Blockade of the PD-1/PD-L1 pathway is a novel and attractive treatment approach to many cancers, including malignant glioma, because of its potential to restore anti-tumor immunity.
Sylvia C. Eisele, MD, PhD
Immunotherapies have emerged as new and promising treatment strategies against a variety of cancers and are currently intensely investigated in many tumors, including malignant gliomas. Glioblastomas and other high-grade gliomas are known to create an immunosuppressive tumor micro-environment. This pathway therefore represents an attractive treatment target for this still incurable disease, and the PD-1/PD-L1 pathway appears to play a particularly prominent role.
Therefore, immune checkpoint inhibition by targeting this pathway is an especially attractive therapeutic modality for highgrade gliomas that continue to carry a dismal prognosis despite the best available treatment options to date. More recently, encouranging preclinical data from animal models have resulted in the design and initiation of several early phase clinical trials investigating the treatment efficacy of PD-1 and PD-L1 blocking agents in patients with high-grade gliomas.
This review will provide an overview the PD-1/PD-L1 pathway and the immunosuppressive micro-environment in malignant gliomas and will summarize the pre-clinical results and ongoing early phase clinical trials targeting the PD-1/PD-L1 pathway in these incurable tumors.
Glioblastoma is the most common malignant primary brain tumor, with approximately 11,000 patients diagnosed annually in the United States.1,2 Despite optimal surgical and medical treatment, prognosis remains poor, with 3-year overall survival rates of under 10%. Novel and more effective therapies are therefore needed. Among the many immunotherapeutic modalities in development, immune checkpoint inhibition has gained particular attention following its recent success in treatment of malignant melanoma. Just a few months ago, the US Food and Drug Administration (FDA) granted accelerated approval of pembrolizumab, a monoclonal antibody (Mab) targeting the programmed death (PD)-1 receptor on T cells, based on durable treatment responses among malignant melanoma refractory to previous treatments.3 Moreover, pembrolizumab was recently demonstrated to be superior to ipilimumab, a Mab-targeting CTLA-4, with regard to efficacy and toxicity.4 In addition, the FDA also approved nivoluâ€¨mab, another PD-1 blocking Mab for the treatment of advanced melanoma based on the interim results of a phase lll trial, which was recently completed and published.5 In gliomas, encouraging pre-clinical data in several categories of immunotherapies have translated into early phase clinical trials. This review will provide an overview of the mechanisms of action and will detail the pre-clinical and early clinical data of treatments targeting the PD-1/PD-L1 pathway for malignant glioma.
The Immunosupressive Tumor Micro-Environment
Glioblastomas and other malignant gliomas create a hostile, immunosuppressive micro-environment, which effectively attenuates the body’s immune attack against cancer cells. To this end, gliomas secrete immunosuppressive factors such as TGF-beta,6 IL-10,7 and CCL-2 (MCP-1),8,9 and recruit immunosuppressive immune cells such as regulatory CD4-Foxp3+ T cells (T regs)10 and myeloid-derived suppressor cells (MDSC)11 to the tumor. In addition, malignant glioma cells express surface molecules such as FAS ligand (FasL),12,13 B7-1/B7-2, and PD-L1/PD-L214 that, when bound to their respective receptors (Fas, CTLA-4, and PD-1) on tumor-infiltrating lymphocytes (TILs), alter and dampen their executive function.15,16 One particularly prominent immunosuppressive mechanism exploited by malignant gliomas is enhanced signaling through the PD-1/PD-L1 pathway.
The PD-1 receptor is expressed on immune cells including CD4+ and CD8+ T cells in peripheral tissues upon recognition of a target antigen via the T-cell receptor (TCR).14,17,18 The PD-1 receptor has two ligands, PD-L1 (B7-homologue 1, B7-H1) and PD-L2 (B7-homologue 2, B7-H2). While PD-L1 is found on both hematopoietic cells and non-hematopoietic cells, including cells in peripheral tissues and various tumors,14,19 PD-L2 expression is restricted to reactive immune cells such as dendritic cells and macrophages.20 In a normal and healthy environment, crosstalk between PD-1-positive immune cells and PD-L1-expressing antigen presenting cells (APC) or healthy tissue cells, provides an important mechanism to modulate and attenuate the immune response against self-antigens. This mechanism is therefore crucial to maintaining self-tolerance and to prevent autoimmunity.15,18 Binding of PD-L1 to PD-1 activates a downstream signaling cascade, which triggers reduced executive T-cell function and proliferation. Previous studies have described T-cell exhaustion and anergy when persistently expressing PD-1.21-23 In contrast, stimulation of the PD-1/PD-L1 pathway on CD4+Foxp3+ T cells (T regs) leads to recruitment of additional regulatory T cells,24 which can further reduce â€¨local immunoreactivity.
David A. Reardon, MD
In summary, engagement of the PD-1/PD-L1 pathway within the tumor micro-environment leads to impaired executive CD4+ and CD8+ T-cell function but increases the activity of regulatory T cells, therefore reinforcing the immunosuppressive microenvironment. Therapeutic strategies to release this “break” on the tumor-targeted immune response via PD-1 or PD-L1 blocking agents therefore is currently the subject of intense investigations in several cancers, including glioblastoma.
PD-1 and PD-L1 Expression in Gliomas
Wintterle et al25 initially demonstrated that a high percentage of glioma cell lines express PD-L1 and that co-culture of PD-L1 expressing glioma cells with CD4+ or CD8+ lymphocytes reduces immunoreactivity as measured by decreased cytokine production (IFN-gamma, IL-2, IL-10) and diminished expression of markers of T-cell activation, such as CD69.25 Parsa et al26 subsequently demonstrated that expression of B7-H1 (PD-L1) in glioma cell lines is enhanced by Akt activation within the mTOR/PI3K pathway that was linked to loss of function of the tumor-suppressor gene phosphatase and tensin homolog (PTEN), which occurs frequently in human glioblastomas. In these experiments, indirect activation of Akt due to loss of PTEN resulted in increased expression of B7-H1 protein.26 In other experiments, the presence of IFN-gamma induced expression of B7-H1 in PTEN deficient gliomas, which in turn was associated with an increased apoptotic rate of tumor infiltrating lymphocytes.27 In addition, B7-H1 expression in human gliomas appears to correlate with tumor grade and the highest B7-H1 expression levels are found in grade IV gliomas.28 Bloch et al found that tumor-associated macrophages (TAM) secrete IL-10, which can also induce B7-H1 expression in glioma cells.29
A recent comprehensive study by Berghoff and colleagues28 investigated PD-1 and PD-L1 expression in 135 glioma specimens and found prominent expression of PD-L1 in up to 88% of newly diagnosed glioblastomas.30 The surrounding non-tumor bearing brain showed no or only faint PD-L1 expression. In addition, they were able to correlate the level of PD-L1 expression with glioblastoma subtype defined by gene expression array. Specifically, high PD-L1 expression was more common in the mesenchymal, neural, and classical subtypes whereas low PD-L1 expression was seen in the proneural and G-CIMP subtypes of glioblastoma.30
Thus, increasing levels of PD-L1 expression in glial tumors are associated with higher tumor grade and unfavorable glioblastoma subtypes. In addition, PD-L1 expression appears to be associated with both extracellular cues such as immunocytokine expression (IL-10, IL-2, and IFN-gamma) and intracellular cues such as mTOR-PI3k pathway activation associated with PTEN deficiency.
PD-1 and PD-L1 Blockade in Pre-clinical Studies
Only a few studies have evaluated PD-1 or PD-L1 blockade in pre-clinical glioma models. In one study, C57BL/6 mice bearing gliomas were treated with sham therapy (control group), radiotherapy alone, anti-PD-1 blockade alone, or radiotherapy combined with anti-PD-1 blockade. Mice treated with the combination of radiation and anti-PD-1 blockade had a median survival of 52 days compared with untreated animals in the control group, which had a median survival of 26 days. Up to 40% of the animals treated with the combination therapy survived more than 90 days. When re-challenged by flank injection of tumor cells, animals who had responded to previous therapy did not demonstrate tumor regrowth, suggesting development of an immunological memory in the tumor cells. This treatment response was associated with both an increase in cytotoxic T-cell function and reduction in regulatory T-cell function.31 In a recent study by Wainwright and colleagues,32 C57BL/6 mice were treated with CTLA-4 and/or PD-L1 antibodies following intracranial glioma cell (GL261) injection. After 35 days, >75% of the treated animals were alive, whereas all of the untreated animals in a control group had died.30 Our group recently demonstrated that treatment of immunocompetent mice harboring intracranial glioblastoma tumors with PD-1 or PD-L1 blocking Mabs is associated with long-term cure in 50% and 25%, respectively. Intracranial tumor re-challenge failed to result in tumor growth, which is further supporting the development of a tumor-specific immunologic memory among long-term survivors.33
Clinical Experience with PD-1 and PD-L1 Blockade in Glioblastoma
Immune checkpoint inhibition targeting PD-1 and PD-L1 is currently bring intensely investigated for many hematological and solid malignancies. Robust response rates with acceptable toxicity profiles have been demonstrated for melanoma,4,5,34-36 renal cell carcinoma, non-small cell lung cancer (NSCLC),37,38 diffuse large B-cell lymphoma,39 follicular lymphoma,40 and Hodgkin’s lymphoma.41 Based on this success, the high level of PD-L1 expression among glioblastomas, and the encouraging pre-clinical data discussed above, there are now several ongoing clinical trials investigating immune checkpoint inhibition for glioblastoma patients. Besides therapies targeting CTLA-4 (ipilimumab), several trials are under way or will open soon that investigate the efficacy of therapeutic blockade of the PD-1/PD-L1 pathway. The following section highlights some of the agents being evaluated in ongoing or planned trials for glioblastoma patients.
Nivolumab (BMS-936558) is a monoclonal IgG4 antibody that specifically blocks PD-1, and its efficacy has been evaluated in several cancers including melanoma, NSCLC, prostate cancer, renal cell cancer, and colorectal cancer.5,36,37 In an initial dose-escalation study, durable responses were observed in 28% of patients with melanoma, 27% of patients with renal cell cancer, and 18% of patients with NSCLC.14 The toxicity profile was acceptable and the most common treatment-related adverse reactions were rash, diarrhea, and fatigue. Grade 3 and 4 drug-related adverse events occurred in only 14% of patients; however, grade 5 treatment-related pneumonitis was noted in 1% of patients.37 Since this pilot study, nivolumab has been or is currently being investigated in several clinical studies as monotherapy or in combination with other immunotherapies across solid and hematological malignancies, including melanoma,34,36,42,43 metastatic renal cell cancer,44 and Hodgkin’s lymphoma.41 Nivolumab was recently granted accelerated approval by the FDA for recurrent, advanced melanoma.5 For patients with recurrent glioblastoma, a randomized open-label phase ll/lll study has recently initiated accrual that investigates the efficacy of nivolumab compared with bevacizumab. In a separate cohort of this study, the safety and tolerability of nivolumab alone or in combination with ipilimumab is investigated. This study opened in February 2014 and is projected to complete accrual in June 2017 [NCT02017717].
Another multicenter phase l study will investigate ipilimumab, nivolumab, and the combination of ipilimumab and nivolumab in addition to temozolomide for newly diagnosed glioblastoma. This study is expected to initiate accrual soon [NCT02311920]. Pidilizumab or CT-011 is a humanized monoclonal Ig-â€¨G-1 kappa antibody targeting PD-1 that has been investigated previously in hematological malignancies. In a small phase l study, escalating doses were evaluated in 17 patients with different hematological malignancies and only minor drug-related toxicities (generalized weakness, flushing) were noted.45 In a phase ll, open-label multicenter trial investigating pidilizumab following autologous stem cell transplant (ASCT) in diffuse large B-cell lymphoma (DLBCL), an objective response to treatment was observed in up to 51% of patients.39 The drug was well tolerated, with the most common adverse events being neutropenia (26%), fatigue (25%), and diarrhea (17%). Encouraging results were also obtained in an open-label, nonrandomized phase ll study investigating pidilizumab in combination with rituximab (targeting CD20) for relapsing follicular lymphoma.40 This rather indolent disease is characterized by a long period of stability, and recurrence/progression is typically attributed to failure of immune surveillance. In this small study, Westin et al demonstrated objective responses in 66% of patients, with 52% of patients achieving a complete response. Interestingly, prolonged progression-free survival with treatment was predicted by a prominent T-cell activation signature in tumor biopsy samples.40 Although these results require validation in placebo-controlled randomized studies, PD-1 blockade appears capable of restoring immune surveillance against cancer cells with improved outcome in these patients. Pidilizumab is currently being investigated in several solid tumors, including melanoma [NCT01435369], metastatic colorectal cancer [NCT00890305], hepatocellular carcinoma [NCT00966251], and prostate cancer [NCT01420965]. A current phase l/ll study is investigating the efficacy of pidilizuâ€¨mab in recurrent malignant glioma and diffuse intrinsic pontine glioma. The study began in November 2013 and is expected to complete accrual in November 2015 [NCT01952769].
Pembrolizumab (MK-3475) is a monoclonal IgG4-kappa isotype antibody against PD-1, which was recently shown to successfully treat advanced melanoma refractory to previous treatments with the CTLA-4 targeting antibody ipilimumab and/or BRAF inhibitors.3 In this multicenter, international open label phase l study, 173 patients were treated with pembrolizumab at two different doses (2 mg/kg and 10 mg/kg) and response rates were achieved in 26% of patients. Although drug-related adverse events were common and occurred in 81% of patients, they were mild in general, and grade 3 and 4 drug-related toxicities occurred in only 12%. Fatigue was the most common adverse event, occurring in 3%. Most recently, pembrolizumab was shown to be superior to ipilimumab with regard to efficacy and toxicity in the treatment of advanced melanoma.4
The success of MK-3475 has translated into phase lll clinical trials investigating efficacy of pembrolizumab in NSCLC [NCT02142738, NCT02142738, NCT01905657], melanoma [NCT01866319], urothelial cancers [NCT02256436], and head and neck cancer [NCT02252042].
For glioblastoma, 3 clinical studies have recently opened or are expected to start accrual soon. NCT02313272, a phase l study investigating pembrolizumab in combination with bevâ€¨acizumab and hypofractionated stereotactic re-irradiation in recurrent high-grade gliomas, has opened accrual in January 2015. NCT02311582, a phase l/ll study investigating the safety and efficacy of pembrolizumab in combination with MRI-guided laser ablation in recurrent malignant gliomas, will start enrolling patients soon. A randomized phase ll study of MK-3475 versus MK-3475 plus bevacizumab for recurrent glioblastoma patients will open in February 2015 and is conducted through the Ben and Catherine Ivy Foundation clinical trials consortium [NCT02337491].
BMS-936559 (MDX-1105), a monoclonal IgG4 antibody specifically targeting PD-L1, was evaluated for its efficacy in several cancers following previous chemotherapy or immunotherapy. When tested in several cancer types, durable response rates were observed in 17% of patients with melanoma, 10% of patients with NSCLC, 6% of patients with ovarian cancer, and 12% of patients with renal cell cancer.46 Drug-related adverse events were noted in 39% of patients, the most common of which was infusion reactions, noted in 10% of patients. Only 5% of patients experienced grade 3 or 4 adverse events. Plans for further clinical development of this agent are unclear at this point.
MPDL3280A, a human IgG1 anti-PD-L1 monoclonal antibody, is currently being evaluated in several phase 1 clinical trials in solid cancers, including NSCLC, melanoma, renal cell carcinoma, and lymphoma, including a randomized phase lll trial in NSCL [NCT02008227]. An ongoing phase l study was recently amended to include an arm for recurrent glioblastoma patients [NCT01375842].
MEDI4736, a human IgG1 monoclonal antibody targeting PD-L1, is currently being evaluated for efficacy in melanoma, NSCLC, colorectal cancer, and head and neck cancers. A phase l/ll study of MEDI4736 for glioblastoma patients sponsored by the Ludwig Cancer Institute started accrual in early 2015 and includes separate arms for patients with newly diagnosed glioblastoma, patients with recurrent glioblastomas without previous treatments with bevacizumab and patients with recurrent glioblastomas who have received prior treatment regimens including bevacizumab [NCT02336165].
MSB0010718C, an IgG1 monoclonal antibody targeted against PD-L1, is currently being investigated in Merkel cell carcinoma, NSCLC, gastric cancers, and colorectal cancers. At present, there are no clinical trials with this agent for glioblastoma patients.
Blockade of the PD-1/PD-L1 pathway is a novel and attractive treatment approach to many cancers, including malignant glioma, because of its potential to restore anti-tumor immunity. A growing number of clinical trials support its ability to achieve durable treatment responses across a wide spectrum of malignancies with an overall favorable toxicity profile. Promising pre-clinical data have encouraged the investigation of PD-1/PD-L1 targeting agents in malignant gliomas, and several early-phase clinical trials are under way or will open soon. One of the future challenges will be to integrate immunotherapies into the current standard of care and to develop well-tolerated combinatorial regimens in order to achieve maximal synergistic effects for malignant glioma patients. In addition, there is a critical need for informative biomarkers to monitor treatment response and failure.
ABOUT THE AUTHORS
Affiliations: Center for Neuro-Oncology, Dana-Farber Cancer Institute and Brigham and Women’s Cancer Center; David A. Reardon, MD, Center for Neuro-Oncology, Dana-Farber Cancer Institute and Brigham and Women’s Cancer Center (SCE, DAR).
Corresponding author: David A. Reardon, MD, Center for Neuro-Oncology, Dana-Farber Cancer Institute, 450 Brookline Ave, Boston, MA 02215-5450.