Adjuvant Therapy in Melanoma: New Combination Cytokine and Other Strategies

Introduction

Melanoma is the leading cause of death among all types of skin cancer and is the fifth most common type of cancer in men in the United States. The incidence of melanoma is increasing; in 2008, the incidence of this disease is expected to exceed 60,000, with 8000 deaths.[1] Fortunately, up to 95% of patients will present with either local or regional disease, which is potentially curable.[2] Although the causes of this increased incidence of melanoma are debated, it is thought that increased exposure to ultraviolet radiation interacts in some manner with genetic factors to initiate melanoma.[3]

Histologically, melanoma is characterized by proliferation of transformed melanocytes.[4] These long-lived cells are normally present at the dermo-epidermal junction and impart pigment to the skin; they are typically resistant to DNA damage and apoptosis, a characteristic of melanoma as well. The principal subtypes of cutaneous melanoma are superficial spreading, nodular, lentigo maligna, and acral lentiginous; rarer noncutaneous primary sites include the mucosal membranes and the pigmented epithelium of the eye. Most types of melanoma are more common among whites, although acral lentiginous melanoma occurs with equal frequency in nonwhites.

Several pathologic and clinical factors are predictive of the risk of recurrence and death in persons with melanoma. The clinical factors include age, sex, location, and lymph node or distant organ involvement, whereas pathologically, the tumor depth (Breslow depth), ulceration, mitotic rate, and presence of microsatellites can determine the risk for recurrence or spread.[5] The recent revision of the American Joint Committee on Cancer staging system for melanoma incorporates many of these factors. Some pathologic features, such as involvement of skin layers (Clark levels) and the presence of tumor regression, have not stood up to rigorous multivariate analysis and are no longer used clinically to determine the risk for recurrence or the current staging classification.[6] Over 60% of melanomas have mutations in the B-type RAF-1 kinase (BRAF).[7,8] About 80% of these mutations are found at exon 15, at a single amino acid residue, usually a substitution for valine by glutamic acid, V599E (now referred to as V600E). This mutation causes increased kinase activation and signaling through the mitogen-activated protein (MAP) kinase pathway.[9,10] Surprisingly, this event occurs with high frequency in benign nevi as well as in melanoma.[11] An additional 15% to 20% of melanomas have mutations in N-Ras, which lies upstream of BRAF.[12] Both of these genes code for proteins that are part of the ERK-MAP kinase pathway. A small proportion of melanomas, especially those originating in sun-damaged skin, the palms and soles, and mucous membranes, have mutations or amplifications in the c-kit gene.[13]

Recent reports suggest that integration of molecular profiling with pathologic and clinical features may result in a better prognostic profile and help to more precisely determine an individual's risk for recurrence.[14] It has been suggested, for example, that the melanoma arising from sun-exposed skin in older persons is more likely to have c-kit mutations than the truncal melanoma arising in younger patients, which is more likely to have BRAF mutations.[14]
Surgical excision with wide margins and sentinel lymph node biopsy (for a melanoma whose depth is ≥ 1 mm) is the treatment of choice for primary melanoma. Several large clinical trials have helped to define the width of margins needed to resect melanoma and reduce the risk for recurrence.[15-17]

Currently, melanoma of ≤ 1 mm thickness is resected with surgical margins of 1 cm, whereas melanoma ≥ 2 mm is resected with margins of 2 cm, and intermediate-thickness melanoma is resected with 1- to 2-cm margins. Mapping the draining sentinel lymph node with use of a tracer and careful histologic examination of this node is now a standard of care for melanoma ≥ 1 mm in thickness.[18] High-risk melanoma, defined as stages IIB, IIC, and III, often recurs after excision and is associated with a 40% to 80% chance of death.[2] Given the high risk for recurrence (and subsequent mortality), there is an obvious need for reduction of recurrence risk. This review will address several treatment approaches that have been employed as adjuvant therapy following primary surgical excision.

Interferon Therapy

The interferons (IFNs) can produce antitumor effects through upregulation of the immune system. In humans, IFNs are structurally divided into 2 classes: type I and type II. The type I family, which is used clinically for the treatment of several diseases, includes alpha (alfa), beta, and omega; the type II family has only 1 member, IFN gamma. The various subtypes are further subdivided for pharmaceutical preparations, eg, IFN alfa-2a, IFN gamma-1b.[19] IFNs can stimulate both the innate and adaptive arms of the immune system; for example, they can enhance major histocompatibility complex (MHC) class I antigen presentation, increasing innate immunity and maturation of dendritic cells and leading to enhanced adaptive immunity. They also have antiangiogenic and direct cytotoxic effects on some malignant cells. However, the exact mechanism by which they produce antitumor effects is unclear.
Two large randomized trials have shown that high-dose IFN-alfa-2b significantly reduces the risk for recurrence compared with observation alone in patients with resected cutaneous melanoma.[20,21] A third large randomized trial has shown a significant improvement in overall survival with high-dose IFN-alfa-2b compared with the GM2 ganglioside conjugated to keyhole limpet hemocyanin (KLH) melanoma vaccine in patients with resected stage IIB-III melanoma.[22] In the first trial, ECOG 1684, patients with thick primary melanoma (≥ 4 mm depth) or lymph node-positive disease were randomized to receive IFN-alfa-2b 20 MIU/m2/day intravenously for 4 weeks followed by 10 MIU/m2 subcutaneously 3 times per week for 48 weeks or to observation. Patients who received IFN-alfa-2b had improved relapse-free survival (RFS) (5-year RFS 37% vs 26%; P = .0023) and overall survival (OS) (5-year OS 46% vs 37%; P = .0237) compared with the observation arm. This led to the approval by the US Food and Drug Administration of IFN-alfa-2b for the treatment of high-risk melanoma after surgical resection. Because this high dose appeared to be effective, the next ECOG trial was conducted to determine whether a lower dose of IFN might be as effective as that studied in ECOG 1684. Therefore, ECOG 1690 randomized patients with thick or node-positive melanoma to 1 of 3 arms: (1) high-dose IFN-alfa-2b (20 MIU/m2/day intravenously for 4 weeks followed by 10 MIU/m2 subcutaneously 3 times per week for 48 weeks) for 1 year; (2) low-dose IFN-alfa-2b (3 MIU/m2 3 times per week subcutaneously for 2 years); and (3) observation. Five-year RFS in the 3 arms was 44%, 40%, and 35%, respectively. The difference between high-dose IFN and observation was statistically significant (P = .05), but the difference between low-dose IFN and observation was not. Of note, no improvement in overall survival was noted between either of the treatment arms and observation.
The GM2 ganglioside vaccine had shown promising results in small phase 1 and 2 trials and was brought forward for further clinical testing.[23,24] The gangliosides are complex carbohydrates found on the outer cell membrane that can be as immunogenic as protein antigens. The resulting trial, ECOG 1694,[22] was different from the 2 prior studies because it compared high-dose IFN to GM2-KLH/21 vaccine in patients with resected stage IIB and III melanoma. High-dose IFN improved both RFS (hazard ratio [HR] 1.49; P = .00045) and OS (HR 1.38; P = .023) compared with GM2 vaccine, and this trial was halted early by the data safety monitoring committee. The benefit attributed to IFN was seen in patients with node-positive disease as well as those with node-negative disease.

Because the findings of these trials and others conducted by European and American investigators were inconsistent, several meta-analyses have been performed to assess the data globally. An authoritative meta-analysis performed by Wheatley and colleagues[25] examined 12 randomized controlled trials and found that RFS was highly significantly improved with IFN-alfa by 17% (HR 0.83; P = .000003). The benefit on OS associated with IFN-alfa trended toward a 7% improvement but was not statistically significant (HR 0.93; confidence interval [CI], 0.85-1.02; P = .1). The investigators subsequently have updated their meta-analysis and shown a slightly better outcome for IFN-alfa in terms of OS.
Inasmuch as IFN-alfa has a short half-life necessitating frequent injection and resulting in side effects from rapid variation in blood levels, it has been modified to reduce clearance and increase its half-life.[19] One such modification is the addition of a 12,000-dalton polyethylene glycol chain, which results in pegylated IFN (PEG-IFN). In the treatment of viral hepatitis, PEG-IFN appears to be as efficacious as IFN with reduced toxicity.[26] To determine whether PEG-IFN was effective in melanoma, the European Organization for Research and Treatment of Cancer (EORTC) conducted a large trial, EORTC 18991, which randomized more than 1300 patients with node-positive melanoma postoperatively to either PEG-IFN for a maximum of 5 years or observation. The 4-year RFS was 45.6% and 35.9% (P = .01) in the PEG-IFN and observation groups, respectively. No difference in OS was noted.[27] The greatest benefit appeared to be in patients with microscopic nodal disease. Overall, patients appeared to tolerate this treatment relatively well.
The major issue limiting the use of IFN in patients with melanoma has been its toxicity. IFN is associated with a characteristic and sometimes debilitating constellation of signs and symptoms that can make its use challenging.

Commonly, patients experience fatigue, myalgia, anorexia, altered taste sensation, fevers, chills, loss of concentration and short-term memory, and other neuropsychiatric abnormalities. Significant depression has been reported in some patients. Laboratory abnormalities associated with use of IFN include elevations in transaminase levels, neutropenia and thrombocytopenia, increased creatinine levels, and anemia. Commonly, patients are dehydrated and need to be encouraged to increase their fluid intake and maintain their activity levels. Given the modest benefit of IFN in terms of reducing the risk for recurrence and death from melanoma, its significant toxicity profile has made it an option for only a select subpopulation of patients at high risk for melanoma recurrence.[19]
Granulocyte-Macrophage - Colony-Stimulating Factor

Although granulocyte-macrophage - colony-stimulating factor (GM-CSF; sargramostim) was initially identified as a colony-stimulating factor for myeloid cells in the bone marrow, it was also found to induce differentiation of dendritic cells (DCs).[28] The DCs, sometimes referred to as antigen-presenting cells, display tumor antigens to the immune system in the appropriate context and are increasingly recognized as a vital afferent arm of the immune system. Both quantitative and functional defects in DCs have been described in patients with cancer. These defects may contribute to "tumor escape" of immune surveillance. Specifically, GM-CSF increases the mobilization, differentiation, and function of DCs[29,30] and hence enhanced antigen presentation to CD4 and CD8 T cells.
GM-CSF appeared to be effective in increasing tumor-specific immunity when used as an adjuvant for vaccines.[30] In the metastatic disease setting, GM-CSF has been injected intralesionally and intrahepatically (with use of a hepatic artery catheter) and administered as an inhalational treatment. A GM-CSF-producing adenovirus appears to induce regression of metastatic melanoma when injected into lesions.[31] GM-CSF inhaled as an aerosol resulted in regression of pulmonary metastasis from melanoma in a phase 1 study.[32] In a phase 1 study of patients with metastatic melanoma, a 32% response rate was noted in patients with liver metastases following hepatic arterial infusion of large doses of GM-CSF.[33]
On the basis of these data, Spitler and colleagues[34] evaluated the role of GM-CSF in patients with high-risk (stage IIIB, IIIC, or IV) resected melanoma. Patients received GM-CSF 125 µg/m2 subcutaneously daily for 2 weeks every month for 1 year. The median survival of patients in this study was 37.5 months and far exceeded the median survival of 12.2 months in historical controls. The limitations of the study included the lack of a placebo arm and heterogeneity between the historical control and study groups. Overall, GM-CSF was well tolerated.
To confirm the findings of this trial and to examine the effect of GM-CSF on DCs, Daud and coworkers conducted a study with essentially the same dose and schedule in a similar patient population.[35,36] OS and RFS were 65 months and 5.6 months, respectively. GM-CSF treatment caused an increase in mature DCs, first identified after 2 weeks of treatment and normalizing by 4 weeks. Patients with decreased DCs at baseline had significant increases in DC number and function compared with those with "normal" parameters at baseline (Figure 1).[36] No change was observed in the number of myeloid-derived suppressor cells (MDSCs). Early recurrence (< 90 days) correlated with a decreased effect of GM-CSF on host DCs, compared with late or no (evidence of) recurrence (Figure 2).[36] Therefore, greater increase of DCs was associated with remission or delayed recurrence. Although this study lacked a control group, the survival data for the GM-CSF-treated patients does appear impressive. The benefits of GM-CSF appear to be greater in those with reduced DCs, but inasmuch as these analyses were exploratory, further investigation is warranted.
GM-CSF is currently being actively investigated in combination with other immunologic agents and chemotherapy. GM-CSF with the anti-CTLA-4 antibody ipilumimab has been explored in a phase 1 trial[37] (L. Fong, personal communication) and has shown some promising activity in prostate cancer. In addition, it has been explored as a component of combination therapy[38-40] and also as a maintenance or consolidation regimen in patients with metastatic melanoma who have benefited to some extent from chemotherapy.[41] There is also an ongoing trial with high-dose IL-2 in combination with GM-CSF (J. Lutzky, personal communication).

Anti-CTLA-4 Antibodies

The use of monoclonal antibodies to inhibit cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4) is a novel strategy in melanoma. These antibodies have been found to induce tumor regression and improve long-term survival in tumor-bearing mice.[42-44] Early preclinical studies showed that CTLA-4 serves as a natural braking mechanism for T-cell activation, allowing a return to homeostasis after an immune response. Inhibition of CTLA-4 upregulates several downstream targets, including T helper (Th) 1- and 2- produced cytokines (ie, interleukin 1 and 2) and cell cycle machinery (CDK-4, cyclin D3), leading to a more robust antitumor immune response. Two anti-CTLA-4 antibodies are currently in phase 2 and 3 trials: ipilimumab (also known as MDX-010) and tremelimumab (also known as CP 675,206).
Prolonged, but sometimes delayed, responses have been seen in patients with melanoma who have received either of the anti-CTLA-4 antibodies.[45,46] Similar to the effect seen with IFN therapy, immune-related adverse events correlate with more positive outcomes in patients with metastatic melanoma who receive anti-CTLA-4 therapy, including prolonged time to relapse.[47] Recently, it has been reported that tremelimumab was no better than standard dacarbazine (DTIC) or temozolomide for the up-front treatment of metastatic melanoma.[48] Results of an ongoing trial of ipilimumab and DTIC vs DTIC alone in the setting of metastatic melanoma are expected soon. Because anti-CTLA-4 antibodies appear to be effective in the phase 2 setting in metastatic disease, they are also being considered for use in the adjuvant setting. Currently, an ongoing EORTC study is examining ipilimumab vs observation in patients with high-risk melanoma, and combinations of anti-CTLA-4 antibody with GM-CSF and IFN are promising for evaluation because different parts of the immune system may be stimulated.

Vaccines

Tumor cells can express novel antigens or quantitatively different antigens compared with "normal" or nondividing cells. Harnessing the immune system to destroy malignant cells by recognizing these antigens is a major goal of tumor immunotherapy. Tumor antigens upregulated in melanoma include MART-1 (Melan A), gp100, and tyrosinase; other investigators have focused on the cancer-testis antigens, which are upregulated in tumor cells but present only in human germ cells in the body. Although early studies of therapeutic vaccines against melanoma showed some promise, [49,50] randomized, prospective, placebo-controlled clinical trials have failed to prove a benefit. An allogeneic melanoma vaccine not only failed to improve either DFS or OS, but the study arm actually showed worsened survival.[51] Similarly, although the GM2 ganglioside vaccine showed promise in phase 2 trials,[23] a recent randomized study showed worsened survival in the vaccine population compared with an observation group.[52] Although these signals may be confined to these particular vaccines and not necessarily to peptide or DC vaccines, caution is needed when interpreting vaccine studies given these data.

Chemotherapy

Historically, DTIC has been the standard chemotherapy for patients with high-risk melanoma.[53] The use of chemotherapy in the adjuvant setting in melanoma has been tested in randomized trials and has shown no benefit.[54-56] At this time, chemotherapy should not be administered in the adjuvant setting outside of a clinical trial. Whether chemotherapeutic agents may be effective when used in combination with immune-modulating agents is still under investigation.

Other Novel Therapies

As more details are elucidated about the biology of melanoma, new pharmacologic targets are being identified. Among these are immunologic targets such as programmed death (PD)-1, anti-CD40, and anti-41BB (CD137) antibodies, which are currently in phase 1 and 2 trials.[35] PD-1, a member of the tumor necrosis factor family, has shown promising results in murine trials and is currently in phase 1 investigation. CD40 is a receptor expressed on B cells and DCs; it is bound by CD40 ligand found on activated T cells. Phase 1 trials[57] have demonstrated a modest dose-related benefit (0.3 vs 0.2 mg/kg) that has prompted ongoing phase 2 trials. Also under investigation is adoptive cell transfer after lymphoid depletion (homeostatic lymphoid proliferation).[58] A very interesting proof-of-concept study showed that CD4 T cells directed against NY-ESO-1 antigen were able to produce a response in melanoma.[59]
Targeted agents directed against a growth pathway specific to melanoma are also in development. Sorafenib, a multikinase inhibitor with multiple targets including VEGF, has had modest activity when studied as monotherapy, with a 19% rate of stable disease.[60,61] Src inhibitors such as dasatinib, VEGF kinase antagonists such as sunitinib, and c-kit inhibitors such as imatinib are all under investigation in melanoma.[62].Although these agents are currently not known to be effective in the metastatic setting, they may hold promise for early-stage disease.

Conclusion

Currently, only a few adjuvant therapies show any effectiveness in terms of decreasing the mortality rate in patients with high-risk melanoma. IFN is the only agent that has been proven to reduce recurrences after surgery, although its benefit on survival is likely very modest. However, this modest benefit must be weighed against the toxicities, and hence candidates for adjuvant IFN therapy must be chosen carefully.
Among the newer agents, GM-CSF shows promise: two phase 2 studies have shown that GM-CSF administered subcutaneously may be associated with prolonged survival. Currently a major randomized trial is under way to determine the degree, if any, of this benefit. The combination of GM-CSF with other immune modifiers, such as anti-CTLA-4 antibody or IFN, may be the next step in clinical trials, although no trial data are yet available with these combinations. Vaccines have an extensive history in melanoma, but to date no vaccine preparation has shown a survival or recurrence benefit in melanoma, and recent data even indicate a potential for harm. Other agents are on the horizon, including targeted therapies directed toward the specific mutations present in certain melanomas, which could be used in the adjuvant setting. Some exciting cell-based therapies have been shown to work in proof-of-concept clinical trials; more data are awaited prior to studying these agents in the adjuvant setting.

This activity is supported by an independent educational grant from Bayer HealthCare Pharmaceuticals.