Therapy for metastatic melanoma: the past, present, and future

In 2002, it was discovered that cutaneous melanoma is a molecularly heterogeneous disease with approximately 40% to 60% harboring an activating mutation in the gene encoding for the serine/threonine kinase protein kinase B-raf (BRAF) with 90% of the mutations resulting in a substitution of valine for glutamate at amino acid 600 (V600E) [5‐8]. Mutated BRAF leads to constitutive activation of the mitogen-activated protein kinase pathway (MAPK) that in turn increases cellular proliferation and drives oncogenic activity. Given the relatively high incidence of mutant BRAF as well as its oncogenic potential, investigators have long sought to selectively inhibit mutated BRAF. Earlier attempts to inhibit BRAF in patients with melanoma with sorafenib were largely unsuccessful secondary to the poor sensitivity of sorafenib to selectively target mutant BRAF that led to intolerable off-target side effects through inhibition of wild-type BRAF and other off-target effects [9‐13]. Recently, highly selective BRAF inhibitors capable of silencing mutant BRAF (V600E) with little effect on wild-type BRAF have emerged (Table 1). In a phase 1 study, the first of these selective BRAF inhibitors, vemurafenib, demonstrated substantial tumor regression in 81% of patients with metastatic melanoma who had a BRAF (V600E) mutation and received the recommended phase 2 dose [13, 14]. The follow up phase 2 (BRIM2) study of previously treated patients demonstrated a confirmed response rate (RR) of 53% with a 6.8 month median duration of response [15]. Finally, a phase 3 randomized control trial (BRIM3) of previously untreated patients compared vemurafenib to dacarbazine demonstrating improvements in RR (48% versus 5%), progression free survival (5.3 versus 1.6 months), percent of patients alive at six months (84% versus 64%) with a 75% reduction in risk of death [16]. A second BRAF inhibitor, GSK2118436, showed similar efficacy in a phase 1/2 study although OS data are not yet mature [17]. In addition, 10% to 30% of patients with a BRAF mutation have a non-V600E mutation with the most common non-V600E mutation being V600K which is present in 5% to 20% of melanoma patients with a BRAF mutation [7, 18]. Both vemurafenib and GSK2118436 have shown activity in V600K mutant melanomas and while vemurafenib is not currently approved for patients with V600K mutations, further studies are examining its efficacy in non-V600E mutant patients [16, 19]. Finally, both vemurafenib and GSK2118436 have been tested in patients with brain metastasis with apparent activity in the brain although the number of treated patients remains small at present [19, 20]. In summary, both highly selective BRAF inhibitors, vemurafenib and GSK2118436, have demonstrated excellent clinical activity with a high response rate and low toxicity in patients with BRAF V600E mutations but, unfortunately, both therapies are limited by a relatively short duration that averages around six months. The most important toxicity related to BRAF inhibitors is accelerated growth of cutaneous squamous cell carcinomas (SCC) and keratoacanthomas through paradoxical activation of MAPK signaling occurring in approximately 20% of patients [21]. Fortunately, these SCC are easily removed and cured through local excision.

Melanoma is characterized as one of the most immunogenic tumors due to the presence of tumor infiltrating lymphocytes in resected melanoma, occasional spontaneous regressions, and clinical responses to immune stimulation. The immunogenicity of melanoma has led investigators to study novel immune strategies to overcome tumor immune evasion. One mechanism by which T cells self-regulate their activation is through expression of cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4). CTLA-4 functions as a negative co-stimulatory molecule for the T cell, and therapies that antagonize CTLA-4 remove the brakes from the T cell leading to a net effect of T cell hyper-responsiveness [22]. Ipilimumab is a fully human IgG1 monoclonal antibody that blocks CTLA-4, thereby increasing T-cell activity and promoting antitumor activity [23]. Two phase 3 randomized clinical trials have evaluated ipilimumab in metastatic melanoma [23, 24]. In the first trial of patients with previously treated unresectable stage III or IV melanoma, ipilimumab demonstrated an improved overall survival versus glycoprotein 100 peptide vaccine (gp100) (10.1 versus 6.4 months) [24]. In the second phase 3 trial in previously untreated patients with metastatic melanoma, ipilimumab plus dacarbazine demonstrated improvement in OS versus single agent dacarbazine (11.2 versus 9.2 months) [23]. In both phase 3 studies, the response rate, complete response (CR) and partial response (PR) was only 10% to 15% and the disease control rate (CR, PR, and stable disease (SD)) was approximately 30%. In addition, the improvement in percent of patients alive at one and two years is consistently 10% better than the non-ipilimumab containing arms (Table 2). While the response rate and improvement in OS in ipilimumab is relatively modest, the toxicities of the therapy, including immune-related enterocolitis, hepatitis, and dermatitis, are highly manageable [24].

The recent success of both vemurafenib and ipilimumab instilled hope into physicians and patients with metastatic melanoma; however, the limitations of both therapies emphasize the importance of developing novel treatment strategies. One of the major challenges in overcoming the limitations of these novel therapies is both increasing the duration of response to BRAF inhibitors and improving the response rate of ipilimumab. We will discuss multiple strategies to address the limitations of vemurafenib below.

c-kit, also known as CD117, is a receptor tyrosine kinase that is mutated in approximately 20% of acral, mucosal, and chronically sun-damaged skin [45]. The ligand for KIT is stem cell factor (SCF) and binding of SCF to c-kit induces activation of downstream signaling pathways that are involved in mediating growth and survival signals within the cell including the P13K-AKT-mTOR pathway and the RAS-RAF-MEK-ERK pathway. KIT has been implicated in the pathogenesis of several cancers including acute myeloid leukemia and gastrointestinal stromal tumors (GIST) [46‐49]. Unlike in GIST where c-kit mutations tend to be deletions or insertions in exon 11, c-kit mutations in melanoma occur at multiple sites along the gene including both the juxta-membrane domain at exon 11 and exon 13 and the kinase domain at exon 17 and are usually point mutations that do not correlate with KIT copy number or CD117 expression [50, 51].

Imatinib, an oral tyrosine kinase inhibitor with known activity against c-kit activated tumors, was tested in three phase 2 studies of patients with melanomas that harbor c-kit mutations [52‐54]. The first study enrolled patients with metastatic melanoma who expressed at least one protein tyrosine kinase (c-kit, platelet derived growth factor receptors, c-abl, or abl related gene) demonstrating a response in only one patient, interestingly in the patient who had the highest level of c-kit expression [52]. Of note, mutations in c-kit were not required prior to entry in this study. In the second study, 28 patients with c-kit mutations and amplifications with advanced unresectable melanoma arising from acral, mucosal, and chronically sun-damaged sites were treated with imatinib mesylate 400 mg orally twice daily in six-week cycles until disease progression or unacceptable toxicity [54]. Durable overall response rate (ORR) was 16% with a median time to progression of 12 weeks and median OS of 46.3 weeks. The third study enrolled 43 patients with c-kit mutated metastatic melanoma treated with imatinib who demonstrated an ORR of 23.3% and median OS of 14 months, with 51% alive at one year [53]. Please see Table 3 for a summary of these three trials. In summary, the responses to imatinib were primarily demonstrated in patients harboring mutations in c-kit at exon 11 or exon 13. Activity of dasatinib in c-kit mutated melanoma has already been demonstrated and additional trials with imatinib, nilotinib, and dasatinib are currently ongoing [43, 50, 55]. As is the case with the BRAF inhibitors, combinations of c-kit inhibitors with cyotoxic agents, immunotherapies, and other targeted therapies are underway.

One mechanism by which melanoma is thought to evade the immune system is through tumor expression of programmed death ligand 1 (PD-L1). PD-L1 is a negative regulator of the immune system that acts through binding of the PD-1 present on activated lymphocytes and PD-L1/PD-1 interaction causes immune tolerance through apoptosis of the activated lymphocyte [66‐68]. MDX-1106 is a genetically engineered fully human immunoglobulin G4 monoclonal antibody specific for human PD-1 [69]. A phase 1 study of anti-PD-1 antibody, MDX-1106, demonstrated single agent responses in a variety of previously treated, refractory solid tumors including melanoma with few treatment-related immune toxicities [69]. The clinical experience with anti-PD1 treatment is limited but encouraging and a dose escalation study of the combination of MDX-1106 and ipilimumab is currently recruiting [43]. In addition to antibodies that target PD-1, phase 1 trials of anti-PD-L1 are also underway.

The presence of tumor infiltrating lymphocytes (TIL) in resected melanoma samples is one of the reasons melanoma is often characterized as an immunogenic malignancy. Attempts to isolate, expand and infuse TIL for the treatment of cancer is termed adoptive cell therapy (ACT) and has shown promise for the treatment of patients with metastatic melanoma [70‐73]. Early studies using TIL and IL-2 produced a response rate of 34%. Later studies using nonmyeloblative lymphodepletion with cytotoxic chemotherapy, with or without total body irradiation, induced response rates as high as 72% [70‐75]. One of the limitations of ACT is that TIL expansion is not possible for all patients although it was recently reported that TIL is successfully generated in > 60% of all patients with melanoma and this figure is even higher in patients who did not receive prior chemotherapy [76]. In order to circumvent the process of tumor resection and isolation of TIL, investigators have developed methods to genetically modify autologous peripheral T cells to express a T cell receptor that targets melanoma antigens. This approach has induced a response rate in a subset of patients with melanoma and a similar approach has also yielded results in patients with refractory chronic lymphocytic leukemia [77]. In summary, ACT for the treatment of melanoma is a promising, albeit resource consuming method for the treatment of metastatic melanoma.

With the approval of ipilimumab, the role of HD IL-2 remains in question. While the clinical benefit of HD IL-2 is roughly similar to ipilimumab, the toxicity is worse and tolerability is less. Multiple strategies have attempted to identify biomarkers to predict clinical benefit of HD IL-2. Recently, a retrospective study found that patients with an NRAS mutation have a higher likelihood to respond to HD IL-2 [78]. In addition, a prospective study of patients with metastatic melanoma and renal cell carcinoma found that elevated pre-treatment serum VEGF and fibronectin are inversely correlated with the response rate to HD-IL-2 [79]. Both of these findings require further study at additional centers before changing clinical practice. Multiple efforts to increase the response rate of HD IL-2 while maintaining the duration of response have been attempted. The addition of a peptide vaccine (gp-100) to HD IL-2 demonstrated an improved objective response rate (16% versus 6%, P = 0.03) and overall survival (17.8 versus 11.1 months, P = 0.06) [80]. There are plans to improve upon this finding by using more potent vaccines.

Finally, it is becoming more evident that patients who progress on HD IL-2 do derive clinical benefit from ipilimumab. In the phase 3 study of previously treated patients, patients who progressed on prior HD-IL-2 received a similar benefit to ipilimumab as those who did not [24]. Similarly, a retrospective study analyzing patients who progressed on HD IL-2 and then received ipilimumab had a response rate (19%) and OS (12 months) to ipilimumab that was similar to previously reported historical controls [81]. Given the lack of approved agents and the limited but quantifiable benefit, we believe there still remains a role for HD IL-2 in patients who are fit enough to receive it.