Identification and Characterization of Tumor Initiating Cells in Various Mouse Mammary Tumor Models

Breast cancer is not a single disease as it can be classified into different subtypes according to cellular composition, morphology, proliferative index, genetic lesions and therapeutic responses. The molecular and cellular mechanisms underpinning tumor heterogeneity remain a central question in the cancer biology field. To explain the multitude of breast cancer phenotypes, it has been proposed that tumor-initiating cells (TICs) might originate from different cells within the mammary lineage. Further, different oncogenes might elicit distinct phenotypes in a given cell, adding to the complexity of the disease. Single oncogene amplification is observed in human breast tumors. For instance, the human epidermal growth factor receptor type-2 (HER2) proto-oncogene (also called ErbB2, or Neu) is amplified in about 20% of human breast cancers and associated with aggressive phenotypes and poor prognosis. Similarly, 16% of breast cancers have c-Myc gene amplification. Yet, it remains largely disputed which mammary cell type responds to a specific initiating oncogenic mutation. We decided to take a forward approach by testing two contrasting potent oncogenes, Neu (murine form of HER2) or c-Myc, that are frequently overexpressed in breast cancer patients for their capacity to transform different cells of the mammary lineage. A number of transgenic mouse mammary tumor models have been generated that mimic human breast cancers. Previous studies reported that TICs of Neu driven tumors are the luminal progenitor cells. These studies made use of a mouse model that constitutively overexpressed the Neu oncogene in mouse mammary glands under the MMTV promoter. In contrast, c-Myc overexpressing tumors show a heterogeneous mixed phenotype that is composed of both luminal and myoepithelial cells. Due to the presence of bi-lineage derived cells, we hypothesized that the TICs of c-Myc tumors are the bi-potent stem cells. To clarify the identity and characteristics of TICs, we employed tractable mouse models, Tet-On-Neu/MMTV- rtTA and Tet-On-Myc/MMTV-rtTA, in which we can induce the overexpression of an oncogene at any given time point by administration of doxycycline. The possibility to induce overexpression of an oncogene at adulthood (8-9 weeks after birth) closely recapitulates the timing of somatic mutations acquired by breast cancer patients. In order to demonstrate the direct contribution of distinct breast epithelial cellular lineages to Myc and Neu driven tumorigenesis, we took a FACS approach to separate bi-potent adult stem cells from lineage-committed progenitor cells and differentiated cells. We transplanted the purified cell populations, then induced oncogenes by administration of doxycycline and observed tumor formation in vivo. 9 These tumors obtained from the transplantation experiments were compared with the natural arising primary mammary tumors from Tet-On-Neu/MMTV-rtTA and Tet-On- Myc/MMTV-rtTA by their histological and molecular features. In parallel, we also monitored the clonal growth of these sorted cells in vitro after oncogene overexpression using organotypic 3D cell culture assays. Taken together and partially in contrast to our initial hypothesis, Neu induced tumors that are composed of luminal cells can originate not only from the luminal-committed progenitors but also from the bi-potent stem cells. Moreover, luminal committed progenitors were able to give rise to c- induced tumors that are composed of both luminal and myoepithelial lineages. We currently validate our findings by employing additional cell surface antibodies to closer define and sub-fractionate mammary cellular lineages and by an in vivo lineage tracing approach. Our results suggest that Neu and c-Myc TICs can arise from different cellular subtypes in the mammary gland. Notably, progenitor cells seem to be able to establish tumors that consist of basal and luminal cells. These results add to the growing notion that genetic predisposition directs cell fate towards distinct breast cancer phenotypes.