Background
Approximately 70–80% of breast cancers express estrogen receptor alpha (hereafter referred to as ER) and they depend on estrogen signals for continued growth [
1]. Hence, endocrine therapies targeting ER using tamoxifen or aromatase inhibitors (AIs) are the first-line adjuvant therapies offered to patients with ER-positive (+) breast cancer [
2]. However, whatever the hormonal strategies are used, intrinsic or acquired resistance may occur [
3,
4]. Increasing evidence indicates that estrogen-independent growth develops in a significant proportion of patients, leading to hormonal therapy resistance. Several molecular mediators of estrogen-independent growth have been described, such as regulation of microRNAs [
5], growth factor signaling crosstalk [
6,
7], and mutation of ER [
8].
Long non-coding RNAs (lncRNAs) are currently defined group of transcripts that are more than 200 nucleotides in length. Despite lacking of the protein-coding potential, emerging evidence indicates that lncRNAs may serve as master gene regulators through various mechanisms, and thus, the dysregulation of lncRNAs expression is often linked to diverse human diseases including cancer [
9,
10]. Although we are only beginning to understand the function of lncRNAs, increasingly evidence have led the belief that lncRNAs may regulate gene expression through interacting with DNA, RNA, or protein. Recent studies have implicated some lncRNAs in endocrine resistance in breast cancer [
11]. For instance, a recent report showed that the lncRNA HOTAIR is upregulated in tamoxifen-resistant ER+ breast cancers and thus contributed to tamoxifen resistance [
12]. More recently, studies from other groups suggest that lncRNA UCA1 confers tamoxifen resistance through different signaling pathways [
13,
14]. However, the underlying mechanism of how lncRNAs regulate intrinsic hormonal resistance and estrogen-independent growth is still far from being understood.
LincRNA regulator of reprogramming (linc-RoR) was originally identified to be able to regulate reprogramming of iPSCs [
15]. A subsequent report demonstrates that linc-RoR may serve as microRNA sponge to regulate the expression of the core transcription factors Oct4, Sox2, and Nanog in human embryonic stem cell self-renewal [
16]. Our previous study reports that linc-RoR inhibits p53 translation by interacting with heterogeneous nuclear ribonucleoprotein I (hnRNP I) in response to DNA damage [
17], and further studies suggest a link between linc-RoR and oncogene c-Myc in cancer progression [
18]. Of interest, it is reported that linc-RoR is dramatically overexpressed in triple negative breast cancer (TNBC), and the expression of linc-RoR promotes cell invasion via miR145/ARF6 pathway [
19]. However, little is known as to whether linc-RoR have any role in estrogen-independent growth of ER+ tumors.
The present study demonstrates that linc-RoR functions as a modulator to promote estrogen-independent growth and tamoxifen resistance of breast cancer cells. Linc-RoR expression is upregulated in response to estrogen deprivation. While knockout of linc-RoR abolishes estrogen-independent growth of MCF-7 cells, rescue assays reverse the phenotype. Importantly, linc-RoR confers MAPK/ERK pathway activation by decreasing the stability of phosphatase DUSP7 which is the negative regulator of ERK, suggesting a novel role for linc-RoR in regulation of MAPK/ERK signaling in response to estrogen deprivation.
Methods
Reagents
Sources of primary antibodies were: pERα (16 J4), ERα (D6R2W), pERK1/2 (D13.14.4E), ERK1/2(137F5), pMEK1/2(41G9), MEK1/2(D1A5) and MKP-2 (D9A5) from Cell Signaling (Danvers, MA, USA); MKP-1(D-3) and MKP-3 (F-12) from Santa Cruz (Dallas, TX, USA); DUSP7/MKP-X(26910–1-AP) and GAPDH (60004–1-Ig) from Proteintech (Rosemont, IL, USA). Secondary antibodies conjugated with IRDye 800CW or IRDye 680 were purchased from LI-COR Biosciences (Lincoln, NE, USA). PCR primers were purchased from IDT (Coralville, IA, USA). U0126 and NSC95397 were purchased from Cell Signaling and Santa Cruz (Dallas, TX, USA), respectively. TAM was obtained from Sigma-Aldrich (St. Louis, MO, USA).
Cell culture
MCF-7 parental, gRNA control and RoR KO cells were maintained in RPMI 1640 medium supplemented with 10% FCS. For estrogen deprivation treatment, these cells were cultured in phenol red-free RPMI medium containing 5% charcoal stripped FCS (E2-free medium). All the medium contain 2 mM L-glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin.
LncRNA profiling
For lncRNA profiling, we used Human Disease-Related LncRNA Profiler, as described previously [
20]. Total RNA was extracted using Direct-zol™ RNA MiniPrep (Zymo Research) and cDNA synthesis was carried out using RevertAid™ Reverse Transcriptase (Thermo Fisher) with random primers. Delta-delta Ct values (regular medium versus E2-free medium) were used to determine their relative expression as fold changes, as previously described [
21].
Western blot
Cells were harvested, and proteins were extracted and quantified as previously described [
5]. Protein samples were separated in a polyacrylamide SDS gel before transferring to PVDF membrane. After probing with a primary antibody, the membrane was incubated with a secondary antibody labeled with either IRDye 800CW or IRDye 680. Finally, signal intensity was detected by the Odyssey Infrared Imaging System (LI-COR Biosciences, Lincoln, NE, USA). The intensities of band were quantified by imageJ software (version 1.6.0, Windows, NIH).
Quantitative RT-PCR (qRT-PCR)
PCR was performed using a standard SYBR Green method (Bio-Rad, Hercules, CA, USA) as previously described [
22]. We used primers linc-RoR-RT-5.1A and linc-RoR-RT-3.1A; pS2-RT-5.1 and pS2-RT-3.1; CXCL12-RT-5.1 and CXCL12-RT-3.1; c-Myc-RT-5.1 and c-Myc-RT-3.1; cyclin D1-RT-5.1 and cyclin D1-RT-3.1 to detect the mRNA level of these genes (Additional file
1: Table S1). GAPDH was used as an internal control. Delta-delta Ct values were used to determine their relative expression.
KO of linc-RoR by CRISPR/Cas9
We used a dual gRNA approach to knock out exon 4 of linc-RoR (RoR E4) by CRISPR/Cas9 system [
23,
24]. Briefly, dual gRNA and donor vector were co-transfected into MCF-7 cells. One week later, the transfected cells were subject to puromycin (1 μg/ml) selection; and surviving cells were sorted by FACS based on GFP signal into 96-well plates and then expanded in 12-well plates. Potential clones were further verified by genomic PCR and qRT-PCR.
Cell proliferation assay
Cell proliferation assays were carried out by MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assays, as previously described [
25].
In silico data analysis
Correlations between breast cancer patient survival and DUSP7 expression (probe: 214793_at) was analyzed by KM plotter (
http://kmplot.com). Auto select best cutoff was chosen in the analysis. A total of 3951 breast cancer samples were split in high and low groups according to the cutoff value, respectively. The hazard ratio with 95% confidence intervals and log rank
P value was calculated and significance was set at
P < 0.05. The expression of DUSP7 in breast cancer grouped by ER status was identified from TCGA databases in Oncomine (Compendia Biosciences,
www.oncomine.org) using the analysis of “Dataset view” as previously reported [
26].
Statistical analysis
All statistical analyses were performed using the GraphPad Prism Version 6.0 program. The continuous variables are summarized as mean and standard error of mean (S.E.M.) unless stated. The two-sample t test was used to compare the mean of a continuous variable between two samples. The Satterthwaite’s t test was used for mean comparison when the variances in two samples were unequal. Association between two categorical variables was evaluated by using the Fisher’s exact test. Relapse-free survival (RFS) curves were calculated with the Kaplan-Meier method and were analyzed with the log-rank test. All P values were two-sided and P values <0.05 were considered as significant.
Discussion
Estrogen-independent activation of ER through MAPK/ERK pathway was reported previously and its role in endocrine therapies resistance has been well documented, however, the precise underlying mechanism of its regulation is still largely unknown. In this study, we identified and characterized linc-RoR as a novel regulator of MAPK/ERK signaling cascade and thereby may have critical roles in ligand-independent growth of breast cancer cells. Several lines of evidence support this notion. First, linc-RoR is induced by estrogen deprivation; linc-RoR KO results in suppression of cell growth under estrogen deprivation and increases the sensitivity of MCF-7 cells to tamoxifen. Second, linc-RoR KO abolishes the activation of ER and its downstream target under estrogen deprivation. Third, linc-RoR activates the MAPK/ERK cascade upon estrogen deprivation, whereas this activation is dramatically repressed in linc-RoR KO cells. Forth, re-expression of linc-RoR in KO cells restores the estrogen deprivation-induced activation of ER and ERK. Fifth, as an important regulator of ERK, DUSP7 is less stable (shorter half-life) in linc-RoR KO cells than in vector control cells.
ER is the primary modular protein responsible for most effects of estrogen on breast cancer cells. The classical model of ER activation involves the binding of estrogen to ER; on the other hand, ER can also be activated in ligand-independent manner. It is well known that phosphorylation of ER is important for its estrogen-independent activation and endocrine therapy response in breast cancer. The serine118 residue of ER is a well-studied site which is known to be activated by MAPK/ERK in response to estrogen deprivation and thus facilitates the interaction of ER and coactivators, leading to the ligand-independent activation of ER and resistance to endocrine therapy [
34,
35]. Work in many laboratories has reported the phosphorylation of ER at Ser118 is upregulated in tamoxifen-resistant MCF-7 cells [
36,
37]. Importantly, low level of phosphorylation of ER at Ser118 also is associated with significantly improved disease-free and overall survival [
38]. In this study, we demonstrate that E2-free treatment induces the activation of ER at Ser118 and results in the upregulation of ER target genes. However, both ER activation and upregulation of its target gene are abrogated by linc-RoR KO. Importantly, re-expression of linc-RoR in KO cells substantially restores the activation of ER as well as its target gene expression under estrogen deprivation. These findings suggest that linc-RoR might be required for estrogen-independent activation of ER.
Mitogen-activated protein kinase (MAPK) pathways are evolutionarily conserved kinase modules involved in the regulation of fundamental cellular processes such as cell growth, proliferation, survival, migration and apoptosis [
39,
40]. A number of studies have hinted that the aberrant activated MAPK/ERK cascade plays a critical role in many aspects of tumorigenesis. Nearly 50% of human malignancies exhibit unregulated ERK signaling [
41]. In breast cancer, the upregulated MAPK/ERK signaling is correlated with poor survival in triple-negative breast cancer patients [
42]. In addition, the MAPK/ERK pathway also influences chemotherapeutic drug resistance to doxorubicin and paclitaxel in breast cancer cells [
43]. In consistence with these results, our analysis of ERK and ER activation also suggests that the MAPK/ERK signaling cascade is activated upon estrogen deprivation, leading phosphorylation of ER at Ser118. ERK inhibitor U0126 dramatically alleviates estrogen deprivation-induced ER phosphorylation at Ser118, suggesting that the estrogen deprivation-induced ER activation is dependent on ERK activation.
The role of lncRNAs MAPK/ERK pathway has been reported in literature. Li et al. showed that lncRNA CASC2 suppresses proliferation of gastric cancer cells by inactivating ERK and JNK [
44]. Moreover, lncRNA MALAT1 is able to promote the proliferation and metastasis of gallbladder cancer cells by activating the ERK/MAPK pathway [
45]. However, up to date there is no evidence that any lncRNA is involved in estrogen-independent activation of MAPK/ERK signaling. We show that although there is no difference in the phosphorylation of ERK between gRNA control and linc-RoR KO cells under regular condition, the activation of ERK is considerably lower in linc-RoR KO than in gRNA control cells under E2-free condition. In addition, re-expression of linc-RoR in KO cells reverses the phosphorylation of ERK. This result highlights the critical role of linc-RoR in estrogen deprivation-induced MAPK/ERK pathway activation.
It is well known that ERK is positively regulated by MEK and negatively regulated by MKPs. Since there is no difference in MEK activation between gRNA and linc-RoR KO cells either in regular or E2-free condition, we explore the possible involvement of MKPs in linc-RoR-mediated ERK activation in response to estrogen deprivation because the loss of MKPs has been shown to cause the constitutive activation of ERK signal pathway [
46‐
48]. In support of these findings, we show that the downregulation of ERK activation in linc-RoR KO cells is restored by MKP inhibitor NSC95397 under E2-free condition. In particular, DUSP7 appears to play a critical role estrogen deprivation-induced ERK activation in MCF-7 cells. Despite a report that high expression of DUSP7 may play a role in development of leukemia [
49], other studies also suggest that a low level of DUSP7 is associated with progression of kidney cancer [
50]. Here we provide evidence that DUSP7 is downregulated in response to estrogen deprivation and linc-RoR KO significantly increases the half-life of DUSP7, possibly resulting in the inactivation of ERK. Since p53 can induce DUSP7 [
51] and linc-RoR can strongly suppress p53 translation [
17], it is our speculation that linc-RoR/p53 axis may play a role in regulation of ERK phosphorylation. Nevertheless, the detailed mechanism remains to be determined yet.
Finally, the clinical data analysis also supports the significance of DUSP7 in breast cancer. For instance, DUSP7 level is lower in ER+ breast tumors than in ER- breast tumors. Particularly, the low expression of DUSP7 is correlated with poor prognosis (Fig.
6b). Therefore, it would be interesting to determine in future whether the linc-RoR/DUSP7/ERK axis may serve a target for therapeutic intervention.