YAP, CTGF and Cyr61 are overexpressed in tamoxifen-resistant breast cancer and induce transcriptional repression of ERα

Estrogen receptor α (ERα, encoded by ESR1) is a hormone-activated nuclear transcription factor that regulates gene expression and plays important roles in breast cancer progression. ERα is expressed in ∼70% of breast cancers, making them sensitive to endocrine therapies, including tamoxifen (Harvey et al., 1999). ERα-positive breast cancer patients have better prognosis than patients with ERα-negative breast cancer due to the benefits they can receive from endocrine therapies targeting ERα (Early Breast Cancer Trialists’ Collaborative, 2005). However, although tamoxifen therapy has contributed to a 25–30% decrease in breast cancer mortality, 30% of 5-year tamoxifen-treated patients develop tamoxifen resistance within 15 years (Early Breast Cancer Trialists’ Collaborative, 2005). Therefore, delineating the mechanism of tamoxifen resistance could be a breakthrough in resolving a significant clinical problem for breast cancer treatment.

The molecular mechanisms of tamoxifen resistance include loss of ERα, expression of truncated forms of ERα, post-translational modification of ERα, dysregulation of ERα co-activators and increased receptor tyrosine kinase signaling pathways, such as the epidermal growth factor receptor (EGFR) pathway (Musgrove and Sutherland, 2009). Because tamoxifen acts as an ER antagonist, de novo resistance to it is primarily caused by a lack of ERα expression (Jaiyesimi et al., 1995; Musgrove and Sutherland, 2009). Studies have reported that ERα is downregulated in tumors with acquired resistance to tamoxifen (Dowsett and Haynes, 2003; Encarnacion et al., 1993; Gutierrez et al., 2005; Johnston et al., 1995). Johnston et al. demonstrated that 17–28% of breast tumors with acquired resistance to tamoxifen did not express ERα (Johnston et al., 1995). In addition, a study using tissues from 39 breast cancer patients revealed that 17% of tamoxifen-resistant tumors showed loss of ERα expression (Gutierrez et al., 2005). Loss of ERα expression could be due to enhanced methylation at the ERα promoter region (Iwase et al., 1999). A recent study reported that ERα is negatively regulated in tamoxifen-resistant MCF-7 cells by direct binding of Slug (also known as SNAI2) to the ERα promoter, which causes transcriptional repression of ERα, suggesting that direct modulation of ERα transcription could be part of the tamoxifen resistance mechanism (Li et al., 2015).

The transcriptional co-activator Yes-associated protein (YAP; also known as YAP1) is a major downstream effector of the Hippo signaling pathway (Wang et al., 2009). In breast cancer, YAP and the bona fide YAP target genes connective tissue growth factor (CTGF; also known as CCN2) and cysteine-rich angiogenic inducer 61 (Cyr61; also known as CCN1) are involved with cancer progression (Zhao et al., 2008). Guo et al. demonstrated that YAP promotes cell proliferation and survival through increased phosphorylation of Akt family proteins (Guo et al., 2019). Xie et al. used 44 breast patient samples to show that CTGF and Cyr61 overexpression correlates with an advanced stage of breast cancer (Xie et al., 2001). Interestingly, forced expression of Cyr61 induced estrogen-independent growth, increased invasion of MCF-7 cells in vitro, enhanced tumorigenesis in vivo and led to downregulation of ERα (Tsai et al., 2002).

In this study, we found that YAP, CTGF and Cyr61 were upregulated, and ERα was downregulated, in patients with recurrent breast cancer following tamoxifen treatment. In addition, there was a negative correlation between YAP, CTGF and Cyr61 expression and ERα expression. We examined the molecular mechanism of tamoxifen resistance using a tamoxifen-resistant (TR) cell line. The TR cells showed loss of ERα with a concomitant increase in EGFR and YAP activity. We found that the increased YAP activity was driven by EGFR signaling and that CTGF and Cyr61, downstream effectors of YAP, caused transcriptional repression of ERα by directly binding to the promoter regions of ERα. We also found that extracellular CTGF bound to integrin αv, resulting in activation the FAK/Src/NF-κB p65 signaling axis and leading to increased glycolysis, which eventually led to enhanced cell growth and migration of the TR cells.

By performing a survival analysis using the Kaplan−Meier Plotter tool (http://kmplot.com/analysis/) for human samples as described in the Materials and Methods, we found that ER-positive, tamoxifen-treated breast cancer patients with high YAP, CTGF and Cyr61 expression showed lower survival rates than those with low expression (Fig. 1A). In addition, CTGF expression was inversely correlated with the survival of ER-positive recurrent breast cancer patients in the METABRIC patient cohort dataset (Fig. 1B). In the same dataset, CTGF and Cyr61 were upregulated in patients who received hormone therapy (Fig. 1C). Also, we found that YAP, CTGF and Cyr61 expression was inversely correlated with ERα expression (Fig. 1D). Interestingly, that inverse correlation was increased in patients who received hormone therapy (Fig. 1D). A gene set enrichment analysis performed on expression data from METABRIC showed that YAP, CTGF and Cyr61 expression correlated negatively with genes related to estrogen responses (Fig. 1E). Immunohistochemical analysis of breast cancer tissues derived from 27 breast cancer patients [ER+, HER2− (HER2 is also kwon as ERBB2)] who received tamoxifen treatment and for whom recurrence occurred in 11 patients among them, revealed that YAP, CTGF and Cyr61 were upregulated while ERα was downregulated in the recurrent patients treated with tamoxifen (Fig. 2A). Also, ERα expression was inversely correlated with the expression of YAP, CTGF and Cyr61 (Fig. 2B). Observing a negative correlation between ERα and YAP, CTGF, and Cyr61 in publicly available datasets and patient tissues, we hypothesized that YAP, CTGF and Cyr61 might be involved in tamoxifen resistance through interference with ERα signaling. To further confirm our hypothesis, we used a tamoxifen-resistant MCF-7 cell line. We first treated MCF-7 tamoxifen-sensitive (TS) and tamoxifen-resistant (TR) cells with 4-hydroxytamoxifen (4-OHT), which is the active metabolite of tamoxifen, to confirm the tamoxifen-resistant phenotype of the TR cells. MTT assays, colony formation assays and cell cycle analyses confirmed the tamoxifen-resistant phenotypes in the TR cells compared to the TS cells (Fig. S1). We observed upregulation of YAP signaling through increased expression and decreased S127 phosphorylation of YAP (S127 phosphorylation of YAP leads to cytoplasmic retention of YAP; Zhao et al., 2010). CTGF and Cyr61, YAP downstream targets, were also overexpressed in the TR cells (Fig. 3A,B). The expression of ERα and its target genes [cathepsin D (CTSD) and pS2 (also known as TFF1)] were downregulated in the TR cells (Fig. 3B). Through a cell fractionation assay, we observed decreased ERα expression and increased YAP nuclear translocation in the TR cells (Fig. 3C). We also observed decreased ERα expression and increased YAP, CTGF and Cyr61 expression in the TR cells using immunofluorescence microscopy (Fig. 3D). Reporter assays confirmed that ERα-dependent transcriptional activity (MERE-luc) was downregulated while YAP-dependent transcriptional activity (GTIIC-luc) and CTGF and Cyr61 promoter activities were upregulated in the TR cells (Fig. 3E).

Because we observed that YAP was upregulated and ERα was downregulated in the TR cells, consistent with our analyses of public datasets, we next investigated the effect of YAP on ERα expression. When we overexpressed YAP in the TS cells, ERα was downregulated, whereas YAP inhibition using verteporfin in the TR cells led to the recovery of ERα levels (Fig. S2A). Also, we found that YAP induced tamoxifen resistance in the TS cells, whereas YAP inhibition increased sensitivity to tamoxifen in the TR cells (Fig. S2B,C). Altogether, these results indicate that upregulated YAP signaling induces tamoxifen-resistant phenotypes in MCF-7 cells by negatively regulating ERα.

We next sought to further elucidate the role of CTGF and Cyr61, downstream effectors of YAP, in the phenotypes of the TR cells. When we administered 4-OHT to the CTGF- (CTGF si) and Cyr61-knockdown (Cyr61 si) TR cells and performed MTT and colony formation assays, we observed increased sensitivity to 4-OHT (Fig. 4A,B). In addition, flow cytometric cell cycle analyses revealed that CTGF and Cyr61 knockdown resulted in a dose-dependent increase in the sub-G1 and G1 populations following 4-OHT treatment (Fig. 4C). Using colony formation assays, we observed that neutralizing extracellular CTGF and Cyr61 using CTGF- and Cyr61-specific antibodies increased tamoxifen sensitivity in the TR cells (Fig. 4B, right panel). Because ERα was downregulated in the TR cells, we examined the ERα levels in the CTGF- and Cyr61-knockdown cells and found increased ERα mRNA and protein expression. Neutralizing extracellular CTGF and Cyr61 also increased ERα (Fig. 4D). To further elucidate the effects of CTGF and Cyr61 on ERα-mediated signaling, we performed luciferase assays to measure the ERα transcriptional activity. CTGF and Cyr61 knockdown and neutralization resulted in increased responsiveness to E2 (Fig. 4E). Interestingly, we reproduced these results in two triple-negative breast cancer (TNBC) cell lines, MDA-MB-231 and Hs578T. Knockdown of CTGF and Cyr61 increased sensitivity to 4-OHT, ERα expression and E2 responsiveness, even in those ERα-negative TNBC cell lines (Fig. S3). These data collectively suggest that CTGF and Cyr61 confer resistance to tamoxifen by downregulating ERα. Because CTGF and Cyr61 knockdown conferred tamoxifen sensitivity to the TR cells, we overexpressed CTGF and Cyr61 in the TS cells to further investigate their role in tamoxifen resistance. MTT and colony formation assays revealed that CTGF and Cyr61 overexpression induced tamoxifen resistance (Fig. S4A,B). In addition, cell cycle analyses showed that 4-OHT did not induce G1 arrest or apoptosis in cells overexpressing CTGF and Cyr61 (Fig. S4C). Treatment with recombinant human CTGF and Cyr61 (rhCTGF and rhCyr61) increased resistance to tamoxifen (Fig. S4B, right panel). Overexpression of CTGF and Cyr61 and treatment with rhCTGF and rhCyr61 downregulated both ERα expression and ERα activity (Fig. S4D,E). Using another ERα-positive cell line, T47D, we confirmed that CTGF and Cyr61 decreased sensitivity to 4-OHT, ERα expression and E2 responsiveness (Fig. S5). These results, which are consistent with the results seen upon CTGF and Cyr61 knockdown in the TR cells, indicate that CTGF and Cyr61 decrease ERα levels and induce tamoxifen resistance.

The above results indicate that in the course of obtaining tamoxifen resistance, YAP-mediated CTGF and Cyr61 transcription is increased, and CTGF and Cyr61 overexpression leads to downregulation of ERα. Therefore, we next investigated the mechanism that leads to the increase in YAP signaling. Because increased expression and downstream activation of growth factor receptors, such as EGFR, HER2, and insulin-like growth factor 1 receptor are associated with tamoxifen resistance, we examined EGFR expression and downstream signaling activity (Musgrove and Sutherland, 2009). We found that EGFR was overexpressed in the TR cells, and downstream components, such as Akt and ERK family proteins (ERK1/2, also known as MAPK3 and MAPK1, respectively), were also activated in the TR cells (Fig. 5A). Specific kinase inhibitors (ZD1839, an EGFR inhibitor; U0126, an ERK inhibitor; and AI#8, an Akt inhibitor) were administered to the TR cells to further determine the signaling pathway leading to CTGF and Cyr61 overexpression (Fig. 5B; Fig. S6A). ZD1839 and U0126 downregulated YAP and increased YAP phosphorylation at S127, leading to downregulation of CTGF and Cyr61 at the transcriptional level (Fig. 5B). These results demonstrate that increased EGFR/ERK signaling leads to increased YAP activity and overexpression of CTGF and Cyr61 in the TR cells. To further elucidate the mechanism by which CTGF and Cyr61 downregulate ERα, we hypothesized that they might have nuclear functions that lead to the transcriptional repression of ERα. Although CCN proteins have primary functions as secreted proteins, some previous studies have indicated the possibility of nuclear transport (Planque et al., 2006; Wiesman et al., 2010). Wahab et al. showed that CTGF, after binding to the cell surface, is internalized to the cytosol and translocated to the nucleus through endocytosis (Wahab et al., 2001). Consistent with those reports, we found CTGF and Cyr61 in the nucleus of MCF-7 TR cells by confocal imaging and cell fractionation assays (Figs 3D and 5C).

To determine whether CTGF and Cyr61 negatively regulate ERα at the transcriptional level by directly binding to the ERα promoter region, we first sought to find the exact genomic location in the ERα promoter region that was negatively regulated in the TR cells. To that end, we generated four reporter constructs with different ERα promoter regions (A, B, C, and F-luc) that are known to be highly used in breast cancer (Kos et al., 2001; Reid et al., 2002). Luciferase assays with those reporter constructs revealed that the activity of A, B, and C-luc was decreased in the TR cells, whereas F-luc activity did not differ between the TS and TR cells (Fig. 5D, lower left panel). Therefore, we next performed chromatin immunoprecipitation quantitative PCR (CHIP-qPCR) with ERα promoter region-specific primers to determine whether CTGF and Cyr61 bind directly to those promoter regions. We found that both CTGF and Cyr61 bind to ERα promoter regions A and B but not C and F (Fig. 5D, lower right panel). To further dissect the binding domains of CTGF and Cyr61 on the ERα promoter regions, we used a sequence-based prediction method to predict the putative DNA-binding domains of CTGF and Cyr61 (Yan and Kurgan, 2017). Interestingly, we found a conserved serine-threonine-arginine (STR) domain in the predicted DNA-binding domains of CTGF and Cyr61 (Fig. 5E, upper panel). After confirming that site-directed mutagenesis of that STR domain did not impair the translocation of CTGF and Cyr61 to the nucleus, we performed luciferase assays with mutant proteins (Fig. 5E; Fig. S6B). STR mutation of CTGF and Cyr61 relieved the inhibitory effect of those proteins on ERα transcription (Fig. 5E, middle left panel). Also, rhCTGF and rhCyr61 treatment decreased activity in the ERα promoter regions in the TS cells, whereas knockdown and neutralization of CTGF and Cyr61 increased activity in the ERα promoter regions in the TR cells (Fig. 5E, lower panels and middle right panel). These data collectively suggest that CTGF and Cyr61 translocate into the nucleus and bind directly to ERα promoter regions A and B through the conserved STR domain, resulting in negative regulation of ERα at the transcriptional level.

Because we observed direct binding of CTGF and Cyr61 to the promoter regions of ERα and elucidated the protein domain that is critical for binding, we next tried to find the binding motif on the promoter regions of ERα. When we aligned promoter regions A and B, we observed a consensus GCCTCTA sequence in both regions (Fig. 5F, upper panel). We confirmed that the GCCTCTA motif is a putative transcriptional factor binding site using a bioinformatical approach (Farre et al., 2003). Therefore, we generated GCCTCTA motif deletion mutants of promoter regions A and B and performed luciferase assays. We observed that the inhibitory effects of CTGF and Cyr61 on those regions were diminished in the deletion mutants, indicating that the GCCTCTA motif is critical for CTGF and Cyr61 binding (Fig. 5F, middle and lower panel).

Beyond the role of CTGF and Cyr61 in the negative regulation of ERα, we tried to determine further roles for those proteins in the oncogenic properties of the TR cells. Because secreted CTGF and Cyr61 are known to relay intracellular signals related to cell proliferation by binding on cell surface integrins, we checked the status of integrin-related signaling molecules (Jun and Lau, 2011). Focal adhesion kinase (FAK, also known as PTK2) and Src, a kinase downstream of the integrins, were activated in the TR cells (Fig. 6A) (Mitra and Schlaepfer, 2006). Also, phosphorylation, nuclear translocation, and transcriptional activity of NF-κB subunit p65 (also known as RELA), a downstream component of Src, were increased in the TR cells (Fig. 6A,B) (Lee et al., 2007; Lluis et al., 2007). Among the diverse roles of NF-κB p65 reported in previous literatures, we focused on its function in transcriptional regulation of Glut3 (also known as SLC2A3) (Kawauchi et al., 2008; Watanabe et al., 2012; Zha et al., 2015). We found that among the Glut family proteins known to be associated with cancer, Glut3, but not Glut1, Glut2 and Glut4 (also known as SLC2A1, SLC2A2 and SLC2A4, respectively), expression, was increased in the TR cells (Fig. 6A,B; Fig. S6C) (Adekola et al., 2012). Also, Glut3 was found to be overexpressed in recurrent breast cancer samples from tamoxifen-treated patients and Glut3 expression was positively correlated with YAP, CTGF, and Cyr61 expression (Fig. S7). By administering specific inhibitors of YAP (VP, verteporfin), Src (PP2) and NF-κB p65 (celastrol), we found that Glut3 was transcriptionally upregulated through the YAP/FAK/Src/NF-κB p65 axis in the TR cells (Fig. 6E). Contrary to our expectation that both CTGF and Cyr61 would activate FAK/Src/NF-κB p65-mediated transcription of Glut3, only CTGF, but not Cyr61, overexpression activated the signaling axis (Fig. 6C). Also, rhCTGF, but not rhCyr61, treatment induced FAK/Src/NF-κB p65-mediated Glut3 expression in the TS cells, and CTGF, but not Cyr61, neutralization downregulated FAK/Src/NF-κB p65-mediated Glut3 expression in the TR cells (Fig. 6D, left two panels). By administering rhCTGF to the integrin αv-knockdown TS cells and neutralizing integrin αv using integrin αv-specific antibodies in the TR cells, we found that secreted CTGF relays signal through integrin αv (Fig. 6D, left two panels). This was further supported by co-immunoprecipitation assays showing that CTGF physically binds to integrin αv (Fig. 6D, right panel). Because we found that Glut3 was upregulated through the CTGF/FAK/Src/NF-κB p65 axis in the TR cells, we assessed alterations in glycolytic phenotypes, such as glucose uptake, ATP production and lactate production rates. Consistent with the results showing an increase of glycolytic phenotypes in the TR cells, CTGF and Glut3 overexpression increased glycolytic phenotypes in the TS cells, and CTGF and Glut3 knockdown in the TR cells showed the opposite effect (Fig. 6F). These results suggest that secreted CTGF, but not Cyr61, relays signal through the integrin αv/FAK/Src/NF-κB p65 axis, which leads to transcriptional upregulation of Glut3 and increased glycolysis.

After establishing the mechanistic link by which CTGF affects glucose metabolism, we examined the functional consequences of the increase in glycolytic phenotypes. Because increased glucose consumption and metabolism contribute to the growth and metastasis of cancer cells (Liberti and Locasale, 2016; Han et al., 2013), we first checked the proliferation rates of the TS and TR cells and found that TR cells had increased proliferation compared to the TS cells (Fig. 7A). In addition, we tested whether the CTGF/FAK/Src/NF-κB p65-mediated upregulation of Glut3 increased cell proliferation in the TR cells. Forced expression of CTGF and Glut3 as well as rhCTGF treatment increased cell proliferation in the TS cells (Fig. 7A, middle panels). Also, knockdown of CTGF and Glut3 and neutralization of CTGF, Glut3 and integrin αv decreased cell proliferation in the TR cells, whereas Cyr61 knockdown and neutralization had no effect (Fig. 7A, lower panels). The increased cell proliferation mediated by CTGF and Glut3 overexpression and rhCTGF treatment in the TS cells was due to increased cell cycle progression (Fig. 7B, top panels). Consistent with this, decreased cell proliferation following CTGF and Glut3 knockdown and treatment with CTGF, Glut3, and integrin αv antibodies was caused by G1 arrest (Fig. 7B, lower panels). Furthermore, treatment with YAP, Src and NF-κB p65 inhibitors phenocopied the effects of CTGF and Glut3 inhibition by inducing cell cycle arrest and decreasing cell growth in the TR cells (Fig. 7C). To confirm the effect of CTGF/Glut3 axis on cell growth, proliferation assays and cell cycle analyses were also performed on T47D cells. We observed that CTGF and Glut3 overexpression and rhCTGF treatment induced cell proliferation through increased cell cycle progression, consistent with the results from the TS cells (Fig. S8A,B).

Because tamoxifen resistance in breast cancer is known to be associated with increased motility and invasiveness, we conducted cell migration assays with the TS and TR cells (Hiscox et al., 2004). The TR cells showed increased cell migration in transwell migration assays (Fig. 8A). Next, to assess the effects of CTGF/FAK/Src/NF-κB p65-induced Glut3 expression on cell migration, we performed cell migration assays under various experimental conditions. Ectopic expression of CTGF and Glut3 and rhCTGF treatment enhanced cell migration in both TS and T47D cells, whereas knockdown of CTGF and Glut3 and neutralization of CTGF, Glut3 and integrin αv reduced cell migration in the TR cells (Fig. 8A; Fig. S8C). Also, treatment with YAP, Src and NF-κB p65 inhibitors decreased TR cell migration (Fig. 8B). Our data collectively suggest that CTGF and Cyr61 contribute to tamoxifen resistance by decreasing ERα expression and that CTGF enhances cell growth and migration by upregulating Glut3 through the FAK/Src/NF-κB p65 signaling axis (Fig. 8C).

In ER-positive breast cancer patients, tamoxifen is commonly considered as the first-line therapy because it provides significant clinical benefits in reducing tumor progression. However, acquired resistance following long-term treatment can cause tumor relapse. In this study, we demonstrated that CTGF and Cyr61, downstream effectors of YAP, are overexpressed in tamoxifen-resistant breast cancer and lead to acquired tamoxifen resistance through transcriptional repression of ERα. Mechanistically, CTGF and Cyr61 bind to the promoter regions of ERα, negatively regulating ERα transcription, and CTGF also induces Glut3 expression, which increases glycolytic phenotypes, cell proliferation and migration. Knockdown of CTGF and Cyr61 sensitized the TR and TNBC cells to tamoxifen treatment, and knockdown of CTGF reduced cell proliferation and migration in the TR cells. Conversely, overexpression of CTGF and Cyr61 increased tamoxifen resistance, and overexpression of CTGF increased cell proliferation and migration in the ER-positive cells.

Because tamoxifen exerts its therapeutic effects by acting as an ER antagonist, loss of ER is associated with the development of tamoxifen resistance. A lack of ER expression is known as the dominant mechanism of intrinsic resistance to tamoxifen (Saxena and Sharma, 2010). Transcriptional repression of ER and population remodeling of heterogenous ER-positive tumors into ER-negative tumors are potential mechanisms that could cause the loss of ER in ER-positive breast cancer (Chang, 2012). Interestingly, previous studies have demonstrated the negative correlation between EGFR and ER (Rimawi et al., 2010; Tsutsui et al., 2002; Zhou et al., 2009). Forced EGFR activation in ER-positive breast cancer downregulates ER expression and promotes estrogen-independent growth (Moerkens et al., 2014; Stoica et al., 2000). Also, ERK activation in ER-positive breast cancer leads to a decrease in ER expression, whereas ERK inhibition in ER-negative breast cancer increases ER expression (Bayliss et al., 2007; Oh et al., 2001). A recent study has shown that EGFR activation leads to transcriptional downregulation of ERα and that EGFR inhibition increases sensitivity to tamoxifen in vivo (Jeong et al., 2019). In line with previous research, we found that EGFR-mediated ERK activation led to increased YAP signaling, followed by CTGF and Cyr61 overexpression, which led to the loss of ER.

Several studies have demonstrated the crosstalk between Hippo and ER signaling. Ablation of LATS (herein referring to both LATS1 and LATS2), the kinase upstream of YAP and TAZ (TAZ is also known as WWTR1), results in stabilization of ERα in breast cancer (Britschgi et al., 2017). Zhu et al. found that YAP and TEAD physically interact with ERα and bind to ERα enhancers using an in vivo proximity-dependent labeling technique (Zhu et al., 2019). Interestingly, G protein-coupled estrogen receptor (GPER; also known as GPER1), stimulated by agonists such as E2 and 4-OHT, activates YAP/TAZ signaling and increases cell proliferation and migration of breast cancer cells (Zhou et al., 2015). Similarly, CTGF is induced by E2 and 4-OHT treatment in SkBR3 cells, which lack ER expression, through GPER signaling, which promotes cell proliferation and migration (Pandey et al., 2009). These studies indicate that 4-OHT acts as an agonist for GPER and that GPER activation can lead to activation of YAP and induction of downstream genes, including CTGF and Cyr61. In this regard, our results showing that YAP, CTGF, and Cyr61 are overexpressed in tamoxifen-resistant breast cancer cells through EGFR/ERK signaling can further explain GPER-mediated YAP activation since GPER is known to transactivate EGFR (Filardo et al., 2000). Furthermore, the novel role of CTGF and Cyr61 as transcriptional repressors of ERα adds an additional dimension to previous studies demonstrating GPER-mediated induction of CTGF and Cyr61 as a mechanism leading to tamoxifen resistance.

Also, Cheng et al. demonstrated that CTGF physically interacts with ER and diminishes the transcriptional activity of ER (Cheng et al., 2011). However, electrophoretic mobility shift assays revealed that CTGF did not directly inhibit the binding of ERα to the estrogen response element, implying that the physical interaction between CTGF and ERα did not downregulate ERα-mediated transcriptional activity (Cheng et al., 2011). They also found that CTGF downregulated the expression of ERα at the mRNA level, consistent with our results (Cheng et al., 2011). Furthermore, they showed that the regions 188 to 349 of CTGF, which includes the STR domain, is responsible for those effects, supporting our results (Cheng et al., 2011).

Increased aerobic glycolysis in cancer cells correlates with increased cancer progression by promoting cell proliferation, migration and survival (Liberti and Locasale, 2016; Vander Heiden and DeBerardinis, 2017). Several studies have demonstrated a correlation between tamoxifen resistance and glucose metabolism. Woo et al. demonstrated that increased aerobic glycolysis is linked with tamoxifen resistance, and Fiorillo et al. reported that enhanced mitochondrial activity drives tamoxifen resistance (Woo et al., 2015; Fiorillo et al., 2017). Recently, a study reported a connection between extracellular matrix (ECM) remodeling and glucose metabolism. In that study, hyaluronan digestion by hyaluronidase triggered glycolysis by upregulating Glut1, which led to increased migration, suggesting an interesting link between ECM components and glycolysis (Sullivan et al., 2018). Here, we have shown that extracellular CTGF binds to integrin αv and activates FAK and Src kinases, which are integrin-mediated intracellular signaling molecules. Consistent with previous studies, showing evidence that Src directly activates NF-κB p65, we found that CTGF/FAK/Src signaling activates NF-κB p65 in the TR cells, followed by NF-κB p65-mediated transcriptional upregulation of Glut3 (Lee et al., 2007; Lluis et al., 2007).

In summary, our findings reveal that CTGF and Cyr61 promote tamoxifen resistance by transcriptionally downregulating ERα and that CTGF induces cell proliferation and migration through the FAK/Src/NF-κB p65 axis. Also, our study provides a partial explanation of the mechanism that leads to the loss of ER in recurrent breast cancer that was originally ER positive (Heitz et al., 2013). In vivo evidence from clinical datasets also revealed that tamoxifen resistance is correlated with expression of YAP downstream effectors CTGF and Cyr61. Taken together, our results indicate the clinical potential of CTGF and Cyr61 as therapeutic targets for tamoxifen-resistant breast cancer.

MDA-MB-231, Hs578T and T47D human breast cancer cell lines were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). All cell lines were maintained in Dulbecco's modified Eagle's medium (DMEM; 10-013-CVR, Corning, NY, USA) containing 10% fetal bovine serum (FBS, Youngin Frontier, Seoul, Korea), 100 U/ml penicillin (15140-122, Corning) and 100 µg/ml streptomycin (15140-122, Corning). The cells were incubated at 37°C in 5% CO₂ in a humidified atmosphere. MCF-7 tamoxifen-sensitive (TS) and tamoxifen-resistant (TR) cells were kindly provided by Prof. Keon Wook Kang (College of Pharmacy, Seoul National University, Seoul, South Korea). MCF-7 TR cells were established using previously reported methods (Choi et al., 2007). Briefly, MCF-7 cells were washed with phosphate-buffered saline (PBS) and maintained in Phenol Red-free DMEM, containing 10% charcoal-stripped FBS with 0.1 μM 4-hydroxytamoxifen (4-OHT). The cells were continuously exposed to 0.1 μM 4-OHT for 2 weeks, and the 4-OHT concentration was gradually increased to 3 μM over 9 months. The initially decreased cell growth rate of the 4-OHT-exposed MCF-7 cells increased gradually, indicating the development of tamoxifen resistance.

Antibodies against the following proteins were used in this study (indicated dilutions are for western blots when not specified): actin (sc-8432, 1:500), β-tubulin (sc-5274, 1:500), CTGF (sc-365970, 1:200 for immunofluorescence staining and 1:100 for immunoprecipitation), Cyr61 (sc-374129, 1:500 for western blots, 1:50 for immunohistochemistry, and 1:200 for immunofluorescence staining, and 1:100 for immunoprecipitation), ERK1 (sc-94, 1:2000), ERα (sc-8002, 1:1000 for western blots, 1:200 for immunohistochemistry, and 1:200 for immunofluorescence staining), GAPDH (sc-47724, 1:1000), Glut3 (sc-74497, 1:250 for western blots and 1:50 for immunostaining), Histone H1 (sc-34464, 1:500), integrin αv (sc-9969, 1:500 for western blots and 1:100 for co-immunoprecipitation), NFκB p65 (sc-8008, 1:500), p-ERK1/2 (Y204) (sc-7383, 1:1000) and YAP (sc-101199, 1:1000 for western blots, 1:50 for immunohistochemistry, and 1:200 for immunofluorescence staining) (Santa Cruz Biotechnology, Dallas, TX, USA); Akt (#9272, 1:1000), FAK (#3285, 1:1000), Myc tag (#2276, 1:1000), p-Akt (S473) (#9271, 1:500), p-EGFR (Y1068) (#2234, 1:500), p-FAK (Y925) (#3284, 1:500), p-NFκB p65 (S536) (#3033, 1:500), p-Src (Y416) (#2101, 1:500), p-YAP (S127) (#4911, 1:1000) and Src (#2109, 1:500) (Cell Signaling Technology, Beverly, MA, USA); EGFR (ab52894, 1:1000) (Abcam, Cambridge, UK); CTGF (500-P252, 1:1000 for western blots and 1:100 for immunohistochemistry) (PeproTech, Princeton, NJ, USA). Horseradish peroxidase (HRP)-conjugated secondary antibodies were: anti-rabbit IgG (#7074, 1:2000) and anti-mouse IgG (#7076, 1:2000), obtained from Cell Signaling Technology; and anti-goat IgG (sc-2020, 1:2000), obtained from Santa Cruz Biotechnology. Isotype control antibodies were: normal mouse IgG (sc-2025) was obtained from Santa Cruz Biotechnology; and normal rabbit IgG (#2729) was obtained from Cell Signaling Technology. Recombinant human CTGF (GTX48458-PRO) and Cyr61 (GTX48189-PRO) protein were purchased from Genetex (Alton Pkwy Irvine, CA, USA). 4-Hydroxytamoxifen (4-OHT, H7904), β-estradiol (E2758), celastrol (C0869), RNaseA (R6513), and verteporfin (SML0534) were purchased from Sigma-Aldrich (St Louis, MO, USA). Hoechst 33342 (H1399) and propidium iodide (PI; P3566) were purchased from Thermo Fisher Scientific (Waltham, MA, USA). Akt Inhibitor VIII (124018), PP2 (529573) and U0126 (662005) were purchased from Calbiochem (St Louis, MO, USA). ZD1839 (PKI-GFTB) was purchased from Biaffin GmbH & Co. KG (Kassel, Germany).

Control siRNA (sc-37007) and CTGF (SC-39329), Cyr61 (sc-39331), Glut3 (sc-41218) and integrin αv (sc-29373)-specific siRNAs were purchased from Santa Cruz Biotechnology. The pFlag-YAP, GTIIC-luc, and CTGF-luc constructs were generously provided by Professor Eek-hoon Jho (The University of Seoul, Seoul, Republic of Korea). The pGLB-MERE construct, which contains double consensus estrogen response element was provided by Dr El-Ashry (Georgetown University, Washington DC, USA). The NFκB-luc plasmid was previously described (Ju et al., 2013). The CTGF, Cyr61, and Glut3 expression vectors were generated by PCR-amplifying the coding sequence of CTGF, Cyr61 and Glut3 and cloning into the pcDNA3.1 myc-his A vector (Invitrogen, Carlsbad, CA, USA) using HindIII and XhoI restriction enzyme sites. The Cyr61-luc was generated by cloning the PCR-amplified Cyr61 promoter region into the pGL3-basic vector using KpnI and SacI restriction enzymes. The ERα promoter region luciferase constructs, A-, B-, C- and F-luc, were generated by cloning the PCR-amplified ERα promoter regions, +1 to +163 bp, −321 to −170 bp, −1977 to −1859 bp, −1117140 to −1117015 bp (relative to ERα transcription start site), into the pGL3-basic vector, respectively, using XhoI and HindIII restriction enzymes sites. The primers used for cloning were as follows: pcDNA3.1-CTGF forward, 5′-CCCAAGCTTCCATGACCGCCGCCAGTAT-3′ and reverse, 5′-CCGCTCGAGTGCCATGTCTCCGTACATCTT-3′; pcDNA3.1-Cyr61 forward, 5′-CCCAAGCTTCCATGAGCTCCCGCATC-3′ and reverse, 5′-CCGCTCGAGGTCCCTAAATTTGTGAATGTCATT-3′; pcDNA3.1-Glut3 forward, 5′-CCCAAGCTTATGGGGACACAGAAGGTCACC-3′ and reverse, 5′-CCGCTCGAGGACATTGGTGGTGGTCTCCTT-3′; pGL3-Cyr61(Cyr61-luc) forward, 5′-CGGGGTACCGCCACTGTGGGTATTAATTTG-3′ and reverse, 5′-GGAGAGCTCTGTGGCGCGCAG-3′; pGL3-A (A-luc) forward, 5′-CCGCTCGAGGGAGCTGGCGGAGGG-3′ and reverse, 5′-CCCAAGCTTCTGGAAAAAGAGCACAGCCCG-3′; pGL3-B (B-luc) forward, 5′-CCGCTCGAGAGGCAGCACATTAGAGAAAGC-3′ and reverse, 5′-CCCAAGCTTCTTTACTTGTCGTCGCTGCTG-3′; pGL3-C (C-luc) forward, 5′-CCGCTCGAGTCACACACTGAGCCACTCGCA-3′ and reverse, 5′-CCCAAGCTTCCCTGCTGGATCAAGAACGTC-3′; and pGL3-F (F-luc) forward, 5′-CCGCTCGAGCCAAAACTGAAAATGCAGGCT-3′ and reverse, 5′-CCCAAGCTTCTTGAAGAGAAGATTATCACTCAG-3′.

CTGF and Cyr61 mutant constructs and the deletion mutants for the ERα promoter region luciferase constructs A- and B-luc, were generated using the Q5 site-directed mutagenesis kit (New England BioLabs, Ipswich, MA, USA), according to the manufacturer's protocol. The mutagenic primers used for site-directed mutagenesis were as follows: CTGF mutagenesis forward, 5′-AGCCGTTACCAATGACAACGCCTC-3′ and reverse, 5′-GCAGCGATGCCCATCCCACAGGT-3′; Cyr61 mutagenesis forward, 5′-AGCCGTTACCAATGACAACCCTG-3′ and reverse, 5′-GCAGCGATACCAGTTCCACAGGTC-3′; A-luc mutagenesis forward, 5′-ACCTCGGGCTGTGCTCTTTTTC-3′ and reverse, 5′-GACGCAGCGCATGTCCCG-3′; and B-luc mutagenesis forward, 5′-TCCAGCAGCGACGACAAG-3′ and reverse, 5′-TGAGTTTCACGGCCAGGG-3′.

Cells were seeded 24 h prior to the treatment with various reagents. Antibodies were used at 1 μg/ml and equal amount of isotype control antibody was used to treat the control groups. Recombinant human CTGF and Cyr61 proteins, reconstituted in distilled water at 0.1 mg/ml, were used at a final concentration of 100 ng/ml. Tamoxifen (4-OHT), dissolved in ethanol, was used at 1 and 3 μM. β-Estradiol, dissolved in ethanol, was used at 10 nM. Verteporfin, dissolved in DMSO, was used at 3 μM for cell cycle analysis and 1 μM for all other experiments. ZD1839, Akt inhibitor VIII, U0126, and PP2 were dissolved in DMSO and used at 10 μM. Celastrol was dissolved in DMSO and used at 0.5 μM. Equal volume of the solvents was used for the control groups.

Cells were seeded 24 h prior to transfection. Lipofectamine 3000 transfection reagent (Thermo) was used according to the manufacturer's protocol. For transfection of plasmid DNAs to cells seeded on 60 mm dishes, 5.5 μg of plasmid was mixed with 11 μl of P3000 reagent and 11 μl of Lipofectamine 3000, incubated at room temperature for 15 min, and then added to the cells. For transfection of siRNAs to cells seeded on 60 mm dishes, 166 pmol of siRNA was mixed with 17 μl of Lipofectamine 3000, incubated at room temperature for 15 min, and then added to the cells.

Cells were seeded at 10⁴ cells/well in 12-well plates. For proliferation assays performed with expression vector- and siRNA-transfected cells, the cells were harvested and seeded at 24 h post transfection. For proliferation assays performed with reagent treatment, the cells were treated with the reagent for 24 h, and the culture medium was changed to complete DMEM (DMEM containing 10% fetal bovine serum, 100 U/ml penicillin, and 100 mg/ml streptomycin). Every 24 h for 4 days, cells were harvested by trypsinization, resuspended in 1 ml of medium, and counted with a hemocytometer in triplicate.

Cell viability was measured using a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, M2128, Sigma-Aldrich) Thiazolyl Blue assay. Cells were seeded in 96-well culture plates at 5×10³ cells per well. For MTT assays performed with expression vector- and siRNA-transfected cells, the cells were harvested and seeded at 24 h post-transfection. For MTT assays performed with reagent treatment, the cells were pretreated with the reagent for 24 h. 4-OHT was added to the cells after 24 h. After treatment with 4-OHT, MTT was added to the culture medium at a final concentration of 0.5 mg/ml and incubated at 37°C for 3 h. After removing the medium, formazan was dissolved in DMSO (BP231-1, Thermo Scientific), and optical density was measured at 590 nm using a microplate reader (Multiskan EX, Thermo Scientific).

For the two-dimensional colony formation assays, cells were seeded at 1000 cells/well in six-well plates. For two-dimensional colony formation assays performed with expression vector- and siRNA-transfected cells, the cells were harvested and seeded at 24 h post transfection. For two-dimensional colony formation assays performed with reagent treatment, the cells were pretreated with the reagent for 24 h. 4-OHT was added to the cells after 24 h. After 48 h of treatment with 4-OHT, the culture medium was changed to complete DMEM. After 14 days, the cells were washed with PBS and stained with 0.5% Crystal Violet for 30 min. Then, the cells were washed twice with PBS. The images were obtained with a HP Color Laser Jet CM2320nf MFP scanner (Hewlett-Packard, Palo Alto, CA, USA) and the colonies were counted using ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Cells were seeded at 10⁵ cells in 24-well plates. For dual luciferase assays performed with expression vector- and siRNA-transfected cells, the cells were harvested and seeded at 24 h post transfection. After 24 h, cells were transfected with 250 ng reporter constructs and 5 ng pRL-CMV, as an internal control, using Lipofectamine 3000 (Thermo Fisher Scientific). For dual luciferase assays performed with reagent treatment, the reagents were added to the cells at 4 h post transfection. After 24 h, dual luciferase assays were performed using a Dual-Luciferase Reporter Assay System (E1910, Promega, San Luis Obispo, CA, USA) according to the manufacturer's instructions. Luminescence was measured using a GloMax 20/20 Luminometer (Promega). For the detection of MERE-luc activity, the cells were transfected with 250 ng reporter construct and 25 ng pRL-TK, as an internal control, using Lipofectamine 3000 (Thermo Fisher Scientific) and the culture medium was changed to Phenol Red-free DMEM containing 10% charcoal-stripped FBS at 4 h post transfection. After 24 h of hormone starvation, estradiol (10 nM) was treated for an additional 24 h and dual luciferase assays were performed.

In this study, 28 paraffin-embedded tissue samples collected from patients (ER+; estrogen receptor+, HER2−; human epidermal growth factor receptor 2−) who received tamoxifen treatment were included for immunohistochemical analysis. Among the 28 patients, 11 patients (patient # 18 to 28) developed recurrence, while 17 patients (patient # 1 to 17) did not. The use of human tissues in this study was approved by the Institutional Review Board of Samsung Medical Center, Korea (IRB number 2017-11-124). Informed consent was obtained for all tissue donors and all clinical investigation have been conducted according to the principles expressed in the Declaration of Helsinki. No statistical methods were used to determine the sample size. The investigator was not blinded to the group allocation during the experiment and data analysis. All samples were used for analysis and there were no inclusion/exclusion criteria. The slides from patient # 1 were used for staining with isotype control antibodies. The slides from patient # 2 to 28 were used for staining with anti-ERα, YAP, CTGF, Cyr61 and Glut3 antibodies. For immunohistochemical analysis of the tissues, the patient tissue slides were first incubated in a dry oven at 60°C for 1 h followed by deparaffinization in xylene. The slides were rehydrated by two sequential incubations in 100%, 90%, 80%, and 70% ethanol for 5 min each. After rehydration, heat-induced antigen retrieval was performed using a sodium citrate buffer (10 mM sodium citrate, pH 6.0). The slides were then incubated with primary antibodies in PBS with 3% horse serum in a humidified chamber overnight followed by incubation for 30 min with a biotin-conjugated secondary antibody (PK-6101/PK-6102, Vector Laboratories, Burlingame, CA, USA). After washing in PBS, ABC reagent (PK-6101/PK-6102) was applied to the sections and incubated for 30 min. After washing in PBS, a color reaction was performed with 3, 30-diaminobenzidine (DAB, SK-4105, Vector Laboratories), and the slide was washed with PBS. After counter-staining with hematoxylin (HHS32, Sigma) and clearing with a graded ethanol series and xylene, the sections were mounted with Canada balsam (03984, Sigma), observed and imaged at 200× using an Olympus BX50 microscope with a DP72 Digital Camera System (Olympus, Tokyo, Japan). The DAB intensities were calculated using Fiji/ImageJ software, according to a previously reported protocol (Crowe and Yue, 2019; Schindelin et al., 2012). Briefly, color deconvolution was performed with images of the stained tissues. The mean intensity of the DAB channel was measured, and log2-transformed. The relative DAB intensities of the tissues were calculated by normalization to the intensity of isotype control antibody-stained tissues.

Cells were seeded at 5×10⁵ cells on glass coverslips and left to settle for 24 h. For immunofluorescence confocal microscopy undertaken with expression vector-transfected cells, the cells seeded on glass coverslips were transfected with the target vectors. After 24 h, cells were washed with PBS, fixed with 4% paraformaldehyde and permeabilized with 0.2% Triton X-100 for 15 min. Fixed samples were blocked with 3% skim milk in PBS for 1 h, followed by incubation with primary antibody diluted in 1% skim milk in PBS for 2 h. After being washed with PBS, the samples were treated with anti-mouse IgG Cy3 (A10521, Thermo Fisher Scientific). For DNA staining, cells were incubated with Hoechst 33342 dye (1 μg/ml) for an additional 10 min. Immunofluorescence was monitored and imaged with a Nikon C2 Si-plus confocal microscope (Nikon, Tokyo, Japan).

Cells were seeded at 8×10⁵ cells in 60 mm dishes. For cell cycle analyses done with expression vector- and siRNA-transfected cells, the cells were harvested and seeded at 24 h post transfection. 4-OHT was added to the cells after 24 h and left for an additional 48 h. For cell cycle analyses undertaken with reagent treatment, the cells were treated with the reagent for 24 h. The cells were harvested, washed with PBS, fixed in 70% ethanol overnight at −20°C, and incubated in PBS containing 40 μg/ml propidium iodide (PI) and 100 μg/ml RNase for 10 min at room temperature. Cell cycle analyses were performed using BD FACS Canto II (BD Biosciences, Franklin Lakes, NJ, USA), and data were analyzed using FlowJo software.

Cells were seeded at 2×10⁶ cells in 100 mm dishes. Cells were harvested and lysed with cytoplasmic extraction buffer [10 mM HEPES, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol (DTT), 0.5 mM PMSF] for 15 min on ice. The cells were agitated for 10 min at 4°C and NP-40 was added to a final concentration of 0.5%. Samples were centrifuged at 16,060 g for 5 min and supernatants were collected as cytosolic fractions. Nuclear pellets were washed three times with cold PBS and resuspended in nuclear extraction buffer (20 mM HEPES, pH 7.9, 400 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 1 mM PMSF), and homogenates were incubated for 15 min on ice. Nuclear extracts were agitated for 10 min at 4°C and centrifuged at 16,060 g at 4°C. Supernatants were collected as nuclear fractions.

For the detection of secreted proteins, methanol/chloroform precipitation was performed. Briefly, cells were seeded at 8×10⁵ cells in 60 mm dishes. After 24 h, the culture medium was harvested, centrifuged at 805 g for 5 min to remove cell debris. 100 μl of culture medium was used for methanol/chloroform precipitation and the precipitated protein samples were analyzed by western blotting.

Cells were washed once with PBS and lysed with lysis buffer [20 mM Tris-HCl pH 7.4, 0.1 mM EDTA, 150 mM NaCl, 1% NP-40, 0.1% Triton X-100, 0.1% SDS, 20 mM NaF, 1 mM Na₃VO₄, 1× protease inhibitor (Roche, Basel, Switzerland)]. The amount of protein in the cell lysates was determined by standardization with a BCA Protein Assay Kit (Thermo). Equal amounts of proteins (20 μg) were resolved by SDS-PAGE and transferred to nitrocellulose membranes (Whatman, Dassel, Germany). After blocking with 5% skim milk in Tris-buffered saline with 0.5% Tween 20 (TBST) at RT for 1 h, membranes were incubated with primary antibodies overnight at 4°C, followed by 2 h of incubation with HRP-conjugated secondary antibodies at room temperature. Protein bands were visualized with Dyne ECL STAR Western Blotting Detection Kit (DN-250, Dyne, Seongnam, Korea) and SuperSignal West Femto Maximum Sensitivity Substrate (34094, Thermo). The exposed films were scanned using Color LaserJet Pro MFP M281fdw scanner (Hewlett-Packard). The position of molecular mass markers (kDa) are shown on the left of each blot.

Total RNA was isolated using TRIzol reagent (Molecular Research Center). Complementary DNA was synthesized using the TOYOBO RT-PCR kit (FSQ-201, TOYOBO, Osaka, Japan). PCR was performed with the Kapa SYBR qPCR kit (KK4601, Kapa Biosystems) in a Thermal Cycler Dice (Takara, Otsu, Shiga, Japan) according to the manufacturer's protocols. The C(t) values were normalized using GAPDH. The primers used for quantitative real-time PCR (qRT-PCR) were as follows: cathepsin D forward, 5′-CCCGCATCTCCGTCAACAA-3′ and reverse, 5′-GGGTCCCTGCTCAGGTAGAA-3′; CTGF forward, 5′-CTGTGGAGTATGTACCGACGGCC-3′ and reverse, 5′-ATGGCAGGCACAGGTCTTGATGAAC-3′; Cyr61 forward, 5′-GCTGCGGCTGCTGTAAGGTC-3′ and reverse, 5′-GGCGCCGAAGTTGCATTCCA-3′; EGFR forward, 5′-ATAGTCGCCCAAAGTTCCGTGAGT-3′ and reverse, 5′-ACCACGTCGTCCATGTCTTCTTCA-3′; ERα forward, 5′-AGCACCCAGGGAAGCTACTGTT-3′ and reverse, 5′-TGAGGCACACAAACTCCTCTCC-3′; GAPDH forward, 5′-GGCAAATTCCATGGCACCGTCAAGG-3′ and reverse, 5′-GCCAGCATCGCCCCACTTGATTTTG-3′; Glut1 forward, 5′-AAGCTGACGGGTCGCCTCATG-3′ and reverse, 5′-CTCTCCCCATAGCGGTGGACC-3′; Glut2 forward, 5′-GTGGCTTGGGGACACACTTGGAAGA-3′ and reverse, 5′-GGAACCAGGCCTGAAATTAGCCCAC-3′; Glut3 forward, 5′-GGCTGTCACTGGTGGCTGCTTTATG-3′ and reverse, 5′-GAGAGTGCCAAAGGCACCCCG-3′; Glut4 forward, 5′-TTGCTCCCACTCACCTGCGG-3′ and reverse, 5′-GACAGAAGGGCAGCAGGACCAG-3′; pS2 forward, 5′-TGGCCACCATGGAGAACAAGG-3′ and reverse, 5′-GGCGTGACACCAGGAAAACCAC-3′; YAP forward, 5′-CACCCACAGCTCAGCATCTTCG-3′ and reverse, 5′-TGGGGTCCTGCCATGTTGTTG-3′.

Cells were seeded at 2×10⁶ cells in 100 mm dishes. Cells were harvested and lysed with lysis buffer. For immunoprecipitation, 500 µg of total cell lysate in 1 ml of lysis buffer was precleared with 30 µl Protein G Sepharose 4 Fast Flow 50%(w/v) slurry (17061801, GE Healthcare, Piscataway, NJ, USA) for 2 h at 4°C. After centrifugation at 1000 g, supernatants were incubated with primary antibody at 2 µg/ml, followed by incubation with 30 µl Protein G Sepharose 4 Fast Flow 50% (w/v) slurry for 3 h at 4°C. Immunoprecipitates were washed three times with lysis buffer. After removal of the supernatant, beads were resuspended in SDS sample buffer and boiled for 10 min. Samples were resolved by SDS-PAGE and analyzed by western blots.

For chromatin cross-linking, cells (3×10⁷ cells) were harvested, washed once with PBS, and incubated in 1% formaldehyde for 10 min at room temperature. The formaldehyde was quenched by adding glycine to a final concentration of 125 mM and incubating for 5 min. Then, the cells were washed twice with PBS and lysed in a lysis buffer (10 mM Tris-HCl pH 8.0, 10 mM NaCl, 0.2% NP-40 and 1× protease inhibitor), and the resulting lysates were centrifuged for 5 min at 2370 g. Precipitated nuclei were treated with 200 U Micrococcal Nuclease (New England Biolabs, Hitchin, Hertfordshire, UK) at 37°C for 15 min, resuspended in nuclei lysis buffer (50 mM Tris-HCl pH 9.0, 10 mM EDTA, 1% SDS and 1× protease inhibitor), and diluted in IP dilution buffer (20 mM Tris-HCl pH 8.0, 150 mM NaCl, 2 mM EDTA, 0.01% SDS, 1% Triton X-100 and 1× protease inhibitor). The lysates were then disrupted with a sonicator (Dr Hielscher, GmbH, Germany) on ice to obtain DNA fragments of 200–1000 bp in length (2 pulses, 10 s on/50 s off at 70% amplitude). Digested chromatin was immunoprecipitated with CTGF and Cyr61 antibody overnight at 4°C with rotation. The immunocomplexes were collected with Protein G Sepharose 4 Fast Flow 50% (w/v) slurry for 4 h at 4°C with rotation and washed sequentially as follows: once with washing buffer 1 (20 mM Tris-HCl pH 8.0, 150 mM NaCl, 2 mM EDTA, 0.1% SDS and 1% Triton X-100), once with washing buffer 2 (10 mM Tris-HCl pH 8.0, 0.25 M LiCl, 1 mM EDTA, 1% NP-40 and 1% deoxycholate), once with 0.1× TE buffer (1 mM Tris-HCl pH 7.6 and 0.1 mM EDTA). Then, the chromatin was eluted by incubating the beads twice in elution buffer (0.1 M NaHCO₃, 1% SDS) at 30°C for 15 min each. The eluted chromatin was reverse cross-linked overnight in 250 mM NaCl at 65°C with RNase (final concentration of 166.6 µg/ml), and DNA extraction was performed using phenol/chloroform and ethanol precipitation. The precipitated DNA was assessed by qRT-PCR with the primers as follows: ERα promoter region A forward, 5′-CGTCCTGGGACTGCACTTGCTC-3′ and reverse, 5′-AGAGCACAGCCCGAGGTTAGAGG-3′; ERα promoter region B forward, 5′-GCCCCTGGATCCGTCTTTCG-3′ and reverse, 5′-TGTCGTCGCTGCTGGATAGAGG-3′; ERα promoter region C forward, 5′-CGCACATGCGAGCACATTCC-3′ and reverse, 5′-CCTGCTGGATCAAGAACGTCTTTCC-3′; and ERα promoter region F forward, 5′-CCAAAACTGAAAATGCAGGCT-3′ and reverse, 5′-CTGAGTGATAATCTTCTCTTCAAG-3′.

Glucose uptake level was assessed using a Glucose Uptake Cell-Based Assay kit (Cayman Chemical, Ann Arbor, MI, USA). Cells were seeded at 1×10⁴ cells per well in 96-well Nunc^™ F96 MicroWell^™ Black Polystyrene Plates (Thermo Fisher Scientific). For glucose uptake assays performed with expression vector- and siRNA-transfected cells, the cells were harvested and seeded at 24 h post transfection. After 24 h, the level of glucose uptake was measured according to the manufacturer's instruction. Relative fluorescence units were examined at 485–535 nm using a microplate reader (Varioskan Flash, Thermo Fisher Scientific).

Cellular ATP levels were measured using a CellTiter-Glo^® Luminescent Cell Viability Assay kit (Promega). Cells were seeded at a density of 5×10³ cells per well in 96-well plates. For ATP assays performed with expression vector- and siRNA-transfected cells, the cells were harvested and seeded at 24 h post-transfection. After 24 h, the cellular ATP levels were examined according to the manufacturer's instruction. The luminescence was measured using a microplate reader (Varioskan Flash, Thermo Scientific).

Lactate production levels were determined using a Lactate Assay kit (Biovision, Milpitas, CA, USA). Cells were seeded at 8×10⁵ cells in 60 mm cell culture dishes. For lactate assays performed with expression vector/siRNA transfected cells, the cells were harvested and seeded at 24 h post-transfection. After 24 h, the culture medium was changed with DMEM without FBS, followed by 8 h incubation. Culture medium was collected from each sample and centrifuged at 805 g for 5 min to obtain the supernatant without cell debris. Lactate assays were performed according to the manufacturer's protocol, and the colorimetric density was assessed at 570 nm using a microplate reader (Multiskan EX, Thermo Scientific).

Cell migration assay was performed using 8 μm pore size transwell chambers (3422, Corning, Manassas, VA). The lower chamber was filled with complete DMEM. Cells (5×10⁴ cells) were suspended in DMEM without FBS and planted into the upper chamber. For Transwell migration assays performed with expression vector- and siRNA-transfected cells, the cells were harvested and seeded at 24 h post transfection. After 16 h, the cells stained by Crystal Violet dye on the bottom surface of the polycarbonate membranes were imaged with a Nikon SMZI8 microscope (Nikon). The intensity of the migrated cells was measured with ImageJ software.

Survival curves of ER+, tamoxifen-treated breast cancer patients were acquired through the Kaplan–Meier Plotter database (http://kmplot.com/analysis/) (Nagy et al., 2018). The auto select best cutoff option was used to split the patients. CTGF and Cyr61 expression comparison in clinical samples and co-expression analysis of YAP, CTGF and Cyr61 with ERα were performed using cBioportal (http://cbioportal.org/), based on expression data (Expression log intensity levels from Illumina Human v3 microarray) from the METABRIC dataset (Cerami et al., 2012; Gao et al., 2013; Pereira et al., 2016). Also, survival analysis of ER+, recurrent breast cancer patients according to CTGF expression was acquired using the METABRIC dataset. Gene set enrichment analysis (GSEA) was performed with the expression data from the METABRIC dataset using annotations in the Molecular Signatures Database (MSigDB) H collection (Mootha et al., 2003; Subramanian et al., 2005). For each comparison, enriched genes following YAP, CTGF and Cyr61 amplification were ranked by the log2 fold-change (mean in amplified group/mean in unaltered group) for input into GSEAPreranked (1000 permutations were performed). Default settings were used for all other GSEA parameters.

All results were confirmed in at least three independent experiments. Data are presented as mean±s.d. All statistical analyses were conducted using GraphPad Prism version 6.01 for Windows (GraphPad Software, La Jolla California, USA). The F test and Brown–Forsythe test were conducted to compare the variances. Statistical comparisons of the data were performed using either two-tailed unpaired t-test, one-way ANOVA (post hoc Dunnett's test or Tukey's test), or two-way ANOVA (post hoc Dunnett's test or Tukey's test). A value of P<0.05 was considered statistically significant (*P<0.05; **P<0.005; ***P<0.0005).

The peer review historyis available online at https://journals.biologists.com/jcs/article-lookup/doi/10.1242/jcs.256503