Research in the laboratory is focused on the molecular mechanisms that control carcinoma progression and metastasis. Our research is concentrated in three areas: (1) Understanding the collaborative impact of paracrine and systemic signaling on tumor growth and progression. (2) Identification of mechanisms through which paracrine and systemic signals can induce epithelial cells to enter into a mesenchymal/stem-cell state. (3) Ellucidating the complex molecular mechanisms that regulate carcinoma invasion and metastasis.
Invasive human breast cancer
Invasive mouse mammary tumor
Experimental lung metastasis
Stromal-epithelial and systemic interactions during tumor progression. Our research has revealed that several mesenchymal cell types that form the tumor-associated stroma are important in supporting carcinoma cell growth within the tumor. Thus, mesenchymal stem cells and myofibroblasts in the tumor-associated stroma can release paracrine signals that impart invasive and metastatic powers to nearby carcinoma cells. At the same time, primary carcinomas release endocrinal signals that impinge on the bone marrow and spleen and induce the formation of several types of inflammatory cells that may then be recruited by tumors, via the general hematogenous circulation, to help the tumors to rapidly assemble a highly functional, tumor-supporting stroma. Some of this work has revealed that primary tumor cells that release G-CSF into the circulation evoke neutrophilia in the general circulation. This yields, in turn, a state in which circulating tumor cells that have been released into the circulation are protected while in the circulation from attack and elimination by natural killer (NK) cells; their extravasation into the parenchyma of distant tissues is also facilitated.
We have also been interested in the effects of localized wound-healing on the outgrowth of tumors in anatomically distant sites. Thus, it is known that women whose primary breast cancers have been removed surgically often develop a spike of metastatic relapses between 6 and 18 months after the original surgery. We speculated that post-surgical wound-healing is responsible for these relapses driven by previously disseminated micrometastatic deposits. Accordingly, we used a model of wound healing in one flank of an immunocompetent mouse and implanted antigenic breast cancer cells in the contralateral flank. These tumors usually grow out briefly before most are then eliminated by the adaptive immune system. As we found, this immune attack is thwarted by localized inflammation at a distant site, indicating systemic immunosuppressive effects of wound healing that may be responsible suppressing the observed attack by the adaptive immune cells, which may be help to explain the phenomenon of post-surgical metastatic relapses in breast cancer patients.
Paracrine and systemic signals that induce epithelial cells to enter into the mesenchymal/stem-cell state. The epithelial-mesenchymal transition (EMT) represents a cell-biological program that enables both normal and neoplastic epithelial cells to acquire mesenchymal cell attributes, such as motility, invasiveness, and a resistance to apoptosis. At the same time, our work has revealed that both normal mammary epithelial cells as well as their neoplastic derivatives that are forced experimentally through an EMT program acquire many of the attributes of normal and neoplastic stem cells. This holds implications for the pathogenesis of metastasis, since carcinoma cells that have acquired mesenchymal attributes enabling them to physically disseminate also acquire the self-renewal trait that allows them to serve as tumor-initiating cells and thus as founders of macroscopic metastatic colonies at distant sites.
Future anti-cancer therapies may be focused on eliminating the more mesenchymal tumor-initiating cells, also termed cancer stem cells (CSCs), within carcinomas, thereby reducing their malignant phenotypes. However, our work demonstrates bidirectional plasticity between the various distinct phenotypic states that are created by the EMT program within carcinomas. More specifically, the successful therapeutic elimination of CSCs may be followed by the formation de novo of CSCs that are created by the dedifferentiation of more epithelial, non-CSCs within existing tumors. This indicates that future anti-cancer therapies that aim to cure carcinomas must be designed to eliminate both CSCs and their non-CSC derivatives.
Our work in a transgenic model of pathogenesis of mammary adenocarcinoma has revealed that one transcription factor, termed Slug, is responsible for choreographing the stem-cell state in the basal cells of the normal mammary gland; indeed, down-regulation of Slug results in depletion of normal mammary stem cells. In contrast, in mammary carcinoma cells, the paralog of Slug, termed Snail, is responsible for the initiation and maintenance of the CSCs in adenocarcinomas of the mammary gland. Moreover, these adenocarcinomas and associated CSCs are initiated in an epithelial cell layer in the mammary ducts distinct from that harboring normal mammary SCs, indicating that CSCs do not derive directly from normal mammary SCs.
Our work also indicates that carcinoma cells activating the EMT program rarely if ever complete an EMT program but instead advanced partway through the EMT program, advancing to a hybrid epithelial/mesenchymal state in which CSCs also reside. There is significant plasticity between the various phenotypic states between the fully epithelial and fully mesenchymal states. Recent work has begun to reveal the transcription factors that stabilize the residence of carcinoma cells in various epigenetic states between the fully epithelial and fully mesenchymal states.
Effects of the EMT program on the immunobiology of mammary carcinoma cells. We begun to examine the immunological microenvironment of epithelial vs. mesenchymal mouse mammary carcinoma cells. Extensive work of others has shown that more mesenchymal cells have elevated resistance to a variety of chemotherapeutic regimens. We have accordingly examined the responsiveness of mammary carcinoma cells to elimination by another anti-neoplastic treatment – checkpoint immunotherapy. To begin, we have found that cytotoxic T cells are largely excluded from tumors with a more mesenchymal phenotype and those that have entered are functionally exhausted. Moreover, immunosuppressive regulatory T cells are abundant in the mesenchymal tumors whereas they are largely absent from the epithelial tumors. Most importantly, the epithelial tumors are effectively eliminated by anti-CTLA4 checkpoint immunotherapy, whereas the mesenchymal tumors are unaffected by this therapy. Moreover, this work demonstrates that tumors that contain only 10% mesenchymal carcinoma cells with the remainder being epithelial carcinoma cells are largely protected from checkpoint immunotherapy, providing one possible explanation of why many types of carcinoma cells are unresponsive to checkpoint immunotherapy. This work is currently being extended to ovarian carcinoma cells.