E2F1 in Melanoma Progression and Metastasis

lated and inactive in melanoma cells from metastatic lesions ( 28 ), E2F1 is expressed at considerably higher levels in metastases than in primary tumors of melanoma patients ( 13 , 29 ). Based on these fi ndings, we have investigated the role of E2F1 in cancer pro-gression and characterized its molecular ac-tion in different human melanoma cell lines. To determine the effect of endoge-nously high E2F1 levels ( Figure 1, A , left) on invasiveness, highly metastatic mela-noma cell lines SK-Mel-103 and SK-Mel-147 (a gift from Dr M. Soengas, Department of Dermatology, University of Michigan, Comprehensive Cancer Center) as verifi ed by xenograft experiments ( 30 , 31 ) were infected with an adenoviral vector express-ing gene-specifi c small hairpin RNA (shRNA), shown to effi ciently knock down E2F1 ( Figure 1, A , right), and tested for migration behavior using Boyden chamber and scratch assays. In both cell lines, cells expressing shE2F1 had reduced cell inva-sion and motility compared with cells expressing control shRNA ( Supplementary Figure E2F1-induced invasiveness, emphasizing the importance of the E2F1 1 , available online). Growth rates of SK-Mel-147 cells expressing E2F1 shRNA or control shRNA and nontransfected cells were similar, as measured by viability assay. Loss of E2F1 expression had no inhibitory effect on G1-to-S phase progression, indi-cating that E2F1 is not required for the proliferation of metastatic melanoma cells. Because epithelial-to-mesenchymal transi-tion is considered a key step for progression to metastatic stage ( 32 ), and loss of E-cadherin is frequently found in meta-static tumors ( 33 ), we examined whether suppression of E2F1 is associated with alterations of E-cadherin. We observed an increase of membranous E-cadherin expression in SK-Mel-147 cells when E2F1 activity was blocked, suggesting its involvement in E2F1-mediated invasion ( Supplementary Figure 1 , available online). SK-Mel-103 and SK-Mel-147 cells expressing either E2F1 shRNA or control shRNA were implanted subcutaneously to generate localized xenografts or intravenously to induce disseminated tumors. Although both cell lines, irrespective of whether they expressed control shRNA or E2F1 shRNA, formed primary tumors that grew at similar rates ( Figure 1, B ), E2F1-knockout cells Tumor metastasis continues to be the most important problem in the field of cancer. Patients who present with metastatic disease or those who develop metastases after successful management of the primary tumor carry a universally grave prognosis. To improve treatment outcomes for these patients, a broader understanding of the determinants of this process is necessary. The E2F1 transcription factor is con-sidered to be the “fi nal frontier” of the G1-to-S phase boundary that is tightly regulated through the retinoblastoma protein (Rb). Inactivation of Rb liberates E2F1 from the suppressive complex, which, in turn, induces continuous expression of target genes whose products promote cell cycle progression ( 1 , 2 ). Moreover, E2F1 can suppress differentiation of a variety of cell types, including keratinocytes and chondrocytes ( 3 , 4 ). Ectopic expression of E2F1 results in neoplastic transformation of rodent cells ( 5 , 6 ), and fi ndings from transgenic models indicate that increased E2F1 activity is associated with tumor de-velopment in several tissues ( 7 – 9 ). In addi-tion, deregulation of E2F1, either by overexpression or Rb inactivation, induces DNA damage and thus could contribute to cancer by causing mutations ( 10 , 11 ). Abnormalities in E2F1 gene expression and/or