The undeniable success story of ASCO 2012 was Bristol Myers-Squibb's (BMS) anti-programmed death-1 (PD-1) antibody, nivolumab (BMS-936558). Phase Ib data suggested nivolumab use was associated with rapid and durable responses in heavily pre-treated patients across multiple tumor types; namely malignant melanoma, renal cell carcinoma (RCC) and non-small cell lung cancer (NSCLC). Furthermore, the drug exhibited an exceptional early safety profile; not only relative to other immunotherapies, which are typically associated with a considerable toxicity burden, but relative to all other agents used to treat unresectable malignant melanoma.
PD-1 blockade is a novel mechanism of checkpoint inhibition that inhibits PD-1 receptor activation, or PD-1 ligand (PD-L1 or PD-L2) binding to the PD-1 receptor, or the B7.1 receptor (also known as CD80). Similar to cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4) activity, upregulated PD-1 expression on the surface of tumor-infiltrating lymphocytes dampens T-cell activation, thus preventing T-cell-induced apoptosis of the tumor cells.
Mario Sznol provided an update to the nivolumab Phase Ib data presented last year in heavily pretreated unresectable melanoma patients (Abstract CRC9006). Objective response rate (ORR) for patients receiving the 3 mg/kg dose the dose selected for further clinical development was 41 percent (compared to 31 percent across all dose levels).While ipilimumab [Bristol Myers-Squibb's Yervoy], a CTLA-4 inhibiting immunotherapy, often takes several months for responses to be observed in patients, 45 percent of nivolumab-responding patients showed a minimum of 30 percent tumor assessment 8 weeks following the start of treatment. As of March 2013, the median duration of response was 24.0 months, and 58% of responding patients continued to show tumor shrinkage. Median overall survival was shown to be 20.0 months, at the 3 mg/kg dose level. The median PFS for patients receiving nivolumab at the 3 mg/kg dosing level was 9.7 months. Three treatment-related deaths from pneumonitis were reported, although these adverse events occurred across multiple tumor types (with no deaths occurring in the melanoma arms of the study).
Phase I data from another PD-1-targeting agent were also presented (Abstract 9009) lambrolizumab (Merck's MK-3475) which recently featured in the headlines after the agent was awarded breakthrough designation by the FDA. The interim ORR across multiple dose levels was reported to be 38 percent (in heavily pretreated unresectable malignant melanoma patients), although the highest percentage (52 percent) was observed in those patients receiving 10 mg/kg every two weeks. Furthermore, 10% of patients in the 10 mg/kg twice weekly arm exhibited a complete response. Physicians consider the agent to have an acceptable safety and tolerability profile, like nivolumab, whereby the incidence of grade 3-4 adverse events occurred at a rate of 12.6 percent. There was one reported patient death from suspected grade 2 pneumonitis (a 96 year old male).
The third agent discussed was Roche/Genentech/Chugai's MPDL3280A (Abstract 9010). Unlike nivolumab and lambrolizumab, MPDL3280A targets PD-L1 (i.e. the ligand), rather than the receptor. Safety data revealed no treatment-related deaths, and no pneumonitis was reported. Grade 3-4 treatment-related adverse events were observed in 14 percent of patients. The ORR in heavily pretreated malignant melanoma patients was 29 percent.
Questions persist on the optimal mechanism of action for inducing PD-1 blockade. Blockade of the receptor PD-1 is associated with dual-antagonism of both ligands for the receptor (PD-L1 and PD-L2). PD-L1, however, is still capable of inducing T-cell inhibition through activation of the co-inhibitory B7.1 receptor; also expressed on T-cells. PD-L1 blockade, in contrast, prevents PD-L1 activation of the B7.1 and PD-1 receptors. However, PD-1 receptor activation is still possible, through PD-L2-mediated T-cell inhibition. The roles of PD-L2 and B7.1 in T-cell anergy require more thorough evaluation, although early safety and efficacy data provide an intriguing possibility of combining the two therapeutic approaches, thus blocking the entire pathway altogether.
Additional questions also persist. For example, is PD-1 blockade better than CTLA-4 inhibition Will this novel class of PD-1 pathway blockers displace ipilimumab as the standard of care in the treatment of unresectable malignant melanoma. How should these therapies be optimally used, or sequenced To what extent do documented responses off-therapy arise in the drug-treated population Are these off-therapy responses as durable as those observed in those patients that continue to receive therapy These questions will likely be addressed by the four Phase III studies already planned or ongoing for PD-1 antagonists in unresectable malignant melanoma.
The optimal treatment regimen for PD-1 inhibitors has also yet to be identified. Should they be used as monotherapies, or in combination with other agents such as other immunotherapies or potentially targeted therapies in the BRAF mutation-positive unresectable melanoma patient population. What is the optimal combination. Are there safety and tolerability issues associated with using the drugs in combination. Ipilimumab use, for example, is associated with a high incidence of colitis and other immune-related adverse events. Will adding another immunotherapy further compound these toxicities, or will we see benefits from a safety standpoint: in the same way that adding the MEK inhibitor, trametinib [GlaxoSmithKline's Mekinist] in combination with the BRAF inhibitor dabrafenib [GlaxoSmithKline's Tafinalar] appears to result in a reduced incidence of secondary tumors (e.g. squamous-cell carcinomas) compared to dabrafenib monotherapy. Early data for combinations of anti-PD-1 therapy suggest that there is hope for improving survival for patients with advanced melanoma without inducing dramatically higher levels of toxicity.
Jedd Wolchock presented data on a Phase I study in which nivolumab was administered in combination with ipilimumab (Abstract 9012). The optimal dose was identified to be 1 mg/kg nivolumab and 3 mg/kg ipilimumab. The one-year survival rate in patients receiving the concurrent regimen at the selected dose was 82 percent, and the ORR was 53 percent. Interestingly, the nature of the responses appeared to be different from those of the monotherapies: responses were typically more rapid, and at the dose chosen for further study, all responding patients achieved deep or complete responses. There were no treatment-related deaths, although treatment-related grade 3-4 adverse events were reported in 53 percent of patients. These were most commonly clinical chemistry abnormalities that were easily managed lipase elevations (13 percent), AST elevations (13 percent), and ALT elevations (11 percent)that physicians considered to be acceptable. Intriguingly, no new safety signals were reported that would not be expected with either of the two agents being used as monotherapies.
Data were also presented from an initial safety and efficacy study in which the PD-L1 antagonist MPDL3280A was administered in combination with vemurafenib [Roche/Genentech/Daiichi Sankyo/Chugai's Zelboraf]. Grade 3 rash and AST/ALT elevations (up to grade 3) were reported in all patients by the time they had received their fourth cycle of therapy all of which resolved following vemurafenib dose interruption or reduction and a complete response was experienced by one of the first three patients. Physician optimism surrounding the MPDL3280A/vemurafenib combination was somewhat tempered by the safety data. While physicians stressed that both sets of data are far too early to draw meaningful conclusions, both appear to be encouraging.
PD-L1 Immunopositivity as a Predictive Biomarker
Physicians typically consider vemurafenib (as well as the recently approved agents dabrafenib and trametinib) to hold an advantage over ipilimumab given the presence of a biomarker which is predicative of drug efficacy: activating mutations on the intracellular protein BRAF, which can therefore be targeted to arrest tumor cell proliferation. The presence of this predictive biomarker therefore means that patients harboring this mutation will be eligible to receive a BRAF/MEK inhibitor in the first-line setting for unresectable malignant melanoma. Physicians were hugely optimistic when it was suggested at ASCO 2012 that PD-1 overexpression could predict response to nivolumab therapy. While this certainly may be the case: tumor membrane staining at 5 percent using an automated system was associated with response (p=0.004); some individuals with PD-L1 negative tumors still exhibited responses (Abstract 9011). As a result, there is insufficient rationale to deny patients access to PD-1 blocking therapy if tumor biopsies are PD-L1 immunonegative.
Undeniably, further work needs to be undertaken to investigate the possibility of stratifying patients to predict whether or not they will respond to PD-1 therapy. However, the story does not appear to be as clear-cut as that for the BRAF and MEK inhibitors.
The Promise of Immune Checkpoint Therapy
A key feature of PD-1 blockade is the promise the drugs hold not just in melanoma but in other cancer indications as well. Other clinical data presented at ASCO 2013 suggest similarly impressive responses in NSCLC and RCC, although PD-1 targeting agents are also being trialed in hematological malignancies, as well as other solid tumors (including gastric cancer). The early clinical data appears to be hugely compelling, and all agents have clear benefits over the limited treatment options in pretreated advanced malignancies, and potentially over some first-line treatment options as well. Moreover, the combinations of immunotherapies, and combinations of immunotherapies with small molecule targeted agents offer exciting exploratory avenues, in addition to the further proof-of-principle of immune checkpoint therapy.
Furthermore, given the identification of additional immune checkpoint targets (e.g. the co-inhibitory receptors TIM-3 or LAG-3), physicians express optimism that similar deep and durable responses can be observed following administration of novel immune checkpoint therapies in multiple, if not all, malignancies. The overriding consensus at ASCO 2013 was the belief that immune checkpoint therapy can displace targeted therapies as the spearhead of cancer therapeutic development, and that immune checkpoint therapy could establish itself as the new pillar in the fight against cancer. Inevitably, more twists and turns lie in the road ahead, although physicians are hopeful that PD-1 pathway blockers will deliver on their early promise. These are exciting times to be working not only in melanoma, but in cancer as a whole.
Samuel Mentzer is an analyst with the Oncology team at Decision Resources.