Search

The highly metastatic amelanotic C8161 human melanoma line was found to exhibit complete dominance of its undifferentiated and metastatic phenotype in multiple somatic cell hybridization studies designed to bypass the presence of potential tumor suppressor genes. In a three armed approach involving somatic cell fusions of C8161 with recipient lines of greater differentiation, different lineage, and different tumorigenicity status, the metastatic and undifferentiated phenotype of C8161 was promiscuously dominant. In somatic cell hybrids produced between the C8161 and a group of non-metastatic human melanoma lines which exhibited melanocyte differentiation markers including S100, HMB-45, NKI/C3, and melanin, the fusions were uniformly metastatic and undifferentiated. In somatic cell hybrids of C8161 and MCF-7 the fusions exhibited an estrogen independent and unresponsive, estrogen receptor (ER) negative, and highly metastatic phenotype. In fusions between C8161 and HMS-1, an immortalized 'benign' human myoepithelial line which produced an abundant extracellular matrix (ECM) and high levels of protease and angiogenic inhibitors including maspin, tissue inhibitor of metalloproteinase-1 (TIMP-1), alpha1-antitrypsin (alpha1-AT), protease nexin II (PN-II), thrombospondin-1 and soluble basic fibroblast growth factor (bFGF) receptors, the hybrids showed complete absence of matrix, absent maspin expression, markedly decreased protease inhibitor and angiogenic inhibitor production, high levels of proteases and angiogenic factors, and a highly metastatic phenotype. In our somatic cell fusions, the human-human hybrids represented true and complete fusions and not hybrid clones selected for by loss of dominant-acting growth suppressor genes. This finding was supported by detailed comparative genomic hybridization (CGH) studies, Q-banding karyotype analysis, and autofusions of representative clones. The purposeful creation of inherently unstable human-murine fusions between C8161 and B16-F1 where loss of putative suppressor loci would be expected, resulted in fusions exhibiting decreased growth and non-metastatic behavior with progressive chromosomal loss. Neither p53, nm23, DNA methyltransferase, activated ras, fibroblast growth factor-4 (FGF-4), or epidermal growth factor receptor (EGFR) mediated the acquisition of the metastatic or undifferentiated phenotype within the C8161-human fusions. These studies are the first studies ever to successfully transfer the complete metastatic phenotype by somatic cell fusion and support the presence of a new high level regulatory pathway(s) involving dominant trans-acting factors which act pleiotropically to regulate an undifferentiated and highly metastatic phenotype.

The EWS/FLI1 fusion protein is created by the translocation between chromosomes 11 and 22 that appears in most Ewing's sarcomas. This chimeric protein has been demonstrated to be an aberrant transcription factor. Genes up regulated by EWS/FLI1 but not by full-length FLI1 were identified by representational difference analysis (RDA). We have characterized a novel gene, EWS/FLI1 activated transcript 2 (EAT-2) that was cloned from a murine cDNA library using a differentially expressed RDA fragment. EAT-2 expression is seen within 4-8 h of EWS/FLI1 induction. Its expression correlates with transformation of NIH3T3 cells by chimeric proteins related to EWS/FLI1 but not by unrelated genes. EAT-2 is expressed in normal murine tissues and contains a unique but biochemically functional SH2 domain. An homologous sequence in the human genome has been identified and mapped to chromosome 1q22. Human EAT-2 transcripts were identified by reverse transcriptase-polymerase chain reaction (RT-PCR) in Ewing's sarcoma cell tumour cell lines. EAT-2's unique structure and correlation with transformation make it a candidate for playing a role in the transformation of NIH3T3 cells and the oncogenesis of Ewing's sarcoma.