Epithelial Polarity, Organization and the Angiogenic Switch

Institution: University of California, San Francisco
Investigator(s): Nancy Boudreau, Ph.D. -
Award Cycle: 2004 (Cycle 10) Grant #: 10IB-0157 Award: $75,000
Award Type: IDEA
Research Priorities
Biology of the Breast Cell>Biology of the Normal Breast: the starting point



Initial Award Abstract (2004)
Breast cancer can be considered a defect in patterning of the breast. Unlike normal breast cells which are organized in spherical structures designed to secrete milk into a central opening or lumen, breast cancers are disorganized, lack this spheriodal pattern and instead begin to invade into the adjacent tissue. More recently it has been recognized that in order for these disorganized tumor cells to survive and eventually spread or metastasize, they must somehow entice new blood vessels to form nearby to supply blood carrying oxygen and essential nutrients. What has remained a mystery however, is why tumor cells and not normal breast epithelial cells are capable of inducing the new blood vessels to form.

Our central hypothesis is that the loss of the breast tissue organization observed early in tumor development acts as the trigger to allow breast epithelial cells to begin producing factors required for stimulating new blood vessels to appear. However, if the tumorigenic breast cells can be experimentally manipulated to resume a normal breast tissue organization, this will also prevent recruitment of new blood vessels and therefore limit the potential for growth and metastatic spread of the tumor.

We will use a tissue culture model which allows normal breast cells to form three-dimensional spheriodal structures by providing a recombinant basement membrane (rBM), which they normally would attach to in the human breast. We can then compare organized normal cells with the disorganized cells which are not provided with rBM. Although tumor cells do not form organized structures when given the rBM, we can force them to form organized colonies, by blocking excess cell surface receptors with monoclonal antibodies. We can therefore compare the potential of organized/disorganized pre-malignant l and tumor cells to be able to induce new blood vessels and impact tumor progression and growth.

Although increased metabolic demand of tumor cells can trigger the production of factors which act to bring new blood vessels to the tumor mass, it seems likely that earlier events including loss of tissue organization also acts to turn on these programs even before there is an obvious expansion and increased demand by larger tumor masses. This may indicate that anti-angiogenic treatment could be considered for use as an early preventative measure to stop tumor cell expansion.


Final Report (2007)
Overview: The progression from pre-malignant to invasive breast cancers is dependent upon generating a new vascular supply. A variety of changes in the breast tumor environment have been linked to increased production/release of factors that encourage blood vessel development (angiogenic factors). The tumor cells have been shown to play a central role in generation of these angiogenic factors, including VEGF. Increased MAPK, PI3K, altered integrin a6?4 signaling and activation of ? catenin, all characteristic of breast tumor cells, can contribute to increased production of angiogenic factors. However, each of these pathways is attenuated in polarized and organized normal breast cells or can be reduced if tumor cells are forced to adopt a polarized, organized acinar structure.

Aims and Hypothesis/questions addressed: We predicted that correct polarity and 3-D tissue architecture suppresses expression of angiogenic factors, whereas disorganized normal or breast tumor cells display increased transcription of key angiogenic mediators.

Determine whether polarization and/or three-dimensional acinar organization of pre-malignant breast epithelium suppresses expression of key angiogenic mediators including VEGF, PDGF and MMP-9. We evaluated expression of VEGF, PDGF and MMP¬9 and observed a reduction in expression of both VEGF and MMP-9 in organized polarized breast epithelium compared to disorganized tumorigenic breast epithelial cells. The levels of PDGF however were not linked to the organization or polarity of the cells. We did show however that Ang-1 a vascular stabilizing agent is increased in polarized organized normal epithelium compared to the tumorigenic counterparts.

Determine whether breast tumor cells forced to adopt an organized and/or polarized morphology show reduced expression of angiogenic factors? We showed that restoring expression of a key morphoregulatory transcriptional mediator, HoxD10 which is lost in aggressive tumors, could revert tumor cells to a polarized phenotype. This normalized phenotype was accompanied by a reduction in VEGF expression and a concomitant increase in Ang-1, a vascular stabilizing gene.

Evaluate whether organized or disorganized epithelial cells influence angiogenesis in vivo. To address this aim we compared the in vivo angiogenic potential of breast epithelial cells transfected with HoxD10 which induces a polarized morphology and those transfected with HoxA10, which does not induce a polarized morphology. We observed that the HoxAl 0 expressing cells induced a much greater degree of angiogenesis and increased vascular density in vivo.

Plans for continuation of the project: We observed that the ability of Hox D10 to revert tumor cells to an organized phenotype requires cooperative interactions with the ?4 integrin. Thus we will explore these interactions of HoxD10 and integrin signaling further using ?4 integrin mutants which interfere with 3DIrBM derived signaling. We will also investigate whether retention of ?¬catenin in junctions is also required for this attenuation of angiogenic potential. In addition we plan determine whether the accompanying reduction in ERK signaling observed with HoxD10 cells cultured in 3D1rBM is required for the reduction in VEGF mRNA expression. In addition we will identify the promoters of downstream target genes which are accessed and bound by HoxD10 in tumor cells cultured in 3DlrBM and establish how they contribute to acquiring and maintaining normal tissue organization.