Introduction

At the 2002 American Association for Cancer Research's Annual Meeting, 'global profiling of cancer' was among the top three most active areas of ongoing research for attendees, and to date, more than 2000 papers have been published using microarray analysis. The rapid development of this field may be attributed to the encouraging and exciting results from the earliest microarray experiments (Schena et al., 1995; DeRisi et al., 1996), as well as the significant advancements in the technology during the past decade. The microarray formats have become more comprehensive in terms of the number of genes printed on the microscope slides (from 100 cDNAs in 1996 to 40 000 cDNAs in some arrays today). As a consequence of the size and complexity of these data, collaborations between scientists in previously disparate fields, such as genetics, mathematics, molecular biology, and clinical medicine, have become necessary in order to identify the relevant relationships present in the increasingly large data sets. Subsequently, the higher resolution views of gene activity have allowed a variety of human cancers to be 'classified' at the molecular level. Studies are beginning to emerge that use these novel classification models to predict clinical outcome and to propose new molecular targets for therapy (Alizadeh et al., 2000; Dhanasekaran et al., 2001; Khan et al., 2001; MacDonald et al., 2001; Rosenwald et al., 2002; Shipp et al., 2002; Sorlie et al., 2001; van't Veer et al., 2002). Although such predictions overlap with those produced by traditional clinical data, in some cases certain predictions may be derived only from studies using gene expression patterns (Alizadeh et al., 2000; Rosenwald et al., 2002; Shipp et al., 2002; van't Veer et al., 2002). For the diagnosis and treatment of patients with melanoma, this type of genetic information would be a significant advance.

Expression profiles and cutaneous melanoma

The molecular pathogenesis of melanoma remains a mystery. Specifically, the important genetic changes responsible for melanoma progression are poorly characterized, and the heterogeneity of the clinical course of the disease is unexplained. As a consequence, melanoma continues to be an unpredictable cancer. Given its potential for aggressive growth, and its refractoriness to available chemotherapeutic agents, it is no surprise that long-term survival for patients with metastatic melanoma has not improved since the 1970s (Ries et al., 2001). Facing such limited treatment options and bleak survival statistics, patients with melanoma are frequently willing to tolerate toxic and experimental therapies as their only hope for extended life or potential cure.

Incorporating genomic technology into melanoma research is a valuable strategy to identify novel disease-related genes. The importance of applying genomic approaches to the study of cutaneous melanoma is highlighted by the recent publication from the ambitious Sanger Institute's Cancer Genome Project that showed that the gene B-RAF is widely mutationally activated in melanoma (Davies et al., 2002). Although the current review is focused on genome-wide measurements of gene expression in melanoma, this B-RAF publication deserves special mention, as this new information may expedite the development of new pharmacologic therapies for melanoma (Figure 1).

Figure 1.

Proposed B-RAF signal transduction. B-RAF acts downstream of Ras in the MAPK pathway. Extracellular signals are represented in pink. The interaction between the G-protein-coupled receptor pathways and the MAPK cascade is unclear in the melanocyte, but in other cell types B-RAF is involved

Full figure and legend (43K)

WNT5A – a new disease gene for melanoma progression

Since metastatic melanoma does not have fixed histopathological subclasses, it is an attractive target for expression profiling studies aimed at identifying molecular subclasses. A survey of expression patterns from melanoma tumors and cell lines using cDNA microarrays showed discernible subclasses of melanoma specimens (Bittner et al., 2000). Combinations of analytic methods were applied to first group the samples without any presuppositions (unsupervised class discovery) and then to rank the genes based on the difference of behavior of the genes in the classes (supervised linear discriminant analysis). Many of the genes that created the distinct classes were important for cell motility and invasive ability. Specifically, WNT5A was identified as the gene that best defined the new subclasses of tumors (Figure 2). In vitro analysis of melanoma cell lines differing in WNT5A expression levels revealed that increased expression correlated with the increased motility and invasiveness of the cell. Although the role of WNT5A signaling in melanoma progression had not been suggested previously, these results hint at the possibility that the clinical course of the patients whose tumor specimens has decreased WNT5A expression levels would have less aggressive tumors and longer survival, while the subset of tumors with overexpression of WNT5A would have highly aggressive tumors and poor clinical outcome.

Figure 2.

A supervised algorithm, weighted gene analysis, was used to identify a list of genes that discriminate between the subclasses of melanoma tumors. The top 22 genes obtained by the analysis are listed in order of decreasing weight, and WNT5A was identified as the gene that best discriminates between the new subclasses of tumors

Full figure and legend (52K)

To examine further the possibility that WNT5A is an important signaling molecule in metastatic melanoma and therefore of possible clinical significance, Weeraratna et al. (2002) extended the results of the original microarray experiments and demonstrated that WNT5A expression levels can have a significant effect on the motility characteristics of the cell. Using several in vitro and in vivo methods, they showed that overexpression of WNT5A in human melanoma cells resulted in cells with greater invasive ability as compared to parental melanoma cells that have low basal levels of WNT5A, and that the invasive ability is likely mediated through specific protein kinase C pathways (PKC), which are thought to be associated with cytoskeletal organization and invasion. As additional support for the idea that this pathway is responsible for the invasive phenotype of the melanoma cells, the authors demonstrated that inhibition of this pathway by desensitization of the WNT5A receptor, Frizzled5, by an antibody that interfered with activation by WNT5A, resulted in a decrease in the activation of the presumptive PKC pathway and the inhibition of in vitro motility and invasion phenotype of the melanoma cells. Moreover, the encouraging results from the in vitro experiments were mirrored in a small series of tumor samples. WNT5A protein expression in human melanoma biopsies directly correlates with increasing tumor grade while inversely correlating with patient survival (Figure 3).

Figure 3.

(a) WNT5A protein expression in human melanoma biopsies correlates directly with increasing tumor grade, and inversely with patient survival. WNT5A negativity is represented by blue, moderate positivity for WNT5A is shown in orange, and strong WNT5A staining is represented in red. The majority of benign tumors (NEVI) stain negative for WNT5A. WNT5A staining of primary melanomas is heterogeneously distributed, while most of the metastatic melanoma tumors demonstrate strong WNT5A staining. (b) Tissues were stained with the antibody to WNT5A, and secondary-stained with Cy-5 (red staining) and counterstained with DAPI (blue staining). The distribution of the protein among tissue and cell types was observed. WNT5A is not expressed in the majority of nevi (B, C). In primary melanomas, the staining was more heterogenous where some tumors stained only focally positive for WNT5A (D) and others stained strongly throughout the tumor (E). Melanoma cells in vertical growth phase were strongly positive for WNT5A at the leading edge of the tumor (F). Metastases exhibited strongly positive staining (G), which was specific to the melanoma cells and not surrounding cell types such as lymphocytes (H). Upon increased magnification of cells exhibiting giant cell morphology, a phenotype that is highly associated with malignancy, WNT5A staining is strongly positive, as it usually is in these cases (I). In this particular biopsy, the time to metastasis was 17 months, and the patient did not survive. Interestingly, in a histopathologically similar tumor there was a very focal WNT5A positivity, with mostly negative staining throughout the tumor (J), the patient did not present with metastases for over 9 years, and is still alive today

Full figure and legend (68K)

These studies of WNT5A suggest that this Wnt family member may play an important role in the metastatic progression of melanoma cells. In addition, the direct correlation between WNT5A expression and the motility phenotype suggests that the WNT5A ligand/FRIZZLED5 receptor pair may be a viable therapeutic target for the inhibition of melanoma tumor progression.

Rho C – a potential antitumor target

A clever study by Clark et al. (2002) used oligonucleotide arrays to identify novel genes that were specific for melanoma metastases. Using a parental cell line with weak metastatic potential and in vivo selection scheme to derive a robust metastatic cell line, the gene expression profiles of the subsequent metastatic pulmonary tumors were compared to the primary tumors. A total of 32 of the approx 7000 genes analysed had enhanced expression levels (>2.5-fold) in the metastases as compared to the primary tumors. The most interesting genes expressed in the highly invasive tumors included genes that regulate cytoskeletal organization and cell migration, an observation shared by the study of Bittner et al.

Surprisingly, one gene, Rho C, enhanced the motility and invasive capability of poorly metastatic cells so markedly that it appears to be a sufficient stimulus for metastasis on its own. In addition, the enhanced motility and invasive ability of the melanoma cells could be suppressed by introducing a dominant inhibitor to the Rho signaling pathway. RHO C is a member of the Rho GTPase family of proteins that are important regulators of several cellular processes, including the actin cytoskeleton, cellular adhesion, motility, vesicle transport, cell-cycle progression, cytokinesis, and transcription (Ridley, 2001). The specific effector proteins that RHO C activated to enable the melanoma cells to exit the blood and colonize the lungs in this mouse model are unclear, but recent studies have demonstrated a relation between the Rho proteins and PKC family of serine–threonine kinases. Notably, RHO C is known to activate PKD (PKC ), which is associated with cytoskeletal changes and could potentially increase the motility of the cell (Yuan et al., 2000; Palmantier et al., 2001). Taken together with the WNT5A data, the merging of these signaling pathways to activate PKC provides a plausible mechanism for the enhanced cellular motility found in both data sets of highly invasive melanoma cells. Another possible explanation for this finding is suggested by the recent finding that Rho C overexpression induces the increased expression of angiogenic factors in breast epithelial cells (Sahai and Marshall, 2002).

Recent studies have linked Rho C overexpression to pancreatic ductal adenocarcinoma and inflammatory breast cancer, but the study by Clark et al. is the first to connect overexpression of Rho C to melanoma progression. In addition, it illustrates the potential of gene expression profiling to identify new contributors to melanoma. It is predicted that the Rho proteins will make valid therapeutic targets (Sahai and Marshall, 2002).

Mitf–Bcl-2 – an experimental antisense therapy

A recent article by McGill et al. (2002) used oligonucleotide arrays to identify the transcriptional targets of the microphthalmia gene, Mitf, a transcription factor essential for the melanocyte lineage and subsequently expressed in most primary melanomas (King et al., 1999). One of the identified targets was Bcl-2, a potent antiapoptotic factor, whose expression level was shown to be regulated in melanocytes by Mitf and critical for melanocyte and melanoma survival. These data provide good evidence that the Mitf–Bcl-2 pathway is important for survival of melanoma cells and may help to explain why melanoma tumors are resistant to cytotoxic chemotherapeutic agents. This study provides support for Mitf–Bcl-2 as a potential target for therapeutic intervention. In fact, a phase III clinical trial is underway in Europe that uses BCL-2 antisense oligonucleotides with decarbazine to treat advanced melanoma (Jansen et al., 2000).

Potential for treatment predictions

Predicting treatment outcomes for patients living with cancer remains one of the most challenging aspects of patient care and an area clearly in need of advancements. To date, one study has utilized gene-expression profiling in melanoma to search for molecular predictors of treatment response (Wang et al., 2002). Cytokine therapy with high-dose interleukin-2 (IL-2), pioneered by Steven Rosenberg at the National Cancer Institute, has caused dramatic responses and durable remissions in a minority of patients with metastatic melanoma, for whom standard therapy had failed (Rosenberg et al., 1985,1987; Lotze et al., 1986). Although the toxic side effects associated with the administration of high-dose IL-2 are significant, it remains the only therapeutic agent that can evoke a cure for metastatic disease. Unfortunately, the mechanism of the immune attack on melanoma tumors remains unclear, and clinical characteristics are insufficient to predict the minority of patients who will benefit from this treatment.

To identify predictors of immune responsiveness, Wang et al. (2002) compared the gene expression patterns of subcutaneous melanoma metastases before and after various immunotherapy treatments (the majority of which included systemic IL-2 administration) that were associated with a complete response, partial response, stable disease, or no response to treatment. The overall gene-expression patterns from the melanoma tumors did not demonstrate subclasses based solely on clinical response to immunotherapy.

One interesting observation of the data was that in the majority of melanoma tumors, the gene-expression patterns in two tumor samples from the same patient – sampled before and after treatment – were more similar to each other than to any other of the tumors in the data set. This observation suggests that the global molecular signature of a metastatic tumor is largely retained following a treatment regimen. One interpretation of this observation is that the metastatic tumors that were sampled following treatment (or following a recurrence) represent tumor cells that survived the treatment and re-established the metastatic process. Of interest, Perou et al. (2000) have observed a similar finding in breast cancer. Remarkably, the gene-expression profiles of primary breast tumors appeared most similar to the metastatic tumors from the same patient than either was to the other primary breast tumors in the study. Again, this observation may represent the regrowth of the primary tumor from microscopic cells left behind following surgery or cells that survived adjuvant treatment.

For gene-expression analysis, especially for cutaneous melanoma, this observation is important because one of the limitations for performing expression analysis (and other genomic assays) is tissue scarcity for early-stage disease. Generally, the pathologist uses the whole specimen in order to make the histologic diagnosis of a primary melanoma, thus making the primary tumor unavailable for further investigations. However, given the data presented above, much can be learned about the genetics of the primary tumor from studying its metastatic counterpart.

Conclusion

The thousands of hybridization experiments carried out thus far have identified many new and interesting genes that are proposed as central to the pathophysiology of human cancer. Time and testing will reveal the accuracy of the assessment of biologic relevance and the clinical utility of these genes. Using gene-expression profiling to study cutaneous melanoma is beginning to yield new insights into the molecular genetics of this unpredictable and aggressive cancer. The results will be important for the complete understanding of melanoma tumors and equally important for the development of new treatments. In the future, gene-expression studies may be incorporated into clinical trials with the hope that a patient's genomic information will lead to tailored treatment regimes.

References

1. Alizadeh AA, Eisen MB, Davis RE, Ma C, Lossos IS, Rosenwald A, Boldrick JC, Sabet H, Tran T, Yu X, Powell JI, Yang L, Marti GE, Moore T, Hudson Jr J, Lu L, Lewis DB, Tibshirani R, Sherlock G, Chan WC, Greiner TC, Weisenburger DD, Armitage JO, Warnke R, Levy R, Wilson W, Grever MR, Bryd JC, Botstein D, Brown PO and Staudt LM. (2000). Nature, 403, 503–511. | Article | PubMed | ISI | ChemPort |
2. Bittner M, Meltzer P, Chen Y, Jiang Y, Seftor E, Hendrix M, Radmacher M, Simon R, Yakhini Z, Ben-Dor A, Sampas N, Dougherty E, Wang E, Marincola F, Gooden C, Lueders J, Glatfelter A, Pollock P, Carpten J, Gillanders E, Leja D, Dietrich K, Beaudry C, Berens M, Alberts D, Sondak V, Hayward N and Trent J. (2000). Nature, 406, 536–540. | Article | PubMed | ISI | ChemPort |
3. Clark EA, Golub TR, Lander ES and Hynes RO. (2002). Nature, 406, 532–535.
4. Davies H, Bignell GR, Cox C, Stephens P, Edkins S, Clegg S, Teague J, Woffendin H, Garnett MJ, Bottomley W, Davis N, Dicks E, Ewing R, Floyd Y, Gray K, Hall S, Hawes R, Hughes J, Kosmidou V, Menzies A, Mould C, Parker A, Stevens C, Watt S, Hooper S, Wilson R, Jayatilake H, Gusterson BA, Cooper C, Shipley J, Hargrave D, Pritchard-Jones K, Maitland N, Chenevix-Trench G, Riggins GJ, Bigner DD, Palmieri G, Cossu A, Flanagan A, Nicholson A, Ho JW, Leung SY, Yuen ST, Weber BL, Seigler HF, Darrow TL, Paterson H, Marais R, Marshall CJ, Wooster R, Stratton MR and Futreal PA. (2002). Nature, 417, 949–954. | Article | PubMed | ISI | ChemPort |
5. DeRisi J, Penland L, Brown PO, Bittner ML, Meltzer PS, Ray M, Chen Y, Su YA and Trent JM. (1996). Nat. Genet., 14, 457–460. | Article | PubMed | ISI | ChemPort |
6. Dhanasekaran SM, Barrette TR, Ghosh D, Shah R, Varambally S, Kurachi K, Pienta KJ, Rubin MA and Chinnaiyan AM. (2001). Nature, 412, 822–826. | Article | PubMed | ISI | ChemPort |
7. Duggan DJ, Bittner M, Chen Y, Meltzer P and Trent JM. (1999). Nat. Genet., 21, 10–14. | Article | PubMed | ISI | ChemPort |
8. Jansen B, Wacheck V, Heere-Ress E, Schlagbauer-Wadl H, Hoeller C, Lucas T, Hoermann M, Hollenstein U, Wolff K and Pehamberger H. (2000). Lancet, 356, 1728–1733. | Article | PubMed | ISI | ChemPort |
9. Khan J, Wei JS, Ringner M, Saal LH, Ladanyi M, Westermann F, Berthold F, Schwab M, Antonescu CR, Peterson C and Meltzer PS. (2001). Nat. Med., 7, 673–679. | Article | PubMed | ISI | ChemPort |
10. King R, Weilbaecher KN, McGill G, Cooley E, Mihm M and Fisher DE. (1999). Am. J. Pathol., 155, 731–738. | PubMed | ISI | ChemPort |
11. Lotze MT, Chang AE, Seipp CA, Simpson C, Vetto JT and Rosenberg SA. (1986). JAMA, 256, 3117–3124. | PubMed |
12. MacDonald TJ, Brown KM, LaFleur B, Peterson K, Lawlor C, Chen Y, Packer RJ, Cogen P and Stephan DA. (2001). Nat. Genet., 29, 143–152. | Article | PubMed | ISI | ChemPort |
13. McGill GG, Horstmann M, Widlund HR, Du J, Motyckova G, Nishimura EK, Lin YL, Ramaswamy S, Avery W, Ding HF, Jordan SA, Jackson IJ, Korsmeyer SJ, Golub TR and Fisher DE. (2002). Cell, 109, 707–718. | Article | PubMed | ISI | ChemPort |
14. Palmantier R, George MD, Akiyama SK, Wolber FM, Olden K and Roberts JD. (2001). Cancer Res., 61, 2445–2452. | PubMed | ISI | ChemPort |
15. Perou CM, Sorlie T, Eisen MB, van de Rijn M, Jeffrey SS, Rees CA, Pollack JR, Ross DT, Johnsen H, Akslen LA, Fluge O, Pergamenschikov A, Williams C, Zhu SX, Lonning PE, Borresen-Dale AL, Brown PO and Botstein D. (2000). Nature, 406, 747–752. | Article | PubMed | ISI | ChemPort |
16. Quackenbush J (2001). Nat. Rev. Genet., 2, 418–427. | Article | PubMed | ISI | ChemPort |
17. Ramaswamy S and Golub TR. (2002). J. Clin. Oncol., 20, 1932–1941. | PubMed | ISI | ChemPort |
18. Ridley AJ. (2001). Trends Cell Biol., 11, 471–477 | Article | PubMed | ISI | ChemPort |
19. Ries LAG, Eisner MP, Kosary CL, Hankey BF, Miller BA, Clegg L and Edwards BK. (2001). SEER Cancer Statistics Review, 1973–1998, Vol. 2001, National Cancer Institute: Bethesda, MD, USA.
20. Rosenberg SA, Lotze MT, Muul LM, Chang AE, Avis FP, Leitman S, Linehan WM, Robertson CN, Lee RE, Rubin JT, Seipp CA, Simpson CG and White DE. (1987). N. Engl. J. Med., 316, 889–897. | PubMed | ISI | ChemPort |
21. Rosenberg SA, Lotze MT, Muul LM, Leitman S, Chang AE, Ettinghausen SE, Matory YL, Skibber JM, Shiloni E, Vetto JT, Seipp CA and Reichert CM. (1987). N. Engl. J. Med., 313, 1485–1492.
22. Rosenwald A, Wright G, Chan WC, Connors JM, Campo E, Fisher RI, Gascoyne RD, Muller-Hermelink K, Smeland EB and Staudt LM. (2002). N. Engl. J. Med., 346, 1937–1947. | Article | PubMed | ISI |
23. Sahai E and Marshall CJ. (2002). Nat. Rev. Cancer, 2, 133–142. | Article | PubMed | ISI |
24. Schena M, Shalon D, Davis RW and Brown PO. (1995). Science, 270, 467–470. | Article | PubMed | ISI | ChemPort |
25. Shipp MA, Ross KN, Tamayo P, Weng AP, Kutok JL, Aguiar RC, Gaasenbeek M, Angelo M, Reich M, Pinkus GS, Ray TS, Koval MA, Last KW, Norton A, Lister TA, Mesirov J, Neuberg DS, Lander ES, Aster JC and Golub TR. (2002). Nat. Med., 8, 68–74. | Article | PubMed | ISI | ChemPort |
26. Sorlie T, Perou CM, Tibshirani R, Aas T, Geisler S, Johnsen H, Hastie T, Eisen MB, van de Rijn M, Jeffrey SS, Thorsen T, Quist H, Matese JC, Brown PO, Botstein D, Eystein Lonning P and Borresen-Dale AL. (2001). Proc. Natl. Acad. Sci. USA, 98, 10869–10874. | Article | PubMed | ISI | ChemPort |
27. van't Veer LJ, Dai H, van de Vijver MJ, He YD, Hart AA, Mao M, Peterse HL, van der Kooy K, Marton MJ, Witteveen AT, Schreiber GJ, Kerkhoven RM, Roberts C, Linsley PS, Bernards R and Friend SH. (2002). Nature, 415, 530–536. | Article | PubMed | ISI | ChemPort |
28. Wang E, Miller LD, Ohnmacht GA, Mocellin S, Perez-Diez A, Petersen D, Zhao Y, Simon R, Powell JI, Asaki E, Alexander HR, Duray PH, Herlyn M, Restifo NP, Liu ET, Rosenberg SA and Marincola FM. (2002). Cancer Res., 62, 3581–3586. | PubMed | ISI | ChemPort |
29. Weeraratna AT, Jiang Y, Hostetter G, Rosenblatt K, Duray P, Bittner M and Trent JM. (2002). Cancer Cell, 1, 279–288. | Article | PubMed | ISI | ChemPort |
30. Yuan J, Slice L, Walsh JH and Rozengurt E. (2000). J. Biol. Chem., 275, 2157–2164. | PubMed |