Building the perfect beast: complex mouse models teach surprisingly simple melanoma lessons
Year: 2009
Type: article
Abstract: When I was a child, I spend endless hours building scale models of almost anything. Typically, it was very frustrating trying to transform lifeless plastic pieces into a product that looked and felt like the real thing. Sometimes, the fit wasn’t correct, or the authenticity was lacking, and sometimes the overall setting just wasn’t right. Occasionally, however, with great effort, all the pieces would fall into place and I would hold in my hand a very impressive representation of the subject of my loving attention. Ironically, many years later, I find that my time is still occupied in the same dogged pursuit. It is, therefore, a pleasure to extol the virtues of the recent papers of Dankort et al. and Dhomen et al., who have ‘built’ what appears to be outstanding, scaled-down representations of human cutaneous malignant melanoma in mice genetically engineered to express in melanocytes mutationally activated BRAF. Expression of this oncogene in the absence of the tumor suppressor phosphatase and tensin homolog deleted on chromosome ten (PTEN) proved to be a particularly effective genetic combination, mimicking two molecular lesions faithfully detected in nearly a third of human melanomas. Surprisingly, little else seems to be required to create a highly malignant, metastatic tumor. BRAF is a serine/threonine-specific protein kinase and a member of the RAS→RAF→MEK→ERK MAP kinase signaling pathway. Genomic data demonstrating activating point mutations in 50 to 70% of human melanomas, predominantly BRAFT1799A encoding BRAFV600E, have provided compelling associative evidence of a critical role in the development of this potentially fatal disease. Remarkably, mutated BRAFV600E is also expressed in up to 80% of human nevi, including benign lesions, suggesting a very early initiating role. Inexplicably, however, initial attempts to validate the importance of BRAFV600E experimentally in mammalian melanomagenesis were far from successful. Inducible in vivo expression of a BRAFV600E transgene in mouse melanocytes produced very few melanomas (Jeong et al., 2008), and ectopic expression of BRAFV600E in human melanocytes was much less effective at generating melanomas in skin grafts compared to the NRASQ61K mutation (Chudnovsky et al., 2005), also frequently detected in human melanoma. Our own data confirmed the relative robustness of NRASQ61K in transforming cultured immortalized mouse melanocytes compared to BRAFV600E (L. Ha and G. Merlino, unpublished data). Indeed, BRAFV600E expression has been shown to induce cell cycle arrest with features of senescence in cultured human melanocytes (Mooi and Peeper, 2006). Now these two new papers show that Cre-mediated activation of a BRAFV600E knock-in allele by tyrosinase promoter-driven expression of a tamoxifen-inducible CreERT2 recombinase is able to reproducibly generate impressive melanocytic lesions. The timing or level of BRAFV600E expression may have been critical, but either way these data constitute irrefutable vindication of a key role for mutated BRAF in melanomagenesis. A relatively minor difference between the results obtained in the two papers concerns the degree to which BRAFV600E can promote melanomagenesis without additional genetic modification. Dhomen et al. report that Cre-mediated activation of a single BRAFV600E allele is sufficient to produce frank melanomas, albeit with a median latency of 12 months and without a metastatic phenotype. Dankort et al. show that while expression of the oncogenic BRAFV600E can strongly incite hyperplastic lesions, it is insufficient to trigger development of a fully malignant melanoma. Notably, Dankort et al. did show that the gene dosage of BRAFV600E significantly influenced the size and pigmentation of benign hyperplastic lesions, in alignment with observed increases in oncogenic BRAF copy number in human melanoma. The reason for the variance between the two results is not clear, although subtle differences in the design of the conditional BRAF alleles, the behavior of the two different tyrosinase-CreERT2 transgenes, or the genetic background of the mice under study could be the culprit. Dankort et al. went on to show that the consequences of expression of BRAFV600E in melanocytes could be greatly amplified through deficiency in PTEN, a tumor suppressor implicated in hindering the development of numerous cancers, including melanoma. The PTEN protein has both lipid phosphatase and protein phosphatase activities: the lipid phosphatase activity counteracts the function of phosphatidylinositol-3 kinase (PI3K) and downstream AKT and has been implicated in regulating cell-cycle progression and apoptosis, while the protein phosphatase activity helps regulate focal adhesion formation, cell spreading, and migration. The consequences of pairing mutant BRAF and PTEN deficiency were striking; mouse melanomas developed with 100% penetrance, short latency (25–50 d), and with metastases to lymph nodes and lungs. Critically, this same genetic combination has been found in approximately 30% of human melanoma, and the percentage is significantly higher if AKT activation downstream of PTEN is included. In contrast, and of equal interest, Dhomen et al. reports that deficiency in INK4A, encoding the cyclin-dependent kinase inhibitor p16INK4A, is not required for BRAF-induced melanomagenesis, although the absence of INK4A does reduce melanoma latency and enhance multiplicity to some extent. They also show that wildtype INK4A is not needed for BRAFV600E-induced nevi formation. A molecular explanation for the impressive efficacy of the BRAF/PTEN combination remains to be elucidated, but these new models may prove to be extremely useful in this regard. Dankort et al. also employed their mouse model to assess the anti-melanoma efficacy of PD325901 and Rapamycin, drugs known to inhibit MEK1/2 and mTorc1, respectively. These protein kinases act in pathways downstream of BRAF and PI3K, respectively, and would be expected to be essential for growth and/or survival of melanomas arising in a BRAF/PTEN model of melanoma. As single agents, each drug was able to prevent the appearance of melanomas in response to BRAF activation and PTEN silencing. Perhaps of greater import, the combination of PD325901 and Rapamycin was able to shrink the size of already existing mouse melanomas significantly. These data indicate at a minimum that this mouse, with its high penetrance and short latency, will be quite valuable as a preclinical melanoma model, and also provide encouragement for future combination treatments that target both the RAS→RAF→MEK→ERK and the PI3K→AKT→mTOR pathways (Bedogni et al., 2006). Naturally, problems persist as we all continue to learn to improve modeling human disease in the genetically tractable mouse. Mouse skin, where most melanocytes are found in hair follicles, is histologically distinct from human skin, in which many melanocytes reside in the keratinocyte-laden epidermis. Consequently, mouse melanomas typically arise from the dermis, while human melanomas tend to arise from the epidermis; these new models have not solved this problem. Tamoxifen induction of CreERT2 is a powerful and versatile tool that affords exquisite spatial and temporal control over the genetic alterations implicated in melanomagenesis, but tends to simultaneously activate legions of melanocytes, rather than individual melanocytes as might be expected for stochastic mutagenesis. Finally, the most potent path to melanoma was found to be through the simultaneous activation/inactivation of BRAF/PTEN, respectively, through the same elegant conditional recombination mechanism. However, such simultaneous genetic damage would rarely if ever happen in patients. These models could be, and I am sure will be, appropriately modified, as it is also becoming exceedingly apparent that the order of genetic and epigenetic changes is crucial to the natural, successful evolution of a tumor. Nevertheless, given that the term ‘perfect model’ is an oxymoron, something I learned long ago in childhood working with bits of molded plastic, one is hard-pressed to quibble over the exactness of these new exciting mouse melanoma models, which will undoubtedly prove to be of great value to future basic and translational melanoma research efforts.
Source: Pigment Cell & Melanoma Research
Author: Glenn Merlino
Institution: National Institutes of Health
Language: English
Topic: Melanoma and MAPK Pathways
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: hybrid
Estimated APC paid: $5,000
OpenAlex ID: W2093924594