A clinical overview of current and emerging therapies for metastatic melanoma
Melanoma is a highly aggressive skin tumour originating from a special type of skin cells called melanocytes (produces melanin, the pigment that gives skin colour). Metastatic melanoma, when the disease spreads to other parts of the body, is one of the most challenging cancers to treat. Indeed, the 10-year survival rate of patients with metastatic melanoma is less than 10 per cent.
Besides surgery, radiotherapy and chemotherapy, new and advanced therapeutic options for malignant melanoma can be divided into two major categories: a) targeted therapy, which generally targets only the cancer cells that contain mutations in a particular gene causing the disease; and b) immunotherapy, which targets and stimulates a person’s own immune system to recognise and destroy cancer cells more efficiently.
A number of frequent driver mutations (ie, genetic errors that control the disease) have been identified in melanoma. These mutations include BRAF, NRAS, KIT, GNAQ, GNA11, NF1 and TERT. Activating mutations in BRAF are found in 50-to-60 per cent of melanomas, with the V600E mutation accounting for more than 80 per cent of the cases with mutated BRAF. Of note, a comprehensive study of six clinical sites from around Ireland totalling 689 patients performed by our laboratory showed that the rate of V600E mutation is significantly lower in the Irish population (circa 25 per cent), compared to the international norm (van den Hurk, Melanoma Research, 2015).
The abnormal proteins created by these mutated genes essentially become stuck in the ‘on’ position in the transcription process, leading to uncontrolled cellular proliferation and impaired cell death. Since mutations in either BRAF or NRAS are signalled by MEK1/2 proteins, MEK inhibition is also emerging as an attractive therapeutic strategy.
Targeted therapy in the form of signal transduction inhibitors, notably BRAF and MEK inhibitors, has proven to be effective in clinical trials. Vemurafenib, a small-molecule BRAF inhibitor, received FDA approval in 2011 as an oral drug — a ground-breaking event at the time.
Vemurafenib binds to and deactivates the defective protein that results from the V600E mutation in BRAF. Vemurafenib induces complete or partial tumour regression in 81 per cent of patients, and additionally provides a significant advantage in progression-free survival (PFS) and overall survival (OS) to unresectable, advanced patients. An alternative small-molecule BRAF inhibitor is dabrafenib, which is administered orally to patients with advanced metastatic melanoma. The drug was FDA-approved in 2013 after showing improved PFS compared to previously-used chemotherapy drug DTIC (dacarbazine).
Sorafenib was the first BRAF inhibitor drug that was clinically developed in patients with metastatic melanoma. Unfortunately, in addition to BRAF, this drug non-selectively inhibited other protein kinases as well as CRAF, VEGF (vascular endothelial growth factor) and PDGF. Therefore, the clinical utility of sorafenib in melanoma has been proven to be very limited.
There are several MEK inhibitor drugs that have been tested in clinical trials for patients with metastatic or unresectable melanoma, including selumetinib (AZD6244), PD-0325901, trametinib (GSK1120212), AS703026, cobimetinib (GDC-0973/XL518) and MEK162.
Of these, trametinib, which inhibits MEK1/2 and is given orally to patients with metastatic melanoma, has been shown to be best at demonstrating improved PFS over DTIC. The drug was FDA-approved in 2013. A small number of melanomas that start on mucosal membrane and acral skin (palms, soles, nail bed) have mutations in the C-KIT gene. Clinical trials are now testing drugs that are known to target cells with changes in C-KIT, such as imatinib (Gleevec), dasatinib (Sprycel) and nilotinib (Tasigna). Of these, imatinib has shown the most promise so far.
While the above-mentioned targeted agents have been heralded as providing a revolution in terms of therapeutic options, many patients show disease recurrence after initial treatment response as a result of drug resistance. Newer studies focus on testing the effect of these promising drugs in combination.
Following improved response over single-drug use, the combination of BRAF and MEK inhibitor therapy (dabrafenib and trametinib) for patients with metastatic melanoma was FDA-approved in 2014. The response rate was 76 per cent; however, a considerable number of patients presented with side-effects. Other signal transduction inhibitors such as PI3K, AKT and CDK inhibitors are being clinically evaluated in combination with BRAF and MEK inhibitors. These approaches demonstrate that inhibiting multiple targets in either the same or separate pathways may be clinically beneficial to patients.
Unfortunately, the options for patients whose tumour does not contain a BRAF mutation are limited. A recent clinical trial found axitinib treatment of advanced BRAF wild-type metastatic melanoma, followed by chemotherapy with paclitaxel/carboplatin, prolonged disease control and survival with acceptable toxicity. Axitinib is a small-molecule, multi-tyrosine kinase inhibitor that targets VEGF signalling. The phosphatidylinositol 3-kinase (PI3K) signalling pathway has emerged as an interesting drug target for the last few years. Clinical evaluations are ongoing for several inhibitors of PI3K (pan-isoform and isoform-specific), dual PI3K/mTOR, AKT, and mTOR (mTORC1 and dual mTORC1/2 inhibitors). CDK inhibitors are another class of drugs that have been evaluated for the treatment of advanced melanoma for the past few years. CDK inhibitors such as palbociclib, dinaciclib and ribociclib have been tested in several early phases of clinical trials and show promising results, either alone or in combination with other inhibitors.
Types of immunotherapy
In 2011, ipilimumab, the first drug belonging to the class of checkpoint inhibitors, was approved for the treatment of advanced melanoma, and immunotherapy exploded as the new weapons against cancer.
Different treatments are grouped under the type of immunotherapy approach: Checkpoint inhibitors, vaccines, adoptive cell therapy and oncolytic viruses.
Checkpoint inhibitors are drugs targeting the so-called immune checkpoints, molecules and pathways normally meant to prevent hyperactivation of the immune system. The main pathways targeted by FDA-approved drugs are the cytotoxic T-lymphocyte antigen 4 (CTLA-4) pathway and programmed death-1 (PD-1) pathways. CTLA-4 is a receptor expressed by activated cytotoxic T cells and regulatory T cells that interacts with B7 ligands competing with CD28. CD28-B7 interaction is needed to complete T-cell activation started by T-cell receptor (TCR) antigen recognition. After trials that showed how ipilimumab could achieve a plateau in survival after three years, with a durable response being kept for more than 10 years, it was approved by the FDA as a first-line treatment for locally-advanced melanoma. Blocking antibodies against the PD-1 pathway include pembrolizumab and nivolumab, directed toward PD-1 itself, and atezolizumab, avelumab, and durvalumab, directed against its ligand PD-L1. PD-1 is a receptor expressed on activated T, NK and B cells, as well as myeloid cells, which interacts with its ligands PD-L1 and PD-L2 to limit T-cell activation. This can ultimately lead to T-cell apoptosis, anergy or exhaustion, defined as a dysfunctional state of T-cells that impairs their ability to kill melanoma cells. Its ligand PD-L1 is expressed by antigen-presenting cells but can be expressed in tumours as a mechanism of escape from the immune response through INF gamma induction or intrinsic oncogenic pathways, eg, PTEN loss. Blocking this pathway could rescue exhausted T-cells and avoid their inactivation by PD-L1-expressing tumours. In consecutive clinical trials (Checkmate-003, Checkmate-066 and Checkmate-037 for nivolumab; Keynote-001, Keynote-002 and Keynote-006 for pembrolizumab), PD-1 inhibitors have demonstrated higher response rates than ipilimumab, with the advantage that toxicities due to these drugs occur less frequently than the ones induced by ipilimumab. This is due to the fact that CTLA-4 is an early checkpoint, while PD1/PD-L1 acts later and is more specific to the anti-tumour immune response. Blocking antibodies against PD-L1 are currently under investigation, with preliminary data demonstrating durable responses, as well as affordable toxicities.
Cancer vaccines are useful to rescue the lack of infiltration in so-called ‘cold tumours’ that are without inflammatory infiltration. One of the reasons for the lack of inflammatory infiltration is a low number of mutations (‘mutational burden’).
Among different types of tumours, melanoma is generally the one with the highest mutational burden, and this partly explains why immunotherapies work so well in this kind of neoplasia. The higher the number of mutations, the higher is the possibility to produce neoantigens, proteins with modifications because of the mutations, tumour-specific and therefore recognisable by the immune system — albeit there are variations in mutational burden from patient-to-patient, and a low neoantigen production can be at the origin of a part of cold tumours. Vaccination towards neoplastic antigens stimulates the broadening of anti-tumour T-cell repertoire, especially in those neoplasia with a low mutational burden. Usually, neoplastic antigens are injected in association with immune adjuvants that favour the onset of a strong immune response. Another recent addition to immunotherapies is oncolytic virus therapy, recently approved for unresectable melanoma. The therapeutic protocol (which is quite complex and extremely expensive) involves the use of a herpes simplex virus type 1, talimogene laherparepvec (T-VEC), that replicates inside the tumour, generating a reaction similar to the one induced by cancer vaccination. These two therapies, in fact, have the common property to elicit an abscopal effect, that is, to induce an effect on lesions that are in different parts of the body, far from the site of injection, through the induction of a systemic immune response.
Lastly, adoptive cell transfer is an immunotherapy based on the classical observation that the number of immune cells generally correlates with the survival of patients. Therefore, immune cells are injected into the patient in order to increase their number. Dendritic cells and T-cells are specifically the most frequent type of immune cells used for this purpose. T-cells harvested from peripheral blood can be also modified before re-infusion by cloning the TCR of cancer-reacting T-cells or designing chimeric antigen receptors (CARs) directed towards tumour antigens. The technique itself requires high levels of technology and re-infusion after isolation; modification and expansion is successful in around half of the patients. Moreover, response rates are around 50 per cent in melanoma for classical adoptive cell therapy that goes up to 90 per cent for adoptive CAR-T cell therapy.
In spite of a significant improvement in survival, immunotherapy, but in particular checkpoint therapy, has presented three main challenges: Low response rates; the definition of ‘response’ itself; and immune-related toxicities.
Need for combination therapies
In clinical trials, only approximately 20 per cent of patients show a response to ipilimumab, while response rates to nivolumab and pembrolizumab are around 30 per cent. On the other side, checkpoint inhibition brings relevant toxicities related to a hyper-reactive immune system. Ipilimumab gives immunological adverse reactions in 70 per cent of the patients, with 25 per cent of these being of severe grade.
These complications are rash/pruritus, GI diarrhoea, colitis, hepatitis, hypophysitis and, more rarely, vitiligo, alopecia, and Stevens-Johnson syndrome. Nivolumab and pembrolizumab have similar toxicities, with the addition of pneumonitis, an adverse reaction peculiar to PD1/PD-L1 inhibition.
Clinical trials with different doses and schemes of administration of these therapies, in order to reduce the intensity of these toxicities, are currently ongoing. Checkmate-004 phase 1, Checkmate-069 randomised phase 2 and Checkmate-067 randomised phase 3 studies evaluated the efficacy of the combination between ipilimumab and nivolumab, registering response rates of 55, 61 and 57 per cent, respectively, but burdened with the price of around 50 per cent of grade 3-4 immune-related side-effects. A way to avoid these collateral effects is to adopt a scheme with consecutive instead of concurrent administration of the two drugs; that has achieved higher response rates (40 per cent) for nivolumab, followed by ipilimumab compared to ipilimumab, followed by nivolumab (20 per cent), without increasing significantly the incidence of high-grade toxicities compared to concurrent therapy.
Some ongoing studies are evaluating combination therapies, not only among checkpoint inhibitors, but also with other immunotherapies, with targeted therapies and with classical chemo and radiotherapy. The main strategy of combination therapy is to associate with other drugs the inhibition of PD-1/PD-L1 axis, due to its great efficacy in monotherapy, together with the lower levels of toxicity. The ‘partner drug’ to couple should depend on the characteristics of the tumour of the patient.
Need for biomarkers predictive of response and resistance
Regarding the clinical course during immunotherapy, the main difference with targeted therapies is that the response occurs later (up to 30 months) because it is a multi-step process involving different cell types; however, responses (when they occur) tend to be durable, due to the ability of the immune system to keep an immunological memory. Four patterns of response have been defined specific to immunotherapy: Immediate response without new lesions; durable, stable disease; flare effect (response after increase of the pre-existing lesions); and response with the appearance of new lesions. Moreover, a third of the patients with partial response will eventually progress, due to the development of resistance to the therapy. The concept of resistance to immunotherapy is difficult to define because of its multifactorial nature. The immune response is a dynamic process, influenced not only by the patient’s characteristics, but also by the treatments received. Therefore, the distinction that has been made between primary resistance (immunotherapy does not work from the beginning), adaptive resistance (the immune system recognises the tumour under immunotherapy but the tumour develops escape mechanisms) and acquired resistance (relapse and progression after initial response) is quite scholastic, because these types share the same mechanisms. These mechanisms can be intrinsic to the melanoma cell, such as activation of the mitogen-activated protein kinase signalling with or without loss of PTEN, expression of the WNT/β-catenin pathway, loss of interferon-gamma signalling pathways, loss of the ability to express tumour antigens, and expression of the IPRES (innate anti-PD-1 resistance) signature. Alternatively, they can be extrinsic from melanoma cells and depend on the immunosuppressive components of the tumour inflammatory microenvironment, such as regulatory T-cells, MDSCs and macrophages with an M2-like polarisation.
For all of the reasons listed above, there is the urgent need to find predictors of response and resistance to immunotherapies. Predictors should be used not only before treatment, in order to personalise the therapy, helping the clinician to choose the better combination of drugs for the patient, but also during the therapy, in order to diagnose the development of resistance at an early stage and shift to another treatment as soon as possible. Monitoring response to immunotherapies is challenging due to the slow evolution of the response.
To early-detect response and resistance, taking into account the dynamic nature of the tumour microenvironment, would require repeatedly sampling the tumour and the peripheral blood during the course of treatment. This approach has been followed in research settings to determine the efficacy of putative predictive biomarkers, comparing them at different moments of the therapeutic protocol (before administration, early therapy, response, progression). The most important conclusion drawn at the moment is that due to the complex landscape of the immune microenvironment, prediction shall be based on a panel of different biomarkers to take into account inherent complexity.
Currently, there are no validated predictive panels, but different parameters have been investigated, alone and in combination. The composition of the immune infiltrate in the tumour must certainly be taken into account, since different types of immune cells have different impacts on tumour progression, as well as the cytokine environment, the expression of immune checkpoint molecules and the presence of enzymes, such as IDO-1. Better responses appear to happen in tumours with high infiltration by CD8+ cytotoxic T lymphocytes next to PD-L1-positive cells. Nevertheless, PD-L1 expression turned out to be an inefficient predictor of response to nivolumab, since responses have been reported also in PD-L1-negative tumours. Higher mutational burden, neoantigen load and T-cell clonality have been correlated with response to PD-1 blockade, but no cut-off was identified that could be used in the clinics to decide whether or not to administer the treatment.
Intriguingly, the gut microbiome of the patient could be a potential biomarker — in particular, Bifidobacterium with PD-1 blockade and Bacteroides with CTLA-4 blockade have shown a synergistic effect in favouring tumour control. The best approach that integrated all of these different putative markers is represented by the cancer immunogram, which at the same time takes into account tumour sensitivity to immune effectors; tumoral burden and neoantigen production; general immune status of the patient; intensity of immune cell infiltration in the tumour; evidence of checkpoint molecule expression; evidence of soluble inhibitors; and evidence of inhibitory tumour metabolism.
At our centre at UCD, we are involved in several EU-funded research programmes focused on melanoma. One of these programmes, SYS-MEL (www.sysmel.com), is investigating several key molecular pathways that drive melanoma progression, with a focus on epigenetically-regulated targets, the apoptosis cascade, and kinase signalling. In this programme, we have validated several protein biomarkers on tissue microarrays (TMAs) and the resulting quantified protein expression data will be used in a proprietary systems modelling-based approach to aid the clinical decision-making for the treatment of melanoma patients.
Another programme is the MEL-PLEX consortium (‘Exploiting MELanoma disease comPLEXity to address European research training needs in translational cancer systems biology and cancer systems medicine’), a Marie Sklodowska Curie action to train 15 early-stage researchers in a network comprised of prominent European universities and private companies.
It aims to develop and validate predictive models for disease progression, prognosis and responsiveness to current and novel (co-) treatment options, and to provide superior and clinically-relevant tools and biomarker signatures for personalising and optimising melanoma therapy (http://melplex.eu/).