Systemic therapy for metastatic malignant melanoma – from deeply disappointing to bright future?

Authors


Paul Lorigan, CRUK Department of Medical Oncology, Christie Hospital NHS Trust, Wilmslow Road, Manchester M20 4BX, UK, Tel.: +44 161 446 3126, Fax: +44 446 3299, e-mail: paul.lorigan@manchester.ac.uk

Abstract

Abstract:  The last decade has seen a considerable improvement in the understanding of the biology of melanoma. Advances have come in the understanding of the importance of critical oncogenes and tumour suppressor genes, epigenetic phenomena, signalling pathways, drug resistance mechanisms, the pivotal role of the local immune system, and the importance of cell–cell and cell–matrix interactions. Many of these pathways and interactions include potentially ‘drugable’ targets. These developments have allowed the identification and/or design of a range of new, targeted therapies. Evaluation of these new drugs has brought a whole new series of challenges. These include indentification of appropriate pre-clinical models, overcoming the redundancy inbuilt in complex biological systems, identification of appropriate molecular and clinical endpoints to show that the drug is hitting the target, how to combine treatments, and new toxicities. For the first time, there is the possibility of personalised treatment for melanoma patients, based on individual host and tumour characteristics. This paper discusses the range of new drugs and targets have been identified, the outcome of clinical trials, and the directions for future advances.

Introduction

The incidence of malignant melanoma continues to rise. In the USA, the American Cancer Society predicts that there will have been 62 190 new patients and 7910 deaths in 2006 (1). The incidence in Germany (10–12/100 000) and the UK (13/100 000) is similar (2,3). With a few exceptions, there is no clear evidence from recent Cancer Registry data that sun avoidance campaigns have had a dramatic effect on incidence (4,5). However, there is evidence that the incidence of thick melanomas is static, with the rise in incidence of melanoma being largely due to an increase in thin melanomas, suggesting earlier detection (2,6). Mortality rates vary by country, gender and ethnic origin – in the UK in 2003 there were 8114 new cases and 1817 (22%) deaths (1,3). The median survival for patients with metastatic melanoma reported in recent phase III clinical trials is 7–9 months (7,8). The last decade has seen considerable effort to improve treatment for patients with advanced disease. We now understand better the mechanisms of chemotherapy resistance, appreciate the importance of signalling pathways, cell–cell and cell–matrix interactions, and the critical role of the immune system. This understanding has allowed development of rational treatment approaches, targeting these areas. But will these result in better outcomes? Is there a bright future for melanoma patients?

Standard chemotherapy

Single agent dacarbazine (DTIC) has been the standard of care for many years. Response rates of 7–13% have been reported in recent large phase III trials, with a further 15–28% having stable disease (7,8). However, few responses are longstanding. Whilst combination chemotherapy results in higher response rates, there is no clear evidence that this improves survival (9,10). High-dose interleukin-2 (IL-2) has a reported response rate of 16% in selected patients, with 6% of patients achieving a complete response and over half of these free of disease at 2 years (11). However, toxicity with high-dose IL-2 is considerable and has limited its use to highly selected patients being treated in specialist centres with experience in this area. Response rates for low-dose IL-2 are just 2–3% (12).

An overview presented at the American Society of Clinical Oncology Annual Meeting in 2003 took account of data from three randomized trials of combination therapy versus biochemotherapy presented at that meeting and concluded that biochemotherapy does not prolong survival, may improve time to progression and response rate, but is associated with more toxicity and expense and cannot yet be seen as a standard of care (13–15). A systematic review by Eigentler et al. (16) of 41 randomized clinical trials showed that whilst some regimens were associated with an increased response rate, none resulted in an improvement in overall survival. A recent meta-analysis of biochemotherapy included 18 randomized trials – 11 trials of chemotherapy ± interferon (IFN) and seven trials of chemotherapy ± IFN and IL-2. Over 2600 patients were entered into the trials, with 555 responses and 2039 deaths. There was a clear benefit for biochemotherapy for response (odds ratio 0.59, 0.49–0.72, P < 0.00001) for both the IFN (0.60, 0.46–0.79, P = 0.0002) and IFN + IL-2 (0.58, 0.44–0.77, P = 0.0001) subgroups. However, there was no benefit in overall survival (0.99, 0.91–1.08, P = 0.9) (17).

Newer chemotherapy

Fotemustine is a chloroethyl-nitrosurea which has shown activity against melanoma, with an indication of activity in patients with brain metastases (18). A randomized phase III study comparing fotemustine with DTIC in 229 patients showed a trend towards a survival advantage (median 7.3 vs 5.6 months, P = 0.067) and delayed time to developing brain metastases (22.7 vs 7.2 months, P = 0.059) (19).

Temozolomide is an imidazotetrazine with a mechanism of action similar to DTIC. Unlike DTIC which requires metabolic activation, temozolomide is spontaneously converted to its active intermediary. A large phase III study comparing three weekly DTIC 250 mg/m2 intravenously on days 1–5 versus four weekly oral temozolomide 200 mg/m2 on days 1–5 demonstrated equivalence for survival, response rate and toxicity, with a trend towards temozolomide being superior for progression-free survival (PFS) and certain quality of life domains (7). As a result of this equivalent outcome, although temozolomide did not receive a licence in advanced melanoma, there is extensive off license use for this indication. The optimum dose and schedule of temozolomide have not yet been defined. Extended schedules may reduce the effects of DNA repair mechanisms, especially those mediated by O6-methylguanine-DNA-methyltransferase (MGMT), and so maximize the effects of treatment (20). Chronic, low-dose administration (75 mg/m2 daily for 6 weeks, repeated every 8 weeks) is associated with significant activity and is well tolerated, but appears to cause selective CD4+ lymphopenia and an increased risk of opportunistic infection (21). A randomized phase III trial of extended schedule temozolomide 150 mg/m2 1 week on/off continuously, versus DTIC 1000 mg/m2 i.v. 3-weekly by the European Organization for Research and Treatment of Cancer (EORTC) and US investigators that completed accrual in March 2007 will give valuable information on the importance of scheduling in treatment with temozolomide.

Temozolomide may have better blood–brain barrier penetration than DTIC and can be combined safely with radiotherapy in patients with primary brain tumors (22). A recent phase II trial reported on 151 melanoma patients with brain metastases who did not require immediate radiotherapy and received temozolomide, reported a 7% response rate with 29% stable disease and a median survival of 3.5 months (23). A study by the EORTC of temozolomide + brain irradiation in patients with brain metastases from malignant melanoma was closed because of accrual problems. There were a number of cases of treatment related encephalitis in those patients receiving full-dose temozolomide and concurrent radiotherapy. Temozolomide is currently being evaluated in combination with a number of new drugs.

Novel treatment approaches

Advances will come with increase in understanding of the molecular events that control cell division, invasion and metastasis, evasion of immune surveillance and resistance to chemotherapy. A number of drugs targeting these areas are currently under evaluation (Table 1). Challenges include identifying targets that are critical to the survival of the malignant cell, developing strategies to overcome the redundancy built in to most biological systems, optimizing sequencing and combination with other treatments, and dealing with the new spectrum of toxicities associated with these new agents.

Table 1.   Targeted therapies under evaluation
DrugTarget
  1. VEGF, vascular endothelial growth factor; PDGF, platelet-derived growth factor; mAb, monoclonal antibody; MTOR, mammalian target of rapamycin; XIAP, X-linked inhibitor of apoptosis.

SorafenibRAF, VEGF-R, PDGR-R
BortezomibProteasome inhibitor
MS-275Histone deacetylase inhibition
ThalidomideAnti-angiogenic, immunomodulatory
Revemid (CC5013)Anti-angiogenic, immunomodulatory
Vitaxin (MEDI 522)αvβ3 integrin mAb
CNTO95αv integrin mAb
Oblimersenbcl-2 antisense
XIAPXIAP antisense
Semaxanib (SU5416)VEGF-R2, kit
BevacizumabVEGF-R mAb

Drug resistance modifiers

Temozolomide exerts its anticancer activity through the methylation of DNA at the O6 position of guanine residues. The DNA repair protein MGMT is known to be important in tumor resistance to temozolomide. O6-(4-bromothenyl)guanine (lomeguatrib) and O6-benzylguanine act as pseudosubstrates that exhibit a greater affinity for the active binding site of MGMT than O6-methylguanine (24,25). Pretreatment with either of these agents may reduce resistance to temozolomide. A randomized phase II study showed no benefit of combining lomeguatrib and temozolomide (26).

Poly(ADP-ribose) polymerase (PARP) is involved in base-excision repair of DNA damage and inhibition of PARP may also sensitize tumors to the effects of temozolomide and other alkylating agents. Preliminary results from a phase II study of AG014699 12 mg/m2 and full-dose temozolomide (200 mg/m2 5× daily 4-weekly) in patients with advanced malignant melanoma showed significant activity (response and stabilization) and significant myelosuppression (27). Other PARP inhibitors including an oral formulation, are currently in clinical trials.

Pro-apoptotic agents

A number of antiapoptotic proteins including bcl-2, Bcl-xL, and X-linked inhibitor of apoptosis (XIAP), are over expressed in melanoma and may confer resistance to chemotherapy (28,29). There is no good evidence that any one of these is a key, central control of apoptosis, so targeting a single pathway may not be effective. Oblimersen is an antisense agent targeted to mitochondrial bcl-2. Results from a randomized phase III trial comparing DTIC combined with oblimersen against DTIC alone in 771 patients showed improved PFS (2.6 months vs 1.6 months, P < 0.01) and response rate (13.5% vs 7.5%, P = 0.007) but no statistical difference in overall survival (9.0 months vs 7.8 months, P = 0.077) (8). Problems with study design and failure to measure tumor bcl-2 expression made these results difficult to interpret. A significant interaction between baseline serum LDH and treatment was noted, with oblimersen significantly increasing survival in patients with a normal LDH. Many ongoing phase III studies now exclude patients with elevated LDH.

The inhibitor of apoptosis (IAP) family includes eight proteins. One of these, XIAP, blocks both the endogenous (mitochondrial) and exogenous (death receptor-related) apoptosis pathways and an antisense oligonucleotide directed at XIAP is in clinical evaluation at present (30). A phase II trial of YM155 administered as a 7-day infusion showed some activity, and studies in combination with chemotherapy are ongoing (31).

Signalling pathways and signal transduction inhibitors

A number of signalling pathways are critical to the survival of the melanoma cell. Understanding of the complexities of these pathways has increased greatly over the last few years and is dealt with in a number of excellent reviews (32,33). Most important of these is the mitogen-activated protein kinase pathway (MAPK) i.e. Ras-Raf-Erk. MAPK is activated in nearly all melanomas and is involved in proliferation, invasion and survival. Consequently, the individual kinases are potential targets for therapy.

There are identified mutations in the RAS/RAF pathway in over 80% of cases of melanoma. The commonest of these somatic mutations is the V600E mutation in BRAF (34–36). Recently agents have become available that, at least in vitro, are effective inhibitors of BRAF. Sorafenib is an oral multi-targeted kinase inhibitor of serine/threonine and receptor tyrosine kinases. It has pronounced in vivo and in vitro effects on tumor cells and vasculature and inhibits several targets relevant to melanoma including vascular endothelial growth factor (VEGF) receptors, platelet-derived growth factor (PDGF) receptors and BRAF (37,38).

Early studies showed no evidence of single agent activity for sorafenib in melanoma (39). A multi-centre phase I open label dose escalation study of carboplatin AUC6 and paclitaxel 225 mg/m2 and sorafenib given on days 3–18 was extended to include more patients with melanoma in view of the extremely encouraging 85% disease control rate (40). The PFS of responders was estimated at 15.2 months and that of patients with stable disease at 6.7 months, with the median PFS being 8.8 months. There was also significant toxicity, predominantly bone marrow suppression with significant neutropenia (71%) and febrile neutropenia (24%). Importantly, there was no evidence that the BRAF status of the melanomas’ predicted for chance of benefit from treatment. The encouraging data prompted the ongoing ECOG 2603 study which randomizes 800 patients to carboplatin AUC6 and paclitaxel 225 mg/m2 with either placebo or sorafenib 400 mg twice daily. A very similar second-line phase III trial (PRISM) in 270 patients who had already had first-line treatment with standard chemotherapy reported no difference in the primary end-point of PFS, being 17.9 weeks for the sorafenib containing arm and 17.4 weeks for the placebo containing arm (HR 0.906, P = 0.492). It is interesting that the PFS was so good in both arms, significantly better than would be expected for patients receiving first-line therapy, indicating that patients fit enough to have second-line therapy for melanoma are a self-selecting group of better prognosis patients (41).

Whilst the initial carboplatin, paclitaxel and sorafenib data are encouraging, they are also consistent with considerable toxicity of treatment. An alternative agent with which to combine sorafenib is DTIC or its analogue temozolomide. There is a theoretical rationale for combining DTIC with sorafenib in that DTIC treatment of melanoma cells in vitro resulted in up regulation of VEGF expression (42). In vivo models indicated that these cells had increased tumor growth and metastatic potential (43). A phase I study combining sorafenib with DTIC 1000 mg/m2 showed that full doses of both drugs were easily achievable and indeed well tolerated (44). Initial data from two phase II studies have been presented. A single arm study in 83 patients showed disease stabilization in 51% and a median PFS of 14 weeks (95% CI 12, 19; 28% censored) (45). A randomized phase II study of DTIC ± sorafenib in 101 patients showed a significant improvement in PFS in favour of the sorafenib arm. The median PFS was 11.7 weeks (95% CI 6.1, 17.9) for the control arm and 21.1 weeks (95%CI 16, 28) for the sorafenib-treated patients. However, there were no differences in overall survival data observed (46) with a trend in the opposite direction (HR 1.154, P = 0.056).

One attractive approach would be a completely oral regimen. In a randomized phase II study of temozolomide and sorafenib in advanced melanoma patients received either extended dose temozolomide (75 mg/m2 orally days 1–6) or standard dose temozolomide days 1–5, in each case with sorafenib 400 mg b.i.d. Separate cohorts were available for patients who had either had brain metastases or prior temozolomide. These phase II data indicate activity of this regimen in those patients who have not previously been exposed to temozolomide. It is important to realize that whilst sorafenib may be an effective anti-angiogenic agent, the indications to date are that at the dosages used, we are unlikely to have inhibited the BRAF activity in melanoma significantly. Thus the important question of whether the RAS/RAF pathway forms an attractive therapeutic target in melanoma has not really been addressed and will be the subject of many future studies.

NRAS is mutated in approximately 15% of melanomas. The only clinical agents to date that affect RAS activity are the farnesyl transferase inhibitors. These agents impair the post-translational modification of the RAS protein and prevent their membrane localization, impairing their activity. Clinical trials in melanoma are ongoing. Activating mutations of BRAF allow the kinase to remain active in the face of upstream inhibition of RAS, further limiting the potential usefulness of RAS as a target in melanoma. MEK is the immediate downstream signalling kinase from RAF. A large, randomized, phase II study of temozolomide ± AZD6244, an oral MEK inhibitor, has completed accrual and the results are awaited.

A number of other pathways are interlinked with MAPK. Mammalian target of rapamycin (MTOR) is a downstream component of the PI3K/Akt pathway and acts through nuclear factor-κB (NF-κB). Temsirolimus (CCI-779) is an inhibitor of MTOR currently in clinical evaluation. A large phase III study in renal cell carcinoma comparing temsirolimus, temsirolimus + IFN, and IFN showed a survival advantage for single agent temsirolimus (47). However, despite evidence of activity in preclinical models, a phase II study in 33 melanoma patients showed no single agent activity (48).

Growth factor receptors

Melanoma expresses a number of growth factor receptors including EGFR, PDGFR and c-kit. Receptor expression has been shown to change with disease progression, and may be differentially expressed on different tumor subtypes (49). This study in 102 primary melanomas showed KIT abberation (mutation and/or increased copy number in 39% of mucosal, 36% of acral and 28% of melanomas on chronically sun damaged skin, but not in any melanomas in skin without chronic sun damage. The role of inhibitors of ckit in these subtypes is under evaluation (50). A phase II trial of erlotinib in 14 patients showed no responses (51).

Proteasome inhibitors

The proteasome is a multi-enzyme complex that serves as a major protein degradation pathway (52). The orderly degradation of cellular proteins is critical for the regulation of signal transduction, transcriptional regulation, response to stress and the control of receptor function. The proteasome controls the levels of proteins that are important for cell-cycle progression and apoptosis in normal and malignant cells including cyclins, caspases, bcl-2 and NF-κB (53). Deregulation of the ubiquitin-proteasome pathway may contribute to tumor progression, drug resistance and altered immune surveillance (54). Boronic acid-derived compounds inhibit the proteasome pathway in a highly specific manner (55). Bortezomib has shown activity in a number of different tumor types and is licensed for the treatment of multiple myeloma. However, a phase II study of bortezomib in melanoma patients was terminated due to the lack of clinical responses (56). A phase II trial of bortezomib in combination with carboplatin and paclitaxel as first-line treatment in patients with metastatic cutaneous and ocular melanomas is ongoing. Tanespimycin is an inhibitor of heat-shock protein 90 (HSP-90). HSP-90 inhibition results in degradation of a number of RAF family proteins, including mutant BRAF. Preliminary results of single agent treatment suggested some activity, but with 29% grade III non-haematological toxicity (57).

Inhibition of tumor suppressor genes

Cancer is associated with epigenetic as much as genetic alterations and the fate of an individual cell may depend on a delicate balance between gene expression and suppression. Epigenetic alterations in cancer cells affect virtually all cellular pathways that have been associated to tumorigenesis, and so drugs acting in this area can have a broad range of activities (58). Epigenetic gene silencing occurs mainly through two mechanisms, histone deacetylation by deacetylases (HDAC), and DNA hypermethylation (56,59–62). Although histone deacetylation has a fundamental role in regulating gene expression, HDAC inhibitors directly affect transcription of only a relatively small number of genes. The majority of these genes are directly or indirectly involved in the regulation of cell growth and survival, providing a mechanistic explanation of the anticancer properties of HDAC inhibitors. Inhibitors of HDAC induce cell cycle arrest, differentiation, and promote the apoptosis of melanoma cells in vitro, and many have potent antitumor activities in vivo. MS-275 is an orally active synthetic pyridyl carbamate HDAC inhibitor. A recent phase II study of single agent treatment randomized 28 patients showed no objective responses, but disease stabilization was seen in 20–30% of patients (63).

The silencing of gene expression through DNA methylation contributes to defects in antigen presentation and apoptosis in melanoma and renal cell cancer. Combination of a hypomethylating agent and immunotherapy may result in improved antigen presentation and so increase the activity. A phase I trial of 5-aza-2′-deoxycytidine (decitabine) plus high-dose IL-2 showed that decitabine augmented haemoglobin F levels and altered DNA methylation and gene expression in peripheral blood mononuclear cells in a dose-independent manner that overlapped with the administration of IL-2. Objective responses occurred in 31% of melanoma patients (64).

Other new drugs

STA 4783 is a novel inducer of HSP-70. A randomized double-blind phase II study of paclitaxel ± STA 4783 in 81 patients with stage IV melanoma reported a doubling of PFS in the combination treatment arm, with subgroup analysis suggesting the benefit was greatest in chemo-naive patients (65). A prospective randomized phase III study has been started.

Anti-angiogenic agents

Angiogenesis is necessary for the growth of primary tumors and metastases. Tumor cells express endothelial markers that do not respond to normal angiogenic control. They recruit other tumor cells by the production of growth factors into the stroma. Melanoma metastases tend to be very vascular and are a potential candidate for anti-angiogenic therapy.

Thalidomide has anti-angiogenic and immunomodulatory properties and has been used successfully in the treatment of Kaposi’s sarcoma, myeloma and renal cell cancer (66). The main toxicities are dose-dependent neuropathy, constipation and skin rash. Recent trials in melanoma that added thalidomide to temozolomide reported a trend towards superior response rates and survival with this combination compared with temozolomide monotherapy (67,68).

Lenalidomide (revemid, CC5013) is a potent analogue of thalidomide that produces T-cell stimulation and has shown single agent activity in melanoma (69). A large phase III trial comparing revemid with placebo as second-line therapy in advanced melanoma was stopped in April 2004 on the advice of the Data Monitoring Committee due to inactivity of the study drug in this setting. Unfortunately, the data remain unpublished. A phase I/II study of lenalidomide in combination with DTIC has established the dose for phase II testing (70).

In patients with metastatic melanoma, soluble VEGF-A and VEGF-R-3 pretreatment levels may prospectively identify high-risk patients with a worse prognosis, and serum VEGFR-1 may be predictive of overall survival (71). Bevacizumab is a monoclonal antibody against VEGF that has shown a significant survival advantage when combined with chemotherapy in advanced colorectal cancer and lung cancer (72,73). A randomized phase II study of bevacizumab with or without low-dose IFN (1 MU/m² subcutaneously daily) recruited 32 patients. Treatment was well tolerated, and the median PFS was 3 months in both arms (74). A phase II study of erlotinib and bevacizumab in 29 patients with metastatic melanoma reported a 9% response rate with a further 22% stable disease (75). A phase II trial of carboplatin, weekly paclitaxel and bevacizumab 10 mg/kg on D1 and D15 in 20 patients showed that the combination was well tolerated with an 8-week PFS rate of 70% (76). The role of bevacizumab as adjuvant therapy in high-risk melanoma is currently being examined in a large phase 3 study being carried out in the UK.

The integrin αVβ3 is involved in angiogenesis, growth and motility, and possibly tumor-induced osteoclastic activity. Vitaxin (MEDI-522) is a humanized monoclonal antibody to αVβ3 integrin (77). The final results of a randomized phase II trial of DTIC ± MEDI-522 are awaited but plans to take this forward to a phase III study have been cancelled. Cilengitide is a cyclic Arg-Gly-Asp peptide that blocks the binding domain on αVβ3 integrin. A phase II study showed no single agent activity in melanoma (78). CNTO 95 is a monoclonal antibody to αV integrins that has shown activity either alone or in combination with DTIC (79,80). An extended phase II study completed accrual in mid-2007. SU5416 (semaxanib) is a selective inhibitor of VEGF receptor-2 and kit receptor tyrosine kinase. A recent phase II study of 31 patients with melanoma showed it to be well tolerated and reported two partial responses (81). However, significant lymphopenia was described.

A number of drugs designed to targeted more than one pathway/receptor (e.g. EGFR, VEGF, etc.) are currently in evaluation in phase II trials. These include sunitinib, AG013736 and vandetanib.

Immunotherapy – vaccines, cytokines, antibodies and cell therapy

The concept of immunotherapy encompasses a broad range of approaches that result in tumor cell kill by activation of components of the immune system in either an untargeted or targeted way (Table 2). There are rational arguments for why cytokines and tumor vaccines should be effective in melanoma, including documented spontaneous remissions, lymphocytic infiltration of tumors, expression of developmental and melanoma-specific antigens on tumor tissue, and responses to biological agents.

Table 2.   Immunotherapy strategies in melanoma
Vaccines
 Vaccine type
  Peptide
   Developmental (e.g. MAGE)
   Melanoma specific (tyrosinase, melan A, gp-100)
  Ganglioside
  Cell lysate
   Autologous (e.g. hsp-96)
   Cell line (e.g. melacine)
  Whole cells
   Autologous
   Cell lines (e.g. canvaxin)
  Gene therapy based
   Peptide
   Cytokines
 Vaccine delivery
  Intradermal
  Subcutaneous
  Intramuscular
  Intravenous
  Intranodal
Adjuvants/immune modulators
 IL-2
 IL-21
 GM-CSF
 CpG motifs
 Anti-CTLA-4 antibodies
 Anti-CD25 antibodies
Accessory cells
 Dendritic cells
  Peptide pulsed
  Ex vivo expansion
 TILs
 Others
Patient conditioning
 Non-myeloablative chemotherapy

Recently, efforts have been focused on presentation of tumor antigens to T cells to stimulate an immune response specific for tumor cells, and on approaches that focus on effector cells – often both strategies are used. However, despite being able to produce large numbers of circulating tumor-specific T cells, there are few reports of useful clinical responses with this approach alone. We are now beginning to understand the critical role of the tumor microenvironment in controlling the interaction between immune effector cells and tumor cells, and the local factors that lead to immune suppression and T-cell dysfunction. Consequent on this, a number of new targeted treatments have been developed and are under evaluation.

Vaccines

Vaccine approaches have been a cornerstone of immunotherapy research, yet despite over 15 years of studies with melanoma vaccines, none has yet been shown to improve outcome. Key questions at present include the identification of new antigens, whether we need to use autologous antigens, efforts to induce and maintain a strong immune response, and how to break immune tolerance at the tumor site. Vaccination may be carried out with whole cells, cell lysates, proteins or specific peptide fragments. New technologies have the potential to rapidly provide a list of genes differentially expressed in malignant cells, increasing the number of possible targets. Gene therapy-based vaccine approaches, usually using viral and/or plasmid DNA delivery are being studied and have delivered genes, such as those encoding GM-CSF and melanoma-derived epitopes (82–84). In theory, autologous vaccines are superior to allogeneic vaccines as tumors can be heterogenous in terms of phenotype and growth kinetics. However, the generation of autologous vaccines is complex, time consuming and requires significant amounts of tissue, and it has been debated whether they are practical for routine use. A recent phase III study of an autologous HSP-based vaccine that randomized 215 stage IV patients on a 2:1 basis to vaccine or physicians choice showed no difference in survival, but subgroup analysis showed a survival advantage for the vaccine in patients with stage M1a disease (626 days vs 383 days) (85).

The results of a large randomized trial using an allogeneic cancer vaccine from three cell lines (Canvaxin) in patients with stage III melanoma or stage IV rendered no evaluable disease (NED) surgically, was closed prematurely on the advice of the Independent Data Monitoring Committee (IDMC) (86). One thousand and one hundred and sixty patients with resected stage III disease and 496 patients with resected stage IV disease (stage IV NED) were randomized to either Canvaxin + BCG or placebo + BCG. There was a survival disadvantage to canvaxin treatment in both studies. The median survival in the stage III study had not been reached, but the 5-year survival was 59% for the canvaxin patients and 68% for the placebo patients. In the stage IV study, the median survival was 32 months for the canvaxin patients and 39 months for the placebo patients, with a 5-year survival of 40% and 45% respectively. A large phase III study of adjuvant ganglioside vaccine GMK in stage II patients, being conducted by the EORTC, was stopped early by the IDMC because of inferior survival in the vaccine arm (87). It is clear that the results of both of these studies are a significant setback to the development of a vaccination strategy in melanoma.

Antigen presentation and dendritic cells

The mechanism of antigen presentation to T cells is now better understood, and the key role of co-stimulatory molecules appreciated. Dendritic cells (DCs) are professional antigen-presenting cells which are central to an immune response and appear ideal effectors for tumor vaccines (88). They can be cultured and expanded ex vivo, allowing antigen loading with one or many different antigens. A recent multi-centre phase III trial compared autologous peptide-pulsed DC vaccination with standard chemotherapy in stage IV melanoma (89). DCs were loaded with MHC class I-and II-restricted peptides. The study was closed after 108 patients had been treated when an interim analysis indicated that there were no significant differences in response rate or overall survival between the two arms.

Anti-CTLA4 monoclonal antibodies

The critically important role of the tumor microenvironment and the role of inhibitory molecules is now better understood. An immune signal will only be generated when an antigen is presented by an MHC molecule and a co-stimulatory molecule, B7.1 or B7.2. Binding of B7 molecules to CD28 provides a signal for T-cell activation. After activation, T cells upregulate CTLA4 which competes for binding to B7, resulting in inhibition of TCR signalling, IL-2 gene transcription and T-cell proliferation. CTLA4 has a critical inhibitory role in T-cell control.

Monoclonal antibodies to CTLA4 can break self tolerance and result in autoimmunity in some tissues. Furthermore, there is evidence of antitumor activity. A phase II study of ipilimumab reported a response rate of 13% in 56 HLA-A*0201 patients treated with the study drug and two modified HLA-A*0201-restricted peptides from the gp100 melanoma-associated antigen. Grade III/IV autoimmune toxicity (immune breakthrough events) were seen in 25% of patients, and responses were more common in this group (90). A review of 139 patients treated in 2 non-randomised studies at a single institution reported that development of immune breakthrough events was associated with an increased chance of response, and suggested that prior therapy with IFNα-2b was a negative prognostic factor (91). Two phase I studies with tremelimumab (CP-675,206) in patients with solid tumors, including 43 evaluable patients with melanoma, reported 18 patients that were alive for more than 12 months (range 13–42+), including 13 patients not achieving an overall response to treatment (92,93). As for ipilimumab, the dose-limiting toxicity was predominantly an autoimmune-mediated colitis and diarrhoea. Both of these drugs are now under further evaluation in the advanced disease and adjuvant studies are under consideration. The antitumor effect of anti-CTLA4 antibody administration appears to be due to increased T-cell activation, rather than inhibition or depletion of T-regulatory cells (94). Among the many unanswered questions with these agents are whether autoimmune breakthrough events correlate with response, whether objective response is required to get a clinical benefit, interactions with previous treatment including vaccines, and whether they will require co-administration of a vaccine in the adjuvant setting.

Other regulators of local immune response

Programmed death-1 (PD-1) receptor is a part of the B7:CD28 family of costimulatory molecules that regulate T-cell activation and tolerance. The affinity of the interaction is intermediate between that of B7:CD28 and B7:CTLA-4, suggesting a further pathway that can be targeted to modulate immune function (95). Unlike CTLA-4, the ligand for PD-1 (PD-L1) can be expressed directly on tumor cells, including melanoma (96). Strategies to block the interaction of PD-1 with its ligand are now being examined.

Other T-regulatory cells are also critical in controlling the local immune response. A distinct population of 5–10% of CD4+ cells constitutively express CD25. These CD4+CD25+Foxp3+ regulatory cells (T-regs) have a role in suppressing both the proliferation and effector functions of immune cells, and are over represented in the lymph nodes in stage III melanoma (97). Subsets of T-reg cells (naturally occurring versus inducible) may differ in their ability to suppress response to melanoma antigens (98). The number of CD4+CD25+ T-regs is upregulated in patients with melanoma and renal cell cancer (99). IL-2 upregulates CD4+CD25+ T-regs and selective inhibition of this IL-2 mediated enhancement of regulatory T cells may enhance the therapeutic effectiveness of IL-2 administration (100). Targeting CD25 may have limitations, as it is also expressed on effector T cells. A number of monoclonal antibodies to CD25 are available, but none have yet been evaluated in this situation (101). The recombinant IL-2 diptheria toxin conjugate Denileukin Diftitox has been shown to reduce regulatory T cells and enhance vaccine mediated T-cell immunity in murine models (102). Ontak® has shown activity in relapsed/refractory non-Hodgkin B-cell and T-cell lymphoma, and is approved for the treatment of cutaneous T-cell cutaneous lymphoma (103,104). However potential activity in other tumors is unclear, with one study showing no benefit in melanoma but another study showing significant activity in renal cell carcinoma (105,106). These is also increasing evidence for the role of immature myeloid cells in the control of T-cell function, and as a potential therapeutic target in melanoma (107).

New cytokines

Adjuvant pegylated IFNα-2a has shown activity in stage III patients with sentinel node positive disease, with a significant improvement in relapse free and distant metastasis free survival compared with no treatment (108). A phase II study of escalating doses of peg-IFNα-2a reported a dose response, with the highest response rate of 12% for patients treated with 450 μg weekly (109).

Interleukin-21 is a pleiotropic cytokine that activates CD8+ T cells and NK cells. Data from recent studies indicate that IL-21 renders human CD4+CD25 T cells resistant to T-reg mediated suppression and suggest a novel mechanism by which IL-21 could augment T-cell activated responses (110). Two phase I studies in melanoma and renal cell carcinoma treated a total of 72 patients. Biological activity was seen with increases in soluble CD25, CD8, CD4 and NK cells. Two patients with melanoma achieved a complete response, with a further three partial responses in renal cell cancer patients (111).

Interleukin-12 promotes TH1-cell differentiation and stimulates the production of IFN-γ from TH1 CD4 and CD8 T cells as well as NK cells. The effect of this and other downstream cytokines elicited by IL-12 is to inhibit tumor cells directly, augment immune cell function and activate antiangiogenic mechanisms (112–115). A number of studies have suggested that IL-12 would be a useful vaccine adjuvant, and a recent peptide-based vaccine study with IL-12 and aluminium hydroxide showed an enhanced immune response (116).

Thymosin α-1 is a 28 amino acid polypeptide that has a number for effects on the immune system and has been evaluated for the treatment of hepatitis B and C, HIV, and as immunotherapy in patients with cancer (117,118). It appears to have an effect on IL-6 and IL-10, may interact with IL-2, and has an effect on DCs. A phase II study comparing thymosin α-1 plus DTIC with or without IFN-α compared with DTIC + IFN-α in 483 patients suggested a survival benefit with addition of thymosin α-1 to either DTIC or DTIC + IFN, with a median survival of 8.8–10.3 months in the thymosin α-1 arms compared with 6.6 months in the DTIC + IFN arm (119).

Other immune modulators

Other molecules are also involved in the regulation of the immune response. Toll-like receptor 9 (TLR9) is expressed on plasmacytoid DCs. Immunostimulatory DNA CpG motifs stimulate TLR9 and act to amplify the effects of peptide-based vaccines and also induce innate immune responses (120,121). Other agents under evaluation include 852A, is a novel TLR7 agonist that stimulates plasmacytoid DCs to produce multiple cytokines involved in a cell-mediated antitumor response (122). STAT3 is important as a transcription factor controlling the production of immunosuppressive cytokines, and inhibition of STAT3 may restore T-cell antitumor activity (65,123).

Adoptive cell therapy

Adoptive cell transfer provides the opportunity to overcome some of these tolerance mechanisms described above, enabling expansion of highly selective antigen-presenting cells or highly reactive T-cell populations, and by manipulation of the microenvironment with an immune adjuvant or non-myeloablative chemotherapy (124). A response rate of 51% has been reported for 35 heavily pretreated patients treated with lymphodepleting conditioning chemotherapy with fludarabine and cyclophosphamide, followed by cell infusion with autologous tumor-reactive, rapidly expanded tumor infiltrating lymphocyte cultures, and high-dose IL-2 (125).

Conclusion

Malignant melanoma has proven to be one of the most difficult tumor types to treat. Unlike most common cancers, the last 20 years have seen no real improvements in systemic therapy for patients with metastatic disease. We know that ‘more is not better’, and we have a good idea which patients are more likely to benefit from treatment – important information but a poor second to effective treatment options. For the common cancers, both adjuvant therapy for surgically resected disease and systemic therapy for advanced disease are well established, and there is a real enthusiasm for ‘personalized’ treatment based on emerging biological prognostic factors. So what about melanoma? The sense is that the last 5 years have seen a rapid increase in our understanding of the complexity’s and redundancies built in to the systems that influence proliferation, invasion and metastasis. More importantly, a number of new drugs have been developed to specifically target these pathways. Well-designed trials are ongoing, with relevant clinical and translational end-points. Whether they will actually work is the first question to answer. Thereafter, we will need to consider how best to use these drugs – individually, in combination, sequencing, etc. There is a sense of optimism that the next 5 years will see a real improvement in the outcome for patients with melanoma.

Conflict of interest

PCL has carried out advisory board work for Schering Plough, Pfizer, Genta Inc and Bristol Myer Squibb. TE has carried out advisory board work for and received honoraria from Schering Plough, Bayer, Pfizer and Aventis/Genta. He has received research support from Schering Plough, Bayer & Pfizer. AH has carried out advisory board work, received honoraria and research support from Onyx Pharmaceuticals, Bayer HealthCare, Schering-Plough, Pfizer, Synta Pharmaceuticals, Schering Germany, 3M Pharmaceuticals and Aventis/Genta.

Ancillary