Abstract

Introduction

The process of wound healing to replace damaged tissue involves epithelial tissue regeneration; a small population of replenishing stem cells gives rise to differentiated progeny that replace the damaged tissue. This characteristic of accelerated epithelial tissue regeneration is shared by the epithelial component in a growing tumor, albeit on a genetically unstable background. In fact, epithelial tumors have been described as ‘wounds that do not heal’ because of the molecular and cellular similarities between the mesenchyme associated with wounds and that of carcinomas [1]. Many solid tumor types, including breast cancer, exhibit a functional hierarchy of cancer cells of which only a small subpopulation of replenishing stem-like cells can give rise to the differentiated cells that comprise the bulk tumor [2–6]. In human breast cancers, these tumorigenic breast cancer stem cells are enriched in cells with a CD44+/CD24-/low/ESA+ phenotype [2].

Other than the ability to seed a tumor in a nonobese diabetic (NOD)/severe combined immunodeficient (SCID) mouse, it is unclear what phenotypic and functional differences distinguish the cells that fuel carcinoma growth from cells that comprise the tumor bulk. Furthermore, it is unclear what mechanisms control the maintenance and survival of these tumorigenic cells. These issues have been difficult to study, in part because of the presupposed lack of appropriate model systems. Use of primary breast cancer cells is considered to be the best means to study tumor repopulation because it is presumed that, upon long-term cultivation in vitro, cancer cells lose the dynamic characteristics of a regenerating tissue [7]. However, experiments with primary tumor cells are costly and difficult to control because of small sample size and the heterogeneous nature of the cellular, genetic, and epigenetic composition among patient tissue samples.

In order to overcome problems associated with procuring and using primary human tissues, continuous breast cancer cell lines have been developed from various sources, including pleural effusions (MDA.MB.231 and MCF7), primary breast cancers (SUM149 and SUM159), primary tumor recurrences (SUM225), and even xenografted metastatic nodules (SUM1315) [8,9]. These cell lines represent polyclonal populations of cells that have adapted to tissue culture conditions but retain many of their phenotypic and genotypic properties over countless passages [9,10]. Genomic approaches have revealed that, like primary tumors, the gene expression signatures of breast cancer cell lines can distinguish luminal from basal subtypes of breast cancer [11–13]. Moreover, cell line derived gene signatures can correctly classify human tumor samples [8,14]; this suggests that, despite their acquired ability to grow in vitro, cell lines continue to share many of the molecular and genetic features of the primary breast cancers from which they were derived.

Breast cancer cell lines will form tumors in immunodeficient mice that recapitulate the histology, progression, and metastatic spread of the disease [15–17]. Therefore, we postulated that cancer cell lines may also retain the cellular hierarchy characteristic of primary breast tumors. In this study, we report that breast cancer cell lines contain a small population of cells that mimic cancer stem cell behaviors. Similar to primary breast cancers, cell line derived tumor-initiating cells are enriched in cells with the CD44+/CD24-/low/epithelial-specific antigen (ESA)+ phenotype. This work validates the use of cell lines to elucidate the unique mechanisms that govern maintenance and survival of tumorigenic breast cancer stem cells.

Although the existence of tumor-initiating cells in solid human tumors is widely accepted, it is unclear whether cancer-derived cell lines contain similar cells. Primary carcinoma cells are presumed to be the only representatives of actual human cancers, because prolonged in vitro culturing is thought to result in loss of crucial properties including hierarchal organization and heterogeneity found in tumors [7]. However, we have found that established breast cancer cell lines posses a small fraction of self-renewing tumorigenic cells with the capacity to differentiate into phenotypically diverse progeny. This observation suggests that cell lines are excellent models in which to study cancer stem cells.

We found that CD44+/CD24-/ESA+ cells exhibit properties of self-renewal in vitro, form tumors from very few cells, divide slowly, and are selectively resistant to chemotherapy, all of which are hallmarks of cancer stem cells. In light of recent studies that suggest that presence of CD44+/CD24- cells does not correlate with clinical outcome [33] or with distant metastasis [17], we also found that cell lines with few CD44+/CD24- cells were not necessarily less tumorigenic than cell lines with very high numbers of CD44+/CD24- cells. However, it is important to note that in luminal cell lines that are nearly 100% ESA+ (MCF7 and SUM225), sorting cells for the phenotype CD44+/CD24- is sufficient to enrich for tumorigenic cells. In contrast, in basal cell lines with high percentages of CD44+/CD24- cells (MDA.MB.231, SUM159, and SUM1315), sorting for ESA+ cells is sufficient to enrich for breast cancer-initiating cells. Finally, in cell lines that exhibit two distinct CD44/CD24 populations (for example, SUM149), all three markers are required to enrich correctly for tumor-initiating stem-like cells.

Tumor progression is thought to result from the genetic instability within cancer cells that selects for clonal expansion and evolution of more aggressive tumor behaviors over time. However, despite continual genetic and phenotypic drift, breast cancers do not radically change their phenotype, even after many years or various therapeutic regimens. Likewise, cancer cell lines do not radically change their phenotype over many passages [9]. Although it is indeed possible to select clonogenic variants within a cell line that exhibit different genetic and biological properties from the parental population, it is widely recognized that single cell cloning is extremely inefficient, yielding few clones with the capacity to expand and propagate indefinitely. Moreover, single cell clones derived from the same cancer cell line are not equally tumorigenic or metastatic [34–36]. This observation supports the notion that not every cell within a cell line can self-renew or is equally tumorigenic, further illustrating that cell lines possess a functional hierarchy and heterogeneity akin to primary tumors.

During the course of characterizing the hierarchal dynamics of cells in culture, we observed that the percentage of CD44+/CD24-/ESA+ cells within a line varied depending on various conditions. For example, cell density, the presence or absence of growth factors, and even the frequency of passaging influenced the percentage of CD44+/CD24-/ESA+ cells detected in a cell line. These observations suggest that the dynamic nature of cell lines is regulated by their immediate microenvironment.

Conclusion

Although the existence of tumor-initiating cells in solid human tumors is widely accepted, it is unclear whether cancer derived cell lines contain similar cells. Here, we show that human breast cancer cell lines exhibit a cellular hierarchy that is characteristic of primary breast tumors, in which a small population of cells with the CD44+CD24-ESA+ phenotype enriches for tumor-initiating cells that self-renew in vitro and give rise to phenotypically diverse progeny. The fact that a small population of these breast cancer stem-like cells survives chemotherapy is consistent with the notion that the surviving tumorigenic cells are responsible for the 40% recurrence rate of invasive breast cancer. Therefore, it will be of great interest to determine whether CD44+/CD24-/ESA+ cells are directly targeted by alternative therapies such as monoclonal antibody treatment or novel small molecule inhibitors. Based on our results, using breast cancer cell lines to target CD44+/CD24-/ESA+ cells for drug discovery offers a highly promising, reproducible, and cost-effective means to identify therapies that prevent self-renewal or force depletion of tumorigenic breast cancer stem-like cells.