Immune checkpoint blockade has generated dramatic responses in certain types of human tumors. However, the response of breast tumors has been largely limited. We have previously demonstrated that the residence of breast cancer cells in the epithelial or mesenchymal phenotypic states can itself be used as an important determinant of the success or failure of immune checkpoint blockade. The Dongre Lab is focused on understanding the mechanisms by which the EMT program drives resistance to anti-tumor immunity. We work with novel models and cutting-edge techniques right at the interface of cancer biology and immunology.
1. Differential susceptibility of epithelial and quasi-mesenchymal tumors to anti-tumor immune attack.
By using novel, preclinical murine models of epithelial and quasi-mesenchymal breast tumors, we have observed that these two tumor types assemble dramatically different tumor microenvironments and differ in their susceptibility to anti-CTLA4 therapy. Specifically, epithelial tumors are infiltrated by CD8+ T-cells and are sensitive to anti-CTLA4 therapy. In sharp contrast, quasi-mesenchymal tumors recruit immunosuppressive Tregs and M2-like macrophages and exclude exhausted CD8+T-cells to the periphery. Importantly, quasi-mesenchymal tumors are resistant to anti-CTLA4 therapy. (Read our paper - Cancer Research 2017). We are actively investigating the underlying mechanisms that dictate these differential responses of epithelial and quasi-mesenchymal tumors to multiple forms of immune checkpoint blockade therapy. To this end, we are using our preclinical mouse models in conjunction with CRISPR/Cas9 editing, flow cytometry, and imaging techniques.
2. Cross protective effects of quasi-mesenchymal cells in mixed tumors.
In mixed tumors comprised of both epithelial and quasi-mesenchymal cancer cells, minority populations (10%) of quasi-mesenchymal cells can cross-protect the vast majority (90%) of their epithelial neighbors residing in the same tumor from immune attack (Read our paper - Cancer Research 2017). The mechanisms underlying these cross-protective effects remain unclear. Understanding these mechanisms is particularly important as most human breast cancers contain minority populations of mesenchymal cells which can dictate the response of the tumor as a whole to immune checkpoint blockade. We are using spatio-temporal approaches to understand how quasi-mesenchymal carcinoma cells exert cross-protective effects in mixed tumors.
3. Sensitizing quasi-mesenchymal tumors to immune checkpoint blockade therapy.
By taking a series of transcriptomic approaches, we have found that quasi-mesenchymal carcinoma cells express multiple cell-intrinsic immunosuppressive factors relative to their epithelial counterparts. Most strikingly, while abrogation of carcinoma cell-intrinsic CSF1 or SPP1 can partially sensitize quasi-mesenchymal tumors to anti-CTLA4 therapy, abrogation of CD73 leads to complete sensitization of otherwise refractory tumors to anti-CTLA4 (Read our paper - Cancer Discovery 2021). We are developing strategies to target drivers of resistance in order to potentiate the efficacy of immune checkpoint blockade therapy in quasi-mesenchymal breast tumors.
4. Potentiating the efficacy of immune checkpoint blockade at the site of metastasis.
A large majority of breast cancer-related deaths can be attributed to the metastasis of carcinoma cells from the primary tumor to various distant organ sites. To this end, targeting CD73 in combination with anti-CTLA4A immune checkpoint blockade therapy leads to a dramatic reduction in metastatic burden (Read our paper - Cancer Discovery 2021). Given the importance of the EMT program in regulating the process of metastatic colonization and immunosuppression in the primary tumor, we are developing a variety of in vivo models to test whether the EMT program directly results in immunosuppression at the site of metastasis. We are also studying the molecular details underlying the subsequent responses of metastases to immune checkpoint blockade therapy.
Illustrations designed using BioRender.com