CA-170 activity for PD-L1. The authors show no direct binding between CA-170 and PD-L1. In addition, CA-170 fails to rescue the activation of PD-1/PD-L1-blocked Jurkat T cells [142]. Therefore, the further investigation of CA-170 in clinical trials are necessary (Table 2).

Inflammatory signaling, oncogenic, miRNA and metabolic pathways regulation of PD-L1 expression

IFN-γ related inflammatory signaling pathway regulation

In the process of PD-L1 transcription, translation and post-translational modification, amount of the PD-L1 has been shown to be regulated by various factors. Among these regulation manners, the inflammatory signaling regulation on PD-L1 is the most vital one. In 2002, Dong et al. first reported that PD-L1 is abundant in human cancerous tissue of lung, ovary, colon and melanomas, though normal tissues hardly express PD-L1, and the pro-inflammatory cytokine IFN-γ upregulates PD-L1 on the surface of tumor cells [56]. After that, regulation of PD-L1 by IFN-γ was characterized in various tumors as well as in immune cells and normal tissues [143]. Glioma cell lines constitutively express PD-L1 mRNA and protein and IFN-γ treatment strongly enhances PD-L1 expression [143]. More than that, results from Brown showed that monocytes cannot generally express either PD-L1 or PD-L2 but their mRNA expression can be induced by IFN-γ [144]. This regulation is also verified by transcriptome analysis in neutrophil that IFN-γ but not interferon alpha (IFN-α) or interferon beta (IFN-β) induces expression of PD-L1 in neutrophils and suppresses lymphocyte proliferation [145]. In renal tubular epithelial cells, PD-L1 is an epithelial antigen that can be induced by IFN-γ to suppress T cell response [146] (Fig. 5).

Fig. 5.

Overview of the born, banish and death of PD-L1 protein. PD-L1 gene is transcribed (mRNA), translated (protein), glycosylated and then transported to cell membrane (green line). PD-L1 is regulated by ① IFN-γ related pathway (red line), ② aberrant oncogenic pathway (yellow line), ③ miRNA regulation (purple line) and ④ metabolic pathways (dark blue line); and death or secretion by ⑤ proteasome degradation pathway (light blue line), ⑥lysosomal degradation pathway (gray line) and ⑦ exosomal secretion pathway (dark grey line). The solid line is the main biological process and IFN-γ regulatory pathway of PD-L1, and the dotted line is several regulators that mediated PD-L1 expression

Mechanically, the regulation of IFN-γ to PD-L1 is through its classical Janus kinase-signal transducer and activator of transcription (JAK–STAT) pathway. IFN-γ activates both common and distinct STAT complexes, which induce the expression of a series of transcription factors [147]. Among them, interferon regulatory factor 1 (IRF1) binds to PD-L1 promoter to induce PD-L1 expression [148]. Therefore, IFN–γ-JAK1/JAK2–STAT1/STAT2/STAT3–IRF1 axis plays an indispensable role in regulating PD-L1 expression among the reported regulation of PD-L1 [149].

Up to now, IFN-γ is generally considered as the most prominent inducer of PD-L1, and expression of PD-L1 can represent IFN-γ signaling and T cell activity in some settings. However, its regulation is not always consistent. Some reports have suggested that regulation of PD-L1 expression through IFN-γ signaling is distinct in tumor and immune cells. Report from Noguchi suggested that induction of PD-L1 in tumor cells is IFN-γ dependent, while this induction is only partially IFN-γ dependent in tumor associated macrophages [60]. Later studies have shown that IFN-α and IFN-β, also belonging to type I interferons, can also induce PD-L1 expression in melanoma cells, endothelial cells, monocytes and dendritic cells [149]. There may be more complex mechanism in the regulation of IFN-γ to PD-L1. Therefore, it is important to study the different effects of IFN-γ on PD-L1 expression in distinct types of cells (Fig. 5).

In addition to direct regulation, many factors can regulate PD-L1 expression through IFN-γ indirectly. LPS, another inflammatory factor, can induce PD-L1 expression through toll-like receptor 4 (TLR4) and nuclear factor kappa B (NF-κB) in type I interferons dependent manner [150]. Hence, inhibition of the NF-κB pathway suppressed IFN-γ mediated PD-L1 upregulation in melanoma cells, which also suggested a possible crosstalk between IFN-γ and NF-κB signaling [151] (Fig. 5).

Transforming growth factor β (TGF β), an anti-inflammatory cytokine, also has a regulatory effect on PD-L1 expression. TGF β is negatively correlated with PD-L1 in systemic lupus erythematosus and suppresses expression of PD-L1 in healthy monocytes, which is on the contrary to tumor necrosis factor alpha (TNF-α), a PD-L1 expression maintainer in lupus monocytes [152]. Similar results were revealed in renal proximal tubular epithelial cells, where TGF β downregulated the basal and the IFN-γ-mediated PD-L1 expression [153]. Nevertheless, in transplant tolerance, blockade of TGF β can decrease PD-1 and PD-L1 expression in T cells and rescue CD8⁺ T cell anergy and this regulation is also associated with regulatory T cell expansion in dendritic cells of lung cancer microenvironment [154, 155]. The report that simultaneous blockade of the PD-L1 and TGF-β pathways elicits potent and superior antitumor activity relative to monotherapies in multiple mouse models, which provides the possibility that TGF-β may perform as a drug target for PD-1/PD-L1 targeting immunotherapy [156]. In addition, dual inhibition of LSD1 and TGFβ can significantly increase CD8⁺ T-cell infiltration and cytotoxicity when combined with PD-1 blockade, indicating LSD1 and TGFβ as significant targets in cancer immunotherapy [157] (Fig. 5).

In addition, interleukin-12 (IL-12) can mediate PD-L1 regulation depending on different conditions. It positively regulated PD-L1 in endothelial cells and monocyte-derived macrophages by inducing the secretion of IFN-γ and subsequent activation of NF-κB signal pathway, while IL-12 suppresses its expression in THP-1 derived macrophages [158]. Combination of oncolysis and secretion of IL-12 and PD-L1 blocking antibody can augment the anti-tumor activity of HER.2 CAR-T cells and this funding may be deeply suited for clinical application [159](Fig. 5).

These data above indicated that PD-L1 expression can be regulated by numbers of inflammatory factors, and most of them are associated with IFN-γ related pathway. Whereas some regulatory mechanisms are still not clear and distinct in different cells, which indicates that the other multiple mechanisms and factors may be involved in regulatory progresses between inflammatory factors and IFN-γ (Fig. 5) (Table 3).

Oncogenic pathway

Because of the vital function in tumor immune escape and immunotherapy, abnormal regulation of PD-L1 in cancers attracts the interest of scientists. Studies in this area not only reveal mechanisms of PD-L1 gene regulation, but also offer evidences and strategies on combinations of immune checkpoint therapies with oncogenic signaling pathways targeted therapies.

In these years, numbers of oncogenic transcription factors have been identified to regulate PD-L1 transcription. Related to above-mentioned IFN–γ-JAK1/JAK2–STAT1/STAT2/STAT3–IRF1 axis, STAT3 can act on the promoter of PD-L1 directly to increase its transcription and expression in human lymphoma and HNSCC cells [160, 161]. Similarly, NF-κB, a family of transcription factors that are activated in cancers, can also bind to PD-L1 promoter in NSCLC cells, monocytes and breast cancer cells, and directly upregulate PD-L1 transcription [162, 163] (Fig. 5).

Myc proto-oncogene (MYC), an oncogene that contributes to tumorigenesis and is upregulated in approximately 70% of human cancers, can induce PD-L1 expression in multiple mouse and human tumor cell models, including melanoma, NSCLC, leukemia, lymphoma, and hepatocellular carcinoma (HCC) [160]. Specifically, MYC expression is positively correlated with PD-L1 expression in NSCLC [164]. Further chromatin immunoprecipitation (ChIP) analysis revealed that MYC can act on PD-L1 promoter, suggesting that PD-L1 expression can be directly regulated by MYC at the transcriptional level. In prostate cancer, inhibitor of MYC can promote immune cell infiltration and potentiates anti-PD1 immunotherapy, suggesting the potential of MYC inhibitor as anti-cancer immunotherapeutic agent [165] (Fig. 5).

In cancers, because of the aberrant need for oxygen supply, hypoxia environment is a key feature in most tumors. Therefore, series of hypoxia-inducible factors are activated to respond to this oxygen deficiency [166, 167], and expression of PD-L1 is also induced by hypoxia-inducible factors. Both hypoxia inducible factor-1α (HIF-1α) and hypoxia inducible factor-2α (HIF-2α) can interact with the HRE (hypoxia response element) in the promoter region of PD-L1 and directly regulate PD-L1 expression [168]. Guo et al. showed that HIF-1α plays a central role in upregulating PD-L1 expression partially dependent on the activation of the NF-κB pathway in NSCLC cells [169]. Consistent results of PD-L1 regulation by HIFs have also been reported in pulmonary adenocarcinoma, breast cancer and prostate cancer cells [170, 171]. Meaningfully, HIF-1α together with PD-L1, CD8⁺TIL was reported as useful prognostic biomarkers for patients of HNSCC [172] (Fig. 5).

Phosphatidylinositol 3-kinase (PI3K) signaling pathway plays a key role in survival, proliferation, metabolism and mobility of cancer cells and its downstream protein kinase B (AKT)–mTOR can be activated by type I and type II interferons and control interferon-dependent PD-L1 mRNA translation [173, 174]. Abrogation of PI3K–AKT signaling pathway genetically or pharmacologically can decrease IFN-γ induced PD-L1 expression in colon cancer cells, NSCLC cells and other cancer cells in an IFN-γ independent manner. These crosstalk between PI3K–AKT and PD-L1 provides evidence for combinatorial targeted therapy and immunotherapy for the treatment of cancers [175–181] (Fig. 5).

Dual specificity mitogen-activated protein kinase–extracellular signal-regulated kinase (MEK–ERK) pathway, a generally activated signal pathway in human cancers, can regulate PD-L1 through crosstalk with inflammatory signaling. Liu and their group showed IFN-γ induced PD-L1 transcription is inhibited when cells are exposed to MEK inhibitors in multiple myeloma and breast cancer cells [182, 183]. Transcription factor AP-1 (JUN), a primary target of MEK signaling, also cooperates with STAT3 in transcriptional regulation of PD-L1 expression [184, 185]. Mechanism study showed that suppression of MEK can decrease PD-L1 expression through inactivation of JUN and STAT3 in melanoma cells and several NSCLC cells [186, 187]. Rat sarcoma (RAS)–MEK signaling also partially regulates expression of PD-L1 by stabilizing its mRNA in human lung and colorectal cancer [188]. In conclusion, MEK is a mediator for many regulatory factors in regulating PD-L1. The latest study showed that ERK inhibitor increased killing of activated CD8⁺T cells to intrahepatic cholangiocarcinoma cells and promoted anti-PD-1/PD-L1 immunotherapy, which indicated ERK as a potentially novel target for cancer immunotherapy [189].

EGFR and Kirsten rat sarcoma viral oncogene (KRAS) are two common carcinogenic factors in human cancers. Activation of EGFR and KRAS drives PD-L1 expression through their downstream pathways. Accumulation of EGFR can activate EGFR/STAT3/PD-L1 signaling pathway in melanoma and lung cancer cells [190, 191]. In glioblastoma, EGF can directly upregulate PD-L1 expression or maintain PD-L1 stability by increasing COP9 signalosome 6 (CSN6) expression [192]. In addition, EGFR can induce PD-L1 expression through EGFR/ERK/c-Jun pathway in renal cancer with glutamine shorted condition [193]. Inhibition of EGFR can sensitize the PD-1 blockade therapy in syngeneic breast cancer mouse model [12]. In some KRAS mutant NSCLC cells and lung adenocarcinoma, silence of KRAS suppressed ERK activation and in turn diminished PD-L1 expression [186, 194] (Fig. 5).

CDK5, which is a key regulator of cell apoptosis and survival and always abnormally expressed in many cancers, can induce PD-L1 mRNA and protein expression through IFN-γ signaling pathway. CDK5 can increase interferon regulatory factor 2 (IRF2) and interferon regulatory factor 2-binding protein 2 (IRF2BP2) abundance through hyperphosphorylation of IRF2BP2 and this regulation then resulted in IRF1-mediated induction of PD-L1 [195]. Combination of shCDK5 and anti-PD-L1 showed a more significant inhibitory effect in lung adenocarcinoma [18]. These results suggested that PD-L1 is associated with many oncogenic pathways, and these pathway inhibitors may also regulate immunity to play antitumor role. Therefore, the study on the immune regulation mechanism of such inhibitors can also become potential directions (Fig. 5) (Table 3).

miRNA regulation

miRNAs have also been revealed to play a critical role in the transcriptional regulation of PD-L1. In 2009, Gong et al. found IFN-γ can negatively regulate miR-513, and promoted PD-L1 expression at the transcriptional level by decreasing the direct binding of miR-513 to the 3’ UTR of PD-L1 mRNA in human cholangiocytes [196]. Then, various miRNAs were identified to downregulate PD-L1 with similar manner. Induced by TNF-α and IFN-γ, miR-155 can bind to two sites of 3'-UTR of PD-L1 to downregulate its expression in human dermal lymphatic endothelial cells [197]. miR-34a, reduces PD-L1 mRNA level in acute myelocytic leukemia (AML) as well as NSCLC cells [198, 199]. miR-142-5p and miR-25-93-106b perform downregulation function in PD-L1 mRNA in pancreatic cancer [200, 201]. miR-138-5p restrains PD-L1 expression in colorectal cancer [202]; miR-217 targets to PD-L1 and inhibits laryngeal cancer metastasis [203]; miR-152 and miR-200b target PD-L1 mRNA and suppress its expression in gastric cancer cells [204]; miR-17-5p decreases PD-L1 mRNA in melanoma [205]; miR-15a, miR-16 and miR-193a-3p result in downregulation of PD-L1 mRNA in malignant pleural mesothelioma [206]. These miRNAs are correlated with numbers of PD-L1 regulators and complicatedly influenced PD-L1 expression and functions (Fig. 5).

Metabolic pathways

Most tumor cells can acquire energy from glycolysis in the presence of hypoxia. Many studies have shown that the glucose metabolism pathway can regulate the expression of PD-L1. In renal cancer cells, PD-L1 is upregulated through the EGFR/ERK/c-Jun pathway when glucose is deficient. In turn, upregulation of PD-L1 can increase the level of glycolysis mediated by 6-Phosphofructo-2-kinase/fructose-2,6-bisphosphatase 3 (PFK-2/FBPase 3 or PFKFB3) [207]. Consistent with the above, glycolysis can increase PD-L1 expression in breast cancer, osteosarcoma, and ovarian cancer [208]. More meaningfully, higher immunotherapy response and favorable survival were found in immunotherapy of highly glycolytic tumors [208]. Pyruvate kinase M2 (PKM2) is a rate-limiting enzyme that converts phosphoenolpyruvic acid to pyruvate in glycolysis process, which participates in the metabolic reprogramming in activated immune cells and in tumor cells. In primary macrophages, PKM2 silence or inhibition can suppress LPS-induced PD-L1 expression by regulating binding of PKM2 and HIF-1α to hypoxia-response element (HRE) sites of PD-L1 promoter. In dendritic cells, T cells, and colon carcinoma cells, inhibitor of PKM2 can also decrease expression of PD-L1 [209].

There is extremely low level of glutamine in tumor tissues and that in the central part of the tumor tissue is lower than peripheral part [210]. The expression of PD-L1 in renal cancer is associated with level of glutamine in renal cancer. Glutamine deprivation can induce PD-L1 expression by activating EGFR/ERK/c-Jun pathway in renal cancer cells, which can be rescued by glutamine recovery or EGFR/ERK/c-Jun inhibition [193]. In human lung cancer cells and the mouse colon cancer cells, blockade of glutamine utilization can cause reduced cellular GSH and upregulation of PD-L1 protein levels by impairing sarcoplasmic reticulum Ca²⁺-ATPase activity, which promotes the antitumor efficacy of T cells when combined with anti-PD-L1 therapy [211] (Fig. 5) (Table 3).

Degradation and secretion of PD-L1

Proteasome degradation of PD-L1

After PD-L1 was identified as the ligand of PD-1 and its crucial function was discovered in cancer immune escape, degradation of PD-L1 arouses great interest among researchers. As mentioned in PD-L1 ubiquitination, cullin3-SPOP, TRIM21, β-TrCP, ARIH1 as well as MARCH8 and STUB1 can catalyze PD-L1 polyubiquitination and induce PD-L1 degradation through proteasome [17–21] (Fig. 5).

Lysosomal degradation

Besides proteasome degradation, PD-L1 in cytoplasm was also identified to be degraded by lysosomes. In 2017, CKLF-like MARVEL transmembrane domain containing 4 (CMTM4) and CKLF-like MARVEL transmembrane domain containing 6 (CMTM6), two closely related members of the CMTM family, were identified as two stabilizers of PD-L1 that can inhibit PD-L1 lysosomal degradation pathway, while the specific mechanism of CMTM4 in lysosomal degradation is unclear. On the other hand, CMTM6 is mostly located at the cell surface and interacts with PD-L1 and maintains PD-L1 stability. Mezzadra et al. also indicated that CMTM6 may decrease PD-L1 ubiquitination that is modified by STUB1 to maintain PD-L1 stability [21]. Further research confirmed that CMTM6 is not required for PD-L1 maturation but co-localized with PD-L1 at the cell membrane and recycling endosomes. As PD-L1 in cytoplasm can be transported by recycling endosomes to membrane for recycle and others were grown into late endosomes and degraded by lysosomes, CMTM6 can bind to PD-L1 in cell surface and induce efficient endocytic recycling of PD-L1, which prevent PD-L1 from being targeted for lysosome-mediated degradation [33]. HIP1R, another opposite regulator of PD-L1 through lysosomal degradation, can target PD-L1 to catalyze its lysosomal degradation by its conserved C-terminal domain. In this process, ALG-2 interacting protein X (ALIX) and endosomal sorting complex required for transport III (ESCRT-III) are involved to deliver PD-L1 to the lysosome degradation. By analyzing HIP1R motifs binding to PD-L1, researchers designed a chimeric peptide (PD-LYSO) that efficiently interacts with PD-L1 and induces its lysosomal degradation in cancer cells, which can promote antitumor immunity [212]. These studies reveal that lysosome mediated degradation of PD-L1 is a significant pathway in PD-L1 protein stability and the formation, recycle, maturity, and degradation of endosomes performed key roles in PD-L1 expression. These regulatory proteins may be used as immunotherapy targets for cancers. Designing of inhibitors targeting CMTM4/CMTM6 or development of more HIP1R analogs or activators would be of great significance to promote immunotherapy (Fig. 5) (Table 3).

Exosome secretion

In addition to being degraded or recycled, PD-L1 can also be secreted by exosomes as vesicles. Exosomes are subsets of extracellular vesicles with diameter of 30–150 nm, which contain proteins, metabolites and nucleic acids as cargos that can be delivered into recipient cells to exert specific function [217]. In 2012, mesenchymal stem cells (MSCs), which were defined as multipotent non-hematopoietic progenitor cells with unique immunosuppressive capacities, have been shown to express variety regulatory molecules, such as PD-L1, membrane-bound TGF-β, galectins and so on. Moreover, MSCs-derived exosomes were first identified to harbor PD-L1, and deliver PD-L1 membrane molecules to target cells. These results suggest that PD-L1 can be harbored into exosomes membrane by endocytosis [218]. Not only PD-L1 protein, PD-L1 mRNA is harbored in exosomes, suggesting that PD-L1 mRNA may be packaged into exosomes during the formation of the exosomes [219, 220]. Metastatic melanoma-derived extracellular vesicles, mostly in form of exosomes, can carry PD-L1 on their surface, and IFN-γ stimulation can increase the amount of PD-L1 on these vesicles. Similar to PD-L1 in cancer cells, exosomal PD-L1 can also suppress the function of CD8⁺ T cells and accelerate tumor growth, and it may be a potential blood-based biomarker for immunotherapy [35, 221]. Moreover, elimination of tumor exosomes by mutating two exosomal secretory genes Ras-related protein Rab-27A (Rab27a) and neutral sphingomyelinase 2 (nSMNase2) can suppress tumor growth, increase activation of a T cell response in lymph nodes in the TRAMP-C2 and MC38 model, which shows the exosomal PD-L1 immunosuppressive function [222]. Exosomes derived from NSCLC cells, breast cancer cells, glioblastoma stem cell, et al. also harbor PD-L1 protein. In NSCLC, exosomal PD-L1 can inhibit IFN-γ secretion and induce apoptosis of CD8⁺ T cells to promote immune escape and tumor growth [36]. Similarly, PD-L1-containing exosomes that are derived from breast cancer cells suppress tumor immunity in breast cancer tumor microenvironment [223]. Meaningfully, PD-L1 on circulating exosomes can be considered as a promising diagnostic and prognostic marker for pancreatic cancer [224]. Overall, exosomal PD-L1 plays an important role in suppressing tumor immunity, and searching the regulator of exosomal PD-L1 is of great significance for tumor immunotherapy and combination therapy. Exosomal PD-L1 may be a new marker to cancer diagnose and predict PD-L1 antibody drugs response of patients (Fig. 5).

Soluble PD-L1 proteins

Soluble PD-L1 is also an important form of PD-L1 [225]. In lung cancer patients, soluble PD-L1 can be found in the plasma as well as in other liquids, such as the pleural effusion [226]. In function, elevated levels of soluble PD-L1 was reported to be associated with worse prognosis as well as with worse outcomes to checkpoint blockade in renal cell carcinoma and multiple myeloma, which may be due to large tumor burden or an exhausted antitumor immune response [226]. Assessment of soluble PD-L1 also shows promise as a new or additional prognostic tool and biomarker in the management of NSCLC [227].

Conclusions and future directions

Among diverse co-inhibitory factors in immune system, PD-1 and PD-L1 play the most significant roles in normal immune regulation and tumor immune escape. As a therapeutic target in numbers of malignancies, there were lots of studies about PD-L1 and its targeting therapeutic strategies. First, PD-L1 is transcribed and translated to protein structure and followed with its post-translational modification including glycosylation, phosphorylation, ubiquitination, palmitoylation and acetylation to mediate its biological function and stability. PD-L1 distributes variously in tissue, cells and organelles and it expresses in both immune cells and non-immune cells, playing important role in regulating T cell immune and mediating various cancers immune escape by binding to PD-1. There have been many antibodies, peptides and small molecules against PD-1/PD-L1 and some of them have been approved for the treatment of cancers. Otherwise, PD-L1 is regulated by many factors including inflammatory factors, oncogene or pathways as well as miRNAs. Finally, degradation through proteasome and lysosome and secretion by exosome of PD-L1 was exposited. This review may provide a comprehensive understanding of PD-L1 as well as a broad prospect for the search of the important regulator of PD-L1 (Table 3).

Though PD-L1 have become an important target for drug development and there have been various approved drugs and clinical trials targeting it. Various clinical response rate and adverse reactions prevent many patients from benefiting from it [228, 229]. Therefore, a comprehensive understanding of the biological process and regulatory pathways of PD-L1 is important for predicting clinical response rates and avoiding adverse reactions. With more and more drugs targeting PD-1/PD-L1 on the market, it is proved that the potential of PD-L1 is huge as a drug target. However, increasing the number of beneficiaries and improving the response rate are still problems. In recent years, combination trials developed quickly and have taken over new clinical development in PD-1/PD-L1 antibodies [6]. Drug combination can improve drug efficacy, reduce drug resistance and increase the personalization of cancer treatment [230, 231]. Therefore, understanding the various regulatory processes and regulatory factors of PD-L1 can provide more combination drug regimens for cancer treatment. Due to the complex effects of drugs, in-depth understanding of the function of drug targets can provide more theoretical basis and feasible strategy for drug combination.

Although there have been numbers of studies of PD-L1 and cancer therapeutics based on PD-1/PD-L1, and various functions and regulatory pathways have been reported [104, 232, 233], how to apply the biological functions and regulatory factors of PD-L1 to the treatment of tumors still needs further exploration.

Author contributions

D-DS wrote the manuscript. Y-PB, J-RP, L-JZ contributed to critical revision of the manuscript. L-FZ, YG, BW, H-ML, YL, NW contributed to minor revision of the manuscript. H-ML and Y-CZ revised and finalized the manuscript.

Funding

This work was supported by the National Key Research Program (No. 2018YFE0195100); the National Natural Science Foundation of China (No. 82020108030); Youth Supporting Program from Henan Province (No. 2021HYTP060); Basic and Frontier Technology Research Project of Henan Province (No. 212102310313); Basic Research of the Key Project of the High Education from the Education Department of Henan Province (No. 22ZX008); Youth Supporting Program from Zhengzhou University (No. JC202044046).

Availability of data and materials

Not applicable.

Declarations

Conflict of interest

The authors have declared that no competing interest exists.

Ethics approval and consent to participate

Not applicable.

Consent for publication

All the authors listed have approved the manuscript for publication.