No longer an untreatable disease: How targeted and immunotherapies have changed the management of melanoma patients

Abstract

The discovery that BRAF is a driver oncogene in cancer, and complementary improvements in our understanding of the immune system have resulted in new targeted and immune‐therapies for metastatic melanoma. Targeted therapies achieve impressive clinical results in carefully selected patients but the development of resistance seems inevitable in most cases. Conversely, immune‐checkpoints inhibitors can achieve long‐term remission and cures, but in a smaller proportion of patients, and biomarkers to predict which patients will respond are not available. Nevertheless, melanoma has led the evolution of cancer treatment from relatively nonspecific cytotoxic agents to highly selective therapies and here we review the lessons from this paradigm shift in treatment and the opportunities for further improvements in outcomes for melanoma patients.

1. Introduction

Malignant melanoma is the most deadly form of skin cancer. It is estimated that there were over 230,000 cases of melanoma globally in 2012 (http://www.wcrf.org/cancer_statistics/world_cancer_statistics.php), with over 76,000 cases and 9,000 deaths in the US (Siegel et al., 2014), over 103,000 cases and 22,000 deaths in Europe (Ferlay et al., 2013), and over 12,000 cases and 1,500 deaths in Australia (http://www.melanoma.org.au/about‐melanoma/melanoma‐skin‐cancer‐facts.html).

Early stage melanoma (stage I, II and resectable stage III, local lesion, local lymph nodes spread) can be cured by surgery alone, but late stage metastatic disease (unresectable stage III and IV, lesions spread to distant organs) has generally been considered to be incurable (Balch et al., 2009). Left untreated, the advanced disease has extremely poor outcomes with median overall survival of less than a year and 5 year overall survival of less than 10% (Jang and Atkins, 2014). Until 2011, the standard of care for advanced melanoma was the alkylating agent dacarbazine, and in the US, interleukin‐2, high‐dose IL‐2 (HD IL‐2) and interferon‐α‐2b (IFN‐α) were also approved, but these drugs do not provide significant increase in patient survival (Figure 1). Thus, more than 40 years of research had not led to any tangible improvement in patient outcome.

Figure 1.

Progression of advanced melanoma treatment over time. The first FDA approved agent for the treatment of advanced melanoma was the alkylating agent dacarbazine, which was approved in 1975. In 1995 the cytokine IFN‐α was approved, then followed by the approval of another cytokine IL‐2 in 1998. In 2011 two new agents, the BRAF inhibitor vemurafenib and the CTLA‐4 inhibitor ipilimumab were approved, providing a major breakthrough for the treatment of metastatic melanoma. In 2013 the MEK inhibitor trametinib was approved and at the beginning of 2014 a combination of the BRAF inhibitor dabrafenib and the MEK inhibitor trametinib was approved. % = Objective tumour response, *BRAF mutant only.

However, the landscape for melanoma treatment changed in 2011 with the approval of new immune and targeted therapies that extend progression‐free and overall survival and could, in some cases, cure the disease. The monoclonal antibody ipilimumab, an immune checkpoint inhibitor, and the small molecule vemurafenib, a BRAF inhibitor, were the first in class of new generations of therapies that achieve life extension and in some cases cures for metastatic melanoma patients in randomised clinical trials. These drugs represent a landmark in the clinical management of melanoma and both received prompt FDA approval in 2011 for first‐line treatment in melanoma patients. They were the first new drugs to be approved for melanoma in 13 years. Vemurafenib received European approval 6 months later, but it took almost 2 more years for ipilimumab.

These developments have transformed the clinical management of melanoma and improved patient outcomes and the lessons we learn from melanoma will also impact other cancers. However, despite these advances, not all patients with BRAF mutations respond to BRAF inhibitors and the majority of patients who do respond develop resistance after a relatively short period of disease control. Also, only 15–20% of patients respond to ipilimumab, and it is not currently possible to determine which patients will respond and which will not. Toxicity can also limit use of these drugs in some patients. Thus, there is still work to do, but ipilimumab and vemurafenib are the vanguard of a cadre of exciting new drugs currently undergoing testing in hundreds of clinical trials and in addition to vemurafenib, a second BRAF inhibitor, dabrafenib, and a MEK inhibitor, trametinib, have recently received FDA approval, but have yet to be approved in Europe. On their heels are several other similar drugs, so the revolution in melanoma treatment is ongoing and reflecting the enormous strides recently made, Science selected cancer immunotherapy as the 2013 breakthrough of the year (Couzin‐Frankel, 2013).

2. Pre‐2011 therapies

The development of effective treatments for advanced melanoma has been a long hard road. In 1975 the FDA approved the alkylating agent dacarbazine (5‐[3,3‐dimethyl‐1‐triazenyl]‐imidazole‐4‐carboxamide; DTIC) for advanced metastatic melanoma (Figure 1), although objective clinical responses (mostly partial responses) were seen only in 13–20% of patients and durable responses were extremely rare (Eggermont and Kirkwood, 2004). Temozolomide, an orally available DTIC analogue, did little to improve these responses (Middleton et al., 2000) and for the majority of patients durable responses remained elusive even when DTIC or temozolomide were combined with other drugs (Bhatia et al., 2009). A meta‐analysis of 48 head‐to‐head clinical trials with DTIC revealed a weighted average objective response rate (mostly partial responses) of 15.3% for DTIC alone and no increase in survival or response rates with any combination, apart from IFN‐α, which gave at best a modest improvement (Lui et al., 2007). A biochemotherapy (BCT) regimen of cisplatin, vinblastine and DTIC (CVD) with IFN‐α and high‐dose IL‐2 did achieve response rates exceeding 50% in phase 2 trials, but at the price of substantial toxicity, preventing this therapy from becoming standard‐of‐care (Legha et al., 1996). Thus, attempts to improve responses to DTIC were disappointing and for the most part it was used for palliation rather than cure (Figure 1). It would be another 20 years before the FDA approved another treatment for advanced malignant melanoma.

One of the focuses of melanoma research over the years has been immunotherapy. This research strand was sparked by the observation that a small number of patients achieve “spontaneous cures” and these were largely attributed to attack by the patients' immune system on their own tumour. Melanoma became considered to be a highly immunogenic tumour and attempts to modulate the immune system against melanoma became a key challenge, with interleukin‐2 (IL‐2) leading the way. IL‐2 is a cytokine that induces T cell and natural killer cell proliferation and activation, and stimulates production of interferon gamma and tumour necrosis factor by lymphocytes. High‐dose IL‐2 (HD IL‐2) achieved objective tumour responses in 17% of patients and durable responses in ∼6% of patients (Atkins et al., 2000). It received FDA approval in 1995 (Figure 1), but is highly toxic and so is reserved for generally fit, high performance status patients (Alwan et al., 2014). The immunomodulatory and anti‐tumour cytokine interferon alfa‐2b (IFN‐α) also achieved response rates of 15–20% and received FDA approval in 1998 (Figure 1), but this treatment is more effective in early disease and as with HD IL2, toxicity impacted quality of life so its use in advanced disease has been limited (Payne et al., 2014).

3. The breakthrough in melanoma signalling

The breakthrough in melanoma cell signalling occurred in 2002 when it was discovered that the BRAF gene is mutated in about half of melanomas (Chin et al., 2006; Davies et al., 2002), making BRAF the most common driver oncoprotein in melanoma. BRAF is a protein kinase and a component of the RAS/RAF/MEK/ERK signalling cascade (Figure 2A), a pathway that is activated downstream of receptor tyrosine kinases (RTKs) and which regulates cell proliferation, differentiation, survival and death. The most common mutations in BRAF occur at codon V600, and most commonly to glutamate (V600E), lysine (V600K), aspartate (V600D) or arginine (V600R) (http://cancer.sanger.ac.uk/cosmic/). Codon 600 mutations drive BRAF hyper‐activation and lead to constitutive pathway activation (Figure 2B) (Davies et al., 2002; Hingorani et al., 2003; Karasarides et al., 2004).

Figure 2.

The RAS‐RAF‐MEK‐ERK pathway: a therapeutic target in melanoma. A. In normal cells. Receptor tyrosine kinases (RTKs) are activated by binding of their ligands, and they initiate growth signals through activation of various pathways including the RAS‐RAF‐MEK‐ERK MAPK pathway. B. In melanoma cells, the RAS‐RAF‐MEK‐ERK pathway is hyper activated through several mechanisms, including activating mutations in RAS (20% of cases) and BRAF (∼50% of cases), making the cells independent of the RTKs (dotted circles). Constitutively active RAS or BRAF cause sustained activation of MEK, which in turn activates ERK. Activated ERK regulates many cellular processes that are required for cell proliferation and survival. Selective inhibitors of BRAF, MEK and ERK have been developed to inhibit this hyper‐activated pathway at different steps. * indicates a mutation.

It was already known that the small G‐protein NRAS is mutated in about 20% of melanomas, but the discovery of the BRAF mutation focussed attention on this pathway, and it was subsequently shown that the receptor tyrosine kinases KIT and ERBB4, the small G‐proteins HRAS and KRAS, the RAS‐GAP NF1, and MEK1 and MEK2 are also all mutated in a low proportion of samples (http://cancer.sanger.ac.uk/cosmic) (Chin et al., 2006; Curtin et al., 2005; Nikolaev et al., 2012; Prickett et al., 2009). The high incidence of mutations in this pathway presumably explains why MEK and ERK are activated in the majority of melanomas (Meier et al., 2005), although the precise route of signalling can differ in different genetic backgrounds. Thus, whereas BRAF drives MEK/ERK activity in BRAF mutant cells, the closely related protein CRAF appear largely to be responsible for coupling MEK/ERK activation to oncogenic RAS (Figure 2B) (Blasco et al., 2011; Dumaz et al., 2006; Karreth et al., 2011). Moreover, the PI3K/AKT/mTOR pathway can also be altered in melanoma. Allelic loss or reduced expression of PTEN occurs in 20% and 40% of melanomas (Goel et al., 2006; Slipicevic et al., 2005) and a recent whole genome sequencing study in a subset of 121 melanoma samples revealed that ∼44% of samples with BRAF mutations also harboured mutations in, or focal deletion of PTEN. Also notable is that PTEN is altered in ∼4% of melanomas with NRAS mutations (Hodis et al., 2012).

4. Targeted therapies: the early trials

BRAF was rapidly validated as a therapeutic target in melanoma (Hingorani et al., 2003; Karasarides et al., 2004), and several BRAF drug discovery programmes were initiated. The first RAF inhibitor to be clinically tested in melanoma was sorafenib (BAY43‐9006), an orally available small molecule CRAF inhibitor (Lyons et al., 2001). Although the first clinical trial was in unselected patients, retrospective analysis revealed limited efficacy in BRAF mutant patients (Eisen et al., 2006) and when sorafenib was combined with carboplatin and paclitaxel no additional benefit was observed (Hauschild et al., 2009). This caused some to question if BRAF was a good therapeutic target in melanoma (Madhunapantula and Robertson, 2008), but in truth although sorafenib is very active against ^V600EBRAF in vitro, its activity in cells is low, dropping off by over 100 fold (Karasarides et al., 2004). Thus, in patients effective ^V600EBRAF inhibition could not be achieved at maximum tolerated doses (Flaherty et al., 2005). The more selective BRAF inhibitor RAF265 also produced disappointing results because of dose limiting haematologic toxicity (Sharfman et al., 2011) and although the pan‐RAF inhibitor XL281 was well tolerated and showed clinical benefit in 36% of patients, the compound did not progress beyond Phase I/II trials (Schwartz et al., 2009).

The early MEK inhibitors were also disappointing. The first to enter clinical trials was CI‐1040 (PD184352) and although no melanoma patients received the drug, the trial was discontinued due to poor responses and unacceptable toxicities (Lorusso et al., 2005). An analogue of CI‐1040, PD0325901 was evaluated in melanoma patients and achieved partial responses and disease stabilisation in BRAF mutant tumours, validating the pathway as a therapeutic target, but again toxicity prevented development beyond phase I/II (Lorusso et al., 2007). Finally, selumetinib (AZD6244) was combined with temozolomide in BRAF and NRAS mutant patients, but failed to improve overall or progression free survival (Kirkwood et al., 2012).

5. Targeted therapies: the paradigm shift

5.1. BRAF inhibitors

The first BRAF selective inhibitor to be approved for use in BRAF mutant melanoma patients was vemurafenib (PLX4032), an orally available small molecule. A precursor, PLX4720, was developed using a scaffold‐based drug design approach and was over 10 fold more active against ^V600EBRAF than against the wild‐type protein (Tsai et al., 2008). PLX4720 showed good activity in BRAF mutant cell lines and therapeutic efficacy in BRAF mutant melanoma and colorectal xenografts, but vemurafenib displayed better efficacy in xenograft models and better pharmacokinetics, and despite high compound exposures in different pre‐clinical models, it had a very clean toxicity profile (Bollag et al., 2010; Joseph et al., 2010). Based on these exciting pre‐clinical observations, vemurafenib was selected for clinical development, leading to a phase I dose‐escalation clinical trial in ^V600EBRAF mutant enriched melanoma patients. Even in this early clinical trial, vemurafenib displayed evidence of pharmacodynamic biomarker modulation in treatment vs. baseline biopsies specimens, with decreased phospho‐ERK, Ki67 and Cyclin D1 and critically, it induced complete or partial tumour regression in 81% of the patients. Adverse events, especially at the highest doses, included rash, arthralgia, photo‐sensitivity, fatigue and the induction of cutaneous squamous cell carcinoma (cuSCC) and keratoacanthoma (KA) (Flaherty et al., 2010).

A randomized phase III trial showed for the first time a significant improvement in overall and progression‐free survival for melanoma patients treated with a targeted agent, with median progression‐free survival of 5.3 months for vemurafenib and 1.6 months for dacarbazine (DTIC) (Chapman et al., 2011). A further Phase II study aimed to define, in a larger number of patients, the overall response rate in advanced melanoma patients treated with vemurafenib. This study confirmed a response rate of over 50% in patients with previously treated metastatic melanoma bearing V600E or V600K BRAF mutations (Sosman et al., 2012). The latest update for the phase III open‐label study (BRIM‐3) showed a median overall survival with vemurafenib of 13.3 months compared with 10.0 months with dacarbazine (McArthur et al., 2014), confirming that vemurafenib was effective in this patient group. Vemurafenib received FDA approval for use in BRAF mutant melanoma patients in August 2011 and was approved in Europe in December 2011.

GSK2118436 (Dabrafenib), a more potent second generation BRAF inhibitor that has some CRAF activity also demonstrated efficacy against BRAF mutant tumours in pre‐clinical studies (Laquerre et al., 2009). A Phase I trial in patients with solid tumours enriched for previously untreated BRAFV600E melanoma patients with brain metastases demonstrated efficacy and durable responses in some patients (Falchook et al., 2012). A phase III open‐label randomized study confirmed that dabrafenib increased progression‐free survival from 2.7 (dacarbazine) to 5.1 months (Hauschild et al., 2012) with only mild adverse effects, mostly cutaneous squamous‐cell carcinoma, keratoacanthoma, fatigue, pyrexia and neutropenia. Dabrafenib was approved by the FDA for use in BRAF mutant melanoma patients in May 2013 and received European approval in September 2013.

5.2. MEK inhibitors

Although the first MEK inhibitors were disappointing, continued interest in this area led to the development of compounds that are active in BRAF mutant melanoma patients. GSK1120212 (trametinib) is an allosteric MEK inhibitor that does not have the toxicities associated with earlier compounds, possibly because of its higher potency (Gilmartin et al., 2011). In a phase I trial, 10 of 36 BRAF mutant melanoma patients achieved partial response or stable disease (Falchook et al., 2012). The compound was well tolerated and the toxicities, which were manageable and reversible, included rash, dermatitis acneiform, diarrhoea and central serous retinopathy. The phase II trial confirmed that trametinib could achieve tumour regression and disease stabilisation in metastatic ^V600EBRAF or ^V600KBRAF melanoma patients even if they had previously been treated with other BRAF inhibitors (Kim et al., 2013). An open label phase III trial with 322 patients demonstrated progression‐free survival of 4.5 months for the trametinib‐treated patients compared to 1.5 months for the chemotherapy group. Notably, despite allowing crossover between the groups, the 6 months survival for trametinib‐treated patients was 81% compared to 67% for the chemotherapy group (Flaherty et al., 2012). These encouraging results led to FDA approval in advanced ^V600EBRAF or ^V600KBRAF melanoma patients in May 2013 and European approval in April 2014 (Figure 1).

We will comment about new targeted therapies combinations based on BRAF and MEK inhibitors in our section “Combination Therapies”.

6. The importance of RAF dimers

Early studies (Farrar et al., 1996; Luo et al., 1996) demonstrated that enforced dimerization of CRAF could drive its activation, but it took several years for the significance of this observation to become clear. Eight years after the original finding, it was shown that some of the rare BRAF mutant proteins have impaired kinase activity but can nevertheless activate MEK (Wan et al., 2004). This is because although these mutants do not phosphorylate MEK, they form stable dimers with CRAF, stimulating its activity and forcing it to activate MEK. These early studies established that enforced dimerization of RAF can lead to its activation but it is now clear that dimerization is a critical step in the normal activation of the wild‐type proteins (Freeman et al., 2013; Hu et al., 2013).

It has also been established that dimerization plays an important role in cell responses to BRAF inhibitors. Although BRAF inhibitors block MEK/ERK signalling in BRAF mutant cells, they activate MEK in RAS mutant cells, or when RAS is activated by upstream signals (Halaban et al., 2010; Hatzivassiliou et al., 2010; Heidorn et al., 2010; Poulikakos et al., 2011). This is because RAF inhibitors drive BRAF/CRAF hetero‐ and homo‐dimerization in the presence of oncogenic or activated RAS. These dimers contain an inactive inhibitor‐bound partner and an inhibitor‐free partner and within these complexes, the inhibitor‐free partner is activated through scaffold functions or conformational changes driven by the inhibitor‐bound partner. Note that although we generally refer to “dimers”, the stoichiometry of components in these complexes is unclear and is likely to be much higher‐order, because ARAF (a third RAF family member) and KSR (a distant relative) are also present in these inhibitor‐induced complexes (Freeman et al., 2013; Luo et al., 1996; Rebocho and Marais, 2013). Critically, paradoxical activation of MEK and ERK can drive tumourigenesis and appears to underpin the development of cuSCC and KA in patients treated with BRAF inhibitors (Mattei et al., 2013; Oberholzer et al., 2012; Su et al., 2012). Furthermore, some patients treated with BRAF inhibitors rapidly develop secondary tumours including leukaemia (Callahan et al., 2012). Thus, it appears that BRAF inhibitors can impose a Darwinian growth‐selective pressure on pre‐existing lesions with RAS mutations or upstream events that activate RAS, driving the growth of previously indolent lesions in the patients who receive these drugs.

7. The persistence of resistance

Importantly, despite the dramatic responses of BRAF mutant melanoma patients to BRAF and MEK inhibitors, the majority of patients who initially respond to these drug relapse with acquired resistance within a year (Chapman et al., 2011; Flaherty et al., 2010; Sosman et al., 2012). Furthermore, 20% or so of patients present intrinsic resistance and do not respond despite presenting tumours with BRAF mutations. In the majority of cases, resistance is associated with reactivation of the MEK/ERK pathway (Figure 3), and in many cases phospho‐ERK levels are highly elevated once resistance is established (Paraiso et al., 2010; Trunzer et al., 2013).

Figure 3.

Mechanisms of resistance to targeted therapies in melanoma. The remarkable clinical efficacy of BRAF inhibitors in mutant BRAF melanoma is limited by acquired and intrinsic drug resistance, with several mechanisms having been described. Upstream the RAS/RAF/MEK/ERK pathway, activation or upregulation of RTKs such as IGF‐1R, PDGFRα and EGFR signalling through SFKs (SRC family kinases). Downstream the RTKs, on the RAS/RAF/MEK/ERK axis, mutations in RAS, amplification of CRAF and BRAF, expression of truncated BRAF, mutations in MEK or increased expression of MAP3K8/COT or MLK1‐4. Downstream the RTKs, on alternative signalling pathways, deletions in PTEN and mutations in PI3K and AKT. * indicates mutations or activation/upregulation.

Notably, the RAF proteins themselves can mediate resistance to RAF inhibitors. Mutant BRAF is amplified in ∼20% of resistant tumours, leading to increased levels of ^V600EBRAF protein that appears to overwhelm the drugs (Shi et al., 2012b). CRAF amplification can also mediate resistance, although the clinical significance of this is unclear (Montagut et al., 2008). Alternative splicing of exons 4–10 of mutant BRAF causes in‐frame deletions that generate shortened forms of ^V600EBRAF (∼61‐kDa) that lack the RAS‐binding domain (Poulikakos et al., 2011). Curiously, these truncated proteins dimerize constitutively with wild‐type BRAF and drive resistance, which is surprising because RAS binding is not required for oncogenicity of full‐length ^V600EBRAF (Davies et al., 2002), and the shortened proteins retain the kinase domain to which the RAF inhibitors bind. Clearly, unravelling how these truncated proteins mediate resistance will be both intriguing and informative. Downstream of RAF, mutations in MEK can drive resistance (Emery et al., 2009) although not all MEK mutants are able to do so (Shi et al., 2012a), and at least in experimental systems, the MEK kinases COT and MLK1‐4 (Johannessen et al., 2010; Marusiak et al., 2014) can drive resistance.

Moving upstream, approximately 25% of resistant tumours carry mutations in NRAS (Nazarian et al., 2010; Trunzer et al., 2013; Van Allen et al., 2013). Increased signalling by RTKs can also mediate resistance, with up‐regulation of PDGFRβ, IGF‐1R, EGFR and MET (the latter two signalling through the SRC family of non‐receptor tyrosine kinases) can also mediate resistance (Girotti and Marais, 2013; Girotti et al., 2013; Nazarian et al., 2010; Shi et al., 2011; Vergani et al., 2011; Villanueva et al., 2010). Notably, EGFR or PDGFRβ upregulation can be mediated by the transcription factor SOX10 via TGF‐β signalling (Sun et al., 2014), and even further upstream, the microenvironment can also mediate resistance, with two studies demonstrating that secretion of HGF (hepatocyte growth factor) by the tumour stromal cells can drive resistance (Straussman et al., 2012; Wilson et al., 2012). Of course, RTKs do not only signal through the RAS‐MEK‐ERK pathway, and much attention has been focussed on the PI3K/AKT/mTOR pathway, because mutations in AKT and PI3KCA, or loss of PTEN can all drive resistance (Atefi et al., 2011, 2013, 2011, 2014, 2014).

Fewer studies have investigated how intrinsic resistance is mediated, but the underlying mechanisms could be similar. We recently reported that intrinsic resistance was mediated by a mutation in GNAQ that sustains ERK activity and a deletion in PTEN that activates PI3–K/AKT pathway (Turajlic et al., 2014). Similarly, mutations in RAC1 and HOXD8 are associated with intrinsic resistance, but these results have not been validated and the underlying mechanisms are unclear (Van Allen et al., 2013).

8. Overcoming resistance

Many mechanisms of resistance have been described, so in order to offer the best care, it is important to develop personalised medicine approaches for individual patients. Several targeted agent combinations are being tested for their ability to overcome resistance and many are based on combinations of BRAF and MEK inhibitors to explore the efficacy of simultaneously inhibiting the pathway at several steps (Table 1). Others will explore co‐inhibition of this and other pathways, such as combinations of BRAF or MEK inhibitors with drugs that target CDK4/6 or the PI3K/AKT/mTOR pathway (Table 1). As already observed, BRAF/MEK inhibitor combinations can extend disease‐free and overall survival compared to BRAF inhibitors alone, but also exciting is the development of selective ERK inhibitors which will provide opportunities to target the pathway at up to three nodes. The ERK inhibitor SCH772984 inhibits the growth of cells that are resistant to BRAF and MEK inhibitors (Morris et al., 2013), and two other ERK inhibitors, MK‐8353 and GDC‐0994 are currently undergoing phase I clinical trials in solid tumours including melanoma (NCT01358331; NCT01875705). Clearly, the progress of new compounds and combinations will be followed with close interest.

Table 1.

Ongoing combination of targeted therapy‐based clinical trials for melanoma treatment.

	Clinical trial number	Details
Combination therapy	NCT01826448	A Phase 1b Open Label, Dose Escalation Study of PLX3397 in Combination With Vemurafenib in V600‐mutated BRAF Melanoma
	NCT01820364	LGX818 in Combination With Agents (MEK162; BKM120; LEE011; BGJ398; INC280) in Advanced BRAF Melanoma
	NCT01781572	A Phase Ib/II Study of LEE011 in Combination With MEK162 in Patients With NRAS Mutant Melanoma
	NCT01928940	Japan PhI/II of GSK2118436 and GSK1120212 Combination in Subjects With BRAF V600 E/K Mutation Positive Advanced Solid Tumors (Phase I Part) or Cutaneous Melanoma (Phase II Part)
	NCT01433991	E7050 in Combination With E7080 in Subjects With Advanced Solid Tumors (Dose Escalation) and in Subjects With Recurrent Glioblastoma or Unresectable Stage III or Stage IV Melanoma After Prior Systemic Therapy (Expansion Cohort and Phase 2)
	NCT01909453	Study Comparing Combination of LGX818 Plus MEK162 and LGX818 Monotherapy Versus Vemurafenib in BRAF Mutant Melanoma (COLUMBUS)
	NCT01512251	BKM120 Combined With Vemurafenib (PLX4032) in BRAFV600 E/K Mutant Advanced Melanoma
	NCT01673737	A Phase I/Ib Trial for the Evaluation of SAR260301 in Monotherapy or in Combination With Vemurafenib in Patients With Various Advanced Cancer
	NCT02065063	A Study to Investigate the Safety, Pharmacokinetics, Pharmacodynamics, and Anti‐Cancer Activity of Trametinib in Combination With Palbociclib in Subjects With Solid Tumors
	NCT01701037	Dabrafenib Alone and in Combination With Trametinib Before Surgery in Treating Patients With Locally or Regionally Advanced Melanoma That Can Be Removed By Surgery
	NCT01689519	coBRIM: A Phase 3 Study Comparing GDC‐0973 (Cobimetinib), a MEK Inhibitor, in Combination With Vemurafenib vs Vemurafenib Alone in Patients With Metastatic Melanoma
	NCT01777776	Safety and Efficacy of LEE011 and LGX818 in Patients With BRAF Mutant Melanoma
	NCT01562899	A Study of MEK162 and AMG 479 in Patients With Selected Solid Tumors
	NCT01519427	Selumetinib and Akt Inhibitor MK2206 in Treating Patients With Stage III or Stage IV Melanoma Who Failed Prior Therapy With Vemurafenib or Dabrafenib
	NCT01271803	A Study of Vemurafenib And GDC‐0973 in Patients With BRAF‐Mutation Positive Metastatic Melanoma
	NCT01390818	Trial of MEK Inhibitor and PI3K/mTOR Inhibitor in Subjects With Locally Advanced or Metastatic Solid Tumors
	NCT01616199	Study of PX‐866 and Vemurafenib in Patients With Advanced Melanoma
	NCT01363232	Safety, Pharmacokinetics and Pharmacodynamics of BKM120 Plus MEK162 in Selected Advanced Solid Tumor Patients

Another approach to overcoming resistance is based on concept of the “drug holiday”. Tumours grow more slowly when mice are treated with cycles of 4 weeks of BRAF inhibitor interspersed with two weeks of “drug holiday” than if treated with drug continuously (Das Thakur et al., 2013). This approach is thought to exploit the heterogeneity of the tumour and the idea that the resistant clones are less fit than the parental cells or develop a dependence on the drugs to which they are resistant. Consequently, during the treatment cycle, the parental clones recede while the resistant clones grow. During the “holiday”, the parental clones recover but the resistant clones recede. Thus, over the cycle, one population of cells is always receding and the other is growing resulting in an overall delay in growth (Figure 4). In line with this, upregulation of the EGFR in BRAF inhibitor resistant melanoma cells induces the hallmarks of oncogene‐induced senescence and causes the cells to adopt a slow‐growth phenotype, but when these cells are exposed to a BRAF inhibitor, they proliferate (Sun et al., 2014). Similar responses occur when the cells are exposed to TGF‐β. Thus, melanoma EGFR_high cells are enriched by the drug but diminished during the holiday, whereas EGFR_low cells respond in the opposite direction.

Figure 4.

Overcoming resistance by the “drug holiday” regimen. A. Continuous dosing of BRAF inhibitors arrests proliferation and growth of sensitive cancer cells (green line) but at the same time triggers the onset of resistance in a subset of cells (red line), allowing the resistant cells to grow and replace the sensitive cells in the tumour, until the clinical evidence of resistance emerges. B. Intermittent dosing of BRAF inhibitors puts different selective pressure on the two cell populations. During dosing (ON) the drug arrests proliferation of the sensitive cells (green line), but the resistant cells grow (red line). During the ‘holiday’ (OFF) the parental cells return, but the resistant cells retreat. Cycling through the drug ON and OFF phases extends the period of response in the patients.

The clinical relevance of these types of approaches was established in a BRAF mutant melanoma patient who developed a leukaemia driven by NRAS. Intermittent administration of a BRAF plus MEK inhibitor combination appeared to keep the growth of the melanoma in check, while the combination was thought to have overcame the accelerated growth of the leukaemia driven by paradoxical activation of the MEK/ERK pathway (Abdel‐Wahab et al., 2014). Clearly, we have much to learn about how to best use the drugs at our disposal and it seems likely that careful scheduling or sequencing of the drugs will play a key role in achieving the best responses for individual patients.

9. Melanoma immunotherapy

As already observed, the development of immunotherapies for cancer has been a long hard road, with many years of research providing very low yield. For many years the promise of immunotherapy was more dream than reality. Of course, the immediate function of the immune system is to protect us from infectious diseases, because if we do not survive our first few years of life, cancer does not present a significant problem. However, the ability of the immune system in general, and T cells in particular to target tumour cells suggested that it could be harnessed to battle cancer and over the past three decades, the complexities of the immune system have been studied in great depth. This has led to improved insight into the role that cytokines and chemokines play in regulating immune cell functions and into how T‐cells are activated. It is this insight that has allowed the immunotherapy dream to become reality.

9.1. Cytokines

Interferon alpha‐2b (IFNα) is a key modulator of the immune system that, amongst other functions, upregulates antigen presentation and increases antigen recognition by T cells. This suggested that this cytokine could be used to increase tumour cell recognition by the immune system and led to a number of clinical trials. Overall, these trials reported that IFNα could achieves overall objective responses of ∼20% in patients (Creagan et al., 1986, 1986), but responses in stage IV patients are modest and cumulative toxicities limited their use. Nevertheless IFNα and pegylated IFNα received FDA approval for adjuvant therapy in resected high‐risk stage III melanoma in 1995. As monotherapies these agents have had little impact on overall survival, so combination studies with other agents are still ongoing (Table 2).

Table 2.

Ongoing immunotherapy‐based clinical trials for melanoma treatment.

Therapy	Clinical trial number	Details
Immunotherapy only	NCT00094653	MDX‐010 Antibody, MDX‐1379 Melanoma Vaccine, or MDX‐010/MDX‐1379 Combination Treatment for Patients With Unresectable or Metastatic Melanoma
	NCT00972933	Immunogenicity and Biomarker Analysis of Neoadjuvant Ipilimumab for Melanoma
	NCT00289640	Study of Ipilimumab (MDX‐010) Monotherapy in Patients With Previously Treated Unresectable Stage III or IV Melanoma
	NCT00289627	A Single Arm Study of Ipilimumab Monotherapy in Patients With Previously Treated Unresectable Stage III or IV Melanoma
	NCT00636168	Efficacy Study of Ipilimumab Versus Placebo to Prevent Recurrence After Complete Resection of High Risk Stage III Melanoma
	NCT00324155	Dacarbazine and Ipilimumab vs. Dacarbazine With Placebo in Untreated Unresectable Stage III or IV Melanoma
	NCT01515189	Phase 3 Trial in Subjects With Metastatic Melanoma Comparing 3 mg/kg Ipilimumab Versus 10 mg/kg Ipilimumab
	NCT00495066	Compassionate Use Trial for Unresectable Melanoma With Ipilimumab
	NCT00729950	Study of MDX‐010 in Subjects With Unresectable Stage III or Stage IV Malignant Melanoma
	NCT01696045	Phase 2 Study of Ipilimumab in Children and Adolescents (12 to < 18 Years) With Previously Treated or Untreated, Unresectable Stage III or Stage lV Malignant Melanoma
	NCT01355120	THE IPI ‐ Trial in Advanced Melanoma: Melanoma Patients With Advanced Disease
	NCT01715077	Study of Ipilimumab in the Immune System
	NCT00928031	Long‐term Data Collection for Subjects in MDX‐010 Studies
	NCT01693562	A Phase 1 Study to Evaluate MEDI4736 (anti PD‐L1)
	NCT01704287	Study of MK‐3475 (anti PD‐1) Versus Chemotherapy in Participants With Advanced Melanoma
	NCT01866319	Study to Evaluate the Safety and Efficacy of Two Different Dosing Schedules of MK‐3475 Compared to Ipilimumab in Participants With Advanced Melanoma (MK‐3475‐006 AM1)
	NCT01295827	Study of MK‐3475 in Participants With Progressive Locally Advanced or Metastatic Carcinoma, Melanoma, or Non‐small Cell Lung Carcinoma (P07990/MK‐3475‐001)
	NCT01375842	Study of the Safety and Pharmacokinetics of MPDL3280A Administered Intravenously As a Single Agent to Patients With Locally Advanced or Metastatic Solid Tumors or Haematologic Malignancies
	NCT01721772	Study of BMS‐936558 vs. Dacarbazine in Untreated, Unresectable or Metastatic Melanoma (CheckMate 066)
	NCT01621490	Phase 1 Biomarker Study of Anti‐PD‐1 BMS‐936558 in Advanced Melanoma
	NCT00441337	A Study of MDX‐1106 in Patients With Selected Refractory or Relapsed Malignancies
	NCT00730639	A Phase 1b Study of MDX‐1106 in Subjects With Advanced or Recurrent Malignancies
Immunotherapy combined with targeted therapies	NCT01940809	Ipilimumab With or Without Dabrafenib, and/or Trametinib in Treating Patients With Melanoma That is Metastatic or Cannot Be Removed By Surgery
	NCT01767454	Study of Dabrafenib ± Trametinib in Combination With Ipilimumab for V600 E/K Mutation Positive Metastatic or Unresectable Melanoma
	NCT01673854	Phase II Safety Study of Vemurafenib Followed by Ipilimumab in Subjects With V600 BRAF Mutated Advanced Melanoma
	NCT01400451	Ph I Ipilimumab Vemurafenib Combo
	NCT02027961	A Phase 1 Open‐label Study of Safety and Tolerability of MEDI4736 (anti PD‐L1) in Combination With Dabrafenib and Trametinib or With Trametinib Alone
	NCT01656642	A Study of The Safety and Pharmacology of MPDL3280A Administered in Combination With Vemurafenib (Zelboraf®) in Patients With Previously Untreated BRAFV600‐Mutation Positive Metastatic Melanoma
	NCT01721746	A Study to Compare BMS‐936558 to the Physician's Choice of Either Dacarbazine or Carboplatin and Paclitaxel in Advanced Melanoma Patients That Have Progressed Following Anti‐CTLA‐4 Therapy (CheckMate 037)
Immunotherapy combined with immunotherapy	NCT01708941	Ipilimumab With or Without High‐Dose Recombinant Interferon Alfa‐2b in Treating Patients With Stage III‐IV Melanoma That Cannot Be Removed By Surgery
	NCT01274338	Ipilimumab or High‐Dose Interferon Alfa‐2b in Treating Patients With High‐Risk Stage III‐IV Melanoma That Has Been Removed by Surgery
	NCT01856023	HD IL‐2 + Ipilimumab in Patients With Metastatic Melanoma
	NCT01608594	Neoadjuvant Combination Therapy With Ipilimumab and HighDose IFN‐ë±2b for Melanoma
	NCT01927419	Phase 2, Randomized, Double Blinded, Study of Nivolumab (BMS‐936558) in Combination With Ipilimumab vs Ipilimumab Alone in Subjects With Previously Untreated, Unresectable or Metastatic Melanoma (CheckMate 069)
	NCT01844505	Phase 3 Study of Nivolumab or Nivolumab Plus Ipilimumab Versus Ipilimumab Alone in Previously Untreated Advanced Melanoma (CheckMate 067)
	NCT01024231	Dose‐escalation Study of Combination BMS‐936558 (MDX‐1106) and Ipilimumab in Subjects With Unresectable Stage III or Stage IV Malignant Melanoma
	NCT01866319	Study to Evaluate the Safety and Efficacy of Two Different Dosing Schedules of MK‐3475 Compared to Ipilimumab in Participants With Advanced Melanoma (MK‐3475‐006 AM1)
	NCT01672450	A Study of Intratumoral Injection of Interleukin‐2 and Ipilimumab in Patients With Unresectable Stages III‐IV Melanoma
	NCT01245556	Safety and Efficacy Study of BMS‐908662 in Combination With Ipilimumab in Subjects With Advanced Melanoma
	NCT01783938	Study of Nivolumab Given Sequentially With Ipilimumab in Subjects With Advanced or Metastatic Melanoma (CheckMate 064)
	NCT01480323	A Phase II Study to Evaluate Safety and Efficacy of Combined Treatment With Ipilimumab and Intratumoral Interleukin‐2 in Pretreated Patients With Stage IV Melanoma
	NCT00058279	Monoclonal Antibody Therapy and Interleukin‐2 in Treating Patients With Metastatic Melanoma
	NCT01024231	Dose‐escalation Study of Combination BMS‐936558 (MDX‐1106) and Ipilimumab in Subjects With Unresectable Stage III or Stage IV Malignant Melanoma
	NCT01927419	Phase 2, Randomized, Double Blinded, Study of Nivolumab (BMS‐936558) in Combination With Ipilimumab vs Ipilimumab Alone in Subjects With Previously Untreated, Unresectable or Metastatic Melanoma (CheckMate 069)
	NCT01844505	Phase 3 Study of Nivolumab or Nivolumab Plus Ipilimumab Versus Ipilimumab Alone in Previously Untreated Advanced Melanoma (CheckMate 067)
	NCT01176461	Multiple Class I Peptides & Montanide ISA 51 VG w Escalating Doses of Anti‐PD‐1 ab BMS936558

Interleukin‐2 is another cytokine that, amongst other functions, increases T cell proliferation and maturation into effector T cells and so was also tested for its ability to stimulate anti‐tumour activity. High dose interleukin‐2 (HD IL‐2) achieved objective responses in ∼15% of patients, with durable complete responses in a small proportion (Atkins et al., 2000, 1999). However, HD IL‐2 suffers from multi‐organ complications that require close monitoring and careful management. Nevertheless, it received FDA approval for stage IV disease in 1998 but is largely reserved for younger patients with excellent performance status. Despite these drawbacks, IFNα, and HD IL‐2 established that cell‐free immune‐based treatments could achieve cures, and in some cases in advanced disease patients.

9.2. Immune checkpoint antibodies

A key concepts to emerge from the intense studies of the last decades is that the immune system possesses a number of “immune checkpoints” that ensure that T cells are activated only when and where they need to be (Figure 5). These checkpoints are essential for preventing the induction of autoimmune diseases (Melero et al., 2013), but cancer cells exploit them to evade immune surveillance. Thus, many immunotherapies seek to overcome the checkpoints so that T cells can be freed up to mediate killing of the cancer cells.

Figure 5.

Immune Checkpoints Inhibitors in Melanoma. A. Priming phase. Antigen presenting cells (APC) are specialised cells which present antigens to T cells to regulate T cell activation via a stimulatory signal triggered by the interaction between the MHC receptor and the T cell receptor together with binding of CD80 on the APC and CD28 on the T cells. An inhibitory signal mediated by the interaction between the CD80 on the APC and CTLA‐4 on the T‐Cells modulates the T cell response. Inhibiting CTLA‐4 with antibodies such as ipilimumab induces T cell activation and allows them to recognise and mediate killing of the tumour cells. B. Effector phase. T cells recognise the antigen on cancer cells, but the cancer cell suppress the ability of the T cells to mediate their killing through the interaction between PDL‐1 with PD‐1. Anti‐PD‐1 antibodies such as nivolumab and pembrolizumab, or anti‐PDL‐1 antibodies such as MPDL3280A inhibit this checkpoint, increasing the ability of T cell to target the cancer cells.

9.2.1. CTLA‐4 antibodies

T cell activation is initiated when antigen presenting cells (APC) present antigens to T cells in an MHC restricted manner (Thompson and Allison, 1997). However, to prevent T cells from recognising self‐antigens and initiating autoimmunity, this process is carefully modulated by the checkpoints, and so for an APC/T cells interaction to be productive, a T cell receptor (TCR) called CD28 must bind to CD80 on the APC to drive a positive signal in the T cell that consolidates its activation (Figure 5). Even this step is carefully regulated and during the early interaction, a second TCR called CTLA‐4 is upregulated and competes with CD28 for binding to CD80. In opposition to CD28, CTLA‐4 drives a negative signal that finely‐tunes or dampens T cell activation (Figure 5). Critically, it is this interaction that appears to attenuate T cell responses to cancer cells. In a seminal study in 1996, Allison and colleagues demonstrated that CTLA‐4 neutralizing antibodies could “immunise” mice against a second challenge by the same tumour (Leach et al., 1996), providing the foundation for the development of immune checkpoint agents that target CTLA‐4.

The first anti‐CTLA‐4 antibody to enter the clinic was ipilimumab, a fully humanised monoclonal antibody that binds directly to CTLA‐4. In the phase I trial, it was shown that ipilimumab induced extensive tumour necrosis with lymphocyte and granulocyte infiltration in metastatic melanoma lesions (Hodi et al., 2003). The phase II trial reported objective responses in 5.4% of patients treated with ipilimumab monotherapy, and 14.3% of patients treated with ipilimumab and DTIC (Hersh et al., 2011). A phase III trial reported an overall survival benefit of 10.1 months for patients receiving ipilimumab, compared to 6.4 months for patients receiving a glycoprotein 100 (gp100) peptide vaccine, and at 24 months, survival rates were 23.5 months for ipilimumab compared to only 13.7% for gp100 (Hodi et al., 2010). A subsequent study reported increased survival from 9.1 to 11.2 months for ipilimumab with dacarbazine compared to dacarbazine alone (Robert et al., 2011). Ipilimumab received prompt FDA approval for metastatic melanoma in March 2011 (Figure 1). The 10‐year survival data for ipilimumab are equally impressive, with a 17% survival rate (Schadendorf et al., 2013). The excitement surrounding ipilimumab is confirmed by the number of clinical trials with ipilimumab alone or in combination with all other treatment modalities, including radiotherapy, targeted therapies, conventional therapies, and also other immunomodulators (Table 2). Note however that the anti‐CTLA‐4 antibody tremelimumab failed to achieve significant efficacy in Phase III clinical trial (Ribas, 2010), showing not all anti‐CTLA‐4 antibodies are effective.

9.2.2. PD‐1 antibodies

A second immune checkpoint is mediated by the programmed death‐ligand 1 (PD‐L1; B7‐H1 and CD274) and programmed death receptor‐1 (PD‐1; CD279) signalling pathway. This checkpoint functions when activated (cytotoxic) T cells interact with their target tumour cells. PD‐L1 (B7‐H1, CD274) (Freeman et al., 2000) and PD‐L2 (B7‐DC, CD273) (Latchman et al., 2001; Tseng et al., 2001) are trans‐membrane ligands that are expressed on normal and tumour cells and bind to the PD‐1 expressed on T cells to mark the cells as ‘self’ and prevent activated T cells from killing them. Consequently, even when tumours are highly infiltrated by regulatory T cells, cancer cell expression of PD‐L1 can suppress T cell function and allow the cancer cells to evade the immune system (Ishida et al., 1992; Keir et al., 2008).

That this checkpoint could be exploited was validated when it was shown that expression of PD‐L1 (B7‐H1) reduces tumour immunogenicity, and this could be reverted using PD‐L1 blocking monoclonal antibodies (Dong et al., 2002). It was also shown that PD‐L1 expression was elevated in high‐risk primary melanoma, and was a prognostic marker for poor survival (Hino et al., 2010). Furthermore, it was shown that PD‐1 expression was elevated in T cells from patients with stage IV disease and that PD‐1 expression on tumour‐infiltrating T cells increased during clinical progression (Sznol and Chen, 2013).

Several fully humanised monoclonal antibodies to target the PD‐1/PDL1 axis have been produced. The first evidence of efficacy in a phase I trial was with nivolumab (MDX‐1106, BMS 936558), an anti‐PD‐1 monoclonal antibody, in patients with refractory solid tumours. Although only one colorectal cancer patient achieved a durable complete response, one melanoma patient and one renal cell carcinoma (RCC) patient achieved partial responses, and one melanoma patient and one non‐small cell lung carcinoma (NSCLC) patient achieved significant tumour regression (Brahmer et al., 2010), these data established that PD‐1 blocking antibodies could be effective and were well tolerated. Critically, the three‐year follow‐up data demonstrated prolonged survival in a small number of the patients (Lipson et al., 2013).

Subsequent studies established cumulative response rates of 18% for NSCLC patients, 27% for RCC patients and 28% for melanoma patients (Topalian et al., 2012). Critically, durable responses of over 1 year were observed in almost a third of the patients and a phase I clinical trial in treatment naïve or treatment refractory melanoma patients comparing nivolumab to nivolumab plus a peptide vaccine demonstrated response rates of ∼25%, with some responses lasting for up to 140 weeks (Weber et al., 2013). By treating ipilimumab‐refractory patients, this trial also established that combination or sequencing of nivolumab with ipilimumab was feasible and may be effective.

A second PD‐1 monoclonal antibody, pembrolizumab (MK‐3475) has proven equally effective. A phase I trial with pembrolizumab reported response rates of 38% and no negative effects if the patient had previously received ipilimumab. Again, the responses were durable in the majority of patients and the median survival was more than 7 months (Hamid et al., 2013a). In a recently reported update of the phase I trail in both ipilimumab naïve and refractory patients, the median survival has not been reached but the overall survival is 69% at 12 months and 62% at 18 months, with an ORR of 34% (Ribas et al., 2014), truly astonishing responses in this patient group.

9.2.3. PD‐L1 antibodies

To complement the anti‐PD‐1 agents, antibodies have also been developed against PD‐L1. Critically, PD‐L1 expression increases with melanoma progression and may contribute to escape from the host immune system (Hino et al., 2010). In a phase I study, the antibody BMS‐936559 (MDX‐1105) achieved objective response rates of 6–17% in patients with solid tumours, with durable responses and prolonged tumour stabilisation in some patients (Brahmer et al., 2012). Preliminary results from a phase I trial with a second PD‐L1 antibody, MPDL3280A (RG7446) revealed that some patients observed tumour shrinkage within days of treatment (Hamid et al., 2013b). With this antibody, the objective response rate was 26% (9/35) and the 24‐week progression‐free survival rate was 35%. Finally, preliminary results with MEDI4736, a third PD‐L1 antibody demonstrates that this antibody is also well tolerated and that tumour reduction was observed in NSCLC and melanoma (Lutzky et al., 2014).

9.3. Immunotherapies: the challenges

Although it is early days, the studies above establish that checkpoint antibodies have enormous potential in a range of cancers including melanoma. However a major weakness of immunotherapies is the lack of predictive biomarkers of response; we simply do not yet know how to determine which patients will respond to these therapies. Learning how to decide this will be important if costs are to be kept down for health‐care providers, but also to ensure that patients are not treated with potentially toxic agents unnecessarily (see below). Clearly, this is an area of intense research and recent advances have been made. For example, PD‐L1 expression in tumours combined with T‐cell gene signatures can give an indication of response to MPDL3280A (Powderly et al., 2013).

Alongside the lack of predictive markers is the challenge of learning how to assess patient responses. While some patients respond immediately, others can take months, and in some cases metastatic lesions actually increase in size on computed tomography (CT) or magnetic resonance imaging (MRI) scans before they regress. This may be due to increased immune cell infiltration and a consequent inflammatory response, so may even indicate that the patient is responding. Thus, it may be necessary to re‐evaluate these approaches and the criteria we used to determine response, because the conventional measures of time‐to‐progression or Response Evaluation Criteria in Solid Tumours (RECIST) objectives may be inadequate (Hoos et al., 2010; Wolchok et al., 2009).

Finally, toxicity can be challenging in some patients and may limit treatment. Toxicities range from the mild effects such as rash and fatigue, to the severe effects (grade 3/4) including colitis and diarrhoea. Learning how to manage these will be important if these approaches rolled out into general medical use. Clearly, challenges remain for immunotherapies but with the excitement surrounding these treatments, it seems very likely that these challenges will be met and that these therapies will become more widely available. However, we do need to proceed with care, because as recently observed by the Society for Immunotherapy of Cancer, immunotherapies appear to have unique mechanisms of action and toxicity profiles that will require careful patient selection and clinical management (Kaufman et al., 2013).

10. Combination therapies

Over the last five years, we have witnessed remarkable advances in the development of both targeted and immuno‐therapies for melanoma patients. These approaches have changed the paradigm for treatment of melanoma patients, but not all patients responds and not all patients achieve durable responses. It is therefore natural that combination therapies are being tested in attempts to increase durability of response for individual patients. Many combinations are being tested (1, 2) and here we describe some that serve as exemplars of the potential of drug combinations that may be effective in this patient group.

10.1. BRAF/MEK inhibitor combinations

The availability of both BRAF and MEK inhibitors provided an unprecedented opportunity to test whether inhibiting two nodes in the same pathway would be more effective than inhibiting either node alone. In a phase I/II clinical trial of dabrafenib plus trametinib, 16 of 26 ^V600E/KBRAF patients who had previously relapsed on BRAF or MEK inhibitors alone displayed some level of tumour shrinkage, a confirmed response rate of 19%, and median progression‐free survival of 3.6 months (Smalley et al., 2012). In an open‐label study of 247 ^V600E/KBRAF metastatic melanoma patients, progression‐free survival increased from 5.8 months for dabrafenib alone to over 9.2 months for the combination (Flaherty et al., 2012). Although the phase III data only produced an increase in progression free survival of 8.8 for the BRAF therapy alone to 9.3 months (Long et al., 2014) for the combination in both trials there was a noticeable reduction in adverse events in the combination group with a particularly noteworthy reduction in the incidence of cuSCC from 19% to 7%. This is presumably because the MEK inhibitor blocks paradoxical activation of the pathway driven by the BRAF inhibitor.

We posit that the combination reduces toxicity because the BRAF inhibitor side effects are driven by paradoxical hyper‐activation of the pathway in normal cells, whereas the MEK inhibitor side‐effects are driven by inhibition of the pathway in normal cells. Thus, these agents work against each other in normal cells reducing the side effects, but in the cancer cells they work together to suppress signalling all the harder and improve response. It is curious therefore that the combination induces more pyrexia than either drug alone, suggesting that some of the side‐effects may not be driven by the drugs working against each other in all cells. The problem of pyrexia aside, these clinical data demonstrate that the combination not only improved responses, but also generally reduces toxicity and based on these unprecedented results, the FDA approved the combination for advanced BRAF mutant melanoma patients in January 2014. However, despite these improvements, it is noticeable that most patients relapse even on the combination and the mechanisms of resistance appear to be similar (the same?) to those that mediate resistance to BRAF inhibitors alone (Shi et al., 2014a).

10.2. MEK and CDK4/6 inhibitor combinations

While the majority of the studies with BRAF and MEK inhibitors have focussed on BRAF mutant tumours, recent data has shown that a MEK and CDK4/6 inhibitor combination is effective in NRAS mutant melanoma. CDK4 (cyclin dependent kinase 4) and CDK6 are cell cycle kinases that phosphorylate and inhibit the tumour suppressor RB, allowing the cell cycle to progress. In NRAS mutant melanoma, MEK inhibition induces low levels of apoptosis, but does not induce cell cycle arrest, and this is insufficient to block tumor growth (Kwong et al., 2012). However, using a systems biology approach it was shown that in combination with a CDK4/6 inhibitor, an MEK inhibitor could induce tumour regression in pre‐clinical models of NRAS mutant melanoma (Kwong et al., 2012). These data were the basis for a phase I clinical trial to combine the MEK inhibitor binimetinib and the CDK4/6 inhibitor LEE011 in advanced melanoma, with extremely promising preliminary results showing clinical benefit (complete response + partial response + stable disease) in 86% of patients (Sosman et al., 2014). We eagerly await the mature data from this trial.

10.3. Combination immunotherapies

As mentioned above, the results reporting that MK‐3475 (anti‐PD‐1) was effective in patients who had relapsed on ipilimumab (anti‐CTLA‐4) (Hamid et al., 2013a) demonstrated that progression on anti–CTLA‐4 therapy does not preclude treatment with anti–PD‐1 therapy and therefore that sequencing or even combining different immune checkpoints agents could be beneficial. Indeed, a phase 1 clinical trial of combination nivolumab (anti‐PD‐1) with ipilimumab has shown improved patient outcomes without an escalation of toxic effects. What is exciting about this trial is that at the maximum tolerated dose, 53% of patients with advanced, treatment‐resistant melanoma achieved objective responses, with tumour regression of at least 80% in every patient who presented a response (Wolchok et al., 2013).

These unprecedented results encourage scientists and clinicians alike to consider which combination of checkpoint blockers can be combined for the best effect in individual patients to extend responses it the patients who do respond, and to achieve responses in the 50% or so of patients who do not. This again highlights the importance of predictive biomarkers for immunotherapies and of improving our understanding of which patients will respond and which will not. Clearly, improving our understanding of these issues in melanoma will impact our ability to use these agents effectively in other solid tumours, and the responses in melanoma do not appear to be determined by genotype, suggesting that these agents could have wide utility in many cancers.

10.4. Combination immunotherapy and targeted therapy

In addition to the combinations described above, there is a great deal of interest in combining targeted and immuno therapies. Of course, these drugs are quite different: targeted therapies block the biochemical pathways or mutant proteins that stimulate tumour cell growth and survival, while immunotherapies boost the ability of the immune system to recognise cancer cells. Also, whereas responses to targeted therapies usually occur within weeks, responses to immuno therapies are typically slower and can take by up to 6 months.

This makes it difficult to know how to sequence the drugs. Targeted therapies give a more immediate response, but once patients have relapsed they are likely to have insufficient time and their performance status is generally too poor for them to mount an effective immune response. Indeed, it has been shown that BRAF inhibitor resistant melanoma cells evade the immune system because they upregulate PD‐L1 (Jiang et al., 2013), and there is evidence that patients who fail BRAF treatment also fail to respond to ipilimumab (Ackerman et al., 2012; Ascierto et al., 2014). This argues that immunotherapies should be administered first.

However, not all patients respond to immunotherapies and we are as yet unable to predict which ones will respond. Furthermore, BRAF inhibitors can drive paradoxical activation of ERK in T cells, potentially increasing their activity against tumour cells (Callahan et al., 2014). They also decrease expression of CCL2 and increase infiltration of CD8+ T cells and NK cells into tumours (Knight et al., 2013). These data suggest that BRAF inhibitors could enhance T cell responses, suggesting BRAF inhibitors should be administered before immunotherapy to increase anti‐tumour responses. Moreover, it has been shown that treatment with vemurafenib after treatment with anti‐PD‐1 can cause severe hypersensitivity, multiorgan injury, acute inflammatory demyelinating polyneuropathy and anaphylaxis in some patients (Johnson et al., 2013), again suggesting that BRAF inhibition should proceed immunotherapy. Thus, for some patients with genetically defined tumours for which a therapy with a high likelihood of response exists, it may be preferable to treat with the targeted therapy before the immunotherapy.

In addition to the uncertainties above, it should be remembered that these two classes of drug have quite distinct chemical and pharmacokinetic properties, making it difficult to predict how they will interact in patients. There is also a lack of suitable models in which to rapidly test specific combinations. Targeted therapies are generally tested against human xenografts in immuno‐compromised mice that are unsuitable for testing immunotherapies and immunotherapies are generally tested in mouse models that do not reflect human disease, or only represent a small subset of tumours. Thus, in large part, targeted therapy/immunotherapy combinations are currently tested in human trials with little or no pre‐clinical guidance. The problems inherent in this approach are highlighted by clinical results reporting that concurrent vemurafenib and ipilimumab cause severe hepatotoxicity (Ribas et al., 2013), and concurrent dabrafenib, trametinib and ipilimumab causes severe toxicity, including colitis and intestinal perforation (Puzanov et al., 2014). In contrast, concurrent dabrafenib and ipilimumab are well tolerated (Puzanov et al., 2014), as are concurrent administration of MPDL3280A (anti‐PD‐L1) and vemurafenib (Hamid et al., 2013b).

These latter data provide the hope that combinations of targeted and immuno‐therapy can work, but they reinforce the need for pre‐clinical models in which to test combinations, even with approved drugs.

11. Concluding remarks

Recent insights into the genetics, biology and immunology of melanoma have initiated a new era of rapidly evolving targeted and immune‐based treatments for melanoma. After decades of unsuccessful attempts to find more effective treatments, we now have several active drugs. Clearly, there are challenges to determine how to schedule, sequence and combine these drugs to achieve the best outcome for individual patients. There is also an urgent need for models to test combinations and for the development of predictive biomarkers, but the possibilities that these drugs provide are extraordinary. It is clear that new concepts will emerge from on‐going basic, translational and clinical studies, but the belief that melanoma is an essentially untreatable disease has been dispelled and the promise of further improvements in patient outcomes will surely be realised.

Financial support

Cancer Research UK (15759/A12328), The Wellcome Trust (100282/Z/12/Z).

Conflict of interest

Authors declare no conflict of interest.

Girotti Maria Romina, Saturno Grazia, Lorigan Paul, Marais Richard, (2014), No longer an untreatable disease: How targeted and immunotherapies have changed the management of melanoma patients, Molecular Oncology, 8, doi: 10.1016/j.molonc.2014.07.027.