The MCF10 Model of Breast Tumor Progression

In their landmark 1990 publication, Soule and colleagues reported the establishment of the first nontransformed, human mammary epithelial cell line derived from normal breast tissue, designated MCF10 (1). In this work, the authors described the derivation and characterization of mortal (MCF10M) and immortal adherent (MCF10A) and floating (MCF10F) lines, which were derived from a mastectomy performed on a premenopausal, parous female with fibrocystic breast disease. Characterization of the parental MCF10 lines in the inaugural study showed that MCF10 cells lacked amplification, rearrangement, or mutational activation of oncogenes located on chromosomes 17 and 11, including HER2, int-2, and Ha-ras. Further supporting their nonmalignant nature, the MCF10 parental lines failed to form tumors utilizing protocols with and without hormonal supplementation. The spontaneous derivation of the immortal MCF10 lines marked the first time that normal mammary epithelial cells grown from tissue devoid of malignant or benign lesions were maintained in cell culture without any genetic or chemical transformation. The development and widespread use of MCF10 lines have enabled numerous discoveries in normal mammary epithelial cell biology as well as the fundamental alterations that occur during the progression toward malignancy.

Because of the establishment of MCF10, the use of its most well-known parental line, MCF10A, has allowed breast cancer researchers to model key aspects of normal and aberrant breast epithelial biology. Such biology includes three-dimensional (3D) structural organization and polarity, secretion of and engagement with extracellular matrix, regulation of luminal structures, and functional consequences of common genetic alterations found in malignancy. Importantly, MCF10 sublines, including MCF10AneoT, MCF10AT, MCF10DCIS, and MCF10CA1, have provided indispensable model systems for the characterization of early stage breast lesions and their malignant progression. In this commentary, we discuss seminal work using MCF10 cells to model oncogenic alterations that lead to luminal filling, the functional consequences of common genetic drivers of breast malignancy, and the in situ to invasive breast carcinoma transition (Fig. 1).

Since the pioneering publication by Mina Bissell and colleagues describing a novel method of culturing human breast epithelial cells in 3D (2) and demonstrating that contact with basement membrane has a dramatic impact on cellular phenotypes, breast cancer researchers have utilized this model system to gain insight into the structural and functional characteristics of normal and malignant breast epithelium. In this now widely used culture model, single-cell suspensions of MCF10A cells are embedded in Matrigel (extracellular matrix from mouse Engelbreth–Holm–Swarm sarcomas, which closely resembles the basement membrane) and plated atop a prepolymerized Matrigel layer to avoid cell-to-plastic contact. In this context, single MCF10A cells will proliferate and organize into 3D polarized acini with hollow lumens, which undergo growth arrest within seven to twelve days. MCF10A acini grown in 3D Matrigel recapitulate normal apicobasal polarity, secrete endogenous basement membrane proteins (including laminin V and type IV collagen), and maintain hollow luminal structures. Although primary normal mammary epithelial cells from reduction mammoplasties also maintain these characteristics when cultured utilizing this method, the ability for MCF10A cells to faithfully recapitulate these normal phenotypes has greatly expanded the accessibility and ease with which experiments utilizing this model can be performed.

Seminal work performed by Brugge and colleagues has demonstrated the power of the 3D MCF10A model in elucidating the drivers of malignant progression. Most notably, they showed that ErbB2 (HER2/neu) is sufficient to reinitiate the proliferation of growth arrested acinar structures and provide cell survival signals leading to luminal filling (3). A key follow-up study utilizing the MCF10A 3D model demonstrated that apoptosis is a main regulator of hollow lumen maintenance (4). This study found that while individual perturbation of cell proliferation or apoptosis pathways (via cyclin D1, HPV 16 E7 or Bcl-2, Bcl-xL, respectively) are insufficient to induce luminal filling, oncogenic dysregulation of both apoptosis and proliferation (e.g., by ErbB2 activation) is sufficient to drive this process. In the two decades since these groundbreaking studies first demonstrated the utility of the 3D MCF10A model, numerous other publications have established the value of these models to contribute to our understanding of normal and malignant mammary epithelial cell biology.

The parental MCF10A line is diploid and genetically stable with a largely normal genome. In addition, it is easily cultured and allows for a consistent, reproducible, and accessible model of breast epithelium. Thus, MCF10A cells have been widely used to characterize genetic drivers of malignant progression. Notably, Bachman, Park, and colleagues have made significant contributions to the functional characterization of aberrant epithelial behavior after genetic perturbation. First, Bachman and colleagues demonstrated that deletion of PTEN in MCF10A cells was sufficient to activate PI3K/Akt and MAPK signaling, drive proliferation, and provide apoptotic resistance to matrix detachment and cell rounding (5). The Bachman group then performed a similar study and demonstrated that genetic deletion of TP53 in MCF10A cells led to EGF-independent proliferation, chromosomal instability, and a heterogeneous capacity for anchorage-independent cell growth and invasion (6). Work by Park and colleagues also used the MCF10A cell line to characterize the functional consequences of the common 185delAG BRCA1 mutation (7). In this study, the authors showed that heterozygous 185delAG BRCA1 mutation led to impaired homology-mediated DNA repair, increased copy number loss and loss of heterozygosity, and proposed that BRCA1 haploinsufficiency may drive breast carcinogenesis by facilitating the accumulation of additional genetic alterations.

As these examples illustrate, the MCF10A cell line has been a useful cell culture model for the characterization of genetic drivers of breast malignancy.

MCF10 sublines have provided indispensable models of breast tumor progression, as they have allowed researchers to recapitulate an entire spectrum of breast lesions ranging from transformed lesions (MCF10AneoT), to premalignant lesions with spontaneous malignant progression (MCF10AT and MCF10DCIS), and fully malignant lesions with metastatic capabilities (MCF10CA1). Comprehensive genetic and transcriptomic characterization of the MCF10 progression series has showed that these cell lines accumulate key driver mutations (including in genes like PIK3CA and TP53) at different stages of progression and recapitulate driver mutations seen in primary breast tumors (8). Therefore, the MCF10 subline series is a useful tool to model the progressive molecular and functional changes that occur during breast tumor progression.

Here, we primarily focus on the MCF10DCIS subline due to its unique ability to consistently form lesions resembling ductal carcinoma in situ (DCIS) when injected into immunocompromised mice, which progress to invasive disease with characteristic loss of myoepithelial and basement membrane encapsulation. The ability to mechanistically characterize the drivers of in situ to invasive progression holds significant clinical relevance, as DCIS comprises 20% to 30% of newly diagnosed breast cancer cases in the United States, and our ability to predict risk of invasive progression is limited.

Miller and colleagues pioneered the development and characterization of breast tumor progression models and have generated several of the MCF10 sublines, including MCF10DCIS (initially called MCF10DCIS.com; ref. 9). They were the first to characterize the effect of MCF10DCIS xenografts on the in vivo stromal compartment and showed that the progression of DCIS-like lesions is accompanied by distinct reorganization of the stroma and dynamic signaling between malignant epithelial and nonmalignant stromal cells (10). Our group has also contributed to the detailed cellular and molecular characterization of this model by demonstrating that a subset of MCF10DCIS cells has bipotential progenitor features capable of generating tumor epithelial and myoepithelial cells positive for EpCAM/luminal cytokeratins and p63/SMA/basal cytokeratins, respectively (11). We also showed that coinjection of normal human myoepithelial cells with MCF10DCIS cells suppressed tumor growth and promoted a DCIS histology, while coinjection of fibroblasts increased tumor weight and promoted an invasive histology. Interestingly, the addition of normal myoepithelial cells overcame the tumor growth and invasive progression-promoting effects of fibroblasts.

The tumor microenvironment and tumor–stromal cell interactions greatly influence breast tumor biology and progression. Thus, the faithful recapitulation of the MCF10 sublines for a spectrum of breast cancer lesions and their progression enables researchers to examine these interactions, albeit within an immunocompromised context.

In 1990, Soule and colleagues developed the parental MCF10 cell line series that would drive major discoveries in breast cancer research over the following three decades. The genetically normal and nontumorigenic nature of MCF10 cells has allowed researchers to leverage this accessible model system to study the progression of normal breast epithelium toward malignancy. In this commentary, we reviewed three distinct areas of research that have propelled our understanding of breast cancer biology including the use of MCF10A as a 3D cell culture model, MCF10A in modeling genetic drivers of malignancy, and MCF10 sublines as models of in situ to invasive breast carcinoma transition (Fig. 1). Although due to space constraints we could only discuss a few major studies utilizing the MCF10 models, we would like to acknowledge the many discoveries that are not covered herein and celebrate the continued value that the MCF10 model provides to the biomedical research community.

K. Polyak reports personal fees from Scorpion Therapeutics, personal fees from Vividion Therapeutics, personal fees from Acrivon Therapeutics, and personal fees from Aria Pharmaceutical outside the submitted work. No disclosures were reported by the other author.

The authors thank members of the Polyak laboratory for their critical reading of the manuscript and useful comments. The authors are funded by the National Cancer Institute R35 CA197623 (to K. Polyak) and the DFCI Helen Gurley Brown Presidential Initiative Award (to J. Puleo).