The Cancer Modeling Section seeks to elucidate the complex molecular/genetic program governing tumor genesis and progression through the development and analysis of genetically engineered mouse models of human cancer. Our efforts in this regard are focused on two tumor types, cutaneous malignant melanoma and the pediatric malignancy rhabdomyosarcoma (RMS). Exposure to UV radiation is a causal agent in the vast majority of melanoma. Retrospective epidemiological data have suggested that melanoma is provoked by intermittent, intense exposure to UV, particularly during childhood. Previously, we tested this hypothesis in transgenic mice in which the receptor tyrosine kinase MET was deregulated by virtue of ectopic expression of its ligand, hepatocyte growth factor/scatter factor (HGF/SF). We discovered that a single neonatal dose of burning UV radiation in these mice was necessary and sufficient to induce tumors reminiscent of human melanoma with shortened latency (Noonan et al., Nature 413: 271-2, 2001). A critical role for the INK4A/ARF locus, which helps regulate the RB and p53 pathways and is widely regarded as a key melanoma suppressor in human patients, was confirmed in our animal model by demonstrating that UV-induced melanoma was significantly accelerated in Ink4a/Arf-deficient mice (Recio et al., Cancer Res. 62: 6724-30, 2002). These results strongly suggest that sunburn is a highly significant risk factor in kindreds harboring germline mutations in INK4A/ARF (Merlino and Noonan, Trends Mol. Med. 9: 102-8, 2003). HGF/SF mice harboring an oncogenic mutation in CDK4, carried by some melanoma-prone kindreds, also exhibited highly accelerated melanoma genesis (Tormo et al., Am. J. Pathol. 169: 665-72, 2006). There has been controversy surrounding the relative risks associated with UVB versus UVA radiation. We used albino HGF/SF transgenic mouse to show that UVB, but not UVA, alone is able to induce the full melanoma phenotype (DeFabo et al., Cancer Res. 64: 6372-6, 2004). Recent data also suggest that UV-induced mouse melanoma can by prevented by SPF15 sunscreen, a finding that addresses a controversy in the field associated with the use of sunscreen for skin cancer prevention. This result demonstrates the potential value of our model system for the development of more effective melanoma prevention strategies. RMS, accounting for up to 10% of all pediatric neoplasms and for more than 50% of pediatric soft tissue sarcomas, is believed to arise from imbalances in skeletal muscle cell proliferation and differentiation. However, molecular pathways associated with RMS are poorly understood, due in part to the lack of an RMS-prone animal model. We have discovered that virtually all HGF/SF transgenic, Ink4a/Arf-deficient mutant mice rapidly succumbed to highly invasive RMS (Sharp et al., Nature Med. 8: 1276-80, 2002). Comparable molecular lesions have also been described for human RMS. These data provide genetic evidence that c-MET and INK4A/ARF pathways represent critical and synergistic targets in RMS pathogenesis, and suggest a rational therapeutic combination to combat this pediatric cancer. A panel of highly and poorly metastatic cell lines generated from RMS tumors arising in our mouse model was used in concert with expression profiling to identify a set of genes associated with enhanced metastasis. Functional in vivo studies confirmed that the cytoskeletal linker Ezrin and the homeodomain-containing transcription factor Six1 have essential roles in determining the metastatic fate of RMS cells (Yu et al., Nature Med. 10: 175-81, 2004). Notably, EZRIN and SIX1 expression levels were also both enhanced in human RMS tissue, significantly correlating with clinical stage. Subsequent analyses showed that the Ezrin gene was in fact a direct transcriptional target of Six1, and indispensable for Six1-induced metastasis (Yu et al., Cancer Res. 66: 1982-9, 2006). Further analysis revealed that Six1 regulates Ezrin expression, at least in part, through epigenetic modification of Ezrin gene chromatin, including inducing changes in histone tail methylation and acetylation. Ezrin appears to represent a very promising therapeutic target for patients with advanced stage RMS. The relevance of the p53- and pRb-tumor suppressor pathways to the development of most, if not all types of cancer is unequivocal. However, critical questions remain concerning the relative roles of specific p53- and pRb-pathway members, and how these roles vary among tumors arising from distinct lineages. In particular, functional differences between p53 and its indirect positive regulator ARF in the suppression of tumorigenesis have not been clarified; p53 is degraded by MDM2, an activity that can be blocked by ARF. Enigmatically, although the most commonly mutated gene in human cancer, p53 does not appear to play a significant role in the development of melanoma. In contrast to p53, ARF is frequently inactivated in melanoma through loss of the INK4A/ARF locus, long associated with heritable susceptibility to melanoma, as discussed above. Using a series of animal- and cell culture-based models of melanoma we have found that oncogene-induced senescence, a barrier against early tumor progression, can be overcome by a deficiency in the tumor suppressor ARF, but not p53, facilitating rapid development of melanoma (Ha et al., Proc. Natl. Acad. Sci. 104: 10968-73, 2007). Accordingly, oncogenic NRAS was found to collaborate with a deficiency in ARF, but not p53, to fully transform melanocytes in vitro. We also discovered that Arf helps regulates p53-independent melanocyte senescence by blocking E2F1 function. Taken together, our data help explain in human melanoma the relative abundance and paucity of mutations in ARF and TP53, respectively, and demonstrate that ARF and p53, although linked in a common pathway, suppress tumorigenesis through diverse, lineage-dependent mechanisms. Thus, therapeutics currently being designed to restore wild-type p53 function may be insufficient to counter melanoma and other malignancies in which ARF holds p53-independent tumor suppressor activity. Although an extensive accumulation of epidemiological evidence supports a fundamental role for UV in melanoma, the specific UV-affected molecular pathways and mechanisms remain largely unidentified, limiting the development of successful prevention/treatment strategies. In fact, few specific UV signature mutations such as C to T transitions have been found in genes thought to contribute to melanoma. These uncertainties suggest that mechanisms other than UV-induced DNA mutagenesis are involved in melanoma initiation. We have proposed that an important component of UV-induced melanomagenic alterations may be epigenetic in nature. To test this and other hypotheses, we have recently developed an experimental mouse model that allows melanocytes, specifically and inducibly labeled with GFP, to be highly purified from disaggregated mouse skin by FACS following UV irradiation in vivo. We have already identified a pattern of persistent gene expression changes in melanocytes isolated from mice exposed to UVB, consistent with UV-induced epigenetic modifications such as chromatin methylation. We anticipate that this in vivo model will provide novel insights into the nature of UV-induced damage, and the mechanisms by which UV provokes melanoma
Showing the most recent 10 out of 14 publications
