Breast cancer stem cells tolerate chromosomal instability during tumor progression via c-Jun/AXL stress signaling.

Chromosomal instability (CIN) is critical for tumor evolution, yet its relationship with stemness is unclear. Here, we describe CIN as a key stress induced during tumor initiation that is uniquely tolerated by breast cancer stem cells in an activated signaling state (aCSCs). While we noted elevated CIN specifically in tumors from aCSCs, this was not intrinsic to these cells, as baseline levels were similar to non-stem cell types. This suggests that CIN is induced during tumor initiation, and that aCSCs can better tolerate this stress. Further, this increased CIN may be transient, as it was only in low-burden aCSC tumors, with levels diminishing in more established disease. Phospho-array profiling revealed specific activation of c-Jun stress signaling in aCSCs, which we hypothesized could induce genes responsible for CIN tolerance. Indeed, we identified AXL as a c-Jun dependent gene enriched in aCSCs that enhances resistance to this stress. Thus, CIN tolerance mediated by c-Jun/AXL signaling may be a defining feature of stemness, contributing to breast cancer progression.

A dataset of chromosomal instability gene signature scores in normal and cancer cells from the human breast.

These data show the relative amount of chromosomal instability (CIN) in a diverse array of human breast cell types, including non-transformed mammary epithelial cells as well as cancer cell lines. Additional data is also provided from human embryonic and mesenchymal stem cells. To produce this dataset, we compared a published chromosomal instability gene signature against publicly available datasets containing gene expression information for each cell. We then analyzed these data with the Python GSEAPY software package to provide a CIN enrichment score. These data are useful for comparing the relative amounts of CIN in different breast cell types. This includes cells representing the major clinical (ER/PR+, HER2+ & Triple-negative) as well as intrinsic breast cancer subtypes (Luminal B, HER2+, Basal-like and Claudin-low). Our dataset has a great potential for re-use given the recent surge in interest surrounding the role of CIN in breast cancer. The large size of the dataset, coupled with the diversity of the cell types represented, provides numerous possibilities for future comparisons.