Intravital imaging of cancer: visualizing the bad guys

Abstract

Cancer-related death is mainly due to metastasis: the ability of tumor cells to colonize distant sites after detaching from the primary tumor, migrating and traveling through the body via the circulation. Understanding the mechanisms underlying this metastatic process is vital to develop drugs that specifically target (the formation of) metastases and prevent further metastatic outgrowth. The behavior of tumor cells in vivo is different than in vitro, because tumor cells in vivo are part of their natural 3D microenvironment, which is crucial for invasion and dissemination of the tumor cells and colonization of distant tissues. Therefore, we studied different types of cancer by visualizing tumor cells in vivo using fluorescent mouse tumor models combined with intravital microscopy (IVM). This state-of-the-art technique is the method of choice for studying cancer biology, since IVM permits visualization of primary tumor onset, progression and different aspects of the metastatic process in living organisms. Using IVM we made several discoveries concerning the mechanisms that are used by tumor cells to disseminate and colonize distant tissue. First, we observed that in two different genetic mouse models of epithelial cancer (pancreas and breast), the tumor cell population that detached from the primary tumor migrated not as a collective sheet (what could be expected of epithelial cells) but as single cells. This is important information necessary for developing drug targets that act upon the specific pathways involved in different types of migration. We also discovered that within breast epithelial primary tumors with high expression of the cell adhesion protein E-cadherin (an epithelial marker), there is a fraction of tumor cells that lacks functional E-cadherin (E-cadLO). We demonstrated that these E-cadLO cells have a mesenchymal expression profile and using IVM we showed that this fraction of tumor cells is migrating and invading into the stroma, in contrast to the immobile bulk E-cadHI tumor cells. We found both E-cadLO as E-cadHI circulating tumor cells in the blood and as single cells in lung metastases. However, all metastases larger than 3 cells were EcadHI, which was caused by a switch back towards an epithelial state. Any proposed metastatic outgrowth advantage of mesenchymal cells is therefore irrelevant, as mesenchymal cells switch back to epithelial cells upon expansion. Another observation we made using IVM was the difference between in vitro and in vivo response of tumor cells to the chemotherapeutic agent docetaxel. While our in vitro results showed an increase in apoptotic tumor cells due to mitotic perturbations (as expected from this type of drug), mitotic progression was not affected in apoptotic cells in vivo. In our study, we could directly compare in vitro and in vivo effects of docetaxel by using the same cell lines in both settings and we could directly see the effect of the drug in the 3D natural environment using IVM, which is clearly essential for extrapolating results of the drug’s mechanism of action to its effect in patients. Discovering the true mechanism of action of a drug is also important for the design of new drugs.