Objectives Cell lines derived from human ovarian and endometrial cancers, and their immortalized non-malignant counterparts, are critical tools to investigate and characterize molecular mechanisms underlying gynecologic tumorigenesis, and facilitate development of novel therapeutics. To determine the extent of misidentification, contamination and redundancy, with evident consequences for the validity of research based upon these models, we undertook a systematic analysis and cataloging of endometrial and ovarian cell lines. Methods Profiling of cell lines by analysis of DNA microsatellite short tandem repeats (STR), p53 nucleotide polymorphisms and microsatellite instability. Results Fifty-one ovarian cancer lines were profiled with ten found to be redundant and five (A2008, OV2008, C13, SK-OV-4 and SK-OV-6) identified as cervical cancer cells. Ten endometrial cell lines were analyzed, with RL-92, HEC-1A, HEC-1B, HEC-50, KLE, and AN3CA all exhibiting unique, uncontaminated STR profiles. Multiple variants of Ishikawa and ECC-1 endometrial cancer cell lines were genotyped and analyzed by sequencing of mutations in the p53 gene. The profile of ECC-1 cells did not match the EnCa-101 tumor, from which it was reportedly derived, and all ECC-1 isolates genotyped as Ishikawa cells, MCF-7 breast cancer cells, or a combination thereof. Two normal, immortalized endometrial epithelial cell lines, HES cells and the hTERT-EEC line, were identified as HeLa cervical carcinoma and MCF-7 breast cancer cells, respectively. Conclusions Results demonstrate significant misidentification, duplication, and loss of integrity of endometrial and ovarian cancer cell lines. Authentication by STR DNA profiling is a simple and economical method to verify and validate studies undertaken with these models.

The mechanism by which activation of common signal transduction pathways can elicit cell-specific responses remains an important question in biology. To elucidate the molecular mechanism by which the Ras signaling pathway activates a cell-type-specific gene, we have used the pituitary-specific rat prolactin (rPRL) promoter as a target of oncogenic Ras and Raf in GH4 rat pituitary cells. Here we show that expression of either c-Ets-1 or the POU homeo-domain transcription factor GHF-1/Pit-1 enhance the Ras/Raf activation of the rPRL promoter and that coexpression of the two transcription factors results in an even greater synergistic Ras response. By contrast, the related GHF-1-dependent rat growth hormone promoter fails to respond to Ras or Raf, indicating that GHF-1 alone is insufficient to mediate the Ras/Raf effect. Using amino-terminal truncations of c-Ets-1, we have mapped the c-Ets-1 region required to mediate the optimal Ras response to a 40-amino-acid segment which contains a putative mitogen-activated protein kinase site. Finally, dominant-negative Ets and GHF constructs block Ras activation of the rPRL promoter, and each blocks the synergistic activation mediated by the other partner protein, further corroborating that a functional interaction between c-Ets-1 and GHF-1 is required for an optimal Ras response. Thus, the functional interaction of a pituitary-specific transcription factor, GHF-1, with a widely expressed nuclear proto-oncogene product, c-Ets-1, provides one important molecular mechanism by which the general Ras signaling cascade can be interpreted in a cell-type-specific manner.

2The pituitary-specific, POU-homeodomain factor GHF-1/Pit-1 is necessary, but not sufficient, for cellspecific expression of prolactin (PRL), growth hormone (GH), and thyrotropin. Combinatorial interactions of GHF-1 with other factors are likely to be required; however, such factors and their mechanisms of action remain to be elucidated. Here we identify Ets-1 as a factor that functionally and physically interacts with GHF-1 to fully reconstitute proximal PRL promoter activity. In contrast, Ets-2 has no effect, and the alternatively spliced GHF-2/Pit-1␤ variant fails to synergize with Ets-1. The Ets-1-GHF-1 synergy requires a composite Ets-1-GHF-1 cis element and is dependent on an Ets-1-specific protein domain. Furthermore, the ancestrally related and GHF-1-dependent GH promoter, which lacks this composite element, does not exhibit this response. Finally, Ets-1, but not Ets-2, binds directly to GHF-1 and GHF-2. These data show that a functional interaction of GHF-1 and Ets-1, acting via a composite DNA element, is required to establish lactotroph-specific PRL gene expression, thus providing a molecular mechanism by which GHF-1 can discriminate between the GH and PRL genes. These results underscore the importance of transcription factors that are distinct from, but interact with, homeobox proteins to establish lineage-specific gene expression.

Cell transformation by the Ras oncogene is mediated by members of the ets gene family. To analyse the mechanisms of regulation, we have studied activation of several ets factors by Ras expression. We show that expression of Ha-Ras strongly activates the Ets1 p68 and p54 isoforms and Ets2 in F9 EC cells. We have mapped the Ras responsive elements of Ets1 p68 to two domains, RI+II and RIII. Mutation of threonine 82 to alanine in RI+II abolishes both Ras activation and phosphorylation by MAP kinase. Threonine 82 is part of a sequence that is conserved in Drosophila Pointed P2, an ets protein that has been shown both genetically and biochemically to mediate Ras signalling in Drosophila cells. We extend the comparison of these evolutionary related proteins by showing that Pointed P2 is activated by Ras in mammalian cells and mutation of the homologous threonine abolishes activation. Furthermore, we show that Pointed P2 resembles Ets1, in that it has conserved sequences in a similar position adjacent to the ets DNA binding domain that negatively auto-regulates DNA binding. These results go towards showing that the Drosophila Pointed and vertebrate Ets1 are evolutionary related proteins that have remarkably conserved Ras regulatory mechanisms downstream from MAP kinase.