Morphologic transformation of human breast epithelial cells MCF-10A: dependence on an oxidative microenvironment and estrogen/epidermal growth factor receptors

MCF-10A cells were purchased from American Type Culture Collection (ATCC; Manassas, VA). MCF-10A cells were also kindly provided by Drs. J.D. Yager (Department of Environmental Health Science, The Johns Hopkins University Bloomberg School of Public Health, Baltimore, MD; Source #1), K. Eckert (Gittlen Cancer Research Institute, Penn State College of Medicine, Hershey, PA; Source #2), M. Planas-Silva (Department of Pharmacology, Penn State College of Medicine, Hershey, PA; Source #3), and M. F. Verderame (Department of Medicine, Penn State College of Medicine, Hershey, PA; Source #4). The laboratories at Penn State University that kindly provided MCF-10A cells had obtained these cells independently of one another from different sources. A custom formulation of PHR-free Dulbecco's Modified Eagle's Medium/Nutrient F12 (DMEM/F12) cell culture medium D231SA, trypsin (0.25%, 1×), trypsin-ethylenediaminetetraacetic acid (trypsin-EDTA; 0.05% trypsin, 0.53 mM EDTA, 1×), L-glutamine (200 mM, 100×), and antibiotic/antimycotic (100×) solutions were purchased from Atlanta Biologicals (Norcross, GA). Horse serum (HS) was purchased from Invitrogen (Carlsbad, CA). EGF was purchased from R&D Systems (Minneapolis, MN). Protease inhibitor cocktail tablets were obtained from Roche Molecular Biochemicals (Indianapolis, IN). All other reagents were purchased as described in the text or from Sigma Chemical Company (St. Louis, MO).

MCF-10A cells were maintained in PHR-free DMEM/F12 culture medium unless otherwise specified. Medium was supplemented with NaHCO₃ (1200 mg/L), CaCl₂ (1.05 mM), 5% HS, insulin (10 μg/ml), L-glutamine (2 mM), antibiotic/antimycotic mixture (1%), EGF (20 ng/ml), HC (500 ng/ml), and cholera toxin (100 ng/ml). Cells were fed twice a week and grown to confluence before subculturing. Briefly, cells were washed once with Dulbecco's Phosphate Buffered Saline (D-PBS) and exposed to trypsin for 15-20 minutes before the action of trypsin was stopped with 20% HS-supplemented medium. Cells were then centrifuged at 100 × g in a tabletop centrifuge for 5 min and the cell pellet was resuspended in medium and transferred to other flasks. All cells were grown in a single chamber water-jacketed humidified incubator and maintained in a 37°C, 5% carbon dioxide (CO₂) atmosphere. The number of passages cells have been propagated in a particular type of medium is indicated in parentheses next to the description of the medium [(i.e. -HC/-EGF (#10)]

MCF-10A cells maintained in PHR-free 5% HS-supplemented, HC and EGF-containing [+HC/+EGF (+/+)] DMEM/F12 medium were subsequently grown for the 5-week morphologic transformation assay in +/+, -HC/-EGF (-/-), -HC/+EGF (-/+), or +HC/-EGF (+/-) DMEM/F12 media supplemented with 0.5% HS and 240 μg bovine serum albumin (BSA)/ml in the absence or presence of PHR. Cells in these eight medium groups were non-treated (NT) or treated with 0.01% ethanol (EtOH) alone or with 0.01% EtOH solution of 1 nM E2 or 1 nM 2-fluorestradiol (2-Fl-E2). Initially, cells were either left untreated or treated with appropriate agents and then plated in triplicate in 6-well plates at a density of 5 × 10⁵ cells/well. Thereafter, cells were maintained in continuous culture for 6 weeks, refed and re-treated once a week, and examined microscopically each week for signs of contact-uninhibited growth and the appearance of morphologically transformed foci. Transformed foci were counted once a week from 1-5 weeks at 4× magnification as they appeared along two perpendicular lines intersecting in the center of each well. To assess the reversibility of phenotypic cell alterations, after five weeks, PHR, HC, and EGF were added back singly or together, to cultures that were lacking these factors, and the number of transformed foci was again determined at week 6. The assay was performed once with duplicates of each treatment analyzed. Some treated cells were plated in poly-D-lysine-coated tissue-culture plates in an attempt to increase detailed microscopic visualization and examination of foci.

Cells used for Western blot analysis included MCF-10A cells (non-treated and treated under various treatment protocols and media conditions) as well as MCF-10A cells acquired from different laboratories and grown in -PHR, 5% HS, +/+ medium. Total cell extracts were obtained by first trypsinizing and pelleting cells as described in section B of Materials and Methods and washing once with D-PBS. Cell lysis buffer [5.0 M EDTA, 150 mM NaCl, 50 mM Tris HCl, 1% Triton X-100, 1% SDS, 50 mM dithiothreitol (DTT), and protease inhibitor cocktail tablets (1 tablet per 10 ml buffer)] was added to each tube at 100 μl buffer per 1 × 10⁶ cells and mixed well to lyse the cells completely. Lysates were transferred to microcentrifuge tubes, incubated on ice for 10-30 min., and centrifuged at 12,000 × g in a microcentrifuge at 4°C for 15 min. The supernatants were collected and stored at -80°C for subsequent analyses. Alternately, Pierce (Rockford, IL) NE-PER Nuclear and Cytoplasmic Extraction Reagents were used as per the manufacturer's protocol for the stepwise separation and preparation of cytoplasmic and nuclear extracts. Protein content was measured using Bradford Reagent. Proteins (25-30 μg) were resolved by SDS-PAGE in 12% SDS-Tris-HCl polyacrylamide mini-running gels and transferred onto nitrocellulose membranes (BioRad Laboratories; Hercules, CA). Membranes were incubated with primary antibodies to ERα, ERβ, or EGFR at a dilution of 1:1000 in 5% non-fat dry milk-Tris Buffered Saline/Tween (TBS/T) buffer at 4°C overnight, followed by incubation with both the appropriate horseradish peroxidase (HRP)-conjugated secondary antibody at a dilution of 1:10,000 and anti-biotin antibody at 1:1000 dilution in 5% non-fat dry milk in TBS/T at room temperature (RT) for 1 h. Protein was detected using the Western Lightning Chemiluminescent Reagent Plus Kit from PerkinElmer (Wellesley, MA) as per the manufacturer's directions. Antibodies (Ab) and controls used were: ERα (62A3) mouse monoclonal Ab, EGFR rabbit polyclonal Ab, and anti-biotin Ab (Cell Signaling Technology; Beverly, MA); ERβ (PA1-313) rabbit Ab, human, recombinant ERα RP-310 and ERβ (long form) RP-312 (Affinity BioReagents; Golden, CO); EGF-stimulated A431 cell lysate (Upstate Biotechnology; Lake Placid, NY). Peroxidase-conjugated Immunopure goat anti-mouse and sheep anti-rabbit IgG's were used as secondary antibodies (Pierce Chemical Company, Rockford, IL)

Cells were trypsinized and pelleted according to the protocol outlined in Section B, and RNA was isolated from cells using the RNAqueous RNA isolation system (Ambion, Inc.; Austin, TX) according to the manufacturer's protocol. Immediately afterwards, contaminating DNA was removed using Ambion's "DNA-free" DNase Treatment and Removal Reagents again as per the manufacturer's directions. The RNA supernatants were transferred to new RNAse-free tubes and stored at -80°C. Prior to use in gene expression studies, the concentration and purity of RNA was determined by aliquoting a small amount of the samples in HPLC-grade, RNase-free dH₂O and measuring absorbance at 260 nm and 280 nm. RNA concentration was calculated using a value of 1A_{260 nm} = 40 μg RNA/ml. and its purity assessed by confirming that the ratio of A₂₆₀/A₂₈₀ was near 2.0. Nonrad-GEArray Kit Pathway Specific Gene Expression Profiling System (SuperArray, Inc; Bethesda, MD) was used for the analysis of gene expression after RNA isolation. The detailed manufacturer's protocol was followed for analysis. Briefly, biotinylated cDNA probes were synthesized from 5-10 μg total RNA by reverse transcription using a PCR thermal cycler and SuperArray reagents. Afterwards, cDNA probes were hybridized using a mini hybridization incubator kit reagents to pathway-specific gene expression array membranes (either human toxicity/stress or estrogen signaling) provided by the manufacturer. Finally, membranes were incubated with alkaline phosphatase (AP)-streptavidin, and chemiluminescent detection was performed with the provided CDP-Star substrate and immediate exposure to x-ray film between 0-5 min. Signal intensities were quantitated (semi-log) using UN-SCAN-IT digitizing software (Silk Scientific; Orem, UT) after the x-ray films were scanned onto a computer. Sample signal intensities were normalized against a housekeeping gene's signal intensities. Each membrane contained two spots for each cDNA analyzed. Means of the intensity (in pixels) of the duplicate spots were used for analysis

Significance of differences between two groups was assessed using one-tailed Student's "t" test assuming unequal variances. One-way ANOVA followed by Dunnet's test was utilized to compare all groups to a control group, while One-way ANOVA followed by Tukey's test was used to compare all groups to each other. For all tests, p < 0.05 was considered significant.

MCF-10A cells were grown for 5 weeks in continuous culture in 6-well tissue culture plates in PHR-containing or PHR-free medium lacking either one, both, or none of the factors HC and EGF. Cells in these four types of media were also exposed once a week to either 1 nM E2 or 1 nM 2-Fl-E2 in 0.01% EtOH, or just 0.01% ethanol (EtOH) as a control. Within five days in culture, cells grown in the presence of EGF [+HC/+EGF (+/+) and -HC/+EGF (-/+)] became confluent, whereas those grown in the absence of EGF [-HC/-EGF (-/-) and +HC/-EGF (+/-)] displayed a slower growth rate. Cells in the +/- medium exhibited the lowest proliferative capacity. Transformed-looking foci were noted to first appear after 13 days in culture and only in PHR-free -/- cultures (Fig. 1e-f). After 5 weeks, large, prominent foci became apparent but again, only in PHR-free -/- cultures (Fig. 2e-f). The average number of foci per well of a 6-well tissue culture plate was significantly elevated to 79.5+6.50 (p < 0.0005) in -/- medium versus 0 in media containing either HC or EGF (Table 1). An important outcome of these experiments was the observation that the presence of PHR inhibits the appearance of transformed foci (Figs. 1 & 2; Table 1). To verify that the piled cells observed in Figs. 1 and 2 were live cells and rule out the possibility of artifactual observation, MCF-10A cells seeded in triplicate at a density of 6.25 × 10⁴ cells/well in PHR-containing or -deficient -/- medium for 3 weeks in 6-well tissue culture plates were stained with neutral red (10 μg/ml) and photographed. Transformed foci in PHR-deficient -/- medium as well as the monolayer underneath were stained red, confirming the viability of cells both in foci and monolayer (Fig. 3b). The fact that foci were stained deep red while the monolayer was light red suggests piling of cells in foci and proves that foci being observed were not artifacts. Cultures grown in PHR-containing -/- medium did not form foci, as demonstrated by the absence of darkly stained piles of cells (Fig. 3a).

In an attempt to correlate the transformative and oxidative capacities of E2 within human breast epithelial cells, the MCF-10A cell line was exposed to 1 nM E2 within the four media types already discussed (+HC/+EGF, -HC/-EGF, -HC/+EGF, and +HC/-EGF). In addition, cells cultured in -/- growth medium were also treated with 2-Fl-E2 to assess the transforming potential of a non-redox cycling estrogen. +/+, -/+, and +/- cultures were already shown to be resistant to transformation, since E2 treatment of these cultures did not induce transformation (Figs. 1 & 2; Table 1) within the five-week treatment protocol. However, transformed foci appeared in E2-treated -/- cultures starting from 13 days of exposure (Fig. 1f). Although at 13 days, foci arising in E2-treated cells were light and less prominent (Fig. 1f) than in non-treated cultures (Fig 1e), by 5 weeks the foci in E2-treated cultures were much denser and larger in size (Fig. 2) than they were at 13 days. While foci sizes in the non-treated cells were still larger in comparison to the E2-treated cultures (Fig. 2e-f) even at 5 weeks, E2 treatment resulted in more numerous foci (Table 1), by over 2-fold in PHR-free -/- cultures in comparison to non-treated controls (p < 0.0005). In contrast to E2, exposure of cells to 2-Fl-E2 in -/- medium did not lead to the appearance of transformed foci at 13 days (data not shown). However, at 5 weeks, transformed foci did appear in 2-Fl-E2-treated -/- cultures within certain wells and/or areas of wells where cellular degradation was apparent (Fig. 4b). In wells/areas of apparent healthy cell growth in 2-Fl-E2-treated cultures, no foci were discernable (Fig. 4a). In contrast, E2-induced foci in -/- cultures formed uniformly among and within wells irrespective of cellular disintegration (Fig. 2f). Nonetheless, the apparently transformed foci in 2-Fl-E2-treated cultures were counted and the total foci number was used as a reliable comparative indicator of transformation potential among treatments. After five weeks of growth, E2 significantly (p < 0.0005) enhanced the transformation of MCF-10A cells in -/- medium in comparison to controls, whereas 2-Fl-E2 was unable to increase the basal transformation rate, as evidenced by the formation of a lower number of foci than in controls (Table 1). E2-treated cells formed over five times the number of foci as EtOH controls in -/- medium (Table 1).

To evaluate whether microenvironmental changes induce irreversible morphologic transformation of MCF-10A cells, PHR, HC, or EGF alone as well as HC and EGF together were re-introduced to -PHR/-HC/-EGF culture medium after the formation of foci in 5 weeks. Microscopic evaluation after 6 weeks revealed no obvious changes aside from indications of slightly smaller foci in cultures where either HC or EGF was added (data not shown). Importantly, only the addition of HC induced a partial (~40%) but significant (p < 0.01) decrease in foci numbers (Table 2). Adding back EGF, by itself or with HC, resulted only in a non significant increase in the number of foci.

Slight microenvironmental alterations were seen to profoundly affect the morphology and growth characteristics of MCF-10A human breast epithelial cells. An intriguing characteristic of foci noted in this cell line was the appearance, once foci began to form, of extensive interconnections among them, forming a lattice-like network (Fig. 5a-c). To enable more detailed visualization of transformed foci, 10 nM E2-treated MCF-10A cells were grown in -/- medium until foci formed, trypsinized, and re-plated in poly-d-lysine culture flasks. After 50 days, foci photographed at 40× magnification enabled the identification of a somewhat polarized egg-shaped structure with a distinct, prominent membrane, perhaps of more than one layer, containing the growing contact uninhibited cells within its boundary (Fig. 5d). Growth characteristics of MCF-10A cells grown in the three different types of medium (+/+, -/-, -/+) were examined. Five passages after reaching confluence in their respective medium type, differences among the cells became apparent. Figure 6 shows subconfluent cultures of +/+, -/- (passage #11), and -/+ (passage #13) cells 24 hours after plating. In contrast to +/+ cells, which exhibit normal subconfluent epithelial growth (Fig 6a), -/- cultures already show the presence of small foci growing atop clusters of monolayer cell growth (Fig. 6b). Extensions among foci at this time have already begun to form. MCF-10A -/+ cultures did not form foci at this early time, but displayed a unique morphology of extended, fibroblast-like cells with respect to the other two types of cultures (Fig 6c). Therefore, even a slight microenvironmental alteration can profoundly affect the growth characteristics of human breast epithelial cells, with serious implications regarding their transformation potential.

The necessity for PHR depletion in our studies indicated a need for ERα-mediated events in the transformation of MCF-10A cells. PHR's known weak estrogenicity and binding to ERα [32, 33] may suppress receptor's binding of E2, thereby preventing the relatively more potent hormonal and/or oxidative properties of E2 from effecting transformation. Although this cell line is categorized as ERα-negative, our results made it necessary to re-examine the ERα status of these cells. A commercially available gene expression array kit (SuperArray; Bethesda, MD) was utilized for the analysis of human estrogen signaling pathway gene expression in non-treated and 1 nM E2-treated (twice a week for 2 weeks) MCF-10A cells grown under various media conditions. Table 3 demonstrates detectable gene expression of both ERα and ERβ among the different growth and treatment conditions for MCF-10A cells assessed. ERα gene expression was consistently higher in comparison to ERβ expression levels. Although no indicative trends and/or upregulation in ER expression due to varying treatments and culture conditions was noted, the data demonstrate persistent, detectable levels, particularly of ERα gene expression, in this cell line.

MCF-10A cells maintained in different growth conditions and those acquired from various laboratories were subjected to Western blotting analysis for ERα, ERβ, and EGFR expression (Fig. 7). Nuclear extracts from 16 samples containing 42 μg protein were electrophoresed onto two 12% SDS-Tris-HCl polyacrylamide gels (gel 1: Fig. 7a, c, e & gel 2: Fig. 7b, d, f) and transferred to nitrocellulose membranes. The two membranes were then probed consecutively with antibodies directed against ERα (Fig. 7a & 7b), ERβ (Fig. 7c & 7d) and EGFR (Fig. 7e & 7f). The first three lanes of both gels contain, respectively, control human recombinant proteins ERα and ERβ from baculovirus-infected Sf9 cells, and total cell lysate from EGF-stimulated A431 cells (positive control for EGFR) (Fig. 7a-f; lanes 1-3 & 9-11). All other lanes contain MCF-10A sample nuclear extracts. Membranes were first probed with antibodies directed against ERα, which detect a 66 kD protein. These blots are depicted in Figure 7a-b and appear to show very light but detectable bands at 66 kD in all MCF-10A sample lanes (Fig. 7a-b; lanes 4-8 & 12-16) corresponding to ERα, as corroborated by staining in this region in the ERα positive control (Fig. 7a-b; lanes 1 & 9). Bands are absent in the lane containing recombinant ERβ protein (Fig. 7a-b; lanes 2 & 10), verifying the absence of non-specific binding. Interestingly, EGFR control cell extracts from A431 cells also appear to indicate the presence of even lighter, but still detectable bands at 66 kD (Fig. 7a-b; lanes 3 & 11). A431 are not known to express ERα, but such a possibility is suggested by observed growth inhibition by the antiestrogen tamoxifen [34] and growth stimulation by the estrogenic compound genistein [35] in A431 cells. The same two membranes were re-probed with antibodies directed against ERβ (Fig. 7c-d). Bands around 60 kD appear to be present in the lane containing the recombinant ERβ (Fig. 7c-d; lanes 2 & 10), which has a molecular weight of 59.2, and in all other MCF-10A samples tested (Fig. 7c-d; lanes 4-8 & 12-16). Bands are not present in the lane containing recombinant ERα (Fig. 7c-d; lanes 1 & 9), again confirming the absence of non-specific antibody binding. A431 control cell extracts do not show the presence of ERβ (Fig. 7c-d, lanes 3 & 11) as they did ERα (Fig. 7a-b; lanes 3 & 11). Re-probing the gels with antibodies directed against EGFR (Fig. 7e-f), which detect the protein at ~175 kD, indicates the presence of bands between 140 and 200 kD, again in all MCF-10A samples (Fig. 7e-f; lanes 4-8 & 12-16) and in the EGF-stimulated A431 cell extracts (Fig. 7e-f; lanes 3 & 11), which serve as a positive control for EGFR. Bands are not seen in lanes containing human recombinant ERα (Fig. 7e-f; lanes 1 & 9) or ERβ (Fig. 7e-f; lanes 2 & 10). These data taken together appear to confirm the presence of ERα, ERβ, and EGFR in all MCF-10A samples tested, even in those obtained from different laboratories, and suggest the possible presence of ERα in the human squamous cell carcinoma line A431. What is even more intriguing in these Western blots is the presence of a band at ~200 kD in all MCF-10A samples and A431 samples that is detected by antibodies directed against both EGFR and ERα (Fig. 7a-b & 7e-f; lanes 3-8 & 11-16). This band appears to be induced in MCF-10A cells grown for 28 passages in HC- and EGF-depleted medium (Fig. 7a & e; lane 6). This band is not seen only in the photograph (Fig. 7f, lane 12), but is present as an extremely faint band in the original x-ray film of the newly acquired MCF-10A cells grown in +/+ medium. While highly speculative, an association between EGFR and ERα in both MCF-10A and A431 cells, which is upregulated in MCF-10A -/- #28 cells is one possible explanation for these results. Interestingly, in blots which were analyzed with ERβ antibodies (Fig. 7c-d), this ~200 kD band is apparent only in A431 (Fig. 7c-d; lanes 3 & 11), MCF-10A -/- #28 (Fig. 7c; lane 6), and MCF-10A -/+ #71 (Fig. 7c; lane 8) cells. Again, one possible explanation for this, which will need to be explored further in order to be confirmed, is the existence of a ternary association of ERα/ERβ/EGFR unique to these particular cells. The absence of detectable ERβ at ~60 kD (Fig. 7c-d; lanes 3 & 11) in A431 cells may be an indication of low levels of ERβ in these cells which preferentially associate with EGFR/ERα complex. Overall, however, Western blot analysis appeared to demonstrate detectable levels of ERα, ERβ, and EGFR in MCF-10A cells grown under various culture conditions in our lab and those obtained from a variety of sources. These data also initiate speculation of a direct association of EGFR and ERα in MCF-10A cells. MCF-10A cells grown long-term in HC/EGF-depleted medium (-/- #28) appear to display a high induction of this complex but with the added presence of ERβ. This putative ternary complex is also slightly induced in MCF-10A cells chronically depleted of HC and propagated for a longer time (-/+ #71), presented in lane 8.

In support of the role of an active ERα component in MCF-10A cell transformation, Table 4 demonstrates that 1 nM E2 treatment of -/- MCF-10A cultures upregulated prolactin (PRL) gene expression by over 6-fold, which was statistically significant (p< 0.05), while it down-regulated progesterone receptor (PR) gene expression by over 2-fold, suggesting hormonal estrogen responsiveness of this cell line. PRL gene expression was also seen to be modulated in the presence of EGF by the long-term depletion of HC, where -/+ (#21) MCF-10A cells exhibited a decrease in PRL gene expression by 4-fold (p < 0.05) in comparison to +/+ cultures (Table 5). Expression of several other genes involved in the estrogen signaling pathway was also modulated in response to the chronic depletion of HC from MCF-10A culture medium (Table 5). Expression of the genes c-fos and c-jun was diminished by nearly 2- and over 3-fold, respectively, while c-myc expression was abrogated entirely. ER-binding fragment-associated antigen9 (EBAG9) and EGF gene expression were downregulated by over 4-fold each, while H-ras by 3-fold.