Breast Cancer Tumor-Stroma Interactions in an In Vivo Model

Institution: Vaccine Research Institute of San Diego
Investigator(s): Per Borgstrom, Ph.D. -
Award Cycle: 2009 (Cycle 15) Grant #: 15IB-0133 Award: $279,336
Award Type: IDEA
Research Priorities
Biology of the Breast Cell>Pathogenesis: understanding the disease



Initial Award Abstract (2009)

A breast cancer tumor is not only composed of tumor cells. There are many other types of cells such as fibroblasts, endothelial cells, and immune cells with which tumor cells intimately interact. Thus, a “tumor” can sometimes contain a majority of normal tissues. It is becoming increasingly clear that these normal cells within and around a tumor are not merely passive bystanders, but rather play an active role in controlling the microenvironment in which the tumor grows. This is true for both the primary tumor and secondary tumors that result from metastatic spread. Breast cancer spreads preferentially to liver, lung, brain, and bone. What is it about these tissues that invite metastatic breast cancer? Unfortunately, the study of the mechanisms governing tissue preference are not very well understood, and to a large extent, this is because it is impossible to “catch metastasis in the act” especially in humans, but even in experimental models. We have devised a method in which we can “seed” cancer cells on various tissues (e.g., “soil”) implanted behind a glass window in a skin-fold of live mice. With this tool, called intravital microscopy (IVM), we can actually watch a tumor develop. For example, when we grow a breast tumor on lung tissue, this is similar to metastasis of breast cancer to the lung. We can actually watch the development of various defining characteristics of the tumor and “catch” it at different stages for thorough molecular analysis.

In this project we will perform microarray (gene expression) analysis of tumor and stroma grown in the IVM model. By using stroma from different tissues, we expect to identify specific genes and gene networks that are differentially induced by the different stroma, possibly revealing differences in the stroma that explain the ability of breast cancer to metastasize to different organs. Next, cells of the immune system, such as macrophages and fibrocytes, undergo epigenetic alterations as they adapt over time to the tumor environment. Using our IVM method we will perform gene expression analysis on cells of the immune system that infiltrate the tumor. By studying gene expression in immune system cells over the time course of tumor development, we expect to be able to identify the major adaptation pathways. Finally, we will use a bioinformatics approach to identify human genes that are homologous to the mouse genes identified in our IVM studies, and explore their expression in human breast cancer tumors.

Direct routes to therapeutic intervention are of immediate importance. However, our understanding of stroma-tumor interaction is in its infancy, and the proposed experiments cannot fail but to expand our knowledge in this potentially clinically important area. We expect some of the genes we discover to explain the propensity of breast cancer to metastasize to some organs more quickly than to others. As such, they serve as potential targets to control breast cancer metastasis.




Final Report (2011)

We have developed a mouse intravital microscopy model for studying the development of metastatic breast cancer, with regard to changes in gene expression that occur during metastasis. This problem is very difficult to study in humans for a number of reasons, and therefore an animal model is needed. Therefore, we developed a pseudo-organ model for breast cancer, which involves the implantation of tumor spheroids (the “seed”) upon a bed of minced organ tissue in a dorsal skinfold chamber on a mouse.

In our first aim, we proposed to study gene expression in the developing tumor and the surrounding tissue in fat pad, lung, liver, and skin. We changed our focus to lung, brain and bone marrow, which are the most problematic and potentially important sites of breast cancer metastasis. We have succeeded in growing these in the dorsal skinfold chamber. We proposed that gene expression analysis would be done on microdissected tumor and stroma from these experimental tumors. However, as we became more familiar with this system, we adopted a simpler, but effective plan. These experiments identified gene expression that connects metastatic growth to induction of neovascularization, as well as other genes that may play an important role in breast cancer metastasis. We have completed microarray experiments for N202 breast cancer cells adapting to the lung are now completing the experiments with brain and bone marrow. We still need to perform stromal microdissections.

Our major result is that N202 tumors adapted to the lung express TNF-alpha which is usually secreted by macrophage to suppress tumor growth. This tumor appears to be using TNF-alpha to attract macrophage, to support neovascularization, which is necessary for significant growth. We will proceed to test genes using real-time PCR, which is much simpler than de novo gene expression discovery, given the limited amounts of material involved. We will proceed with these experiments on N202 cells that we have already trained to grow on brain tissue and bone marrow. In these experiments, interest will be in whether the same or different genes or pathways become differentially regulated in a stable manner upon adaptation.

We have overcome several obstacles. The very small size of the dorsal skinfold chamber tumors has been an obstacle, and we overcame this by developing the robust approach mentioned above. Our microarray experiments identified genes that were hard to interpret in the context of breast cancer, and it turned out that these were from propagating fibroblasts and keratinocytes that escaped our selection strategy. Thus, we learned that cell sorting was a necessary part of this approach.