ERK-Mediated Phosphorylation of EZH2 Regulates HER2 Expression on Long-term Genistein-induced Acquired Endocrine Resistance in Estrogen-Receptor-Positive Breast Cancer Cells

Genistein, a soy isoavones, is the most important phytoestrogens in typical oriental diet. Many studies have shown that genistein at lower concentrations promotes breast cancer cells growth through the estrogen receptor pathway. However, recent research has found that long-term consumption of low doses of genistein results in hormone-independent growth phenotypes of MCF-7 tumors, with increased expression of HER2. Overexpression of HER2 has been causally associated with endocrine therapy resistance in human breast cancer. The mechanism by which prolonged exposure to genistein leads to increased HER2 expression is unclear. Whether genistein-induced HER2 expression is the cause of endocrine resistance remains to be determined. We


Abstract
Background Genistein, a soy iso avones, is the most important phytoestrogens in typical oriental diet. Many studies have shown that genistein at lower concentrations promotes breast cancer cells growth through the estrogen receptor pathway. However, recent research has found that long-term consumption of low doses of genistein results in hormone-independent growth phenotypes of MCF-7 tumors, with increased expression of HER2. Overexpression of HER2 has been causally associated with endocrine therapy resistance in human breast cancer. The mechanism by which prolonged exposure to genistein leads to increased HER2 expression is unclear. Whether genistein-induced HER2 expression is the cause of endocrine resistance remains to be determined.

Methods
We selected the MCF-7 and T47D breast cancer cells model with higher ERα and lower HER2. It was investigated whether prolonged exposure to genistein induced TAM-sensitive breast cancer cells to TAMrefractory cells by increasing HER2 expression. Furthermore, it was explored whether HER2 expression and endocrine resistance were associated with EZH2.

Results
We found that genistein had estrogen-like effect and inhibited HER2 expression during short-term exposure. However, long-term exposure to genistein induced acquire endocrine resistance, because of increased expression of HER2. During long-term exposure to genistein, the continuous activation of ERK1/2 phosphorylated EZH2 at Ser21, resulting in a decrease of lysine 27 trimethylation. As H3K27me3 level decreased, the expression of IL-6 and IL-8 increased, and HER2 level gradually increased, forming a feedback loop of ERK1/2 / EZH2/ IL-6 and IL-8 / HER2.

Conclusions
These ndings indicated that high HER2 expression caused by EZH2 phosphorylation was an important mechanism of endocrine resistance. The study also provided a new insight for genistein-induced acquired endocrine resistance. For breast cancer patients, long-term use of soy supplements has potential health risk. Especially, monitoring dietary exposure to genistein is advisable when treated with tamoxifen.

Background
Approximately 70% of breast cancers express estrogen receptor α (ERα) and merit the use of endocrine therapies, such as the estrogen receptor (ER) antagonist tamoxifen (TAM) [1]. However, ER-positive breast cancer frequently acquires resistance to TAM after long-term treatment, which is a serious therapeutic problem [2,3].
Multiple mechanisms are responsible for the development of endocrine resistance. Compelling evidence suggests that the human epidermal growth factor receptor (HER) family plays a critical role in mediating endocrine therapy resistance [4][5][6][7][8]. The ampli cation of the HER2 locus can overcome the growth inhibitory effects imposed by TAM in ER-positive breast cancers [9]. The mechanism by which HER2 overexpression mediates TAM resistance is the crosstalk between ERα and HER2 initiates intracellular kinase cascades, such as MAPK signaling, promoting growth and progression in breast cancer cells, negating the inhibitory effects of TAM [10].
Phytoestrogens are compounds derived from plants that have estrogenic properties and are abundant in the human diet, particularly Soy. Genistein (GE), a major phytoestrogen in soybeans found in processed foods, induces genomic ER signaling in the developing breast cancer [11,12]. Although no studies in humans have been conducted, some studies have shown that genistein negates the inhibitory effects of TAM in breast cancer cell lines and animal models [13][14][15]. Previous study has demonstrated that longterm exposure to genistein leads to estrogen-independence growth in ER-positive breast tumors and results in increased expression of HER2 [16]. The exact cause of abnormal growth factor signaling caused by genistein is unknown. Genistein may not have estrogenic properties when acting on ERpositive breast cancer for a long time, may cause the development of acquired endocrine resistance in breast cancer, in which HER2 may play an important role.
Enhancer of zeste homolog 2 (EZH2) is frequently overexpressed in human bladder, breast, colon and prostate cancers [17]. Studies have shown that EZH2 speci cally catalyzes trimethylation of histone H3 lysine 27 (H3K27me3), resulting in transcriptional repression and chromatin compaction [18][19][20]. Several studies have reported that epigenetic alterations are associated with endocrine resistance in breast cancer [21,22]. Although roles of EZH2 in driving cancer proliferation and invasion are extensively characterized [23,24], few studies investigate the association of EZH2 with acquired endocrine resistance. Data from a study suggest that low H3K27me3 levels were signi cantly associated with aromatase inhibitor resistance in breast cancer patients [25]. Notably, genistein induced PI3K/AKT nongenomic ER signaling to phosphorylate and repress the histone methyltransferase EZH2. As a result, this signaling reduces levels of H3K27me3 in chromatin [26]. However, it is not clear whether inhibition of EZH2 histone methyltransferase activity is associated with increased HER2 expression and endocrine therapy resistance. Therefore, we selected the MCF-7 and T47D breast cancer cells model with higher ERα and lower HER2.
We have investigated whether prolonged exposure to genistein can induce TAM-sensitive breast cancer cells to TAM-refractory cells by increasing HER2 expression. During long-term exposure to Genistein, the phosphorylation of EZH2 and the levels of H3K27me3 were observed, and their effects on TAM sensitivity were evaluated. Furthermore, it was explored whether HER2 was physically associated with EZH2. The data reported here reveal genistein regulates TAM resistance through EZH2/ H3K27me3/HER2 axis in breast cancer.

RNA interference
The 21-nucleotide duplex siRNAs for EZH2, HER2 and one negative control siRNA were synthesized by Santa Cruz (Dallas, Texas, USA). Transfection was carried out using Lipofectamine 3000 (Invitrogen, Carlsbad, CA, USA) following the manufacturer's instructions. Transfection e ciency was evaluated in every experiment by RT-qPCR 24 hours later to ensure that cells were transfected.
Quantitative Real-Time RT-PCR (qPCR) Total RNA from the two cell lines was isolated using the TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and reverse transcribed to cDNA using the ExScript RT reagent (Abm, Zhenjiang, Jiangsu, China). Realtime RT-PCR was performed using StepOnePlus Real Time PCR System (Roche Group, Basel, Switzerland) with speci c primers for EZH2, HER2, IL-6, IL-8 and GAPDH expression was used to normalize for variance. Real-time uorescence monitoring of the PCR products was performed with SYBR Green I uorescent dye (Abm, Zhenjiang, Jiangsu, China). The expression levels of speci c genes are reported as ratios of expression of GAPDH in the same master reaction.

Western Blotting Analysis
Cells were lysed with whole cell lysis buffer, frozen at −80°C and thawed three times to rupture the cell membranes. Samples of the lysates were incubated for 30 min on ice to lyse the nuclei, and then centrifuged at 12,000 rpm for 20 min at 4°C. Equal amounts of protein (20 μg) were subjected to western blot analysis. Western blotting was performed by standard procedures.

Proliferation Assay
Breast cancer cells were seeded in 96-well plates (4000 cells/well). At the end of the cell treatment, cell proliferation was assayed using Cell Counting Kit-8(CCK-8) (Apexbio, Houston, TX, USA) as per the manufacturer's protocol.

Statistical Analysis
Experimental values are presented as mean ± standard deviation (SD). At least three independent trials were performed for each experiment. Statistical analytical Methods and numbers of data points analyzed for each experiment are described in gure legends. Statistically signi cant analysis by two-way ANOVA with Sidak's multiple comparison test (α = 0.05) was performed in experiments graphed in Figure 1a-

Results
Elevated HER2 promotes to genistein -induced acquired endocrine therapy resistance It has been shown that long-term consumption of low doses genistein promotes MCF-7 tumor growth and results in non-hormone dependent tumors with increased expression of HER2 [16]. Long-term TAM treatment also induces overexpression of HER2 as a resistance mechanism to hormonal therapy [9,27]. The different plasma levels of genistein were consistently observed in Asian and Western populations.
For example, the plasma levels were about 0.02 μmol/L in English and 0.26 μmol/L in Japanese and Korean [28,29]. In this study, 1μM was selected as the study dose of genistein. Using ER-positive breast cancer cells, we investigated whether genistein induces resistance to endocrine therapy. After long-term (3 weeks) genistein exposure followed by a 24h withdrawal, cells were treated with different concentrations of TAM for 48h. Compared with the parallel cells, the MCF-7 cells (Fig. 1a) and T47D cells (Fig. 1b) by prolong genistein stimulus and withdrawal showed resistance to TAM, while T47D cells treated with E 2 remained normal TAM sensitivity (Fig. 1c). Western blot analysis showed that HER2 levels decreased after genistein exposure for 24h ( Fig. S1a and b), which was similar to E 2 (Fig. S1c). HER2 levels were signi cantly increased after 3 weeks of exposure to genistein in MCF-7 cells (Fig. 1d). During long-term genistein exposure, T47D cells were subjected to western blot analysis at 2 days, 1 week, 2 weeks, 3 weeks, respectively. HER2 expression decreased initially, then increased gradually. The levels were higher than that in the parental cells at 3 weeks ( Fig. 1e and g). The HER2 levels increased substantially after TAM treatment for 24h (Fig. S1d) and further increased during prolonged TAM treatment ( Fig. 1f and g).
We investigated whether endocrine resistance is associated with up-regulation of the HER2. Under the action of HER2 siRNA (Fig. 1h), genistein-induced resistant cells recovered their sensitivity (Fig. 1i). All these results indicate that HER2 might play an important role in genistein-induced acquired endocrine resistance. ERK1/2 signal activation caused by long-term exposure to genistein is associated with HER2 elevation The results of preclinical studies have indicated that acquired TAM resistance is associated with increased activities of ERK [30]. Genistein induces phosphorylation of ERK1/2 within a certain amount of time, and ERK activity starts at 3h and peaks at 24 h [31]. In the present study, the 24h exposure of the T47D cells to genistein induced phosphorylation of ERK1/2 (Fig. S2a), as TAM did (Fig. S2b). During long-term genistein exposure, T47D cells were subjected to western blot analysis at 2 days, 1 week, 2 weeks, 3 weeks, respectively. Phosphorylation of ERK1/2 increased after 1 week (Fig. 2a). Phosphorylation of ERK1/2 increased at 3 weeks in MCF-7 cells (Fig. 2b). The results of TAM treatment were the same as those of long-term genistein exposure (Fig. 2c). We observed a similar trend in the changes of HER2 protein levels and ERK1/2 phosphorylation levels after long-term genistein exposure.
EZH2 phosphorylation at Ser21 increases and H3K27me3 levels reduce during long-term genistein exposure EZH2 is an estradiol-regulated gene and its promoter contains functional estrogen-response elements [33]. After TAM treatment for 24h, the mRNA and protein levels of EZH2 were down-regulated, as well as H3K27me3 levels (Fig.S3a). During long-term TAM treatment, EZH2 and H3K27me3 levels were also decreased (Fig. 3a). TAM inhibition of EZH2 expression may be caused by its antagonism against ER.
As a phytoestrogen, genistein promoted the expression of ERα target gene after 24h exposure (Fig.S3b). It also increased the levels of EZH2 and H3K27me3 in MCF-7 cells and T47D cells (Fig.S3c). During longterm genistein exposure, EZH2 protein levels increased at 2 days and 1 weeks, and fell back to the same level as the control at 2 weeks and 3 weeks. Remarkably, the levels of H3K27me3 increased rst, and decreased at 2 weeks and 3 weeks (Fig. 3b). The expression of EZH2 decreased slightly and the levels of H3K27me3 also decreased at 3 weeks in MCF-7 cells (Fig. 3c). Collectively it seems highly likely that decreased trimethylation at H3K27 by long-term genistein exposure was not a result of decreased total levels of EZH2, because EZH2 levels remained unchanged or decreased slightly in response to long-term genistein.
Phosphorylation at Ser21 altered the a nity of EZH2 for its substrate, histone H3, which reduces EZH2 methyltransfease activity [34]. We hypothesized that genistein regulates trimethylation at H3K27 by in uencing the phosphorylation of EZH2. To validate our hypothesis, we st examined and compared the phosphorylation levels of EZH2 at Ser21 between genistein and TAM. After long-term TAM treatment, the phosphorylation of EZH2 at Ser21 was gradually increased (Fig. 3c). During long-term genistein exposure, the phosphorylation of EZH2 at Ser21 decreased rst, and increased at 2 weeks and 3 weeks in T47D cells. Likewise, the phosphorylation of EZH2 also increased at 3 weeks in MCF-7 cells (Fig. 3b). These results suggest that phosphorylation of EZH2 at Ser21 may be responsible for the decrease of H3K27me3 during long-term genistein exposure. Unlike genistein, TAM had a stronger inhibitory effect on H3K27 trimethylation, due to enhancing phosphorylation of EZH2 and suppressing expression of EZH2 by antagonizing ER.
ERK decreases the H3K27 trimethylation through phosphorylation of EZH2, contributing to HER2 expression and endocrine resistance We observed a similar trend in the changes of EZH2 phosphorylation and ERK1/2 phosphorylation after long-term genistein exposure. Accordingly, we evaluated whether phosphorylated ERK1/2 affects EZH2 phosphorylation, H3K27me3 levels. When MEK inhibitor U0126 treated T47D cells for 24h after 3 weeks with genistein, the results showed that inhibition of ERK1/2 phosphorylation downregulated genisteininduced EZH2 Ser21 phosphorylation, and up-regulated genistein-decreased H3K27me3 (Fig. 4a). When U0126 was combined with TAM, the results showed that inhibition of ERK1/2 phosphorylation also downregulated TAM-induced EZH2 phosphorylation at Ser21, up-regulated TAM-decreased H3K27me3 (Fig. 4b). Phosphorylation of EZH2 at Ser21 results in decreased EZH2 activity and H3K27me3 levels [34]. Other data suggest that low H3K27me3 levels are signi cantly associated with resistance to aromatase inhibitors in breast cancer patients [25]. To investigate whether H3K27me3 level was related to the expression of HER2 and the sensitivity to TAM, after exposure to 1 μM genistein for 3 weeks, T47D cells were treated with EZH2 siRNA or GSK-J4, an H3K27 demethylase inhibitor. The results showed that GSK-J4 elevated the levels of H3K27me3, inhibited the expression of HER2 (Fig. 4c), and restored TAM sensitivity in T47D cells long exposed to genistein (Fig. 4d). However, EZH2 gene knockout lowered the levels of H3K27me3, promoted the expression of HER2 (Fig. 4e), and diminished the e cacy of TAM (Fig.4f). Increased phospho-EZH2 was not a result of increased total levels of EZH2, because phosphorylation of EZH2 remained unchanged in response to knocking down EZH2 expression (Fig. 4e). The above results suggested that activated ERK1/2 signaling phosphorylates EZH2 and reduces the trimethylation of H3K27. It has been reported that phosphorylation of EZH2 at Ser21 reduces the methylation for H3K27 [34]. Our ndings revealed that reduction of H3K27me3 was involved in increasing expression of HER2 and reducing the sensitivity of the cells to TAM.
Increased IL-6 and IL-8 induced by genistein is involved in HER2 elevation and endocrine resistance A decrease of H3K27me3 by EZH2 de ciency resulted in derepression of silenced IL-6 and IL-8 in MCF-7 or another ER-positive T47D cells [35]. IL-6 is involved in TAM resistance through the downstream activation of multiple signaling pathways [36][37][38]. Previous study suggests a close link between IL-8 and traditional chemotherapy drug resistance [39]. IL-6 and IL-8 may play critical roles in acquired endocrine resistance caused by genistein. Therefore, we analyzed expression of IL-6 and IL-8 in cells stimulated by prolong genistein. Both IL-6 mRNA and IL-8 mRNA signi cantly increased after 3 weeks of genistein exposure (Fig. 5a). After 24 h treatment, like E 2 , genistein inhibited IL-6 mRNA and IL-8 mRNA in MCF-7 cells (Fig. S4a), while only inhibited the expression of IL-6 mRNA in T47D cells (Fig. S4b). TAM, like prolong genistein, increased the expression of IL-6 and IL-8 mRNA (Fig. S4c and d). Our observations con rmed that knockdown of EZH2 gene elevated expression of IL-6 and IL-8 mRNA (Fig. 5b), and GSK-J4 as an H3K27 demethylase inhibitor decreased expression of IL-6 and IL-8 mRNA (Fig. 5c). We analyzed the involvement of IL-6 and IL-8 in HER2 expression and TAM resistance in breast cancer. Here we demonstrated that exogenous IL-6 or IL-8 (24h treatment with recombinant IL-6 or IL-8) elevated expression of HER2 in T47D cells (Fig. 5d). Results of further investigation indicated that IL-6 or IL-8 diminished the e cacy of to TAM (Fig. 5e and f). Together, these results indicate that the elevation of HER2 promoted by IL-6 or IL-8 may be the cause of TAM resistance induced by genistein and EZH2 inactivation and H3K27me3 level reduction were involved in the enhanced expression of IL-6 and IL-8.

Discussion
Despite endocrine therapy has dramatically improved survival in ER-positive breast cancer patients, resistance to treatment is common, resulting in metastatic relapse that cannot be cured [40,41]. Much evidence has demonstrated that increased growth factor signaling, in particular the HER2 pathway contributes to endocrine therapy resistance [42]. HER2 gene expression can be down-regulated by E 2 in the MCF-7 breast cancer cell line [43] through direct transcriptional repression of the HER2 gene [44]. In the current study, genistein, like E 2 , reduced the levels of HER2 and TAM increased the levels of HER2 after short-term treatment. Therefore, we believe that genistein may directly inhibit the transcription of HER2 gene through ER. After long-term exposure, genistein increased the expression of HER2, which is consistent with previous report in an athymic mice xenograft model [16]. Here we demonstrate that longterm exposure to genistein causes the development of acquired endocrine resistance in ER-positive breast cancer cells, in which HER2 may play an important role.
Many observations have con rmed the link between increased HER2 activity and phosphorylation of the downstream MAPK/ERK pathway [45]. TAM was shown to activate ERK in ER-positive MCF-7 and T47D cells but not in ER-negative MDA-MB-231 cells [46]. Unrestrained MAPK signaling phosphorylates Ser-118 in the ER [47,48], alters the ER association with corepressors of transcription [49], leading to loss of the inhibitory effect of TAM, which is a viable mechanism for MAPK to cause TAM resistance. Our previous study has shown that genistein induces activation of ERK1/2 , starting at 3h and peaking at 24 h [31]. In the present study, phosphorylation of ERK1/2 was high again from week 1 during long-term genistein exposure. Our study found that HER2 gene knockout resulted in decreased phosphorylation of ERK1/2. When inhibiting the phosphorylation of ERK1/2, the levels of HER2 protein were down-regulated. These data suggest that HER2 and ERK1/2 interact in a positive feedback regulation mode, which may be the reason for the unconstrained activation of ERK1/2. Meanwhile, HER2 expression is at a high level. Nonetheless, the data presented imply that MAPK may play a causal role in genistein-induced TAM resistance in ER-positive breast tumor cells.
Our further results suggest that the activated ERK1/2 signaling phosphorylated of EZH2 at Ser21 and down-regulated H3K27me3 level. EZH2 as a part of the polycomb repressive complex 2 (PRC2), possesses histone H3K27-speci c methyltransferase activity [50]. Histone H3K27 methylation by EZH2 is an important mechanism of gene silencing [51]. Previous publications report that EZH2 is an estrogenregulated gene [33,52,53]. Here we found that genistein, as a phytoestrogen, up-regulated the levers of EZH2 and H3K27me3 in ER-positive breast cancer cells. However, after long-term genistein exposure, EZH2 protein levels have not changed signi cantly and the phosphorylation of EZH2 at Ser21signi cantly increased. Phosphorylation of EZH2 at Ser21 dissociates EZH2 from chromatin resulting in decreased EZH2 activity and H3K27me3 levels [34], which is consistent with our results. We speculate that due to increased phosphorylation of EZH2, EZH2 protein levels did not increase as they did with short-term genistein exposure. Recent study has shown that EZH2 was evidently less enriched in TAMR cells [54]. Study has found that that low H3K27me3 level were signi cantly associated with aromatase inhibitor resistance [25]. In the current study, we demonstrated that decrease of H3K27me3 caused by EZH2 de ciency was associated with increased HER2 expression and genistein-induced TAM resistance. On the contrary, increase of H3K27me3 inhibited the expression of HER2 and restored TAM sensitivity in T47D cells long exposed to genistein. E 2 is effective in suppressing TNFα induction of the IL-6 and IL-8 genes in MCF-7 cells [55]. Our ndings indicated that genistein was equally effective in suppressing the expression of IL-6 and IL-8 mRNA, possibly due to its estrogen-like activity. Long term anti-hormone therapy alters the function of ERα to create an in ammatory microenvironment in breast cancer [56]. However, we found that genistein upregulated the expression of IL-6 and IL-8 after long-term exposure. Both IL-6 and IL-8 increased the expression of HER2 and diminished the e cacy of to TAM. These data suggest that the up-regulation of in ammatory cytokine increased the levels of HER2, which may play a role in acquired endocrine resistance induced by long-term exposure of genistein. Unfortunately, the mechanism of IL-6 or IL-8 regulating HER2 expression has not been reported yet and further investigations are required to determine it.
In addition, we explored how in ammatory cytokines is elevated and which factors regulate them. Our data indicated that expression of IL-6 and IL-8 were signi cantly increased upon EZH2 silencing and signi cantly decreased upon H3K27 methylation enhancement. Our study disclosed EZH2 levels and activity were correlated negatively with in ammatory cytokines and that H3K27 methylation may account for the epigenetic repression, which is consistent with previous report [35]. For the rst time, we have found that genistein induced TAM resistance via the EZH2/ H3K27me3/in ammatory cytokine /HER2 axis.
In summary, our work reveals a critical epigenetic program that determines HER2 expression as well as cell fate in response to TAM treatment (Fig. 6). In sensitive cells upon short-term exposure of genistein, the expression of HER2 is suppressed because of its estrogen-like function. However, long-term genistein exposure results in sustained activation of ERK1/2 and therefore induces a inactivation of epigenetic enzymes EZH2 by phosphorylation and decrease of H3K27me3 level, which causes high expression of in ammatory cytokines and HER2. Maintenance of HER2 protein at high level phosphorylates the downstream ERK1/2 and reprograms ERα-dependent transcriptional machinery, which renders acquired resistance phenotypes in breast cancer cells. TAM plays a similar role in this pathway. The difference is that TAM directly increases HER2 expression and inhibits EZH2 expression by antagonizing ER, even in the short term.

Conclusions
Taken together, our ndings provide a compelling foundation for elucidating the endocrine resistance mechanism induced by genistein in breast cancer. The timing of exposure to phytoestrogens may be a key component in determining its effects. This study provides a comprehensive understanding of the health risks of dietary exposure to phytoestrogens in breast cancer patients. Breast cancer patients, especially those on TAM should be cautioned against the long-term use of soy supplements and puri ed products in order to achieve more lasting results.

Consent for publication
All authors agree to the publication of the article.

Availability of data and materials
All data generated or analyzed during this study are included in this published article.

Competing interests
The authors declare that they have no competing interests. Funding