Project 3. Role of Growth Factor Signaling

Insulin-like Growth Factor (IGF) and Hepatocyte growth factor (HGF) Pathways

IGF-I is of interest as elevated levels of circulating "free" IGF-1 (unbound to inhibitory binding proteins) have been shown to be associated with a greater risk for the development of premenopausal breast cancer. In addition IGF-I has been shown to regulate the proliferation and surivival of mammary epithelial cells cultured in petri dishes under standard 2D culture conditions and IGF-1 and its receptor has been shown to be over-expressed in breast cancer.

HGF is involved in the normal development of the mammary gland. However, HGF and its receptor c-Met have been implicated in late stages of tumor cell progression and invasion and HGF treatment of the human tumor cell line MDA-231 cells (but not normal cells) induces invasion. These results suggest that the induction of invasion by HGF requires cooperation with other cellular genes.

Three-dimensional Basement Membrane Culture Model

In vitro three-dimensional basement membrane models allow the evaluation of the biological activities of growth factors and other genes associated with breast cancer in events related to the initiation and progression of breast tumors. This model, similar to one originally developed by Bissell and Peterson involves the use of the immortalized human mammary epithelial cell line, MCF10A cells. These cells undergo a program of morphogenetic events in Matrigel basement membrane cultures leading to the development of growth-arrested acini-like spheroid structures that are composed of a single layer of epithelial cells surrounding a hollow lumen. Such cultures allow the examination of the ability of growth factors and other breast-cancer associated genes to allow cells to escape proliferative suppression, to survive in the lumen, to disrupt apical polarity, and to break down/invade the basement membrane.

PI3Kinase signaling in ER negative breast tumors

In ER- breast cancers other growth and survival signals may be substituting for estrogen. One such signal comes from the PI3 kinase pathway. This pathway clearly plays an important role in both tumors that are null for activity of the PTEN tumor suppressor, a negative regulator of the PI3 kinase, and in tumors which have activated HER2. Particular attention is being paid to the PI3 kinase pathway downstream of the IGF-1 receptor and the HER1,2 receptors. Mutation of the PTEN tumor suppressor in women with Cowden's syndrome predisposes to the development of breast cancer and PTEN +/- mice develop breast cancer. However, PTEN mutations are relatively rare in sporadic breast cancer. This suggests that while the PTEN pathway plays an important role in breast cancer development, other components of the pathway such as HER2 amplification is a more frequent mode of derangement.

We will examine the hypothesis that aberrant expression and/or responsiveness to growth factors is involved in the development and progression of ER- human breast tumors. To accomplish this, we will:

1. Examine the phenotypic changes in acinar morphogenesis induced by several growth factors that have been implicated in breast cancer.
   
   
2. Examine how expression of genes whose function or expression are known to be altered in breast cancer perturb the response of breast epithelial cells to these growth factors.
   
   
3. Establish human breast epithelial cell lines that express several levels of PI3 kinase and AKT activity. These lines will be characterized with respect to resistance to apoptotic stimuli, receptor function, and transformed phenotype. Our working hypothesis is that either too much or too little function of the PI3K pathway can promote cell death. If the lines are not transformed by expression of single activated kinases, a large array of other pathways will be tested for synergy in tumor formation. The transcriptional profiles will be obtained in collaboration with Project 2 and compared with profiles of human breast tumor samples.
   
   
4. In parallel we will engineer mice to express constitutively activated PI3K in breast epithelium and to allow tissue specific knockout of p110a and b genes in breast epithelium. This will allow us to study the role of activation of PI3K both independently and in the context of other transgenes or knock-outs. Once again transcription in tumors will be profiled for comparison to human tumors in collaboration with Project 2.

Joan S. Brugge, Ph.D.-Dr. Brugge is Professor of Cell Biology, Harvard Medical School. She is a member of the National Academy of Sciences and a leader in the field of growth factor signaling. Her lab (http://cellbio.med.harvard.edu/faculty/brugge/) has a major focus on the role of extracellular matrix in determining the behavior of breast cancer cells. She will serve as a co-investigator on Project 3. Email: Joan_brugge@hms.harvard.edu

Myles A. Brown, M.D.-Dr. Brown Associate Professor of Medicine, Dana-Farber Cancer Institute and Harvard Medical School. He is a member of the Executive Committee of the Dana-Farber Cancer Institute Women's Cancer's Program and the DF/HCC Breast Cancer Program. His lab focus is on the role of coregulators in nuclear receptor function. He will serve as Principal Investigator of the COE. In addition he will serve as a co-investigator on Project 6 and a collaborator on Projects 2 and 3. Email: Myles_Brown@dfci.harvard.edu

Thomas M. Roberts, Ph.D.-Dr. Roberts is Professor of Pathology, Dana-Farber Cancer Institute and Harvard Medical School. He is a leader in the signal transduction field and the role of PI kinase pathways in cancer. He will serve as a co-investigator on Project 3 and a collaborator on Core 1. Email: Thomas_Roberts@dfci.harvard.edu

Three-dimensional Basement Membrane Culture Model

Project Update 2003

The following examples serve to illustrate the utility of this model. Figure 3 shows the effects of inducible activation of ErbB2/Her2 in preformed growth arrested mammary acini. Activation of ErbB2, but not ErbB1, results in disruption of acinar organization and formation of large structures containing multiple acini-like units with no luminal space. The cells within these disorganized structures maintain an epithelial phenotype and dependence on anchorage to integrin matrix ligands and are unable to invade the basement membrane. ErbB2 activation also disrupts tight junctions and cell polarity in monolayers of polarized epithelia whereas ErbB1 is unable to induce this effect. Activation of ErbB2, but no ErbB1 or EGF, induces reinitiation of DNA synthesis in growth-arrested acini, suggesting that ErbB2 dimers uniquely activate pathways which overcome the proliferative suppression within acinar structures. In addition, the ability of ErbB2-expressing cells to survive within the acinar lumen indicates that ErbB2 is able to protect cells from apoptotic processes that normally eliminate cells within the lumen. The disorganized acini induced by ErbB2 display properties similar to those of carcinoma in situ tumors in vivo suggesting that ErbB2, but not ErbB1, homodimers possess the ability to induce early stages of carcinogenesis in vitro. Since ErbB2 has been shown to be highly amplified or overexpressed in a specific type of carcinoma in situ (comedo), with a high proliferative index filled lumen and undisturbed basement membrane, our results suggest that the in vitro 3D model may be at least partially predictive of functional activities in vivo.

Figure 3. MCF10A cells expressing a chimeric variant of ErbB2 (containing the extracellular and transmembrane domain of the p75 low affinity NGF receptor, the cytoplasmic domain of ErbB2, and FKBP) were incubated in Matrigel for 12 days (left panel). Acini were then treated with AP1510, a small molecule dimerized drug which binds to FKBP and induced dimerization/activation of ErbB2 (right panel).

In contrast, although cells expressing HPV E7 are highly proliferative and escape normal suppression of cell division within the acini, they are unable to survive in the center of the acini (figure 4). Thus, E7 induced structures are composed of proliferating epithelial cells surrounding a hollow lumen. Cells that are deposited into the center of the acini undergo programmed cell death or apoptosis. Extrapolating to human tumors, one could speculate that loss of Rb function in mammary epithelial cells may be sufficient to induce proliferation of mammary epithelial cells but not sufficient to allow filling of the lumen. Additional mutations that allow cell survival in the lumen may be necessary to allow progression to a more advanced carcinoma in situ phenotype.

Figure 4. MCF10A cells were infected with a retrovirus encoding HPVE7 or an empty retroviral vector. The infected cells were then incubated in Matrigel to allow formation of acinar structures. The upper panels show phase contrast images of the structures, the middle panels show staining for the cell proliferation marker, Ki67, and the bottom panels show confocal images through the middle of several structures to show the hollow lumens.

These results demonstrate that the three dimensional model system can be very useful for examining biological activities of oncogenes that can not be examined in standard two dimensional in vitro transformation assays and should be very useful for elucidating the mechanisms responsible for different phenotypic effects of oncogenes and identification of additional genes that can disrupt normal morphogenesis.

As stated above, we will examine the hypothesis that aberrant expression and/or responsiveness to growth factors is involved in the development and progression of ER- human tumors. The in vitro model we have developed will allow us to examine the effects of single growth factors on mammary cell proliferation, survival and invasion and to examine cooperative effects of different combinations of growth factors with other cellular genes.

PI3Kinase signaling in ER negative breast tumors

Project Update 2003

Preliminary work in rodent cell lines has revealed that the constitutive activation of the pathway gives quite unexpected results. First, while PI3 kinase activation is often anti-apoptotic in the context of receptor signaling, constitutive activation of either PI3 kinase or its key downstream target Akt can be proapoptotic. There is some evidence that the proapoptotic effects are dependent on the cells p53 status. Second we find that activation of the PI3 kinase pathway can result in receptor inactivation. Once again it appears that we can mimic this finding with activated Akt. We have also found that constitutive expression of either our activated PI3K construct or Akt-1 renders Rat-1 fibroblasts tumorigenic in nude mice (figure 5). Interestingly Akt 2 or Akt3 expression is less tumorigenic and gives rise to tumors of distinct morphology compared to Akt-1 tumors. We have also achieved expression of both the activated PI3K allele and activated Akt in telomerase immortalized breast epithelial cells using a retrovirus vector.

Figure 5. Tumors formed by Rat1 cells expressing myr-p110. Nude mice were injected on the left flank with 1X106 Rat-1A cells expressing myr-p110 and on the right flank with an equal number of control Rat-1A cells. This picture was taken 10 days post inoculation.

[back to top]