Ets2 Function in Development and Cancer

Ets2 is one of a family of transcription factors that all utilize a conserved Ets protein domain for DNA binding. Like its Drosophila homologs, pointed and yan, Ets2 is regulated by growth factors and oncogenes that use Ras signaling pathways. The phosphorylation of a single threonine residue in an evolutionarily conserved protein docking domain of Ets2 results in transcriptional activation and the induction of Ets2-dependent genes. We generated several different Ets2 alleles by gene targeting methods. A knockout allele was created by deleting exons coding for the DNA-binding domain and nuclear localization signals. Early mouse placental development is dependent on the function of the Ets2 transcription factor. In addition, we performed a knock-in that replaced the single amino acid that is phosphorylated by mediators of the Ras pathway. This reinforced the importance of Ras pathway activation on Ets2 activity, as was previously deduced from the Neu/ErbB2 activation of Ets2 signaling. We also generated a conditional allele that is inactivated by Cre recombinase activity. Ets2 regulates trophoblast stem (TS) cell self-renewal and thus placental development. One of the genes regulated by Ets2 in TS cells is the Cdx2 homeobox transcription factor. Because Cdx2 was previously implicated as a suppressor in colon tumors, we have begun a study of Ets2’s function in suppressing colon tumors. We found that Ets2 deficient intestinal stem cells have a selective advantage compared to normal stem cells. This results from an increased rate of intestinal cell proliferation and intestinal crypt fission perhaps because of decreased differentiation. This results in an increased sensitivity to colon tumor formation.

Beaker Blog post
Tales from the Intestinal Crypt

An Ets2-Dependent, Stromal Restriction of Mammary Tumors

Mouse mammary tumors caused by the transgenic expression of the Polyoma virus middle T antigen or an activated form of Neu/ErbB2 are dependent upon on Ets2 because decreasing Ets2 by even half limits the appearance of such tumors. This dependence on Ets2 is due to a host effect, since malignant tumors transplanted directly into the fat pad of Ets2-deficient mice grow slower than in wildtype mice, while inactivating Ets2 within mammary tumors but not the stromal has no effect.  We showed that disruption of mammary epithelial polarity is tightly associated with the expression of the eu oncogene. Increased expression of VEGF by Neu-transformed epithelial cells led to a dramatic acceleration of tumor formation and metastasis. Unlike the generally accepted model of metastasis, tumor cells that retain their intercellular cohesion appear to bud off into vascular spaces and may passively lodge in lung capillaries. Contributions of this type of non-invasive metastasis may be an important contribution to the well-documented correlation of increased vascularity and metastasis. This accelerated model of breast cancer may be particularly useful for testing angiogenesis therapeutics because of the increased vessel density and the rapid and predictable progression of the disease.

Mammary Epithelial Stem Cells and Breast Cancer

The discovery of breast cancer stem cells provides an opportunity of developing differentiation therapy to instruct malignant cells to adopt a benign fate. Two recent studies reinforce our interest in this idea. In collaboration with Dr. Alexey Terskikh, we investigated the role of the Maternal Embyronic Leucine Zipper Kinase (MELK) in normal mammary epithelial stem cells and mouse models of breast cancer. Using a transgenic reporter gene for Melk expression, we found that Melk expression is preferentially expressed in proliferative mammary epithelial progenitor cells and tumor cells. Both tumorsphere formation in culture and tumor formation in vivo is suppressed by knocking down Melk expression with Lentiviral-mediated shRNA in MMTV-Wnt1 tumor cells. The development of an inhibitor of the MELK kinase has therapeutic potential.

Support for the development of differentiation therapy came from a collaborative project with Pfizer. Administration of bosutinib, a Src family inhibitor, to mice with mammary tumors caused by the MMTV-PyMT oncogene greatly restricted tumor progression by inducing differentiation of the tumor to epidermal and lactational cell fates without widespread cell death. This is an example of the possibility of restricting cancer by inducing differentiation. In current projects we have isolated tumor-initiating cells in culture from MMTV-Wnt1 tumors and are using high-throughput, high-content image analysis made possible by the Conrad Preybs Center for Chemical Genomics to identify compounds and genes regulating the differentiation of both human and mouse tumor-initiating cells.

Figure 1. The clusters of cells in the middle and bottom of the frame are pre-malignant growths developing in the mammary fat pad of a mouse. The smooth branched structure are normal mammary epithelial tissues. The fluorescent green color comes from the green fluorescent protein driven by the MELK gene promoter.






Figure 2. Contrast mammary tumors tissue in control mice in the left panels and in mice treated with bosutinib, a kinase inhibitor (SKI) that induces differentiation to more normal epithelial tissue and a more normalized blood vessel pattern (lower right).












Figure 3. A colon crypt is shown in the process of dividing into two. Nuclei are stained red. Cells labeled in blue have inactivated the Ets2 gene.















Figure 4. Keratin 7 staining (red) of human ES cells detect the formation of trophoblast derivatives.