Introduction
Drug resistance represents a major obstacle to successful therapy of breast cancer, a leading cause of death among women in Western countries [
1]. It is well known that several ATP-binding cassette transporters, such as MDR (multidrug resistance), MRP (multidrug resistance associated protein), and BCRP (breast cancer resistance protein), are related to the development of drug resistance in breast cancers [
2‐
4]. However, many other proteins – including glutathione-S-transferase [
5], β
2-microglobulin [
6], heat shock protein (HSP)27 [
7,
8], 14-3-3σ [
9,
10], and vimentin [
11] – have also been implicated in breast cancer drug resistance. These findings were based upon studies using various chemoresistant breast cancer cell lines such as adriamycin, verapamil, tamoxifen, vinblastine, and paclitaxel resistant MCF-7 cells. Although some aspects of the mechanisms of drug resistance have been characterized, the highly variable response to chemotherapy in the treatment of breast cancers remains poorly understood. Elucidating these drug resistance mechanisms is essential for improving tumor responses to clinical chemotherapies.
A growing area of interest that may reveal one such mechanism is the association of drug resistance with epithelial-mesenchymal transition (EMT) in cancer. EMT is the process by which adherent epithelial cells convert to motile mesenchymal cells and is essential in embryonic development. However, it appears that aberrant activation of EMT occurs in cancer progression [
12], and is involved in highly aggressive, poorly differentiated breast cancers with increased potential for metastasis and recurrence [
13]. EMT has been linked to resistance to various drugs in cancer, including tamoxifen resistance in breast carcinoma cells [
14], paclitaxel resistance in epithelial ovarian carcinoma cells [
15], oxaliplatin resistance in colorectal cancer cells [
16], gemcitibine resistance in pancreatic tumor cells [
17], cetuximab resistance in hepatoma cells [
18], and erlotinib resistance in non-small-cell lung carcinomas [
19]. The activities of several genes are known to contribute to EMT, including decreased expression of E-cadherin, and increased expression of snail, slug, and δ-EF1 (ZEB1) [
20]. Increased expression of vimentin [
21] and N-cadherin [
22] are also seen in EMT. Evaluation of these markers in a drug-resistant cell line may shed light on the relationship between EMT and drug resistance.
TNF-α is a multifunctional cytokine that elicits a variety of biologic responses, such as inflammation and apoptosis [
23]. Additionally, TNF-α has been shown to induce EMT [
24,
25]. Although TNF-α is not currently an anticancer agent for treatment of human cancers (because of side effects such as normal cell toxicity), low doses of TNF-α can markedly sensitize cancer cells to chemotherapy-induced apoptosis [
26,
27]. We previously demonstrated that MCF-7 cell line variants exhibit differences in sensitivity to TNF-α and apoptosis induced by taxol and doxorubicin [
28‐
30]. Specifically, we demonstrated that apoptosis sensitive MCF-7-N cells (MCF-7 N variant) exhibited distinct differences in cell survival and apoptotic signaling when compared with inherently resistant MCF-7-M cells (MCF-7 M variant) [
28]. We further demonstrated that apoptosis sensitive cells (MCF-7-N) could be driven to a resistant phenotype through prolonged exposure to increasing concentrations of TNF-α, leading to a stable, apoptosis-resistant phenotype (MCF-7-TNR) that was in part dependent upon mitogen-activated protein kinase (MAPK) and nuclear factor-κB signaling [
29]. Gene expression profiling revealed that MAPK kinase (MEK)5 was over-expressed in the TNF-α resistant MCF-7-M cells versus the TNF-α sensitive MCF-7-N cells [
31]. A similar upregulation of MEK5 in resistant cells was independently described in MCF-7-F cells, which developed resistance to the pure anti-estrogen fulvestrant through prolonged growth in fulvestrant-containing media [
32]. These studies demonstrate a potential role for the MEK5 pathway in the regulation of progression to drug resistance in breast cancer.
The MEK5/extracellular signal-regulated kinase (Erk)5 tandem is a component of MAPK cascades that mediate signals from various extracellular stimuli to the nucleus and regulate most cellular processes [
33], including gene expression, proliferation, apoptosis, and motility [
34,
35]. MAPK signaling may also play a role in EMT [
12,
36]. Although MEK5/Erk5 signaling has not been extensively investigated, several studies suggest a role in cancer progression. For example, MEK5/Erk5 signaling has been demonstrated in prostate and breast cancer proliferation and tumorigenesis [
37,
38]. Furthermore, inhibition of MEK5/Erk5 signaling in the MDA-MB-231 cell line, an aggressive breast cancer cell line with an EMT phenotype, induces apoptosis [
39].
Based on these findings, which strongly indicate that MEK5/Erk5 signaling may mediate cancer progression to an aggressive phenotype, we further explored the involvement of MEK5/Erk5 signaling in resistance to apoptosis as well as EMT. To test the role played by MEK5/Erk5 activation in progression of breast carcinoma cells to a resistant phenotype, MCF-7 cells (N variant) were used to stably express a constitutive active MEK5 construct. These MCF-7-MEK5 cells exhibit resistance to TNF-α as compared with stable vector cells (MCF-7-VEC). Proteomic analysis based on two-dimensional electrophoresis (2-DE) and various mass spectrometric techniques has been employed in several studies of drug resistance of breast cancers [
8‐
10,
40]. In this study we used a proteomics approach to define mechanisms of the MEK5 signaling cascade in the regulation of drug resistance. The differentially expressed proteins identified in proteomic analyses were confirmed at the gene expression level using reverse transcription RT-PCR assays. Results using immunofluorescence staining and gene expression analysis were consistent with an EMT phenotype in the MCF-7-MEK5 cells. These findings identify a potential role for the MEK5 pathway in coordinately promoting both an EMT phenotype and TNF-α resistance in breast cancer.
Materials and methods
Stable transfection of constitutively active MEK5
The DNA expression-construct pCMV-HA-CA-MEK5 [
41] was used to derive MEK5 over-expressing variants of MCF-7 (N variant) cell line and selected with G418 (Gibco, Paisley, UK). A total of 5 × 10
6 MCF-7 cells were plated in 10 mm dishes in mammary epithelial growth medium and 10% Dulbecco's modified Eagle's medium, and incubated in 5% carbon dioxide at a temperature of 37°C. The following day, cells were co-transfected with pCMV-HA-CA-MEK5/pCMVdsRED-Neo at a 4:1 ratio using FuGENE (Roche, Nutley, NJ), in accordance with the manufacturer's instructions. Stable transfectants were selected by culturing cells in the presence of 200 μg/ml G418 (drug levels maintained during culture of cell stocks but removed for individual experiments). Medium was removed and replaced with fresh medium every 3 days until visible clones appeared. Individual clones were isolated using a sterile cloning ring coated with petroleum jelly and removed with phosphate-buffered saline (PBS) containing 1 mmol/l EDTA. These clones, called MCF-7-MEK5, were transferred to 24-well plates and allowed to grow as separate clones in mammary epithelial growth medium and 10% Dulbecco's modified Eagle's medium. MCF-7-VEC was established simultaneously as a stable, vector control cell clone. For RNA interference studies we used an Erk5 targeting pSUPER-EGFP PolymeraseIII driven RNA interference system (generously provided by Dr Frank E Jones, University of Colorado, Denver, CO, USA). Optimized hairpin RNA interference sequences for Erk5 were identified using OligoEngine software (Oligoengine, Seattle, WA, USA). The pSuper Erk5 EGFP clone was generated by ligating the annealed oligo with
SalI-
BglII digested pSuper-EGFP vector and confirmed by sequencing.
Western blotting assay for MEK5 and Erk5
MCF-7-MEK5 and MCF-7-VEC cells were characterized by immunoblotting with antibodies against MEK5 and Erk5. Western blots of crude whole cell extracts were performed using standard procedures. About 5 × 106 cells were harvested in sonicated buffer for 30 seconds and collected by centrifugation at 1,000 g for 20 minutes. Fifty micrograms of protein was resuspended in sample loading buffer, boiled for 5 minutes, and electrophoresed on polyacrylamide gel. The proteins were transferred electrophoretically to a nitrocellulose membrane, blocked with PBS-Tween (0.05%) in 5% low-fat dry milk solution at room temperature for 1 hour, incubated with specific antibodies (overnight at 4°C), and then washed in PBS-Tween solution three times. Then the immunoreactive proteins were visualized using an electrogenerated chemiluminescence system (Amersham, Arlington Heights, IL, USA) and recorded by fluorography on Hyperfilm (Amersham), in accordance with the manufacturer's instructions.
Colony survival assay for TNF-α in the established MCF-7-MEK5 cells
Clonogenicity assays were performed as we have described previously [
42,
43]. For TNF-α colony survival assay, MCF-7-MEK5 and MCF-7-VEC cells were plated and treated with different concentrations of TNF-α (0 to 100 ng/ml) for 18 hours. Cells were then cultured in fresh media without TNF-α and observed daily for 1 to 2 weeks. Colonies were fixed, stained with crystal violet and counted. Data are represented as percentage clonogenic survival from untreated control cells (100%) ± standard error of the mean (n = 3).
Immunofluorescence analysis of EMT markers and morphology
The expression levels of an epithelial cell marker (E-cadherin) and a mesenchymal cell marker (vimentin) were assessed by indirect immunofluorescence using specific antibodies (E-cadherin: CS-3195 [Cell Signaling Technology, Beverly, MA, USA]; vimentin: V6630 [Sigma, St. Louis, MO, USA]). The distribution of filamentous actin (F-actin) was visualized using Alexa 488 conjugated phalloidin (Invitrogen, Carlsbad, CA, USA) [
44]. Briefly, MCF-7, MCF-7-MEK5, and MDA-MB-231 cells were cultured in eight-well chamber slides for 48 hours. The cells were fixed in 4% paraformaldhyde/PBS for 10 minutes followed by incubation with the primary antibodies and phalloidin at the desired dilution (E-cadherin: 1:50 dilution; vimentin: 1:50 dilution; phalloidin: 1:100 dilution). Alexa 594 and 488 conjugated secondary antibodies (1:1,500 dilution) were used to detect E-cadherin and vimentin, respectively. The nucleus was stained using DAPI containing VectorShield mounting medium (Vector Laboratories, Burlingame, CA, USA). The digital images were captured using Nikon Eclipse 80
i along with the accompanying program IPLab, version 3.6.5 (Nikon Inc., Melville, NY, USA).
2-DE and image analysis
A total of 1 × 107 cells for each cell line (MCF-7-MEK5 and MCF-7-VEC) were collected and homogenized in lysis buffer (7 mol/l urea, 2 mol/l thiourea, 4% 3-[(3-Cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), and 50 mmol/l dithiothreitol (DTT)) containing protease inhibitor cocktail (Sigma), and sonicated three times for 7 seconds each. The method of Bradford (BioRad, Hercules, CA, USA) was used to detect the protein concentrations, and the samples were stored at -70°C for further analyses.
First dimensional electrophoresis was performed using a Protean iso-electronic focus cell unit (BioRad). Precast 11 cm immobilized pH gradient (IPG) strips with pH ranges 5 to 8 were obtained from BioRad. Lysates were thawed and mixed (3:1) with rehydration solution (7 mol/l urea, 2 mol/l thiourea, 4% CHAPS, 50 mmol/l DDT, 5% Triton ×100, 5% ampholytes, and 0.01% of bromophenol blue). Two hundred microliters (approximately 180 μg) of protein sample mixture was loaded to each of the IPG strips. The strips were rehydrated with the sample mixture and overlaid with a layer of mineral oil overnight (16 hours). Iso-electric focusing was carried out in three steps under gradient mode: 250 V for 20 minutes, linear ramp; 8,000 V for 2.5 hours, linear ramp; and 8,000 V, rapid ramp to reach 20,000 V-hours.
The second dimensional electrophoresis was carried out in a BioRad Criterion electrophoresis cell system. IPG strips were equilibrated in two steps: 20 minutes in 60 mmol/l DTT and 20 minutes in 200 mmol/l of iodoacetamide; both were dissolved in equilibration buffer (6 mol/l urea, 1% SDS, 30% glycerol, and 40 mmol/l Tris-base). IPG strips were sealed on the top of SDS gels (Criterion 8% to 16% Tris-HCl Gel obtained from BioRad) with 0.5% agarose. SDS-PAGE was performed at a constant voltage of 200 V for 55 minutes. Gels were washed with 200 ml de-ionized water three times for 5 minutes each. Bio-safe staining solution (50 ml; BioRad) was added to each gel, and the gels were then placed on a shaker for 2 hours. The staining solution was discarded and the stained gels were rinsed with de-ionized water (3×, 30 minutes per wash). Stained gels were scanned with a Gel Doc-XR image system (BioRad) and analyzed with the PDQuest software (version 8.01) (Bio-Rad, Hercules, CA, USA). The proteins of interest were marked for excision. The spots were excised from the gels by a Quest Spot cutter (BioRad), and digested with trypsin (Promega, Madison, WI, USA) in 25 mmol/l ammonium bicarbonate overnight at 37°C to release tryptic peptides using a Progest digestion system (Genomic Solutions, Ann Arbor, MI, USA). Recovered peptides were dried in a Labconco NewCentriVap concentrator (Labconco, Kansas City, MO, USA).
Liquid chromatography tandem mass spectrometry analyses and data processing
The dried protein digests from the 96-well plates were analyzed using an LC-Nanospray-MSn system (Thermo Scientific, Waltham, MA, USA), equipped with a Finnigan Micromass AS automatic sampling system. A BioBasic C18 PicoFrit Column (75 μm × 10.2 cm, tip 15 μm; New Objective, Woburn, MA, USA) was used for separation. The mobile phases consisted of (A) 0.1% formic acid in water and (B) 0.1% formic acid in acetonitrile, at a linear gradient from 0% B to 35% B within 35 minutes, along with a gradient from 35% B to 98% B within 8 minutes, kept at 98% B for 2 minutes and then back to 0% B within 2 minutes. The total acquisition time was 60 minutes for each run. The tandem mass spectrometry spectra were analyzed against the ipi.human.v3.27 database using SEQUEST software (Thermo Scientific, Waltham, MA, USA), and the results were tabulated for each identified protein.
Reverse transcription polymerase chain reaction
Total RNAs from MCF-7-MEK5 and MCF-7-VEC cells were extracted using a PureLink total RNA purification system (Invitrogen) or RNeasy
® mini kit (Qiagen Sciences, Germantown, MD, USA). The reverse transcription was carried out with a SuperScript first-strand synthesis system (Invitrogen) using Oligo(dT)
12–18 primers or with iscript™ cDNA synthesis kit (Biorad Laboratories, Hercules, CA USA). The primer pairs used to amplify the genes
vim (vimentin),
krt8 (keratin 8),
krt19,
hspa4 (HSPA4),
gstp1 (glutathione-S-transferase P),
gstm3 (glutathione-S-transferase Mu 3) and
ckb (creatine kinase B-type), and the genes involved in EMT including
cdh1 (E-cadherin),
ctnnb1 (β-catenin),
snail,
slug and
δ-ef1 (δ-EF1) were designed using the online tool of Oligo Perfect Designer (Invitrogen), and the endogenous
actb (β-actin) was employed as an internal standard. The primer pairs are provided in the additional materials [see Additional data file
1]. Primer specificity was confirmed by BLAST (Basic Local Alignment Search Tool) analysis. Standard PCR was performed using a Platinum blue PCR supermix kit (Invitrogen). Briefly, denaturation was carried out at 94°C for 2 minutes, and 30 cycles (94°C for 30 seconds, 58°C for 30 seconds, and 72°C for 30 seconds), followed by a 5 minute period for elongation at 72°C. The PCR products were isolated by 1.5% agarose gel electrophoresis, and the bands were visualized by SYBR green staining. For real-time PCR analyses, an iCycler iQ5 real time PCR detection system (BioRad), and a SYBR GreenER qPCR supermix kit (Invitrogen) were used as follows: 50°C for 2 minutes, 95°C for 8 minutes and 30 seconds, and 50 cycles (15 seconds at 95°C, 1 minute at 60°C). The data were analyzed with a normalized gene expression method (ΔΔ Ct) [
45] using the iQ5 Optical System Software (BioRad), and the gene
actb was used as a reference for normalization. All experiments were repeated three times independently.
Discussion
Our proteomic analysis of two MCF-7 human breast cancer cell lines revealed at least seven differentially expressed protein spots in two-dimensional gels, which may contribute to the mechanisms of MEK5/Erk5-mediated TNF-α resistance. These proteins are known to play various roles in cellular processes, including detoxification, proliferation, metabolism, and cytoskeletal organization. It is well known that human cytosolic glutathione-S-transferases, which play important roles in cellular detoxification pathways, are of direct relevance to many drug-resistance tumors [
46‐
53]. At least one of these proteins is also implicated in EMT, suggesting a possible connection between TNF-α resistance and EMT. Both VIM and KRT belong to intermediate filaments, which are considered to be the principle cytoskeletal proteins in mammalian cells [
54]. It is believed that the over-expression of VIM, which is related to poor prognosis in breast cancer patients presenting with metastasis potential [
55,
56], results in a more invasive capacity of breast cancer cells
in vitro and
in vivo [
21,
57,
58]. Kokkinos and colleagues [
21] proposed that the intermediate filament transformed from a KRT-rich to a VIM-rich network in the process of EMT in cancers [
21]. Previous studies have shown that MCF-7 breast cancer cells do not typically express VIM but exhibit strong expression of KRT; the acquisition of VIM expression and the loss of KRT19 expression were associated with adriamycin-resistant MCF-7 cells compared with their parental cells [
11]. Our results show increased VIM expression and decreased KRT19 and KRT8 expression in TNF-α resistant MCF-7 cells.
These findings, consistent with the cytoskeletal reorganization seen in both EMT and drug resistance, provoked further evaluation of the EMT markers snail, slug, δ-ef1, E-cadherin, and N-cadherin in the TNF-α resistant MCF-7 cells. The profound morphologic changes and enhanced invasive capabilities of EMT in various cancers are thought to be regulated by several transcription factors, including SNAI1 (snail), SNAI2 (slug), and ZEB1 (δ-ef1). Increased expression of snail and slug has been reported in invasive compared with noninvasive breast tumors and associated with lymph node metastases [
59]. Increased levels of δ-ef1, also seen in invasive breast tumors, have been correlated with de-differentiation [
60]. These transcription factors, known to regulate EMT in development, have all been shown to repress E-cadherin, which is the primary cell adhesion molecule in epithelial tissue [
61‐
65]. The loss of epithelial cell-to-cell adhesion through decreased expression of E-cadherin is the hallmark of EMT that permits acquisition of a motile phenotype [
65]. Interestingly, N-cadherin has been demonstrated to correlate with increased invasion and migration in breast carcinomas
in vitro, regardless of E-cadherin status [
66,
67]. Further
in vivo studies have shown that N-cadherin enhances metastasis of breast tumors via Erk signaling [
68]. In this study we observed a decrease in E-cadherin expression and an increase in N-cadherin expression in the MCF-7-MEK5 cells. Additionally, slug and δ-ef1, but not snail, were significantly increased in the MCF-7-MEK5 cell line.
Acknowledgements
The study was supported by a Department of Defense grant W81XWH-04-1-0557 (MEB and GW), Office of Naval Research Grant N00014-99-1-0763 (MEB, BSB, JAM, and GW), National Institutes of Health DK059389 (MEB), the Louisiana Cancer Research Consortium (BS, GW, and MEB), National Institutes of Health COBRE grant P20 RR020152 (AMN), and Breast Cancer Specialized Program of Research Excellence Grant (P50 CA58183) from the National Cancer Institute (MEB and RS).
We should like to thank Dr JD Lee for providing the CA-MEK5 plasmid and Drs Frank E Jones and Erik K Flemington for providing reagents and advice on generating shRNA-Erk5 plasmids.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
CZ performed 2-DE and RT-PCR, interpreted data, and drafted the manuscript. AMN performed quantitative PCR for EMT gene expression and drafted the manuscript. WX performed RT-PCR. QZ carried out high performance liquid chromatography-tandem mass spectrometry based protein identification and database search. YT performed Western blot analysis. MB conducted RT-PCR analysis. SE generated the MEK5 stable cells and performed colony assays. YZ designed, generated, and validated pSUPER-shRNA-Erk5. LB conducted quantitative PCR analysis of gene expression. DY and CBW conducted RT-PCR and colony assays. RS, JAM, BSB, TEW, and KPN participated in experimental design and interpretation. BS performed immunofluorescence. MEB and GW co-designed the study, directed research efforts, and critically revised the manuscript.