Advances in the Systemic Treatment of Metastatic Melanoma
Advances in the Systemic Treatment of Metastatic Melanoma
Prior to 2011, the only commercially available agents commonly used to treat metastatic melanoma—including dacarbazine, temozolomide (Temodar), fotemustine, carboplatin, paclitaxel, and interleukin-2—demonstrated limited efficacy, and no study involving these agents had shown an improvement in overall survival. The standard of care for the treatment of metastatic melanoma was radically changed by the subsequent approval of two agents, ipilimumab (Yervoy) and vemurafenib (Zelboraf), both of which improved survival in randomized phase III trials. Within the relatively short time that ipilimumab and vemurafenib have been commercially available, phase II data for the investigational agents nivolumab and MK-3475, for the combination of dabrafenib and trametinib, and for adoptive cell therapy strongly suggest even further improvements in treatment outcomes. Within this rich context of effective agents, the challenge for clinicians and investigators will be to develop predictive biomarkers of response, the optimal sequence of therapy for individual patients, and effective combinations. An additional challenge will be to find the appropriate venue and populations to test promising new agents arising from substantial advances in our understanding of molecular alterations in melanoma cells, of mechanisms of resistance to current agents, and of tumor-host immune interactions.
Before 2011, no systemic treatment for unresectable locally advanced stage III or stage IV melanoma had been consistently proven to increase median survival, and no large studies had compared existing treatments to best supportive care. In large controlled randomized trials, the median survival was consistently in the range of 8 to 11 months.[1-3] High-dose interleukin-2 (IL-2) was approved for the treatment of metastatic melanoma based on durable tumor remissions in approximately 5% of patients, but it can only be administered to those with excellent performance status and normal organ function, and to date it has not been compared with any other standard treatment in a randomized trial.
Current advances in the treatment of advanced disease stem from the identification of two specific driver mutations in subsets of melanoma, BRAF and C-KIT, and from advances in our understanding of mechanisms that control T-lymphocyte activation, proliferation, and function, specifically the immune regulatory checkpoints (Table). Controlled clinical trials of vemurafenib (Zelboraf), which potently inhibits signaling from mutant BRAF, and of ipilimumab (Yervoy), which blocks the immune checkpoint cytotoxic T-lymphocyte antigen 4 (CTLA-4), demonstrated meaningful improvements in median survival.[5,6] Ipilimumab was also shown to produce a durable survival benefit in approximately 10% of patients. As a consequence, both vemurafenib and ipilimumab were approved by the US Food and Drug Administration (FDA) in 2011. Results of clinical trials of monoclonal antibodies designed to block another immune checkpoint, programmed death 1 (PD-1), or its ligand, and of combined inhibitors of mutant BRAF and MEK, suggest even further improvements in outcome for subsets of patients. Additional treatment gains may be achieved over the next 5 to 10 years through the combination of active agents, the introduction of new agents against novel molecular and immune targets, and improvements in technology that will increase the feasibility of adoptive cellular therapy outside of a few highly specialized treatment centers.
Ipilimumab is a human immunoglobulin G1 (IgG1) monoclonal antibody that blocks cytotoxic CTLA-4, a coinhibitory receptor that regulates T-cell activation and the function of T-regulatory cells (Figure). Approval followed presentation of results from a phase III trial that compared ipilimumab, 3 mg/kg every 3 weeks for 4 doses, to ipilimumab in combination with a gp100 peptide vaccine, or to the gp100 vaccine alone in patients who had received at least one prior treatment for advanced disease. Although the objective response rate for ipilimumab in both arms combined was only 7%, median survival for patients receiving ipilimumab in either of the two arms was increased to 10 months, compared with 6.4 months in the vaccine-alone arm. Survival rates at 1 and 2 years were also improved for the ipilimumab arms, from 25% to 44%–46%, and from 14% to 22%–24%, respectively. Recent long-term follow-up from earlier phase II trials of ipilimumab have shown that survival rates remain nearly flat from 3 to 5 years, indicating a long-term benefit for a subset of patients. A second phase III trial was conducted in previously untreated patients, comparing ipilimumab at a dose of 10 mg/kg administered with dacarbazine to dacarbazine/placebo. Although median survival in the ipilimumab plus dacarbazine arm was increased to 11.2 months from 9.1 months, the contribution of dacarbazine to the activity of ipilimumab remains unclear.
There are several unique features of ipilimumab treatment that have been described extensively in prior publications, including the induction of autoimmune/inflammatory adverse events and clinical response in small brain metastases in a subset of patients. Several patterns of systemic tumor response have been observed, including mixed responses, disease progression followed by regression, and prolonged disease stabilization that appears to be associated with patient benefit. The unique patterns of response led to the development of new criteria for assessing clinical response to immune therapy agents. Data from the initial phase III study of ipilimumab also demonstrated that patients treated with ipilimumab or ipilimumab plus vaccine, whose disease progressed after they had achieved stable disease or tumor response at 24 weeks, could achieve a second response or prolonged stable disease with a second induction course of ipilimumab. Indeed, among 31 patients eligible for retreatment, objective response or stable disease of at least 24 weeks was observed in 68%. In contrast, the effect of administering maintenance ipilimumab, for example every 12 weeks, remains unclear.
Several combinations of ipilimumab with other agents may further increase activity and improve outcomes. Promising data, including increased overall response rate, progression-free survival, or complete response rate compared with results in prior trials, were presented for ipilimumab in combination with bevacizumab (Avastin), ipilimumab in combination with high-dose IL-2, and tremelimumab (another anti–CTLA-4 antibody) in combination with interferon alfa.[12-14] A reliable predictive biomarker for response to ipilimumab has not yet been identified.[15-18]
Inhibitors of programmed death 1 (PD-1) or its ligand (PD-L1)
PD-1 is an inhibitory receptor that is upregulated on activated lymphocytes. PD-1 has two known ligands, PD-L1 (also called B7-H1) and PD-L2 (B7-DC), which can be expressed on tumor and stromal cells; PD-L1 expression can be induced by cytokines produced by tumor-infiltrating lymphocytes.[19-21] Several agents targeting either PD-1 or PD-L1 are being developed. In a phase I/II study of nivolumab (BMS-936558, MDX-1106), a human IgG4 monoclonal antibody that blocks PD-1, an overall objective response rate of 31% was observed among 106 evaluable patients with previously treated advanced melanoma.[22,23] An ongoing response was seen in 16 of 23 patients with objective response who were followed at least 6 months from onset of treatment. A similarly high objective response rate of 47% was observed among 83 patients with advanced melanoma who were treated with MK-3475, another antagonist antibody of PD-1. Among the 25 patients in this group who had previously been treated with ipilimumab, MK-3475 produced an objective response rate of 40%. Overall, several complete responses were observed, and most patients were continuing in response with a minimum follow-up of 16 weeks. In a multitumor phase I trial of the anti–PD-L1 antibody BMS-936559, 9 of 52 melanoma patients (17%) achieved a complete or partial response.
Toxicities associated with blockade of the PD-1 pathway have been similar in spectrum but less frequent and less severe than those seen with ipilimumab. Grade 3 or 4 adverse events were observed in only 14% of patients treated with nivolumab and in 9% treated with the anti–PD-L1 agent BMS-936559. Pneumonitis was observed in 3% of patients treated with nivolumab and was fatal in 1%, leading to implementation of early detection and management algorithms in an attempt to reduce life-threatening reactions.
In the nivolumab phase I trial, a strong association was discovered between expression of PD-L1 in pretreatment tumor samples, defined as expression on 5% or more of tumor cells, and response to therapy. Additional data will be required to confirm this association in melanoma. Studies conducted by Taube et al demonstrate that metastatic melanoma lesions that express PD-L1 are almost always associated with the presence of tumor-infiltrating lymphocytes (TIL), while those metastatic lesions without PD-L1 expression generally have no TIL. Increasing the activity of PD-1 blockade may require different approaches in the two subsets of tumors—for example, combining PD-1 blockade with other antagonists of lymphocyte functional suppression in PD-L1/TIL-positive tumors, and combining PD-1 blockade with agents that drive lymphocyte infiltration into tumors for those that are PD-L1/TIL-negative.
Adoptive cell therapy (ACT)
Existing immune therapies attempt to induce or expand tumor antigen–specific immune responses in vivo. An alternate approach is to isolate tumor antigen–specific T cells from the patient, either from peripheral blood or a resected tumor, and expand the cells ex vivo before reinfusing the cells back into the patient. Early studies of ACT in the late 1980s and early 1990s produced limited activity, believed to be a result of the limited persistence of the lymphocytes after adoptive transfer.[27,28] Preclinical models demonstrated that persistence of the cells in vivo after adoptive transfer could be increased if the host was preconditioned with lymphoablating chemotherapy and/or whole-body radiation. Subsequent studies of lymphoablation, followed by transfer of TIL in combination with systemic administration of IL-2, demonstrated high response rates—in the range of 50%.[30-32] In the largest study published to date, approximately 20% of patients achieved durable complete remissions. Responses were observed in patients whose disease was progressing on anti–CTLA-4 therapy, and in a subsequent trial, we are aware of a patient responding after exposure to anti–PD-1 therapy, suggesting that ACT provides an antitumor effect that is non–cross-resistant to the checkpoint inhibitors.
Currently, ACT is applicable to only a select subset of patients who have good performance status and normal organ function, resectable tumors from which cells can be isolated and expanded, ability to travel to one of a few specialized centers studying ACT, and ability to maintain their performance while waiting for cells to expand in vitro for 3 to 6 weeks. Various technological advances may allow export of the technology to multiple centers and increase access to more patients—for example, by reducing the generation time and cost of expanding lymphocytes ex vivo. Better selection of antigen-specific T cells from resected tumors, improved expansion techniques, identification of populations with the greatest potential for in vivo activity, and improved approaches to the support of cell expansion and function after adoptive transfer (perhaps by concurrent administration of other cytokines and checkpoint inhibitors) may produce greater efficacy. Several trials have been conducted using peripheral blood lymphocytes that were genetically engineered ex vivo to express either a tumor-specific T-cell receptor or a chimeric antigen receptor (CAR).[33-36] CARs combine the signal-activating machinery of a T cell and the antigen binding site of a monoclonal antibody. By engineering peripheral blood lymphocytes to confer tumor antigen specificity, the costly and labor-intensive process of harvesting cells from tumors, and the concomitant delay in treatment, could possibly be avoided. Moreover, introducing tumor antigen–specific receptors to peripheral blood lymphocytes may extend therapy options to a larger group of patients. Some of the attempts to administer T cells transfected with CARs or specific T-cell receptors have been associated with unexpected toxicity, and overall response rates are currently lower than those reported with expanded TIL, but advances in the technology can be expected over time.