- © 2003 by American Society of Clinical Oncology
Chemoprevention of Melanoma: An Unexplored Strategy
- From the Boston University School of Medicine, Boston, MA; and Skin Oncology Program, Boston; and Oncology Consultants, Cambridge, MA.
- Address reprint requests to Marie-France Demierre, MD, Skin Oncology Program, Boston Medical Center, 720 Harrison Ave, DOB 801A, Boston, MA 02118; email: mariefrance.demierre{at}bmc.org.
Abstract
The incidence and mortality of melanoma has continued to increase steeply—faster than most other preventable cancers in the United States. Current sun protection strategies have yet to reduce this increased incidence and mortality. Chemoprevention, defined as the use of natural or synthetic agents to delay, reverse, suppress, or prevent premalignant molecular or histologic lesions from progressing to invasive cancer, has become an important area in cancer research. Melanoma, with its associated risk factors and its known precursors or premalignant lesions, should lend itself well to chemoprevention. Prerequisites for this research should include determination of the molecular mechanisms of ultraviolet (UV) melanomagenesis; use of animal models to test candidate prevention agents; use of molecular and histologic markers as surrogate end point markers; collection of epidemiological, basic science, or in vitro data on potential chemoprevention candidate drugs; and selection of a high-risk patient population in which to carry out clinical chemoprevention trials. Preliminary data available in all these areas are reviewed. Possible mechanisms and molecular targets for the chemoprevention of UV-induced melanoma are discussed. This recent information should stimulate research in the chemoprevention of melanoma.
MELANOMA, A preventable malignancy, remains the cancer with the fastest rate of increase in incidence, whereas most rates for other cancers are decreasing.1 In the United States, nationwide public and professional education and sun avoidance recommendations have yet to make an impact.2 Furthermore, some sun exposure seems inevitable. In Australia, where the most comprehensive sun protection programs have been in place for more than a decade, the incidence of sunburns over the prior summer did not change significantly from 1991 to 1998.3 Of additional concern has been the depletion of atmospheric ozone, the principal barrier to ultraviolet (UV) atmospheric penetration, which has continued in a seemingly relentless fashion.4 In view of the increasing incidence of melanoma and the apparently unquenchable desire of many individuals to have tanned skin, new melanoma prevention strategies are needed.
Chemoprevention, a term first used by Sporn et al in 1976,5 may be defined as the use of natural or synthetic agents to reverse, suppress, or prevent premalignant molecular or histologic lesions from progressing to invasive cancer. Putative in situ lesions fall into the spectrum of premalignant lesions included in this definition. Chemoprevention has been conceptually divided into the following categories: (1) primary, preventing initial cancer in the high-risk individual; (2) secondary, preventing cancers in those with premalignant conditions; and (3) tertiary, preventing second primary cancers in patients cured of an initial cancer.6 The recent recommendations of the American Association for Cancer Research Task Force in February 20027 included the following statement: “Precancer or intraepithelial neoplasia is a noninvasive lesion that has genetic abnormalities, loss of cellular control functions, and some of the phenotypic characteristics of invasive cancer, and that predicts for a substantial likelihood of developing invasive cancer. . . . There is an urgent need to rapidly develop new treatment and prevention agents for intraepithelial neoplasia.” Melanoma, with its associated risk factors and its known precursors or premalignant lesions, should lend itself well to chemoprevention.8
PREREQUISITES OF A CHEMOPREVENTION PROGRAM
The landmark United States Food and Drug Administration approval of tamoxifen for use in reducing breast cancer risks spurred interest in chemoprevention drugs. This approval, as well as that of two agents for premalignant lesions (celecoxib for reducing adenomatous polyps in patients with familial adenomatous polyposis, and diclophenac for treating actinic keratosis9), further illustrates key prerequisites needed to develop a chemoprevention program. These prerequisites include (1) determination of the underlying molecular mechanisms of carcinogenesis of the cancer of interest; (2) availability of drugs that can target these abnormalities; (3) availability of animal tumor models that facilitate preclinical trials of efficacy and toxicity; (4) availability of molecular or histologic markers of the carcinogenic process to be used as end points and to help obviate the need for prolonged and costly clinical trials; (5) compilation of data from epidemiologic, basic science, and cancer research literature that can yield candidate prevention agents for in vitro or in vivo testing; (6) access to defined groups at high risk for the disease, increasing the likelihood that cost-effective chemoprevention or risk reduction will be demonstrated; and in some settings, (7) the discovery of genetic markers that identify the early events in the carcinogenic process. Such genetic markers could be critical in helping to define high-risk groups or suggesting molecular chemoprevention targets.
We review the available information that we believe should stimulate interest in the chemoprevention of melanoma.
PUTATIVE UNDERSTANDING OF MOLECULAR MECHANISMS OF MELANOMAGENESIS
Two principal areas have been the focus of research: UV radiation and genetic abnormalities. UV radiation is implicated in approximately two thirds of melanoma cases,10 and its pathogenesis has been reviewed.11 The heritable genetic contribution to melanoma is believed to be 10%, although one Swedish research group estimated it to be as high as 18%.12 Although the exact sequence of events leading to melanomagenesis are not known, it is likely a multistep process of progressive genetic changes, altering cell proliferation, differentiation, and death (Fig 1⇓). In UV radiation–related melanoma, mutations probably render individuals more sensitive to the carcinogenic effects of UV light, whereas in non-UV-induced melanomas (eg, occular, mucosal, acral lentiginous, or congenital melanoma), mutations may relate more directly to cell cycle growth control factors. In both settings, failed DNA repair mechanisms may be an important abnormality regardless of which mutations lead to loss of growth control in melanocytes.
Multistep theory of UV-induced melanoma.
UV Melanomagenesis
UV-B (290–320 nm), which is responsible for most of the erythemogenic, carcinogenic, and immunosuppressive effects of UV radiation,13 affects DNA both directly and indirectly (via the production of reactive oxygen species). In contrast, UV-A (320–400 nm), which is involved in immediate skin darkening, photosensitivity, cocarcinogenicity with UV-B, and probably the immunosuppressive effects of UV radiation,14 acts primarily by the indirect mechanism. The initial event in UV mutagenesis, especially with wavelengths in the UV-B range, is the production of pyrimidine dimers, of six to four photoproducts, or of single-strand DNA breaks.15 When promptly repaired, these lesions do not produce any adverse long-term effect. Specifically, in keratinocytes, p53 apoptotic mechanisms play an important role in this repair process, as demonstrated by the peeling of dead skin cells experienced after sunburn. Unlike the case in keratinocytes, p53 inactivation does not seem to play an important role in the early stages of melanomagenesis. In melanocytes, although UV-induced sublethal injury can promote the expression of genes involved in DNA repair, there seems to be resistance to UV radiation–induced apoptosis.11 Possibly, evolution is allowing a certain amount of UV damage to melanocytes and permitting preservation of their melanin-generating photoprotective role in the skin.11 Ultimately, retained damaged melanocytes would be at risk for subsequent mutations. Theoretically, any agents able to reverse or repair UV-induced DNA damage (eg, DNA repair enzymes16) or capable of promoting the apoptosis of melanoma cells17 should be investigated for their chemopreventive potential.
Genetic Abnormalities in Melanoma
In melanoma, the most commonly identified mutation has been in the tumor suppressor gene family. The identification of loss of heterozygosity (LOH) in the 9p21 chromosome region in melanoma18 permitted the identification of this gene now called CDKN2A/p16ink4a. This kinase-inhibiting gene, CDKN2A/p16ink4 commonly exhibits altered coding or regulatory sequences.18 These include cytosine phosphate guanine island methylation and translational repression mutations in the 5′ untranslated region. Recessive germline CDKN2A mutations have been found in 30% to 50% of members of melanoma kindreds,19 in 8% to 15% of individuals with multiple primary melanomas,20 and in some sporadic melanomas.
Mutations in the cyclin-dependent kinase CDK4, which abrogates p16-CDK4/6 binding, has also been identified in a few melanoma families. However, at least 60% of melanoma families have neither mutation. The possibility of other downstream mutations such as in the CDKN2B/p15ink4b gene has also been hypothesized. One report has suggested that the XRCC3 gene,21 involved in the homologous pathway of double-stranded DNA repair, is mutated in the T allele of exon 7. Because UV-irradiated melanocytes do not induce p53-mediated apoptotic mechanisms and do tolerate greater amounts of photoproducts from UV-damaged DNA, failure of this repair system may be extremely important in melanomagenesis. The antiapoptotic protein Bcl-2 has been reported to be overexpressed in melanoma cells,22 and it would be an appropriate prevention target. The transcriptional apoptosis activating factor (Apaf-1) is involved23 in the Bcl-2 pathway and could be another target. Other transcriptional factors such as the activator proteins (AP-1, AP-2) may be relevant in melanoma; one group showed decreased AP-2 expression associated with increased risk of melanoma metastases.24
p16, Rb, AND Ras PATHWAYS AND MELANOMAGENESIS
The CDKN2A/p16ink4a gene encodes two proteins on alternated reading frames: p16ink4a has tumor-suppressive function via inhibition of the CDK4/6-cyclin D complex, and p14ARF (where ARF represents alternate reading frame) is proapoptotic via activation of the p53 gene (Fig 2⇓). p14ARF may also be relevant to melanomagenesis. Although the majority of inherited melanoma-associated CDKN2A mutations affect p16ink4a, approximately 40% of these mutations affect both p14ARF and p16ink4a, in some instances altering the function of both proteins.25 Data from mouse models provide support for the hypothesis that defects in the ARF pathway contribute to susceptibility to melanoma.26 Indeed, p14ARF-specific mutations have recently been described in a melanoma-prone family.27
CDKN2A encodes two proteins on alternate reading frames: p16ink4a and p14ARF. P14ARF inhibits p53 degradation by HDM2. A second gene, CDK4, an oncogene, inhibits the binding of p16ink4a, aborting its tumor suppressor function. Murine studies have shown cooperation between the ARF and Ha-ras pathways in melanomagenesis.26,28
In the p16 and Rb pathway, the formation of the complex of CDK4/6 and their cofactor cyclin D allow phosphorylation (by adenosine triphosphate) of the retinoblastoma (Rb) protein and its inactivation. Phosphorylated Rb protein disassociates from the E2F transcription factors to which it is bound, in turn allowing cellular S-phase DNA transcription, synthesis, and replication.28 This in turn enables the cell cycle to pass the G1/S checkpoint. Thus Rb, when unphosphorylated, acts as an inhibitor of entry into the S phase of the cell cycle. Mutations of CKN2A/p16ink4a will prevent its inhibitory effect on the cell cycle and thus promote tumor cell replication. Similarly, mutated CDK4, even in the presence of a functional CDKN2A/p16ink4a, prevents binding of p16ink4a and similarly allows the cell cycle to proceed.19 Note that restoration of normal CDKN2A/p16ink4a or CDK4 genes, and expression and reactivation of their respective encoded proteins, could be a chemoprevention strategy. In addition, agents that bind or inactivate any of the E2F transcription factor family (Fig 2⇑) with their dimerization partner (DP1) could also inhibit the cell cycle and prevent incipient tumor cell formation. In summary, several points in the CDKN2A pathways represent potential areas of targets for chemoprevention.
The Ras human genes encode guanosine triphosphate–binding proteins involved in key functions of signal transductions.29 Mutations in the Ras proto-oncogenes are commonly found in human solid tumors, including cutaneous melanoma.30 Among the three types (H-ras, K-ras, and N-ras), the N-ras proto-oncogene is the predominant Ras alteration in cutaneous melanoma.31 The transforming activity of oncogenic Ras is dependent on its anchorage to the plasma membrane, and farnesylation is obligatory for Ras oncogenicity. Thus, the blocking of Ras-mediated signal transduction could be another chemoprevention strategy.
Finally, because genetic aberrations are relevant in melanoma, the promise of microarray technology could further enhance our ability to identify molecular genetic targets for candidate prevention agents. In this regard, the recent discovery of the BRAF gene mutated in 66% of malignant melanomas32 and in human melanomas, and the correlated expression of MITF (a transcription factor essential for melanocytic lineage) with Bcl-233 (an overexpressed antiapoptotic factor in melanoma), highlight the key role that such technology will play.
ANIMAL MODELS OF MELANOMA
Animal models have been a critical component of research to identify chemoprevention agents. Although many animals have spontaneous melanomas,34 with canine melanoma being the most common, in most animals, tumor behavior does not mimic that of human melanoma. The major breakthrough of recent years has been the development of a series of transgenic or chimeric mouse models. These include the hairless Skh-1 mouse35 in which UV carcinogenesis can be studied. The SCID-Beige mouse36 will accept human xenotransplants as well as murine tumors. Knockout mice37 are deficient in a variety of factors. These including repair genes such as Xpa (nucleotide excision repair), p53 (apoptosis), XRCC1 (strand break repair), and so on. Others possess inactivated tumor suppressor genes encoding p21, Rb, or NF1 Ras regulatory proteins (or others), enabling the study of cell cycle control. The hepatic growth factor/scatter factor mouse38 has melanocytes in the epidermis or upper hair follicle (unlike the normal mouse dermal melanocyte location) and consequently develops UV-induced melanocytic lesions. The RFP/RET mouse carries the RET oncogene fused to the mouse39 metallothionein-1 promoter-enhancer and develops melanosis and melanoma with UV exposure. The CXCL1 chemokine mouse40 expresses murine macrophage inflammatory protein-2. When the gene encoding this chemokine is transfected into mice that are p16/p19 deficient, it demonstrates the melanomagenicity of macrophage inflammatory protein-2. Ha-ras mice,41 in which the gene is driven by a mouse tyrosinase promoter, develop spontaneous melanomas and can be used to study potential preventive agents. Current mouse models would allow understanding of molecular mechanisms of prevention and the screening of candidate chemoprevention agents.
BIOMARKERS OF MELANOMA
The success of chemopreventive strategies will be facilitated by the identification, development, and validation of early, reliable biomarkers of carcinogenesis. These may be molecular or tissue markers.8 Biomarkers are further subdivided between risk, susceptibility, or prognostic biomarkers (which estimate the probability of a later event in neoplasia) and response or outcome biomarkers (which measure the response to an intervention).42 Response biomarkers that accurately and reliably predict a later event, such as cancer incidence (eg, preinvasive breast neoplastic lesions and breast cancer), can be applied as a surrogate end point biomarker for the later event.42 In melanoma, melanocytic dysplasia reflects an environment permissive for carcinogenesis. The dysplastic nevus, both a well-documented marker of melanoma risk and a precursor lesion, could be validated as a surrogate end point biomarker.
CLINICAL TRIAL DESIGNS AND CANDIDATE AGENTS
The evaluation of candidate chemoprevention agents or the acquisition of chemoprevention data could be obtained in four clinical settings: (1) surrogate marker trials, (2) adjuvant melanoma trials, (3) nononcologic (unrelated disease) randomized trials, and (4) clinical trials in healthy groups at high risk for melanoma. To date, in the first type of clinical trial, four studies have evaluated the effect of topical tretinoin (retinoic acid analog) on the atypical nevus as a surrogate marker for the chemoprevention of melanoma.43–,46 Of these studies, the two most recent by Halpern et al45 and Stam-Postuma et al46 both demonstrated a significant effect on transformation of clinical appearance (including color, size, and border irregularities), from atypical to banal. Although Halpern’s study showed a statistically significant histologic change toward benignity (for cellularity, cellular atypia, and proliferative cellular nuclear antigen), Stam-Postuma’s study had too few biopsies in the control nevus group for optimal statistical analysis. The interpretation of these surrogate marker studies is uncertain. In light of retinoids’ mechanisms of cell death promotion by apoptosis, induction of differentiation, and inhibition of cell proliferation,47 further studies would be of interest.
In the second type of clinical setting, chemoprevention data gathered in a large, randomized, clinical adjuvant trial, similar to the National Surgical Adjuvant Breast and Bowel Project B16 tamoxifen trial,48 would be helpful. To date, no such trial has been performed in melanoma. A relatively high event rate for the second malignancy should be a key prerequisite for such data collection. Approximately 3% of primary melanoma patients will develop a second melanoma, and almost 10% will experience some second malignancy including melanoma. Although the collected data from an adjuvant trial would not conclusively establish risk reduction efficacy,49 it could yield important clues, similar to the tamoxifen study, that would justify the time and expense of a subsequent chemoprevention trial in high-risk individuals.
In a nononcologic setting, two large, placebo-controlled, randomized trials of the lipid-lowering agents gemfibrozil50 and lovastatin51 in coronary heart disease indicated statistically significant lower melanoma rates in the treatment groups, whereas another phase III study of pravastatin showed no significant differences.52 These observations from the study of an unrelated disease have stimulated interest in lipid-lowering agents and melanoma prevention. The use of lipid-lowering drug prescriptions was examined at Denver Veterans Administration Medical Center53 in a case-control format. Fewer melanoma patients had received such prescriptions than had controls. In Philadelphia, in a nested case-control study, the odds of having been prescribed a lipid-lowering agent were significantly lower in patients who had had a melanoma compared with those who had been assigned the diagnosis of a nevus (odds ratio [OR], 0.72; 95% confidence interval [CI], 0.53 to 0.98).54 Lovastatin, a statin, has been found to arrest tumor and normal cells in the G1 phase of the cell cycle55,56 and induce potent apoptotic responses57 and an antitumor effect in a melanoma murine model.58 Because the inhibition of isoprenoid compounds59 by statins may interfere with the function of the Ras oncogene,60 their role as chemoprevention agents should be further explored.
To date, in the fourth clinical trial setting, no candidate chemoprevention agents have been investigated in healthy individuals at high risk for melanoma. The growing data on potential agents (Table 1⇓) could facilitate the design of future trials.
Possible Mechanisms and Molecular Targets for Chemoprevention of UV-Induced Melanoma
EPIDEMIOLOGIC STUDIES OF NUTRITIONAL FACTORS
A series of studies have examined the relationship between nutritional factors and melanoma. The suggestion that a high consumption of polyunsaturated fat leads to the development of melanoma61 was not confirmed. Several investigators did not find any association with melanoma risk;62–,65 Bain et al66 observed a strong inverse relation between high intakes of polyunsaturated fatty acids (PUFAs) and melanoma in a case-control study in Australia. PUFAs could be important in cell cycle control:67 One research group demonstrated that docosahexaenoic acid could inhibit the growth of cultured metastatic melanoma cells.68
Central to UV mutagenesis is the generation of reactive oxidant species and free radicals.69 Hence, antioxidants (beta-carotene; vitamins A, C, and E; zinc; and selenium) have been studied. To date, these studies have been principally case-control (some nested) studies. Stryker et al70 assessed diet, retinol, alpha-tocopherol, lycopene, alpha-carotene, and beta-carotene as they related to incidence of melanoma. No significant associations with melanoma were observed for higher plasma levels of lycopene, retinol, or alpha-carotene.70 Alcohol consumption was noted to be positively associated with risk of melanoma (x for trend + 2.1, P = .03), although Bain et al66 observed a relatively strong, but not statistically significant, association with alcohol. Except for one study,64 which found alcohol to be protective, others have found no association.62,63,65 In two studies, both vitamin E and zinc from food sources were associated with decreased risk of melanoma.62,70 That both zinc and vitamin E are found in foods with a high fat content is noteworthy in light of the recent data on PUFAs and melanoma cells.68 However, when Mahabir et al71 evaluated the effects of 3 months of vitamin E supplementation on mutagen sensitivity among melanoma outpatients, in a randomized, placebo-control trial, the vitamin did not provide protection against bleomycin-induced chromosome damage.
A large Finnish nested case-control study72 of serum micronutrients reported lower serum alpha-tocopherol and beta-carotene levels in melanoma patients than corresponding controls, but the number of cases was too small to be meaningful. No other association was noted for retinol, retinol-binding protein, or selenium. Selenium may have protective effects against UV damage for skin cells, possibly through cellular antioxidant defenses.73 To date, in one mouse model, mice fed selenium had reduced melanoma metastases,74 whereas one Italian cohort exposed to high levels of environmental selenium had an excess of melanoma cases.75
Epidemiologic studies have supported a protective effect of green tea against skin cancer. Recently, green tea polyphenols were shown to inhibit erythema after UV radiation in humans.76 The treated skin had reduced DNA damage and fewer sunburn cells. The epigallocatechin-3-gallate and epicatechin-3-gallate polyphenolic fractions were most efficient at inhibiting erythema. The mechanisms of action of tea may include a combination of anti-inflammatory, antioxidant, and antiproliferative effects. Green tea may also modulate cytochrome p450 and be involved in the detoxification of enzymes, because green tea stimulates glutathione S-transferase.
Genistein, one of the principal soy isoflavones, is a phytoestrogen that has been associated with decreased incidence of breast and prostate cancers. In vitro, genistein was shown to inhibit the growth of melanoma tumor cell lines77 and also enhance cytotoxic T-cell and natural-killer cell activity.78 Its antioxidant properties could also be photoprotective.79
Although further data are needed on the above dietary factors, one should be aware that genetic make-up could influence the effects of diet on one’s cancer risk, because polymorphisms in specific genes can cause differences in the metabolism and detoxification of dietary factors and drugs.80
OTHER POTENTIAL CHEMOPREVENTION AGENTS
Other potential chemopreventive agents are listed in Table 1⇑.81,82 Difluoromethylornithine (DFMO), a suicide inhibitor of ornithine decarboxylase, a key enzyme involved in the biosynthesis of polyamines required for cell growth and divisions,82 could have chemopreventive activity. Polyamines are increased during tumor promotion, and polyamine inhibition inhibits neoplasia. Polyamine synthesis in the skin can be inhibited by topical application of DFMO.83 DFMO has been shown to have tumor-suppressive properties in several experimental systems, including melanoma (in vitro and in metastatic melanoma).84,85
The cyclo-oxygenase-2 (COX-2) inhibitors have generated interest in chemoprevention.86 The activity of COX results in biologically active prostaglandins, thromboxanes, and prostacyclins, which are believed to be implicated in carcinogenesis.87,88 COX activity is derived from two isozymes: COX-1 and COX-2. COX-1 is constitutively expressed, whereas COX-2 is induced by growth factors, cytokines, and tumor promoters and contributes to both inflammation and neoplasia.89 In melanoma, to date, one group reported COX-2 expression in malignant melanoma in 26 of 28 primary melanoma cell lines.90 In a case-control study, an inverse association of nonsteroidal anti-inflammatory drugs and malignant melanoma among women was reported.91 A decreased relative risk of malignant melanoma (relative risk, 0.45; 95% CI, 0.22 to 0.92; P < .05) was observed in 110 women with melanoma who had a regular intake of common over-the-counter nonsteroidal anti-inflammatory drugs (eg, aspirin and ibuprofen), as opposed to 609 female controls matched for age and place of residence. Note that adjustment for sun exposure did not change the magnitude of the estimate. In a hairless mouse model, a COX-2 inhibitor, celecoxib, significantly lengthened the tumor latency period and reduced tumor multiplicity (UV induced).92 In light of the growing interest in COX-2 inhibitors as chemopreventive agents, prospective randomized trials are needed to test the possible risk-reduction potential of these compounds.
Apomine, a novel biophosphonate ester, acting via the Ras signaling pathway, seems to have preclinical and clinical activity in melanoma.93 Topical application in a mouse model resulted in a 55% reduction in melanoma incidence. In nine patients with chemotherapy-resistant metastatic melanoma, oral administration provided disease stabilization in 2 patients and longer survival in one patient (11%)93.
Perillyl alcohol, a monoterpene isolated from essential oils (eg, lavendin, peppermint, spearmint, and cherries) has been shown to inhibit photocarcinogenesis in a nonmelanoma tumor murine model.94 It also has been shown to have anticarcinogenic and antitumor activity in murine tumor models, via inhibition of Ras farnesylation and a decrease in antigenic Ras protein levels. In a Tpras transgenic mouse model, topical application of perillyl alcohol showed a reduction in levels of Ras protein.41 The inhibition of the Ras signaling pathway by either apomine or perillyl alcohol merits further clinical research.
Finally, several advances have been made in the development of melanoma vaccines.95 Theoretically, given their favorable risk profile and postulated mechanism of action, vaccines could be potential chemoprevention candidates.8
HIGH-RISK PATIENT POPULATIONS
Clinically defined high-risk groups for the development of melanoma are well known. Xeroderma pigmentosum is a hereditary syndrome characterized by a deficiency in either nucleotide excision repair or capacity to replicate damaged DNA.96 In the xeroderma pigmentosum homozygous state, patients have a 1,000- to 5,000-fold increase in the risk of developing melanoma. Systemic administration of tretinoin97 and topical application of a DNA repair enzyme98 have been shown to diminish the incidence of nonmelanoma skin cancer, but no studies have been carried out using melanoma as an end point. Subsets of kindreds with familial clustering of dysplastic nevi or melanoma have an 80% to 100% chance of developing the disease. Although they only make up 5% to 10% of melanoma patients, they would be an appropriate group in whom to test efficacy of risk-reduction agents. Patients with more than 50 benign nevi are also a large group in whom such trials could be considered. Patients with a prior history of melanoma may be candidates for adjuvant clinical trials while chemoprevention data are simultaneously collected using reduction in incidence of a second melanoma as a secondary end point.49 Ideally, information about surrogate molecular markers of extremely high melanoma risk would be helpful. In these marker-positive individuals, therapy would be cost effective and might be targeted at the specific locus responsible for the individual’s risk.
MOLECULAR-TARGETED CHEMOPREVENTION TRIALS
Prospective trials of molecular-targeted agents for melanoma could represent a significant advance in our approach to melanoma chemoprevention (Table 1⇑). Although no such trials have been carried out to date, several agents that showed clinical antitumor activity in melanoma would be appropriate candidate agents. Because farnesylation is obligatory for Ras oncogenicity, farnesyl transferase inhibitors have been developed to block Ras-mediated signal transduction and Ras-induced tumorigenesis. The farnesyl transferase inhibitor SCH66336699 A is an inhibitor of N-ras gene cellular membrane binding and inhibits growth of melanoma cell lines. R115777 is a potent and selective nonpeptidomimetic inhibitor of farnesyl protein transferase100 both in vitro and in vivo. This agent is being investigated in metastatic melanoma.
SU5416101 is a small molecule that inhibits vascular endothelial growth factor receptor-2 and has antiangiogenic properties. It is being studied in melanoma, both with and without thalidomide, an agent that also has antiangiogenic properties.
Bcl-2 is an antiapoptotic factor, overexpressed in melanoma. Genasense102 is an anti-bcl-2 antisense oligonucleotide that downregulates bcl-2 and has antimelanoma activity. Apoptosis, blocked by bcl-2, could be eliminated by use of G3139. Other compounds including antitranscriptional agents wait in the wings for future trials.
In conclusion, in the United States, the incidence and mortality of melanoma has continued to increase despite nationwide sun protection programs. Recent knowledge is providing the opportunity to consider chemoprevention of this disease in its precancerous, in situ, or intraepithelial stages. The chemoprevention of melanoma would complement ongoing sun avoidance strategies.103 Prerequisites for this research include knowledge about the molecular mechanisms of UV melanomagenesis; availability of drugs that can target abnormalities; availability of animal models to test candidate prevention agents; availability of molecular and histologic markers to be used as surrogate end point biomarkers; compilation of epidemiologic, basic science, or in vitro data on potential chemoprevention candidate drugs; and availability of a high-risk patient population in which to carry out clinical chemoprevention trials. The preliminary, albeit incomplete, information now available in these areas should stimulate melanoma chemoprevention research.
Footnotes
-
Supported in part by the Carl J. Herzog Foundation (to M.F.D.).
- Received July 29, 2001.
- Accepted September 16, 2002.
