Traditional treatments for breast cancer fail to address therapy-resistant cancer stem-like cells that have been characterized by changes in epigenetic regulators such as the lysine demethylase KDM4. Here, we describe an orally available, selective and potent KDM4 inhibitor (QC6352) with unique preclinical characteristics. To assess the antitumor properties of QC6352, we established a method to isolate and propagate breast cancer stem-like cells (BCSC) from individual triple-negative tumors resected from patients after neoadjuvant chemotherapy. Limiting-dilution orthotopic xenografts of these BCSCs regenerated original patient tumor histology and gene expression. QC6352 blocked BCSC proliferation, sphere formation, and xenograft tumor formation. QC6352 also abrogated expression of EGFR, which drives the growth of therapy-resistant triple-negative breast cancer cells. Our findings validate a unique BCSC culture system for drug screening and offer preclinical proof of concept for KDM4 inhibition as a new strategy to treat triple-negative breast cancer. .

Triple-negative breast cancer (TNBC) is one of the most aggressive subtypes of breast cancer, with only limited treatment options available. Recently, cancer stem cells (CSCs) have emerged as the potential drivers of tumor progression due to their ability to both self-renew and give rise to differentiated progeny. The CSC state has been linked to the process of epithelial-mesenchymal transition (EMT) and to the highly flexible state of epithelial-mesenchymal plasticity (EMP). We aimed to establish primary breast cancer stem cell (BCSC) cultures isolated from TNBC specimens. These cells grow as tumor spheres under anchorage-independent culture conditions in vitro and reliably form tumors in mice when transplanted in limiting dilutions in vivo. The BCSC xenograft tumors phenocopy the original patient tumor in architecture and gene expression. Analysis of an EMT-related marker profile revealed the concomitant expression of epithelial and mesenchymal markers suggesting an EMP state for BCSCs of TNBC. Furthermore, BCSCs were susceptible to stimulation with the EMT inducer TGF-β1, resulting in upregulation of mesenchymal genes and enhanced migratory abilities. Overall, primary BCSC cultures are a promising model close to the patient that can be used both in vitro and in vivo to address questions of BCSC biology and evaluate new treatment options for TNBC.

Cancers are heterogeneous by nature. While traditional oncology screens commonly use a single endpoint of cell viability, altering the phenotype of tumor-initiating cells may reveal alternative targets that regulate cellular growth by processes other than apoptosis or cell division. We evaluated the impact of knocking down expression of 420 kinases in bi-lineage triple-negative breast cancer (TNBC) cells that express characteristics of both myoepithelial and luminal cells. Knockdown of ERN1 or ALPK1 induces bi-lineage MDA-MB-468 cells to lose the myoepithelial marker keratin 5 but not the luminal markers keratin 8 and GATA3. In addition, these cells exhibit increased β-casein production. These changes are associated with decreased proliferation and clonogenicity in spheroid cultures and anchorage-independent growth assays. Confirmation of these assays was completed in vivo, where ERN1- or ALPK1-deficient TNBC cells are less tumorigenic. Finally, treatment with K252a, a kinase inhibitor active on ERN1, similarly impairs anchorage-independent growth of multiple breast cancer cell lines. This study supports the strategy to identify new molecular targets for types of cancer driven by cells that retain some capacity for normal differentiation to a non-tumorigenic phenotype. ERN1 and ALPK1 are potential targets for therapeutic development.