Highlights

• Summarize the discovery of the human PD-1 gene and its associated pathway.

• Review the current use of PD-1 immune checkpoint inhibition in treating cancers.

• Emphasize challenges to PD-1 immunotherapy and the need for greater understanding of how PD-1 regulates cancer progression.

Abstract

Anti-PD-1 therapy is a novel immune-checkpoint inhibition therapy with tremendous potential in treating refractory/relapsed cancers. The 20 year journey of human PD-1 research went through 3 phases: 1) discovering PD-1 gene structure and genomic organization, 2) understanding the mechanism of PD-1 mediated immune-checkpoint regulatory effects in coordination with its ligands (PD-L1 and L2), 3) and translating our knowledge of PD-1 gene into a robust clinical anticancer approach by targeting the PD-1 immune-checkpoint pathway. The success of human PD-1 gene study reflects the advancement and trends of modern biomedical research from the laboratory to the bedside. However, our journey of understanding the PD-1 gene is not yet complete. Clinical investigation data show a high variety of response rates among different types of cancers to PD-1 immune-checkpoint inhibition therapy, with a range of 18% to 87%. There is no reliable biomarker to predict an individual patient's response to PD-1 inhibitory immunotherapy. Patients can present with primary, adaptive, or even acquired resistance to PD-1 immune-checkpoint inhibition therapy. Furthermore, the emerging data demonstrates that certain patients experience hyperprogressive disease status after receiving PD-1 immune-checkpoint inhibition therapy. In conclusion, PD-1 immune-checkpoint inhibition therapy has opened up a new venue of advanced cancer immunotherapy. Meanwhile, further efforts are still warranted in both basic scientific mechanism studies and clinical investigation using the principles of personalized and precision medicine.

Abbreviation

hPD-1

Human Programmed Cell Death 1

PD-L

program death ligand

TILs

tumor-infiltrating lymphocytes

mAb

monoclonal antibody

NSCLC

non-small-cell lung cancer

RCC

renal-cell carcinoma

UBC

urothelial bladder cancer

TTF

time-to-treatment failure

Keywords

Human Programmed Cell Death 1 (hPD-1) gene

Immune-checkpoint inhibition

Immunotherapy

Hyperprogressive disease status

1. Introduction

The abundant genetic alterations contained within human cancers result in the generation of neoantigens. However, often the endogenous immune response is ineffective to against those neoantigens due to cancer's multiple resistive mechanisms, including adaptive immune resistance (Drake et al., 2006). Intensive research aimed at developing immunotherapeutic approaches to treat cancer has been ongoing, which involve utilization of the body's immune response to treat cancer. Immunotherapy has emerged in recent years to join surgery, radiation, target therapy, and chemotherapy as a mainstay in cancer treatment (Mellman et al., 2011). Human PD-1 immune-checkpoint inhibition therapy has greatly impacted the landscape of cancer immune therapy.

Programmed cell death receptor 1 (PD-1) is currently understood to be an immune-checkpoint receptor mainly expressed by T cells that negatively regulates human immune response. No one in our scientific community could have predicted the eventual importance of this gene when Dr. Larry Finger and I first published the complete gene structure and chromosome location of human PD-1 (hPD-1) gene and illustrated its immune modulating effects in the journal Gene in 1997 (Finger et al., 1997). Over the past two decades our community has accumulated extensive scientific knowledge and clinical experience regarding hPD-1 gene. Thanks to the contributions of many devoted scientists and physicians worldwide, hPD-1 inhibition has been demonstrating continued success in clinical investigations treating various advanced human cancers. Furthermore, the success of PD-1 immune checkpoint inhibition therapy helps establish a platform to develop novel cancer immunotherapy via targeting the programmed cell death pathway, though emerging data demonstrates that individual patients suffering from the same type of cancer may respond differently to hPD-1 inhibition immunotherapy.

2. Discovery and characterization of human program cell death pathway

2.1. Identification of PD-1, its ligands, and its function

In 1992, I joined Dr. Larry Finger's laboratory as a graduate student at the Brander Cancer Research Institute. While searching for my PhD dissertation project we became interested in a short, communication style research article titled "Induced expression of PD-1, a novel member of the immunoglobulin gene superfamily, upon programmed cell death" (Ishida et al., 1992). In a search for genes associated with programmed cell death, this article described isolation of the murine PD-1 gene by cDNA subtraction following induction of apoptosis in two murine T-cell lines. As we sought to obtain greater knowledge of PD-1 gene and explore its biological effects, we eventually identified the complete cDNA sequence, gene structure, and genomic organization of the human PD-1 homologue, observed its differential immune regulatory effects, and reported our findings in this journal Gene in November 1997–20 years ago this year (Finger et al., 1997). Shortly thereafter, the number of PD-1 studies grew enormously. Importantly, PD-1-deficient (Pdcd1−/−) mice were noted to spontaneously develop autoimmune phenotypes, suggesting the negative regulatory role of PD-1 in immune responses and its importance in maintaining peripheral tolerance (Nishimura et al., 1999, 2001).

Naïve T cells require both antigen presentation and a second signal via a co-stimulatory receptor such as CD28 to be activated. In contrast to the positive co-stimulatory signal delivered by CD28 receptors, PD-1 delivers a negative signal when bound to its ligands, PD-L1 and PD-L2 (Freeman et al., 2000; Latchman et al., 2001). PD-1 suppresses T cell activation through the recruitment of phosphatase SHP-2 and the subsequent inactivation of Zap 70, which plays a critical role in T-cell receptor signaling (Okazaki et al., 2001; Yokosuka et al., 2012; Freeman et al., 2000).

PD-L1 and PD-L2 (also referred to as B7-H1 and B7-DC, respectively) are both type I transmembrane proteins. Though PD-L1 is expressed infrequently in normal human tissues outside of the placenta and macrophage-like cells, its expression is upregulated by IFN-γ and other cytokines that are released by activated T cells (Petroff et al., 2002; Spranger et al., 2013). The expression of PD-L1 in peripheral tissues is crucial to prevent immune-mediated damage at the time of an inflammatory response, as the activation of PD-1 inhibits T-cell effector functions that would otherwise be carried out against the target cell.

2.2. PD pathway in regards to tumor immunity

Unfortunately, cells from many different human tumors are also capable of evading host immune surveillance by expressing PD-L1 on their surface. Expression of this immune-checkpoint ligand by tumor cells is thought to be induced by an antitumor immune response (Pardoll, 2012). Tumor-infiltrating lymphocytes (TILs) recognize tumor antigens presented by tumor cells, tumor stromal cells and antigen-presenting hematopoietically derived cells (Chen and Han, 2015). The subsequent release of IFN-γ induces PD-L1 expression in these cells, resulting in an adaptive immune resistance within the tumor microenvironment (Fig. 1). This mechanism of endogenous antitumor immunity is supported by the immunohistochemistry findings that PD-L1 expression is only detected in melanocytes that are adjacent to TILs in human melanoma, along with detection of IFN-γ at the interface of the PD-L1+ tumor cells and the TILs (Taube et al., 2012).

Fig. 1. Neo-antigens expressed on the surface of cancer cells are recognized as foreign, inducing a response in tumor-infiltrating lymphocytes (TILs). This response includes the release of IFN-γ, which results in cancer cell's expression of PD-L1. Binding of PD-L1 to its receptor PD-1 inhibits effector functions of the TIL.

Of note, PD-L1 expression in a smaller fraction of tumors is driven by constitutive signaling pathways that involve PTEN, anaplastic lymphoma kinase, and EGFR mutations rather than relying on induced expression via IFN-γ (Parsa et al., 2007; Akbay et al., 2013; Marzec et al., 2008). Additionally, despite recent emphasis on PD-L1, PD-L2 upregulation has also been reported in certain tumors, most notably certain B cell lymphomas (Rosenwald et al., 2003).

Given the expression of PD-1 on TILs and the upregulation of PD-1 ligands on tumor cells, the basis for PD pathway blockade to invoke an anti-tumor host immune response is evident. Antibody blockade of PD-1 and its ligands leading to enhanced antitumor immunity in mouse models served as proofs of concept for PD-1 immune checkpoint inhibition as a means of human cancer treatment (Dong et al., 2002; Blank et al., 2004; Iwai et al., 2002).

3. Clinical application of PD-1 immune checkpoint inhibition in immunotherapy

3.1. Anti-PD-1 antibodies

Nivolumab, a fully human IgG4 PD-1 immune checkpoint inhibitor from Bristol-Myers Squibb, was the first monoclonal antibody (mAb) targeting PD-1 that demonstrated significant clinical activity against human cancer (Brahmer et al., 2010; Topalian et al., 2012). In its phase I clinical trial enrolling patients with advanced solid tumors, cumulative response rates were 18% among patients with non-small-cell lung cancer (NSCLC), 28% among patients with melanoma, and 27% among renal-cell carcinoma (RCC) patients (Topalian et al., 2012). More recent studies have shown an objective response rate of 30–40% in clinical trials involving patients with melanoma (Topalian et al., 2014; Robert et al., 2015a; Weber et al., 2015). This includes objective responses in 32% of melanoma patients who progressed after treatment with ipilimumab, another immune checkpoint inhibitor targeting CTLA-4. Nivolumab also demonstrated superior survival benefits compared to treatment with docetaxel in both squamous and nonsquamous NSCLC, and achieved an objective response rate (ORR) of 15% in squamous NSCLC patients who had progressed after receiving at least two other regimens (Brahmer et al., 2015; Borghaei et al., 2015; Rizvi et al., 2015a). An ORR of 87% was even reported in a phase I study that involved treatment of relapsed or refractory Hodgkin's lymphoma with nivolumab (Ansell et al., 2015).

Currently, nivolumab is FDA approved to treat unresectable or metastatic melanoma, NSCLC, classical Hodgkin's lymphoma, and RCC among others in a list that continues to grow. Pembrolizumab, a humanized IgG4 mAb against PD-1 produced by Merck, is also FDA approved for treatment of melanoma and NSCLC (Robert et al., 2015a, 2015b; Garon et al., 2015). Other pharmaceutical companies have created additional PD pathway inhibitors that target PD-1, including PD-L2 fusion proteins. These signal inhibitors are involved in clinical studies on at least 20 types of both solid and hematologic malignant tumors (Iwai et al., 2017).

3.2. Anti-PD-L1 antibodies

PD-L1 has also been targeted as a means of PD-1 pathway blockade. The anti-PD-L1 IgG1 mAb atezolizumab has been FDA approved for treatment of patients with urothelial bladder cancer (UBC) and lung cancer. Treatment with atezolizumab resulted in an ORR of 26% in UBC patients, and an ORR of 43% in the subset of UBC patients with PD-L1 + tumors (Rosenberg et al., 2016). In a randomized phase II study comparing atezolizumab to docetaxel in NSCLC, patients treated with atezolizumab achieved an ORR of 15%, and an ORR of 38% in the subset of patients with PD-L1 + tumors (Spira et al., 2015). Additional anti-PD-L1 mAbs are also currently under investigation, including MDX1105, durvalumab, and avelumab. Table 1 includes a selection of immune checkpoint inhibitors and their efficacy as monotherapies.

Table 1. A selection of immune checkpoint inhibitors and their targets.

A selection of immune checkpoint inhibitors and their targets, highlighted clinical benefits as monotherapy against several cancers, and current FDA approval status. Abbreviations: ORR = overall response rate, mOS = median overall survival, NSCLC = non-small cell lung cancer, UBC = urothelial bladder cancer.

4. Challenges to PD-1 pathway-related immunotherapy

4.1. Identifying biomarkers to predict response to PD-1 immune checkpoint inhibition

The PD-1 pathway blockade approach to cancer treatment is predicated upon the enhancement of the endogenous antitumor immune response that results from PD-1 pathway signal inhibition. It is therefore reasonable to conclude that the efficacy of this approach is substantially dependent upon the expression of a PD-1 ligand by tumor cells, which would suggest that lymphocyte inhibition within the tumor microenvironment may be actively occurring via PD-1 signaling. This conclusion is supported by a recent meta-analysis that demonstrated that clinical response to treatment with anti-PD-1/PD-L1 antibodies is significantly associated with PD-L1 expression in patients with malignant melanoma and non-squamous NSCLC (Gandini et al., 2016).

Because only a subset of patients benefit from treatment with PD-1 immune checkpoint inhibition, it is critical to selectively target this population. Currently, one PD-L1 immunohistochemistry (IHC) staining using 22C3 antibody is approved by the FDA as a companion diagnostic for selecting NSCLC patients for pembrolizumab (Ma et al., 2016). Investigations to identify other biomarkers that are predictive of clinical outcome following PD-1 immunotherapy are ongoing, and recently have included IFN-γ related gene signatures, tumor mutation burden, defects of mismatch repair genes, and proximity of cytotoxic T cells to PD-L1 + cells (Ribas et al., 2015; Le et al., 2015; Rizvi et al., 2015a, 2015b; Roszik et al., 2016; Taube et al., 2014). Tumors with higher mutational loads exhibit greater numbers of neoantigens that may induce anti-tumor immune responses, signifying a greater likelihood of patient response to immune checkpoint inhibition (Zhang et al., 2017).

Unfortunately, determining which patients are most likely to respond to immune-checkpoint pathway inhibitors has not been entirely straightforward. Even assays for PD-L1 expression have demonstrated conflicting results in terms of predicting clinical outcome (Ma et al., 2016). For example, clinical trials of PD-1 immune checkpoint inhibition in patients with squamous-cell lung cancer and ovarian cancer did not show a correlation between PD-L1 expression in tumor tissue and clinical effect of the therapy (Brahmer et al., 2015; Hamanishi et al., 2015; Disis et al., 2016). Similarly, Hugo et al. reported no association between high mutational load and tumor response in melanoma patients (Hugo et al., 2016).

An additional challenge to identifying biomarkers that will predict patient response to therapy is the inappropriateness of using standard response criteria to identify the response to immunotherapy (Wolchok et al., 2009; Hodi et al., 2016). Unlike standard therapy, the response to which is generally measured radiographically and with laboratory tests at 6–8 week intervals, immune checkpoint blockade therapy is capable of resulting in delayed tumor responses due to the time needed to allow for the function of effector T cells at the site of the tumor (Topalian et al., 2014; Ma et al., 2016). Tumor pseudoprogression, during which time a tumor appears to enlarge on imaging but ultimately regresses, has also been documented (Chiou and Burotto, 2015).

4.2. Resistance to immunotherapy

In cases where PD-1-mediated adaptive immune resistance is not the primary mechanism of a tumor's immune resistance, or cases where patients are not able to mount a tumor-specific response, tumors may not respond to monotherapy with PD-1 immune checkpoint inhibition. Factors that confer this primary and/or adaptive resistance may be intrinsic or extrinsic to the tumor cells (Sharma et al., 2017). Additionally, many patients exhibit acquired resistance to immunotherapy, in which their cancer responds for a period of time prior to progressing.

Various approaches are being considered to combat resistance to immunotherapy, including the enhancement of endogenous T cell function and the adoptive transfer of antigen-specific T lymphocytes. These TILs can be expanded ex vivo, or can be genetically modified to target tumor antigens either through transduction with physiologic T-cell receptors or chimeric antigen receptors. A large challenge of adoptive cell transfer is the identification of tumor antigens that are not expressed by normal tissue in order to minimize toxicity (Fesnak et al., 2016).

Another approach to immunotherapy resistance is the use of combination therapy, including molecularly targeted therapy that may affect anti-tumor immunity (Sharma et al., 2017). For example, treatment with serine/threonine-protein kinase B-Raf (BRAF) and mitogen-activated protein kinase/extracellular signal-regulated kinase (MEK) inhibitors (dabrafenib and trametinib, respectively) resulted in the upregulation of PD-L1 in a mouse model of metastatic melanoma (Hu-Lieskovan et al., 2015). Superior anti-tumor effect was then observed with dabrafenib and/or trametinib treatment when combined with PD-1 blockade therapy in the same study. Targeted therapy may greatly enhance a patient's response to immunotherapy, but the appropriate time, dosage, sequence, and choice of agent to combine with immune checkpoint inhibition will likely be critical to determine optimal combination strategies (Vanneman and Dranoff, 2012).

4.3. Hyperprogressive disease after PD-1 immune checkpoint inhibition

In recent years, immunotherapy has revolutionized the way in which certain cancers are treated. Despite this, recent evidence suggests that treatment with PD-1 pathway blockade therapy may backfire in certain patients (Ledford, 2017). Anecdotal evidence of rapid disease progression in patients treated with anti-PD-1/anti-PD-L1 mAbs was first reported (Lahmar et al., 2016; Saada-Bouzid et al., 2016). This was followed by a review of the tumor growth rate in 131 patients upon treatment with anti-PD-1 therapies, which revealed that 9% of patients developed "hyperprogressive" disease characterized by accelerated tumor growth (Champiat et al., 2017). This phenomenon, the frequency of which the authors felt could be underestimated in this study, was found to be associated with higher age and worse overall survival.

The clear challenges are thus presented of not only identifying biomarkers that will predict a positive clinical response to immunotherapy, but to predict and thereby avoid a hyperprogressive response in a subset of patients. A recent study of 155 patients treated with anti-PD-1/anti-PD-L1 therapy reported all six individuals with MDM2/MDM4 amplification experienced time-to-treatment failure (TTF) of less than 2 months, and four of the six patients exhibited a hyperprogressive response (Kato et al., 2017). Eight of ten patients with EGFR mutations experienced TTF of less than 2 months, and two of ten demonstrated hyperprogression. Of note, hyperprogression may be limited to anti-PD-1/anti-PD-L1 monotherapy, and may not result from combination therapy that includes PD-1 pathway blockade. Further investigation of this hyperprogressive response to treatment is still needed in order to justify a change in the way in which physicians treat patients.

5. Conclusion

Since the gene structure and genomic organization of hPD-1 gene was described 20 years ago, the scientific community has gained significant knowledge of hPD-1 as an immune-checkpoint protein. The novel hPD-1 inhibition immunotherapy has revolutionized the way in which human cancers are treated and has led to durable responses for some patients who run out of treatment options otherwise. The success of translating hPD-1 immune checkpoint inhibition knowledge into clinical investigation and its application in the field of cancer immunotherapy has benefitted immensely from biomedical technological advances in the last 20 years since hPD-1 gene was first discovered and characterized.

Although hPD-1 blockade therapy has demonstrated clinical efficacy in multiple types of solid and hematologic tumors, the biomarkers that best predict response to this type of therapy are still not fully elucidated. Resistance to immunotherapy is a problem in clinical application, and therapy-related disease hyperprogression is also emerging as an issue in need of further investigation. In this age of personalized medicine and precision medicine, identifying patients who will truly benefit from hPD-1 blockade immunotherapy and achieve meaningful clinical response might be more crucial than previously imagined. Additional research is needed to develop more comprehensive knowledge of hPD-1 mediated immune-checkpoint regulation and to identify the biomarkers and genome signature specific for the patient population who respond well to program death pathway blockade immunotherapy.

Acknowledgments

The authors would thank Ms. Kara Eichelberger for providing useful comments on this manuscript and helping in formation the figure.

This study is supported by: NIDA/FDA research grant to JJP (P50 DA036107), AA&MDSIF research grant to JJP (146818), American Cancer Society research grant to JJP (124171-IRG-13-043-02), and a Pennsylvania State University College of Medicine research grant to JJP.

Authorship contributions

JJP designed this study. JJP and KNB wrote this manuscript.

Conflict of interest disclosures

The authors declare that they have no conflict of interest.