Background
During cancer progression, cancer cells first proliferate in the primary cancer site before the acquisition of the migratory behavior that leads to their spread in the body and ultimately to the development of metastasis. Estradiol (E2) and estrogen receptor alpha (ERα) play pivotal roles during ERα-positive breast cancer progression: E2-ERα signaling contributes to cell growth but prevents metastatic potential by preserving the differentiated status of the cells [
1‐
4]. Although the loss of estrogenic signaling is generally associated with disease aggravation, the process remains poorly understood [
4‐
7]. Indeed, many mechanisms may be involved because growth factors assume the control of cell growth and migratory capacities [
1,
4]. We have previously identified COUP-TFI (chicken ovalbumin upstream promoter transcription factor I) as a promoter of estrogen-independent ERα transcriptional activity in breast cancer cell lines [
8,
9]. Moreover, COUP-TFI was found to be overexpressed in breast tumors and to enhance the proliferation of ER-positive breast cancer cells [
9]. COUP-TFI and COUP-TFII are orphan nuclear receptors that can also act by modulating other nuclear receptors, including ERα, functioning selectively as a co-activator or a co-repressor [
10] to control biological processes linked to cellular growth, migration, or angiogenesis and potentially contributing to cancer progression [
10‐
12]. Particularly, COUP-TFI expression is associated with the migration behavior of various cells during embryonic development. Accordingly, evidence from several studies supports that COUP-TFI and COUP-TFII expression in cancer cells may be associated with a dedifferentiation phenotype, the reactivation of embryonic pathways, and migration behavior, supporting the induction of aggressive characteristics in cancers [
11‐
14]. Although COUP-TFI is suggested to be a potent mediator of cancer progression, little is known about the endogenous targets of this orphan nuclear receptor in breast cancer cells.
The chemokine CXCL12 signaling axis may represent one such axis. This signaling pathway, which is composed of the chemokine CXCL12 (also called SDF-1 for stromal cell-derived factor 1) and its receptors CXCR4 and CXCR7, play pivotal roles in the cell migration, angiogenesis, proliferation, and survival of many cancer cells, including breast cancer [
15,
16]. CXCR4 is typically highly expressed in metastatic cells and supports the privileged homing of these metastatic cells to specific sites where the local secretion of CXCL12 is important, namely the bone, liver, brain, and lung [
17‐
19]. Indeed, reduction or the loss of the local secretion of CXCL12 at the tumor site can induce the emergence of metastatic cells that may spread in the organism toward endocrine sources of CXCL12 [
20‐
22]. Although the pivotal role of the CXCL12/CXCR4 axis in cell motility and consequently in cancer metastasis in several tissues is well established, the contribution of CXCL12
via its receptor CXCR7 is less understood.
CXCL12 signaling may be connected to the phenotypic characteristics modified by COUP-TFI; thus, we hypothesized that COUP-TFI could target this signaling pathway in breast cancer cells. Furthermore, as the entire CXCL12/CXCR4 signaling axis is an endogenous target of E2 and is pivotal to hormonal-induced MCF-7 cell growth [
23], COUP-TFI could achieve the loss of its estrogenic regulation. In the present study, we developed MCF-7 breast cancer cells overexpressing COUP-TFI protein and examined the regulation of CXCL12 signaling axis. We provide evidence that COUP-TFI increases the motility of MCF-7 ERα-positive breast cancer cells by acting on CXCL12/CXCR4 signaling as an endogenous target. The modification of CXCL12/CXCR4 expression by COUP-TFI is mediated by the activation of epithelial growth factor (EGF) and its receptor (EGFR) in MCF-7 cells. These results correlate with the expression profiles of COUP-TFI, CXCL12, and CXCR4 in breast tumors compared to healthy samples.
Methods
Antibodies and reagents
A goat polyclonal antibody against human CXCL12 (R&D Systems AF-310-NA), rabbit polyclonal antibody against CXCR4 (Abcam Inc. ab2074), mouse monoclonal antibody against human CXCR7/RDC1 (R&D Systems clone 11G8; MAB42273), a rabbit polyclonal antibody against COUP-TFI (Abcam Inc. ab11954) and a rabbit polyclonal antibody against HA epitope (Santa Cruz sc-805) were used for the immunofluorescence and western blot assays. A mouse polyclonal antibody against phosphorylated ERK (Santa Cruz sc-7983) and rabbit polyclonal antibody against total ERK (Santa Cruz sc-94) were used for the western blot assays.
The reagents used for treatments (17-β-estradiol (E2), ICI182,780 (ICI), and AMD3100) were purchased from Sigma-Aldrich Co. The recombinant CXCL12 used for the proliferation and migration assays was purchased from R&D Systems (350-NS-050).
Cell culture and treatments
MCF-7 cells were routinely maintained in DMEM (Invitrogen) supplemented with 10% fetal bovine serum (FBS; Biowest) and antibiotics (Invitrogen) at 37°C in 5% CO
2. Stably transfected MCF-7 clones were obtained as previously described [
9]. A pool of two independent control clones and two independent COUP-TFI-overexpressing (COUP) clones were used for this study.
When treatments with steroids were required, the cells were maintained for 24 h in DMEM without phenol red (Invitrogen) supplemented with 2.5% dextran-treated charcoal-stripped FBS (dsFBS) prior to the experiments. The treatments were then performed in phenol red-free DMEM with 2.5% dsFBS and E2 (10−8 M), ICI (10−6 M), or both together for 48 hours; 0.1% ethanol was used as a control (EtOH).
Immunofluorescence
Cells were plated on 10 mm‒diameter cover slides in 24‒well plates (5 × 104 cells per well). After 48 h, the cells were fixed for 10 min in phosphate‒buffered saline (PBS) containing 4% paraformaldehyde. The cells were then permeabilized in PBS containing 0.3% Triton X‒100 for 10 min. The primary antibodies were diluted (1:100) in PBS containing 3% FCS and added to the permeabilized cells, which were incubated over night at 4°C. Dye-conjugated secondary antibodies (1:1000, Alexa Fluor, Invitrogen) were incubated 1 h at room temperature. After mounting in Vectashield® mounting medium with DAPI (Vector), images were obtained using an Imager.Z1 ApoTome AxioCam (Zeiss) epifluorescence microscope and processed with Axio Vision Software.
RT-PCR assays
2.5 × 10
5 cells were cultured in 6-well plates and treated as specified. Total RNA was extracted, at least in triplicate, using the Trizol™ reagent (Invitrogen) according to the manufacturer’s instructions. cDNA was generated by MMLV Reverse transcriptase (Invitrogen) using random hexamers (Promega). Quantitative real-time RT-PCR was performed using the iQ SybrGreen supermix (Bio-Rad, Hercules) and a Bio-Rad MyiQ apparatus. The primers (Proligo Primers and Probes, Boulder, CO, USA) used for the cDNA amplifications in the quantitative RT-PCR experiments are described in Table
1. GAPDH and RNA 18S were used as housekeeping genes to normalize the expression levels of the genes of interest. GAPDH was found to be appropriate for normalisation in cell lines because its expression was not affected by treatments and remained stable in control and COUP clones. For tissues, we first verified the choice of the reference gene as an internal control and its suitability in our study. Four housekeeping genes were tested (GAPDH, HPRT1, TBP and 18S RNA). The stability of these genes across different tissues and tumor grades was assessed using geNorm algorithm [
24]. This software has listed HPRT1 as the best gene but HPRT1 is expressed at very low level in normal tissues and tumors, making it quite difficult to accurately quantify and not enough useful as an internal reference in our study. The second best gene, established by the software in the list, was the 18S RNA. This RNA is expressed similarly at relatively high levels in all tumors and made ideal positive control for our study. Thus, we have chosen 18S for normalization.
Table 1
Sequences of primers used in this study
Chemokine (C-X-C motif) receptor 4 (CXCR4) | GCCTTATCCTGCCTGGTATTGTC | GCGAAGAAAGCCAGGATGAGGAT |
Chemokine (C-X-C motif) receptor 7 (CXCR7) | ACAGGCTATGACACGCACTG | ACGAGACTGACCACCCAGAC |
Chemokine (C-X-C motif) ligand 12 (CXCL12) | CACCATTGAGAGGTCGGAAG | AATGAGACCCGTCTTTGCAG |
Nuclear receptor subfamily 2, group F, member 1 (NR2F1) (COUP-TFI) | TACGTGAGGAGCCAGTACCC | CGATGGGGGTTTTACCTACC |
Epidermal Growth Factor (EGF) | CAGGTAATGGAGCGAAGCTTTCA | GTGCATCGACATAGTTCATTCTTCTTG |
Epidermal Growth Factor receptor (EGFR) | GGAGAACTGCCAGAAACTGACC | GCCTGCAGCACACTGGTTG |
GlycerAldehyde-3-Phosphate DesHydrogenase (GAPDH) | GGGCATCCTGGGCTACACTG | GAGGTCCACCACCCTGTTGC |
18S RNA | GCAATTATTCCCCATGAACG | AGGGCCTCACTAAACCATCC |
Melting curves and PCR efficiency analyses were performed to confirm correct amplification. Each experiment was performed at least three times. Results were expressed according to the comparative Ct method (ddCt) for relative quantification of gene expression. For each sample, the difference (dCt) was calculated between Ct values obtained for target and reference amplicons. Comparative ddCt was then determined using as a reference the dCT calculated for the vehicle control sample (ethanol), and absolute values for comparative expression level were determined as equal to 2-ddCt.
Total proteins were extracted in RIPA buffer (1% NP40, 0.5% Na-deoxycholate, and 1% SDS in PBS) with an anti-protease mixture (Complete EDTA free Antiproteases, Roche) and quantified using the Bio Rad DC protein assay kit. The proteins were diluted in Laemmli buffer and denatured at 95°C; 30 μg of denatured proteins were separated on SDS polyacrylamide gels (10 and 15%), transferred to polyvinylidene difluoride membranes (Millipore), and probed with specific antibodies. The antibodies used for the Western blot assays were diluted 1:2000 (for the detection of COUP-TFI, HA, CXCL12, CXCR4 and CXCR7) or 1:5000 (for the detection of ERK or P-ERK). The detection of the immunocomplexes was performed using an enhanced chemiluminescence system (Immune Star, Bio-Rad Laboratories). For the detection of ERK activity, control and COUP cells were cultured for 48 h in phenol red-free DMEM with 0.5% dsFBS. After EGF (10−9 M for 5 or 10 min) or CXCL12 (200 ng/mL for 5 or 10 min) stimulation, whole-cell extracts were directly prepared in 3× Laemmli buffer. Following sonication, the protein extracts were denatured for 5 min at 95°C and analyzed as detailed above.
FAIRE was performed as described by Eeckhoute
et al. [
25]. Briefly, asynchronously growing MCF-7 cells (60-70% confluence) treated or not for 48 h with 10
−8 M E2 were cross-linked with 1% formaldehyde for 10 min at room temperature. Glycine was added to a final concentration of 125 mM, and the cells were rinsed with cold PBS and harvested. The cells were lysed with a solution of 1% SDS, 10 mM EDTA, and 50 mM Tris-HCl (pH 8.1) containing a protease inhibitor cocktail (Roche) and then sonicated for 14 min (30-sec on/off cycles) using a Bioruptor (Diagenode) set at the highest intensity. The soluble chromatin was subjected to three consecutive phenol-chloroform extractions (Sigma, P3803) and incubated overnight at 65°C to reverse the cross-linking. The DNA was then purified using the MinElute PCR purification kit (Qiagen). The relative enrichment of open chromatin for the
CXCL12, CXCR4 and
CXCR7 genes was quantified by real-time PCR performed using the iQ SybrGreen supermix and a Bio-Rad MyiQ apparatus. The primers used for the quantitative PCR experiments were described previously [
23].
Proliferation assay
A total of 2500 MCF-7 cells clones per well were seeded in 96-well plates and cultured in 100 μL of phenol red-free DMEM/2.5% dsFBS and EtOH or CXCL12 (200 ng/mL) for 7 days. Every 2 days, the medium was removed, and fresh treatments were performed. Proliferation was evaluated using the 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT; Sigma) assay. 10-μL of 5 mg/mL MTT solution was added to 100 μL of culture medium in each well and incubated for 2 h at 37°C. The supernatant was removed, and the formazan formed was dissolved in 100 μL DMSO. The absorbance of each well at 570 nm was measured using a microplate reader (Bio-Rad).
Migration assay
The cells were cultured for 48 h in phenol red-free DMEM with 5% dsFBS prior to the experiments. A total of 50,000 cells were plated in the upper chamber of a BDBiocoat control insert (BD Biosciences) in phenol red-free DMEM/0.5% dsFBS with or without AMD3100 (50 μM) or CXCL12 (200 ng/mL); phenol red-free DMEM/2.5% dsFBS with or without AMD3100 (50 μM) or CXCL12 (200 ng/mL) was added to the lower chamber. The cells were allowed to migrate for 24 h at 37°C, and the non-migrant cells were wiped off the upper chamber with a cotton swab. The insert was then placed in phenol red-free DMEM/2.5% dsFBS with calcein-AM (Invitrogen) for 1 h to stain the cells that reached the lower side of the filter. The migrant cells were then counted in 3 fields from at least 3 inserts per experimental condition.
Ethics statement
Human samples were obtained from the processing of biological samples through the Centre de Ressources Biologiques (CRB)-Santé of Rennes (
http://www.crbsante-rennes.com). We have received written informed consent from all patients for the use of their samples analyzed in this study. The research protocol was conducted under French legal guidelines and approved by the local institutional ethics committee (CPP, Comité de Protection des Personnes Ouest V de Rennes) in accordance with Helsinki Declaration. The collection of samples is reported to the Ministry of Education and Research No. DC-2008-338 which is consistent with the current ethics legislation.
Gene expression in breast tumors
The breast tumor samples used were invasive ductal carcinoma and mostly (> 90%) ER-positive. They were divided into 20 SBR (Scarff-Bloom-Richardson grading system) Grade 1, 20 SBR Grade 2, 19 SBR Grade 3, and 23 non-tumor tissues. All samples used in this study were from fresh frozen tissues. The normal breast tissues were adjacent to the tumors but they are majority unmatched to the tumors. Total RNA was extracted using the RNeasy Mini kit (Qiagen) according to the manufacturer’s instructions, and 1 μg of total RNA was reverse transcribed with M-MLV RT (Invitrogen). Gene expression was assessed by real-time PCR (MyiQ5–Bio-Rad) with 4 ng of cDNA, 150 mM of primers (shown in Table
1), and 1× of iQ™ SYBR® Green supermix from Bio-Rad (Bio-Rad, Hercules, CA, USA). Gene expression was also measured in MCF-7 cells and served to adjust the data from different plates. The data were normalized to the expression of 18S RNA and were analyzed using IQ5 software (Bio-Rad).
Statistical analysis
A statistical analysis was performed using Student’s t-test for most of the presented results. The values are provided as the mean ± standard error of the mean (SEM) and were considered statistically significant at p < 0.05. The statistical analysis for the tumor samples was performed using Minitab 16 software. The data are represented by box plots. The absence of a normal distribution of each gene for each category was verified by the Anderson-Darling normality test, and the non-parametric Mann-Whitney test was chosen to analyze our samples.
Discussion
The contribution of COUP-TFI to cancer progression is poorly understood. Nevertheless, this orphan nuclear receptor is known to participate in many biological processes connected to normal or pathological cell proliferation, survival, or migration (for a review, see [
11]). Previous studies have established that COUP-TFs can modulate estrogen signaling, contributing to phenotypical changes in breast cancer cells [
9,
31,
32]. Moreover, our previous study suggested that the overexpression of COUP-TFI in breast tumor cells may contribute to the loss of the epithelial phenotype and acquisition of mesenchymal characteristics [
9]. In the present study, we identified CXCL12/CXCR4 signaling as an endogenous target of COUP-TFI, which could explain some of these phenotypical deviations. We demonstrated that COUP-TFI overexpression selectively and differentially alters the expression of the CXCL12 and CXCR4 genes: the basal level of CXCL12 was reduced, whereas CXCR4 basal expression was up-regulated. It was also noted that COUP-TFI disturbs the estrogenic regulation of CXCL12 and CXCR4 in MCF-7 cells, supporting the idea that COUP-TFI leads to a loss of E2 dependency in breast cancer cells. Interestingly, our data show that COUP-TFI impacts the chromatin condensation state of the proximal promoters of the
CXCL12 and
CXCR4 genes. These modifications of the chromatin structure are known to correlate with the transcriptional potential of regulatory elements and could also suggest epigenetic modifications induced by COUP-TFI. However, the precise molecular mechanisms of COUP-TFI action on the basal and E2-dependent regulation of CXCL12/CXCR4 remain to be determined.
Cancer progression is frequently associated with growth factor-induced control of cell growth and migration [
33]. Notably, cross-talk between the EGFR family and E2 signaling is often associated with the loss of hormonal control of cancer cell growth and the acquisition of metastatic potential [
34]. CXCR4 induction is one of the identified mechanisms for the growth factor control in cancer cells, which supports the migration of cancer cells [
30]. COUP-TFI has been shown to interact with the MAPK pathway, leading to the activation of ERK activity ([
8] and herein). Here, we established that COUP-TFI overexpression leads to an increase in EGF and EGFR relative expression that could, in part, explain the effect of COUP-TFI in the activation of ERK signaling activity. Our results support that the induction of MAPK activity, presumably
via EGF signaling, is responsible for the constitutive induction effected by COUP-TFI on CXCR4 expression. Moreover, our results show for the first time that CXCL12 expression is negatively regulated by EGF signaling in breast cancer cells. We also observed similar results when the cells were treated with different serum concentrations (data not shown). Interestingly, these data are in good agreement with several studies related to stem cell homing/mobilization in bone marrow that have reported that many growth factors can down-regulate the local secretion of CXCL12, thereby promoting stem cell mobilization toward the peripheral blood [
35‐
37]. Taken together, our results support the idea that COUP-TFI can differentially impact CXCL12 and CXCR4 basal expression by activating the ERK pathway. The constitutive activation of a transcription factor—proposed not to be ERα, given our observation that ICI treatment did not impact CXCR4 expression in COUP clones—by the MAPK pathway could explain the induction of CXCR4 expression. It was previously observed that hypoxia-inducible factor 1 alpha (HIF1-α) is induced by EGFR constitutive signaling, leading to CXCR4 up-regulation [
30].
The chemokine network and CXCL12/CXCR4 signaling in particular, as well as EGFR signaling are involved in many aspects of cancer biology, including growth and metastasis [
7,
34,
38]. Indeed, there are many evidences of the essential role of CXCR4 in the enhanced invasion of several types of cancer [
39‐
41]. Furthermore, the down-regulation of CXCL12 expression by promoter hypermethylation has been associated with increased metastatic potential in mammary carcinoma cells [
22] by the loss of autocrine and paracrine CXCL12 retention at the primary tumor site. Our results demonstrate a higher proliferative response to CXCL12 treatment by COUP clones compared to control clones, which could be due to the higher activation of ERK signaling observed after CXCL12 treatment. We also found that the COUP clones exhibited a better migration behavior than the control clones when migrating toward a serum-complemented medium or in response to a CXCL12 chemotactic gradient. Our results suggest that this higher migration behavior is due to the enhanced CXCR4 expression. In addition, we observed that ectopic CXCL12 added to the upper chamber prior to the migration test hampered the migration of both the control and COUP clones. This finding is in good agreement with a study from Zabel
et al., who suggested that the reduction in CXCL12-triggered migration by the additional CXCL12 within the cells could possibly be explained by the desensitization of CXCR4 or disruption of the chemokine gradient [
42]. Moreover, the down-regulation of CXCL12 was previously reported to be necessary to allow the emergence of metastatic cells
in vivo[
20,
21]. Taken together, our results suggest that the down-regulation of CXCL12 induced by COUP-TFI overexpression could be associated, together with the elevation in CXCR4 expression, with increased migration behavior. In other words, we propose that the opposite action of COUP-TFI on CXCL12 and CXCR4 expression enhances the migration capacity of cancer cells through an increase in sensitivity to exogenous CXCL12 and by limiting the autocrine retention effect of CXCL12. Moreover, enhanced EGFR signaling activity was reported to contribute to cancer progression from various origins through the elevation of cancer cell survival, proliferation, and migration [
38]. Our results support that, by repressing CXCL12 expression and inducing CXCR4 expression, the growth factor regulation of CXCL12 signaling could trigger these effects, as was observed during stem cell mobilization from the bone marrow to peripheral blood [
43].
Our previous immunohistochemistry data indicated that COUP-TFI is overexpressed in cancer compared to normal breast tissues [
9]. We also showed that COUP-TFI expression increased in dedifferentitiated ER-negative breast cancer cell lines compared to differentiated ER-positive cell lines. This was correlated to protein markers of dedifferentiated phenotype, for instance E-cadherin silencing and vimentin expression [
9]. A limitation of our study is that it was only performed in MCF-7 cell line. However, it is of interest to note that COUP-TFI represses in vitro the expression of type VII collagen in different human cell lines [
44]. Moreover, cell contact stability was reported to be affected by COUP-TFI overexpression in fibroblast cells, most likely because of alteration of cell attachment proteins expression [
45]. COUP-TFII has also been reported to be overexpressed in breast cancer epithelia [
12]. COUP-TFII over expression was furthermore associated to poor clinical outcome and to invasive behavior of metastatic cells in lymph nodes [
12]. However, in this study, our quantitative RT-PCRs revealed a significant augmentation of COUP-TFI mRNA expression only in grade 1 tumors, whereas grade 2 and 3 tumors exhibited expression of COUP-TFI mRNA that was similar to that observed in the normal tissues. Although, further investigation, particularly by immunohistochemistry, is necessary to reveal COUP-TFI staining in low and high grade tumor biopsies, the discrepancy between transcript and protein levels, may argue for the consequence of additional control mechanisms besides transcription. This may be attributed to differences in the mRNA and protein turn over or could originate from different translational mechanisms that may selectively stabilize COUP-TFI protein. Indeed, the expression levels of a protein depend not only on transcription rates of the gene, but also on additional control mechanisms, such as nuclear export, mRNA localization and stability, translational regulation and protein degradation [
46]. Deregulation of certain of these mechanisms in cancer cells may explain this discrepancy; however, more investigations will be needed to establish that. Interestingly, our in vitro results showed that COUP-TFI overexpression does abolish E2 control of CXCR4 expression and partially reduces CXCL12 regulation. The expression profiles of CXCL12, CXCR4 and CXCR7 in breast cancer biopsies are almost identical to that obtained when we overexpressed COUP-TFI in MCF-7 cancer cells, suggesting that our in-vitro results might have a clinical relevance. It should be investigated whether increasing the expression of COUP-TFI protein during cancer progression could in fact participate in the development of hormone resistance and favor the growth and migration capacity of tumor cells. Recent studies have reported that the COUP-TFII expression level is increased in several different cancer cells, such as breast, prostate, and ovary cancers [
12‐
14]. These studies have also shown that the overexpression of COUP-TFII is associated with a significantly shorter disease-free survival. Indeed, the overexpression of COUP-TFII in prostate cells promotes tumorigenesis and induces an aggressive metastasis characteristic in tumors by inhibiting the TGF-β-induced growth barrier [
13].
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
AB, GK, GF, FP performed the establishment and characterization of stable cellular clones, immunofluorescence, gene expression analysis and cell growth and migration assays. MD, performed RNA extractions from normal and cancerous breast tissues for quantitative RT-PCR analysis. SL and YLD, designed and carried out the RT-PCR amplification of CXCR4, CXCR7, CXCL12 and COUP-TFI in breast tissue samples and analyzed their expression levels. FG, JL and PT performed the characterization tumoral and non-tumoral mammary gland and defined the SBR Grade of ductal carcinoma in situ. FP designed, supervised and coordinated the study, participated in the design of all the experiments. AB and FP drafted the manuscript. All authors read and approved the final manuscript.