Background
Universal features of breast cancer include a loss of control of cell proliferation and organization, a loss of cell polarity, as well as a loss of cell: cell adhesion and cell: basement membrane adhesion. Monolayer cultures of mammary epithelial cells do not closely mimic these features of
in vivo cell and tissue architecture [
1,
2], and also do not recapitulate the alterations in nuclear structure characteristic of breast tumors [
3]. The development of three-dimensional culture systems for mammary epithelial cells has been an important advance in cell culture models, more closely mimicking
in vivo architecture and are, therefore, more relevant systems in which to elucidate changes in cell growth, cell: cell adhesion, cell: cell junctions, and cell: extra-cellular matrix (ECM) interactions that occur during malignant progression [
1,
4,
5].
MCF-10 cells were derived from a patient with fibrocystic disease and the immortalized MCF-10A line arose spontaneously in culture [
6]. MCF-10 cells are diploid, while the MCF-10A line has a stable, near-diploid karyotype [
6,
7] with modest genetic modifications typical of culture-adapted breast epithelial cells [
8] including loss of the p16 locus [
9]. The cells express normal p53 [
9,
10], they do not form colonies in soft agar, and they do not grow in immuno-compromised mice [
11].
Several transformed lines have been derived from MCF-10A cells. The MCF-10AT line was created by forced expression of activated H-ras [
12,
13]. Implanted as xenografts, MCF-10AT cells formed nodules that progressed from hyperplasia to carcinoma
in situ in 25% of mice [
12,
14]. Further selection for tumor formation in mice generated the series of MCF-10CA1 cell lines that are fully malignant, producing tumors in 100% of immuno-compromised mice transplanted with the cells [
15,
16]. Thus, the MCF-10A series of matched cell lines provides a complete spectrum of cell phenotypes, from normal to pre-malignant to fully malignant.
MCF-10A cells cultured in three-dimensional reconstituted basement membrane culture (rBM) develop important features of normal breast tissue via a well described progression of proliferation, cell cycle arrest, apical-basolateral polarization, and finally, apoptosis to create a luminal space [
9,
17,
18]. In addition, cell nuclei of MCF-10A cells forming acini in three-dimensional rBM culture display a reorganized and differentiated nuclear architecture more characteristic of mammary epithelial cells in tissue than those cultured in monolayer [
3]. The resulting acini of MCF-10A three-dimensional rBM cultures resemble those of normal breast tissue, which are clustered in lobules that connect to intralobular ductules that, in turn, connect to interlobular ducts.
The MCF10A series of cell lines provides a unique opportunity to probe malignant progression induced in a molecularly defined way, in a common cell background. We have characterized their ability to form tissue-like structures in a three-dimensional microenvironment. In contrast to MCF-10A cells, MCF-10ATcells, when cultured in three-dimensional in rBM, form multi-acinar structures, and MCF-10CA1a cells, when also cultured in three-dimensional in rBM, form large hyper-proliferative cell aggregates with altered organization and no lumen. We conclude that pre-malignant cells have lost normal proliferation control but retain normal cell polarity, cell:cell, and cell:basement membrane adhesion. Fully malignant cells, on the other hand, have lost the ability to undergo normal acinar morphogenesis. The application of three-dimensional rBM culture to this unique progression series of cell lines with a matched cell background provides a physiologically relevant in vitro model of breast cancer progression.
Discussion
The MCF-10A progression series grown in three dimensional rBM cultures can be a powerful system for studying human mammary tumorigenesis. All three cell lines derive from a common ancestor and the manipulations that have made them increasingly transformed are well defined. While much can be learned by studying these cells in traditional two dimensional tissue culture, real malignant transformation takes place within a three dimensional tissue. The three dimensional rBM system we have utilized to characterize these cells recapitulates much more of the microenvironment of a tissue and allows for a more detailed investigation of the dynamic and reciprocal crosstalk between the extracellular matrix and nuclear gene expression that may play an important role in real breast tumorigenesis [
1,
3,
5,
17,
18]. In this study, we assessed cellular parameters of tissue formation in cells of the MCF-10A progression series and documented them as a faithful model of mammary malignant progression.
Variants of differing malignancy have been developed from the first normal human mammary epithelial cell line, HMT-3522 [
20], by spontaneous transformation with selection for EGF-independent growth [
21‐
23]). These variants, cultured in rBM, have offered significant insights into tumor progression [
24‐
27]. However, the molecular mechanisms driving these cells toward malignancy were not defined in advance, although some information about the process might be inferred from gene expression studies of the established lines [
28].
MCF-10A cells were manipulated by the forced expression of a mutated H-ras-1, which is the human analog of the Harvey sarcoma virus oncogene. This constitutively active Ras variant, which was originally isolated from the bladder carcinoma cell line T24, has a single base mutation causing an amino acid change at residue 12 (Gly to Val) [
29]. When injected into nude mice, the resulting MCF-10AT cells persisted as xenografts and ultimately developed into carcinomas 25% of the time [
12,
13]. In tissue culture studies, expression of this constitutively activated Ras induced transformation of immortalized NIH3T3 cells and of primary rat and hamster cells from multiple tissues, whereas the normal Ras did not [
29,
30]. While few established breast cancer cell lines have Ras mutations, a majority have Ras overexpression via amplification of the gene [
31‐
33]. More recently, activating mutations in H-ras, K-ras, or N-ras were observed in 8 out of 40 human breast cancer cell lines, with another 4 lines having activating mutations in B-Raf, which is downstream in the Ras pathway [
34]. Although only 5% of breast tumors have activating Ras mutations [
31], it was shown that more than half had a 2- to 6-fold increase in Ras expression and a more than 5-fold activation in downstream MAP Kinase activity [
35]. Thus, Ras signaling may be more commonly activated in breast tumor cells by mechanisms other than mutation. In MCF-10A cells, ErbB1-ErbB2 heterodimerization disrupts acinar tissue structure in three dimensional basement membrane cultures and generates a more invasive phenotype [
36]. The invasive phenotype was reverted by inhibition of the Ras/MAP Kinase pathway, supporting a role for the Ras pathway in the malignant progression of ErbB2 expressing tumors. For these reasons, the ectopic expression of activated Ras in MCF-10A cells is likely relevant to the mechanisms of patient breast tumor progression.
While MCF-10AT cells and normal MCF-10A cells grown in three-dimensional reconstituted basement membrane culture both exhibited growth arrest, this arrest appeared to be delayed in the MCF-10AT cells. These cells also demonstrated a nearly complete lack of apoptosis, resulting in the absence of notable lumen formation. A recently published report indicated that in MCF-10AT cells grown in three-dimensional culture, apoptosis was decreased, but not absent, leading to reduced lumen formation [
37]. MCF-10AT cells were normal or nearly normal in their ability to deposit basement membrane and to form cell:cell junctions. Thus the activation of Ras predominantly affects the proliferation state of the cells in three-dimensional culture, consistent with the idea that these cells recapitulate aspects of breast tissue hyperplasia
in vivo [
38].
In contrast, the malignant MCF-10CA1a cells never exhibited growth arrest and never formed normal acinar structures, consistent with reports comparing primary mammary epithelial cells, primary breast tumor cells, and transformed breast cancer cell lines [
4]. The MCF-10CA1a cells exhibited unchecked proliferation in three-dimensional culture despite the continuous occurrence of apoptosis, suggesting the loss of normal growth controls. In addition, the MCF-10CA1a cells showed an incomplete deposition of basement membrane proteins and abnormal cell:cell contacts. Interestingly, the expression levels of the markers assayed in these experiments were not significantly altered. Instead, the proper structural organization of the proteins within the cells was compromised. These phenotypes are consistent with more advanced stages of tumorigenesis, at which the dynamic relationships between cell:cell interaction and cell:ECM contact are altered [
1,
39].
These progressively malignant cell lines were engineered in the same cell background by a defined series of manipulations. They recapitulate tumor phenotypes in a three dimensional culture system, and this will facilitate future studies of malignant progression. Compared to in vivo studies, organotypic culture systems are easier and less expensive to manipulate, facilitate fluorescence based microscopy imaging, have a reduced complexity of cell types in a simplified microenvironment, and are influenced by a reduced number of uncontrolled factors. This simplicity can be a disadvantage, ignoring features not included in the tissue culture, but it can also be a great advantage for mechanistic studies of malignant changes. Of particular interest is the potential use of these cell lines in three dimensional culture to dissect the signal transduction pathways, and especially the ras pathway, that drive these cells to increasing malignancy.
Methods
Cell Culture
MCF-10A, MCF-10AT (formally known as MCF-10AneoT), and MCF-10CA1a cells were obtained from the Barbara Ann Karmanos Cancer Institute (Detroit, MI). The three cell lines were maintained in monolayer in Dulbecco's modified Eagle's medium-F12 (DMEM/F12) (Invitrogen, 21041025) supplemented with 5% horse serum (Invitrogen, 16050122), 1% penicillin/streptomycin (Invitrogen, 15140122), 0.5 μg/ml hydrocortisone (Sigma, H-0888), 100 ng/ml cholera toxin (Sigma, C-8052), 10 μg/ml insulin (Sigma, I-1882), and 20 ng/ml recombinant human EGF (Peprotech, 100-15).
Three-dimensional Cell Culture
MCF-10A, MCF-10AT, and MCF-10CA1a cells were cultured on plastic tissue culture dishes (Nalge Nunc, Rochester, NY,172958) in either Reduced Growth Factor Matrigel without phenol red or Non-Reduced Growth Factor Matrigel with phenol red (BD Biosciences, San Jose, CA, #3562312 and #356234) following either the overlay or embedded procedures of Debnath et. al. [
9]. Briefly, for overlay cultures, cells were prepared for three-dimensional culture by growing to 20–80% confluency in monolayer and seeding in a single cell suspension on 100 μl of Matrigel in a 35mm plate at 7,000 to 15,000 cells/plate or on 40 μl Matrigel in a 4 well chamber slide (Nunc, Lab-Tek, 177399) at 5000 cells/well. For embedded cultures, cells were grown up to 80% confluence in monolayer culture and plated at 2.5 × 10
4 cells/ml in 300 μl Matrigel, on top of a thin layer of 75 μl Matrigel, in each well of a 4 well chamber slide (see Figure
8A). In addition, some overlay three-dimensional cultures of MCF-10CA1a cells were prepared with cells grown in monolayer under minimal media conditions, consisting of DMEM/F12 with 5% horse serum, 0.029M sodium bicarbonate, 10mM HEPES and 1% penicillin/streptomycin (see Figure
8, panels G and H).
Once plated on rBM, all cultures were incubated at 37°C in a 5% CO
2 humidified incubator for up to 20 days. Three-dimensional cultures were grown in assay media [
9] consisting of DMEM/F12 containing 2% horse serum, 5 ng/ml EGF, and 2% Matrigel (for overlay cultures only). Media was replaced every 2 or 4 days. Morphology was observed every 2 days via phase contrast microscopy. Size of resulting structures were measured every 2 days and analyzed via an analysis of variance (ANOVA) test followed by a Tukey adjustment.
Cryoblock Preparation
MCF-10A, MCF-10AT, and MCF-10CA1a cells were harvested at day 6, 12 or 20 by first washing with warm assay media, followed by fixing with 4% formaldehyde in PBS for 20 minutes at room temperature, then permeabilizing with 0.5% Triton X-100 in PBS for 10 minutes at 4°C. Cultures were rinsed with a glycine rinse (130 mM NaCl, 7 mM Na2HPO4, 3.5 mM NaH2PO4, 100 mM glycine) three times for 20 minutes each at room temperature, followed by a blocking step with TBS-1 (10 mM Tris HCL, pH 7.7, 150 mM NaCl, 3 mM KCl, 1.5 mM MgCl2, 0.05% (v/v) Tween 20, 0.1% (w/v) bovine serum albumin, 0.2% (w/v) glycine) for 30 to 40 minutes at room temperature. Cultures were transferred to 1.5 ml Eppendorf tubes; TBS-1 was aspirated and replaced with 15% sucrose for 15 minutes at room temperature, aspirated, and replaced with 30% sucrose for 15 minutes at room temperature. Sucrose was replaced with Tissue Tek O.T.C. compound (Sakura Finetek, #4583, Torrance, CA) or with TBS Tissue freezing medium (Triangle Biomedical Sciences, H-TFM, Durham, NC) and samples were flash frozen in liquid nitrogen and stored at -80°C.
Immunofluorescence
Antibodies used for immunostaining included β4 integrin (Chemicon, MAB#1964), human Laminin V (Chemicon, MAB#19562), pan-Cadherin(Sigma, #C3678), Ki-67 (Zymed, #18-0191), and cleaved Caspase 3 (Cell Signaling, #9664).
To immunostain three-dimensional samples, a thin layer of Matrigel (10 μl) was used to coat the bottom of each well of a 4-well slide and allowed to gel for at least 20 minutes at 37°C. A 10 or 20 μl sample of cells was layered on top of the thin Matrigel layer and incubated for at least 20 minutes at 37°C. Samples were then washed with warm media, followed by one of three fixing methods. Cells were either fixed and then permeabilized or permeabilized and then fixed using 4% formaldehyde in PBS for 20 minutes at room temperature as the fixative, and using 0.5% Triton X-100 in PBS, 10 minutes at 4°C for permeabilization. Alternatively, for caspase staining (Figure
5) cultures were fixed with methanol: acetone (1:1) for 20 minutes at -20°C. Following all fixation methods, samples were rinsed with a glycine rinse, 3 times for at least 10 minutes each time while rocking gently. This step was followed by a block with IF buffer (130 mM NaCl, 7 mM Na
2HPO
4, 3.5 mM NaH
2PO
4, 7.7 mM NaN
3, 0.1% Bovine Serum Albumin, 0.2% Triton X-100, 0.05% Tween 20) containing 10% goat serum (Invitrogen, #10000C) for 1 hour 30 minutes, followed by a secondary block in IF buffer containing 10% goat serum and 20 μg/ml F(ab')
2 (Jackson Laboratories, #115-006-006) for 30 to 40 minutes. Samples were incubated with primary antibodies in secondary block solution overnight at 4°C. Antibody concentrations used were β4 integrin (1:20), Laminin V and pan-cadherin (1:200), Ki-67 and estrogen receptor (1:100), and caspase (1:200). Samples were washed three times for 20 minutes in IF buffer while gently rocking, followed by a 1-hour incubation with the appropriate Alexa-conjugated secondary antibodies (Molecular Probes) at a concentration of 1:200. Nuclei were counterstained with Draq 5 (Alexis Biochemicals, BOS-889-001-R200) prior to a final wash with PBS and mounting with Prolong Antifade Kit (Molecular Probes, P-7481, Eugene OR).
To immunostain cryo-sections, they were first allowed to warm to room temperature in a humidity chamber for 15 minutes. All subsequent steps, except primary antibody incubation, were carried out at room temperature. The sections were washed with PBS with 0.05% Tween 20 for 20 minutes. Sections were blocked with IF buffer containing 10% goat serum for 1 hour 30 minutes, followed by a second block in IF buffer, 10% goat serum and 20 μg/ml F(ab')2 for 30 to 40 minutes. Sections were washed for 20 minutes in Signal Enhancer (Image-iT FX, Invitrogen, #136933) and incubated with primary antibodies in IF buffer with 10% goat serum overnight at 4°C. Antibody concentrations were as described above. Sections were washed three times for 20 minutes each in IF buffer followed by 1 hour incubation with appropriate secondary antibodies (Molecular Probes) at a concentration of 1:2000. Sections were again rinsed 3 times with IF buffer for 20 minutes each followed by labeling with Draq 5, and a final wash with PBS before mounting with Prolong Antifade Kit. Images were captured using laser scanning confocal microscopy with a Leica SP1.
Acknowledgements
We thank Jean Underwood (UMASS Medical School, Worcester, MA, USA) and Christopher Altamirano (UMASS Medical School, Worcester, MA, USA) for early technique development and Nathalie Cohet (UMASS Medical School, Worcester, MA, USA) for helpful discussions. We thank Stephen Baker (UMASS Medical School, Worcester, MA, USA) for assistance with the statistical test and Valerie Weaver (UCSF, CA, USA) for her comments and suggestions. This work was supported by NIH P01 CA82834.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
KI designed and carried out all of the experiments, and prepared the manuscript. IT carried out some of the experiments. AI and JN conceived and designed experiments and critically reviewed the manuscript.