Intrinsic and Extrinsic Factors Governing the Transcriptional Regulation of ESR1

Abstract

Transcriptional regulation of ESR1, the gene that encodes for estrogen receptor α (ER), is critical for regulating the downstream effects of the estrogen signaling pathway in breast cancer such as cell growth. ESR1 is a large and complex gene that is regulated by multiple regulatory elements, which has complicated our understanding of how ESR1 expression is controlled in the context of breast cancer. Early studies characterized the genomic structure of ESR1 with subsequent studies focused on identifying intrinsic (chromatin environment, transcription factors, signaling pathways) and extrinsic (tumor microenvironment, secreted factors) mechanisms that impact ESR1 gene expression. Currently, the introduction of genomic sequencing platforms and additional genome-wide technologies has provided additional insight on how chromatin structures may coordinate with these intrinsic and extrinsic mechanisms to regulate ESR1 expression. Understanding these interactions will allow us to have a clearer understanding of how ESR1 expression is regulated and eventually provide clues on how to influence its regulation with potential treatments. In this review, we highlight key studies concerning the genomic structure of ESR1, mechanisms that affect the dynamics of ESR1 expression, and considerations towards affecting ESR1 expression and hormone responsiveness in breast cancer.

Introduction

An important predictive and prognostic biomarker found in the majority of breast cancers is estrogen receptor α (ER) [1–3]. ER is a transcription factor of the nuclear receptor superfamily that mediates the effects of estrogen ligands and promotes growth and cell survival [4]. Approximately 60–70% of breast cancers are classified as ER-positive, and the proportion of ER-positive cases is forecasted to increase significantly within the next 20 years, while the number of ER-negative cases is expected to decline for unclear reasons [5]. This forecast highlights the need to understand what controls ER status during disease progression. ER-positive breast cancers are defined histologically as 1% of tumor nuclei with detectable ER protein [6]. Examination of ER expression through mammary tumor development indicate that ER heterogeneity varies throughout cancer progression [7]. However, despite observing a small subpopulation of cells expressing ER, the frequency of ESR1 mRNA detection in recent single-cell sequencing studies on normal mammary gland and breast cancer cells suggest that ESR1 expression is biologically important in both normal mammary gland tissues and breast cancer [6, 8–16].

ER is the primary target of endocrine therapies due to its significant role in ER-positive breast cancer proliferation and survival, but the development of resistance to these therapies inevitably occurs despite their initial effectiveness [17, 18]. A higher level of ESR1 expression in ER-positive breast cancer patients correlates with a significantly longer relapse-free survival time (Fig. 1) [19]. However, clinical studies have shown that metastatic tumors, the major cause of breast cancer patient mortality, generally contain relatively lower levels of ER compared to that of the primary tumor [20]. Furthermore, Dieci et al. compared survival rates between breast cancer patients with tumors that maintained high ER expression or transitioned to low ER expression and observed decreased overall survival rates in the latter group [21]. Because there is a direct correlation between cellular levels of ER and ER transcriptional activity/hormone sensitivity and poorer prognosis in ER-negative breast tumors, decreases in ER expression may be a major mechanism that reduces hormone-dependent control in breast cancers [22–25].

Fig. 1.

Kaplan–Meier plot showing the probability of relapse-free survival over time based on ESR1 expression levels (Affymetrix gene ID ESR1 205225 at). This plot was generated using the KM Plotter tool [9]. The patient cohort was restricted to ER+ breast cancer samples where ER status was derived from gene expression data (n = 3082). Samples were split into high and low expression groups relative to the median expression of ESR1 in the patient cohort [19]

Transcriptional regulation of ESR1, the gene that encodes for ER, is an essential component of maintaining hormone sensitivity in breast cancer cells. While other levels of regulation are known to regulate ER expression (posttranscriptional and posttranslational mechanisms), in vitro and clinical studies involving ER-negative cell lines and patient data, respectively, suggest transcriptional control of ESR1 is the main driver in the dysregulation of ER expression [26–35]. This finding is supported by a modeling study of estrogen regulation of ER expression, which indicates that ESR1 is the primary controller of cellular ER levels and this control of the ESR1 gene is facilitated by the interaction between the phosphorylated ER (a transcriptionally active form of ER) and the ESR1 promoter. [36]. It should be noted that approximately 20–25% of ER-negative tumors still express detectable levels of ESR1 mRNA, which shows that, in this subset of tumors, ER may be regulated by posttranscriptional and/or posttranslational mechanisms [37–39]. To study transcriptional regulation, multiple assays have been employed to assess the activity of regulatory elements associated with a gene, the occupancy of essential transcription initiators such as transcription factors, RNA polymerase II (Pol II) and Mediator, and levels of nascent or immature mRNA. In addition to these intrinsic mechanisms, extrinsic factors are also known to regulate ESR1 expression. Therefore, this review will provide an overview of the ESR1 gene and focus on intrinsic and extrinsic transcriptional mechanisms regulating the ESR1 gene.

Characterization of the ESR1 Gene

The ESR1 gene has been characterized and studied extensively over the past three decades; however, questions remain on how the ESR1 gene is regulated due to its inherent complexity. The human ESR1 gene was first cloned from the ER-positive MCF-7 breast cancer cell line by Walter et al., and their findings were further expanded to cover the entire ESR1 mRNA by Green et al. [40, 41]. Subsequent studies on the cloned ESR1 gene focused on identifying regulatory elements associated with ESR1 [42–48]. Kos et al. provides an overview of these studies and a nomenclature for the ESR1 gene that consolidated the reported regulatory regions (Fig. 2a, b) [49]. It should be noted that some of these regulatory regions are expressed in a tissue-specific context, specifically the T1 and T2 promoters in human epididymis and testes [50]. The ESR1 gene spans approximately 300 kilobases (kb) on chromosome 6 (chr6) and consists of eight regulatory elements. Based on the genomic coordinates described by Kos et al., the transcription start site (TSS) of ESR1 begins on chr6:152128814–152129050 in the hg19 assembly published in the UCSC Genome Browser [51]. This region also contains the common acceptor splice site for all ESR1 mRNA variants (+ 164) to form a single ESR1 mRNA transcript. The coordinates of the TSS and regulatory elements of ESR1 are shown in Fig. 2c.

Fig. 2.

Diagram of the ESR1 gene (chromosome 6). aESR1 is approximately 300 kilobases and consists of several reported regulatory elements with some promoters utilized in a tissue-specific manner and eight exons. The nomenclature and numbering used in this diagram are based on NM_000125 and described by Kos et al. [49]. Diagram is not drawn to scale. bESR1 transcript variants differ by their 5′ UTR but are spliced to the same acceptor splice site shown in a. Shown is a screenshot of ESR1 variants on the UCSC Genome Browser (human assembly hg19) with ESR1 regulatory regions highlighted in light blue (chr6:151,968,286-152,147,486). c Genomic coordinates listed of key sites and regulatory elements of ESR1. Genomic coordinates start from the transcription start site (+ 1) and are based on the ESR1 cDNA sequence in NM_000125 and human genome assembly hg19 [51]

Given that there are multiple regulatory elements of ESR1, it is important to determine how each element is utilized. Using reporter gene assays and quantifying ESR1 mRNA expressed from seven promoter regions, MCF-7, T47D, and ZR-75-1 cells, which are major models of ER-positive breast cancers, primarily utilize the A promoter (also known as the proximal promoter) with less activity detected in the B and C promoters [45, 52–54]. Similarly, in ER-positive primary patient samples, ESR1 transcripts overlapping the A promoter were detected most frequently followed by those overlapping the C promoter [54]. In addition to these promoters, a distal enhancer site (ENH1) is approximately 150 kb from the TSS and has also been reported to regulate ESR1. Given the significance of ER to mammary gland development, Eeckhoute et al. investigated the relationship between ER and GATA3, a transcription factor that is an essential regulator of mammary gland differentiation and morphogenesis in mice [55–57]. They used T47D cells and determined that GATA3 was recruited to ENH1 along with Pol II in order to regulate ESR1 expression. A study from our group also showed that the proteasome inhibitor bortezomib represses ESR1 expression via induction of a repressive chromatin environment on ENH1 in addition to loss of Pol II occupancy on the A promoter [58, 59]. Bailey et al. investigated commonly mutated regulatory elements associated with breast cancer risk and used clustered regulatory interspaced short palindromic repeats (CRISPR) to knock out a genomic region near ENH1. Knockout of this region caused not only repression of ESR1 expression but also several neighboring genes within a 300-kb locus [60]. Overall, these studies show that ENH1 is a major regulatory element responsible for regulating ESR1 expression in addition to the A promoter.

While the A promoter and ENH1 appear to be the most utilized regulatory elements, the activation of other promoters may play a role in the transcriptional regulation of ESR1. deGraffenried et al. performed mutagenesis studies on specific regions of the B promoter, which caused a decrease in promoter activity relative to that of the wild-type construct in a luciferase assay [61]. This result suggests that maximum transcriptional activation of ESR1 may require multiple regulatory elements. Therefore, studies focusing on ESR1 regulation should determine the involvement of multiple regulatory elements in order to fully characterize how ESR1 is regulated.

The compilation of large datasets in resources such as the UCSC Genome Browser and the ENCODE project have greatly expanded our view of the ESR1 gene structure. With the ENCODE project, information concerning the chromatin state, predicted regulatory elements, histone marks, transcriptional activity, and transcription factors of a genomic region for different cell lines are available for analysis. In Bailey et al., bioinformatics analyses of genome-wide association studies combined with ENCODE data for the regulatory environment of ESR1 provided insight on common single-nucleotide variants (SNVs) associated with breast cancer risk [60]. From these analyses, several SNVs were identified on regulatory elements associated with ESR1 expression. Using luciferase reporter assays, these SNVs were found to decrease the transactivation potential of these regulatory elements. Additionally, these SNVs were predicted to alter transcription factor chromatin binding on these elements, notably transcription factors known to regulate ESR1 expression such as GATA3 and SIN3A [26, 55, 62]. Studies like Bailey et al. serve as an example of how large sequencing datasets can better inform studies of the multiple transcriptional mechanisms that coordinately regulate ESR1 [60].

Basal Regulation of ESR1

ER is an essential regulator of physiological processes related to development, survival, and proliferation in breast cells in response to its physiological ligand, 17β-estradiol (E2). Therefore, it is crucial to maintain ESR1 expression. A number of transcription factors have been identified that aid in the maintenance of ESR1 expression in breast cancer cells (Table 1). Early studies identified potential regulators of ESR1 by analyzing interactions with ESR1 promoters via reporter gene and electrophoretic mobility shift assays. By these methods, AP-2γ (previously known as ERF-1) and a protein complex containing Sp1, Sp3, and USF-1 were discovered to associate with either the A or B promoters respectively and maintain ESR1 promoter activity [63, 64]. While the direct mechanism of how these transcription factors maintained ESR1 expression was not further investigated, a number of transcription-associated proteins have since been reported to associate with Sp1 and AP-2γ. These transcription factors are able to interact with metastasis-associated 1 (MTA1) and the co-activator p300 to positively regulate ESR1 expression [65–67]. FOXO3a, a member of the Forkhead transcription factor family, can bind to the A and B promoters in NF639 cells, and a dominant negative mutation affecting the transactivation domain caused loss of ER expression [68].

Table 1.

Transcriptional regulators of ESR1

ESR1 regulator	Pos/Neg regulation?	Confirmed ESR1 genomic location or genomic coordinates (chr6)	Reference(s)
ERUBF-1	Potentially positive	Between B and C promoters (− 893 to − 859)	[166]
pRb2/p130-multimolecular complex	Positive or negative	A promoter (chr6:152,128,814-152,128,940)	[167]
MTA-1	Positive (ER+ cells)/negative (ER− cells)	A promoter/exon 1 (+ 146–+ 461)	[67]
AIB1	Positive	A promoter (+ 135), ENH1 (chr6:151,979,485-151,979,625)	[26]
Cohesin	Positive	Between E1 and F promoters (chr6:152,086,545-152,086,774) and between exons 1 and 2 (chr6:152,445,477-152,445,760)	[78]
GATA3	Positive	B promoter (chr6:152,128,313-152,128,608), E1 promoter (chr6:152,022,992-152,023,729), ENH1 (chr6:151,979,485-151,979,625)	[55, 59, 168, 169]
Sp1/Sp3/USF-1	Positive	B promoter (− 245 to − 182)	[61, 170]
p53	Positive	Between A and B promoters (− 128 to − 40), B promoter (− 350 to − 289), and between B and C promoters (chr6:152,127,935-152,128,177)	[69, 171]
Eleanors/u-Eleanors	Positive	Between E1 and T1 promoters (chr6: 152,080,078-152,424,447)	[172]
AP-2γ	Positive	A–D promoters (− 3500 to + 230)	[59, 173]
BRCA1	Positive/negative	B promoter (chr6:152,128,428-152,128,644), near ENH1 (151,979,263-151,979,349)	[174, 175]
FOXM1	Positive	B promoter (chr6:152,128,376-152,128,588), C promoter (chr6:152,125,599-152,125,948 and chr6:152,126,059-152,126,400)	[176]
Pol II	Positive	A promoter (+ 135, + 60), B promoter (chr6:152,128,599-152,128,838), F promoter (chr6:152,011,555-152,011,924), ENH1 (chr6:151,979,485-151,979,625)	[26, 59, 177]
FOXA1	Positive	A promoter (chr6:152,128,599-152,128,838), B promoter (chr6:152,128,428-152,128,644), ENH1 (chr6:151,979,485-151,979,625)	[59, 168, 178]
ER	Negative	A promoter (+ 135), B promoter (chr6:152,128,399-152,128,674)	[26, 168]
SNAI1	Negative	A promoter (~ + 1), Intron 1	[129]
TWIST	Negative	Between B and C promoters, Intron 7 (chr6: 152,458,034 (hg18))	[131, 132]
SLUG/SNAI2	Negative	~ A promoter, between B and C promoters, and C promoter	[133]
BLIMP1	Negative	C promoter (− 2772 to − 2754)	[111]
SIN3A	Negative	A promoter (+ 135)	[26]
WT1	Negative	B promoter (chr6:152,128,562-152,128,700)	[179]
Repressor complex (HDAC1, DNMTs, MBDPs, MeCP2)	Negative	Exon 1 (chr6:152,129,124-152,129,274)	[108]
FOXO3a	Negative	Between C and D promoters (− 3284 to − 1864)	[68]
PR-B	Negative	Near C promoter (− 1757 to − 1752)	[180]
BARX2	N/A	ENH1 (chr6:151,977,586-151,978,005) and F promoter (chr6:152,011,374-152,011,630)	[181]
ZBTB7A	Positive	Between B and C promoters (− 1381)	[182]
MUC1-C	Negative	Exon 1 (chr6:152,011,575-152,011,670)	[183]
MYOD	Negative	Exon 1 (152,143,011-152,143,084)	[184]
SALL2	Positive	Near F promoter (152,008,290-152,008,462)	[185]
MLL, SET1A	Positive	A promoter (chr6: 152,128,814-152,128,940), F promoter (chr6: 152,011,652-152,011,781)	[186]
CTCF	N/A	B promoter (chr6:152,128,487-152,128,716), F promoter (chr6:152,011,363-152,011,792)	[177]
MYC	N/A	B promoter (chr6:152,128,488-152,128,797)	[177]
ZNF217	N/A	E1 promoter (chr6:152,022,992-152,023,561)	[169]

With the advent of chromatin immunoprecipitation (ChIP), several additional protein complexes and components of the transcriptional machinery were found to occupy specific genomic locations on ESR1. The tumor suppressor p53 forms a complex with transcriptional regulators CARM1, CBP, cJun, and Sp1 to bind to the A promoter and overexpression/knockdown studies showed a direct correlation between p53 and ESR1 expression [69]. GATA3 knockout caused ESR1 and ER protein downregulation, and it is speculated that GATA3 may positively regulate ESR1 via coordination with p300 and Pol II on ENH1 [55].

Early studies on the A promoter using luciferase assays showed weak basal transcriptional activity in the absence of E2, suggesting that additional chromatin components are involved in maintaining basal activity of the ESR1 gene. In coordination with transcriptional regulators of ESR1, an emerging area of study focuses on three-dimensional organization of chromatin. Studying transcriptional regulation of ESR1 often presents genomic information in a linear fashion. However, techniques that assess long-range chromatin interactions show how transcription factors and distal regulatory elements act in concert to regulate gene expression (techniques reviewed in [70]). Evidence of long-range chromatin interactions involving ESR1 was described by Dunbier et al. in which ESR1 appeared to be co-regulated with neighboring genes CCDC170, ARMT1, and SYNE1 in ER-positive breast cancer patient samples [71]. Dunning et al. also utilized genome-wide association studies and found enhancer elements between CCDC170 and ESR1 that co-regulated ESR1, RMND1, and CCDC170 [72]. As described earlier, Bailey et al. also observed co-repression of RMND1, ARMT, CCDC170, and ESR1 following CRISPR knockout of a genomic region that is commonly mutated in breast cancer patients [60]. CCDC170 (Coiled-Coil Domain-Containing Protein 170) and ARMT1 (Acidic Residue Methyltransferase 1) have been implicated in breast cancers by causing Golgi reorganization and promoting cell proliferation, respectively [71, 73, 74]. The functions of RMND1 (Required for Meiotic Nuclear Division 1) and SYNE1 (Spectrin Repeat Containing Nuclear Envelope Protein) in breast cancer are unknown, but RMND1 has been reported to affect mitochondria translation in human subjects [75]. It should be noted that the co-regulation of these genes may not always occur. Yamamoto-Ibusuki et al. analyzed correlations between ESR1 and its neighboring genes in breast cancer patient tissue and found higher CCDC170 expression correlated with ER negativity [76]. Nevertheless, this evidence suggests that the chromatin configuration around ESR1 is a point of regulation that may change with breast cancer risk and progression.

Using DNA motif analysis of this genomic region, Bailey et al. identified transcription factors that could be responsible for regulating this locus. Several of these predicted transcription factors are involved in modulating chromatin structures and in maintaining ESR1 expression [60]. For example, cohesin is a protein complex that commonly interacts with CCCTC-binding factor (CTCF) binding sites and plays a critical role in stabilizing the topological organization of chromosomal loops formed by CTCF [77]. By knocking down the cohesin subunits SMC3 or Mediator, ESR1 expression decreased along with reduced Pol II occupancy on the A promoter [78]. Additionally, CTCF itself has been implicated in affecting ESR1 expression through several potential mechanisms. One mechanism involves interactions between the long noncoding RNA u-Eleanors and CTCF along with other positive coregulators of ESR1 such as GATA3, p300, and CEBPB in MCF-7 cells that underwent long-term E2 deprivation. Knockdown of u-Eleanors caused repression of ESR1, suggesting these noncoding RNAs positively regulate ESR1. Other possible mechanisms involve CTCF directly affecting ER-DNA binding and ESR1 enhancer activity that could lead to changes in ESR1 expression [79, 80]. The importance of long-range chromatin interactions is further emphasized by Fullwood et al. [70]. They performed a comprehensive study showing that ER engages in long-range chromatin interactions across the genomic landscape to regulate genes [70]. In relation to ESR1, further study of the three-dimensional chromosomal architecture of ESR1 will allow us to build a more comprehensive model of how all elements of transcriptional regulation coordinate to regulate ESR1 gene expression.

Negative Regulation of ESR1

While ER is normally tightly regulated in order to modulate its transcriptional activity, disruption of this system has been associated with increased risk of breast cancer development as well as metastasis [81, 82]. Specifically, overexpression of ER has significant consequences related to enhanced proliferation of breast cancer cells, decreased sensitivity to endocrine therapies, and ligand-independent activation of the receptor. Targeted overexpression of ER in transgenic mice resulted in the development of ductal carcinoma in situ, and human breast cancer with higher ER expression directly correlated with the risk of developing atypical ductal hyperplasia [83, 84]. Overexpression of ER in MCF-7 breast cancer cells induced ligand-independent activation of ER and enhanced cell growth [85]. This constitutive ER activity could only be blocked by fulvestrant, a full ER antagonist, whereas tamoxifen was insufficient to do so. These downstream effects of ER overexpression were corroborated with the findings of Tolhurst et al. [86]. Copy number aberrations of ESR1, specifically amplification of the ESR1 gene, have also been detected in some ER-positive breast cancers. While controversial, ESR1 amplification has been shown to correlate with responsiveness to endocrine therapies due to higher ER expression [87, 88]. However, analysis of data from the TCGA Pan-Cancer Atlas and METABRIC using cBioPortal reveals that only a small percentage of breast tumors (~ 3%, n = 2813) contain amplifications of ESR1 (Fig. 3) [89, 90]. Overall, these findings emphasize the importance of regulated control of ESR1 expression.

Fig. 3.

Percent of invasive ductal carcinoma (IDC, n = 2440) and invasive lobular carcinoma (ILC, n = 373) cases with ESR1 genomic alterations. The data are derived from The Cancer Genome Atlas Research Network [Pan-Cancer Atlas; https://www.cancer.gov/tcga) and the METABRIC study [164, 165]

Downregulation of cell receptors is an essential and intricate process for modulating their activities following stimulation with their cognate ligand. ER is normally tightly regulated on multiple levels of regulation in a negative feedback loop upon E2 treatment. This is exemplified by the decrease in ESR1 mRNA levels in response to E2 [35, 91, 92]. We determined that E2-stimulated repression of ESR1 involves a rapid decrease in nascent ESR1 transcripts [26]. Furthermore, repression of ESR1 can be blocked with either tamoxifen or fulvestrant, showing E2 stimulates autoregulation of ESR1 [93]. To mediate ESR1 repression following E2 stimulation, ligand-bound ER binds to ESR1 regulatory elements, specifically the A promoter and ENH1 [26, 55]. ER directly interacts with the repressor complex SIN3A at the proximal promoter [26]. SIN3A is a known scaffold protein for HDACs, and the recruitment of SIN3A was accompanied by a repressive chromatin environment, specifically an increase in H4K20me3 on the A promoter, and a loss of Pol II occupancy [94, 95]. Knockdown of SIN3A prevented repression of ESR1 expression by E2, confirming that SIN3A plays a direct role in E2-stimulated ER autoregulation. This work established a model of ESR1 transcriptional regulation in the ER autoregulatory system.

Dysregulation of the autoregulatory system, specifically mechanisms of transcriptional repression of ESR1, is a significant area of interest due to the potential role of ER downregulation in mediating endocrine therapy resistance. As stated previously, approximately 70% of primary breast tumors are classified as ER-positive, and the majority of these tumors maintain their ER-positive status but exhibit reduced ER expression in metastatic sites relative to the primary tumors [2, 5, 21, 96–103]. For ER-negative breast cancers, hypermethylation of CpG islands on the ESR1 A and B promoters was detected in tumor samples, and the magnitude of methylation on these promoters was directly correlated with poor response to endocrine therapies [104]. Hypermethylation also appears to increase with disease progression [105]. To study ESR1 methylation, the Davidson group utilized ER-negative breast cancer cell lines MDA-MB-231 and Hs578t and observed hypermethylation of the A promoter [106, 107]. Several DNA methyltransferases (DNMT1, DNMT3a, DNMT3b), methyl-binding proteins (MeCP2, MBD1, MBD2, MBD3), and the histone deacetylase 1 (HDAC1) were detected on CpG islands in the A promoter [107, 108]. To determine whether hypermethylation was responsible for loss of ER expression, MDA-MB-231 and Hs578t cells were treated with the DNMT inhibitor 5-aza-2′-deoxycytidine (5-aza), which restored ER expression in both cell lines and caused attenuated cell growth [108]. These cells were also treated with HDAC inhibitor trichostatin A (TSA) with similar results. Re-expression of ER also restored tamoxifen sensitivity to MDA-MB-231 cells, inhibiting their growth [109]. These results suggested that coordination between methylation and deacetylation could facilitate both the generation and maintenance of a repressive chromatin environment on ESR1.

Prior to the recruitment of repressive complexes, signaling pathways and their associated transcription factors are also involved in repressing ESR1 expression. Three major signaling pathways associated with ESR1 transcriptional repression are the NF-κB, ERK/MAPK, and PI3K signaling pathways. NF-κB activation is inversely related to ER expression and constitutive activation of NF-κB is commonly found in ER-negative breast tumors [110]. Overexpression of the NF-κB subunit RelB in MCF-7, ZR-75-1, and NF639 (mouse breast cancer cell line) caused decreases in ESR1 mRNA and ER protein and an increase in the expression of the transcription factor BLIMP1. ZR-75-1 cells were transfected with a BLIMP1 expression vector, and BLIMP1 was found to bind to the ESR1 C promoter and repress ESR1 expression [111]. While not directly tested, BLIMP1 has been shown to recruit HDACs, histone methyltransferases (HMTs), and co-repressor complexes [112–114]. ER-negative tumors also frequently overexpress growth factor receptors, notably epidermal growth factor (EGFR) and insulin-like growth factor receptor (IGF-1R) [115, 116]. Unlike IGF-1R, EGFR has an inverse relationship with ER expression in breast cancer patient samples [117, 118]. Additionally, a recent study examining mammary organoid culture indicate that decreasing EGF in culture media promotes fully differentiated luminal A cell population [119]. The El-Ashry group has extensively studied the effects of hyperactivation of the ERK/MAPK signaling pathway on ER expression, which resulted in decreases in ESR1 mRNA and ER protein in MCF-7 cells expressing constitutively active Raf or MEK [120, 121]. To determine the mechanism behind ER downregulation, an epigenetic compound screen was used to identify potential regulators. Several HDAC inhibitors effectively relieved repression on ESR1 expression in ER-negative cell lines lacking a methylated ESR1 A promoter [121]. Class I HDACs (HDAC1, 2, and 3) were recruited with ERK2 along the ESR1 gene, encompassing the A–D promoters. Knockdown of both MEK and class I HDACs increased ER expression by the greatest magnitude. Activation of the PI3K/mTOR signaling pathway in combination with the p70 S6 kinase (S6K1) has also been linked to decreases in ER expression. Inhibition of mTOR with rapamycin treatment caused overexpression of ER in MCF-7 cells, and restoration of tamoxifen sensitivity in hormone-resistant breast cancer cells in vitro and in vivo [64, 122].

While the common denominator in transcriptional regulation of ESR1 is modifications of the chromatin, clinical trials and in vitro studies indicate that epigenetic therapies have thus far not shown promise. Targeting HDACs and DNA methyltransferases have proven to be minimally effective in treating breast cancer patients with little, if any, improvement in survival rates (reviewed in [123]). One recent clinical trial treated patients with hormone-resistant disease (including triple negative breast cancer patients) with 5-aza and the HDAC inhibitor entinostat in combination with endocrine therapy [124]. Only 1 patient out of 27 ER-positive, endocrine therapy-resistant patients exhibited any response to the combination of epigenetic and endocrine therapies [124]. Interestingly, no patients with triple-negative breast cancer exhibited a discernable increase in ER expression and only a few ER-positive tumors showed a minor increase in ER expression. Furthermore, prolonged treatment of MCF-7 cells with TSA actually causes a decrease in ESR1 mRNA and ER protein expression, specifically causing a decrease in ESR1 mRNA stability through the RNA binding protein HuR [122, 125]. These results suggest that specific targeting, as opposed to global modification of the epigenetic landscape, may be more effective at restoring ESR1 expression and endocrine therapy sensitivity.

Transcription factors are potential targets for relieving transcriptional repression of ESR1. Transcription factors that have been reported to repress ESR1 are listed in Table 1 and presented in Fig. 4. The most notable set of transcription factors that transcriptionally repress ESR1 expression are those associated with epithelial-to-mesenchymal transition (EMT), which include SNAI1, TWIST, and SLUG. All three transcription factors are commonly overexpressed in advanced stages of breast cancer, and their expression is often inversely correlated with that of ESR1 [126–128]. Furthermore, breast tumors containing methylation of the ESR1 promoter often also contain methylation of the E-cadherin gene (CDH1), a gene that is transcriptionally repressed by these three EMT factors [105]. The overexpression of SNAI1 in MCF-7 cells induces an invasive phenotype and represses ESR1 mRNA expression. SNAI1 binds to the A promoter and reduces H3K9Ac (activating histone mark) but does not increase methylation on the H3K9 site (repressive histone mark), suggesting SNAI1 may only partially repress ESR1 expression [129]. While no further details of the transcription mechanism were determined, SNAI1 can also recruit the SIN3A complex with HDAC1/2 on the CDH1 promoter to repress its expression [130]. TWIST overexpression in MCF-7 cells conferred resistance to tamoxifen and repression of ESR1. To mediate transcriptional repression of ESR1, TWIST was recruited with DNMT3B and HDAC1 to the A promoter [131]. Fu et al. expanded on these findings by showing that TWIST interacted with one of the major HDAC complexes, the nucleosome remodeling and deacetylase repressor complex (NuRD) [132]. Using the DMT and HDAC inhibitors 5-aza and valproic acid, the repressive effects of TWIST on ESR1 repression were relieved and tamoxifen sensitivity was restored [131]. SLUG, also known as SNAI2, was found to bind to the ESR1 promoter as well and recruited the histone demethylase LSD1 for demethylating H3K4me2, an activating chromatin mark [133]. Knockdown of SLUG or LSD1 restored ESR1 expression, confirming their role in mediating transcriptional repression of ESR1. Examination of ENCODE data reveals that FOXA1 and GATA3 were commonly found across the ESR1 gene, specifically on the A, B, and E promoters [134]. The ENCODE project also revealed other transcription factors that have not been as extensively studied in the context of ESR1 regulation such as c-Myc and ZNF217. c-Myc is an ER target gene, and its expression is directly related to that of ESR1 mRNA. c-Myc also directly interacts with ER to regulate gene expression [135]. c-Myc and ER both interact on the B promoter, suggesting potential regulation of ESR1 by c-Myc and ER. ZNF217 is a Krüppel-like finger protein that can directly interact with ER and can facilitate recruitment of ER to an estrogen response element (ERE), where ER directly contacts the DNA at a palindromic sequence [136]. Furthermore, knockdown of ZNF217 appears to cause a decrease in ER protein expression, implicating ZNF217 as a positive regulator of ESR1 expression.

Fig. 4.

Diagram of ESR1 with transcription factors and co-factors. Shown are transcription factors involved in repressing (yellow) or maintaining basal ESR1 expression (red). Transcription factors (TFs) that have dual repressive and activating roles are noted in yellow and red. TFs with unknown effects on ESR1 expression are noted in gray. Genomic coordinates shown above each regulatory element are based on the UCSC Genome Browser (human genome assembly hg19) [51]. Diagram is not drawn to scale

Squelching has also been proposed as yet another mechanism of transcriptional regulation of ESR1. Squelching is the process of repressing gene expression by sequestering limited factors that regulate transcription such as co-activators and co-repressors (reviewed in [137]). The mechanism was first described by Gill and Patshne using yeast that expressed derivatives of the GAL4 transcriptional activator. After overexpressing GAL4, the expression of genes lacking a GAL4 binding site decreased [138]. Based on this observation, it was suggested that genes containing GAL4 binding sites may be competing with other genes for transcription factors of a finite amount. Squelching was also reported to occur with steroid receptors [139]. By overexpressing ER lacking the DNA binding and dimerization domain, the transcriptional activity of both glucocorticoid receptors (GR) and progesterone receptors (PR) decreased, suggesting ER may be competing with both GR and PR for co-factors. Analyzing the effects on endogenous receptors, treatment of ER-positive breast cancer cell lines with progestin caused a decrease in ER transcriptional activity, suggesting PR may be utilizing co-factors that are important for ER transactivation. Co-factors related to ESR1 gene regulation that are implicated in a squelching mechanism include SRC-3 and p300 [140, 141]. ChIP-seq datasets of activating histone marks and these co-factors revealed a re-distribution of both SRC-3 and p300 upon ligand activation. He et al. showed that overexpression of SRC-3 prevented the re-distribution of SRC-3 from non-E2-responsive genes, which is consistent with the squelching model [140]. For p300, Guertin et al. showed that p300 is bound to ESR1 in the absence of ligand, promoting basal ESR1 gene expression [141]. When activated, ER recruits p300 to activated genes, resulting in the redistribution of p300 and the loss of p300 at ESR1, leading to repression of ESR1 gene expression [141]. Given the prevalence of co-factors such as SRC-3 and p300 in transcriptional regulation by multiple transcription factors, the squelching mechanism may play a significant role in context-dependent modulation of ESR1 during disease progression.

Transcriptional Regulation of ESR1 by the Tumor Microenvironment

A major contributor to the altered control of ER expression in breast tumor is the tumor microenvironment (TME). Evidence for regulation of ER expression by the TME is apparent based on clinical data showing either repression of ESR1 mRNA expression or downregulation of ER protein in metastatic samples relative to that in the patient-matched primary samples [20, 142]. Extrinsic TME-derived soluble factors (growth factors, cytokines, cholesterols, heavy metals, cell stress-related factors, exosomes, and other) have been reported to affect ER expression via paracrine interactions. Additionally, co-culture with a variety of cell types also induced repression of ESR1 via mechanisms involving signaling pathways and cell stress. Tables 2 and 3 cite the factors and co-culture approaches that have been applied to the study of TME-associated regulation of ESR1. Despite the complexity and multiplicity of these paracrine interactions, regulation of ESR1 expression is a convergence point. This highlights a generalized property that TME-derived factors can promote breast cancer disease progression by diminishing ER expression and endocrine control.

Table 2.

Extrinsic factors affecting ER expression

Factors/signaling pathways	Cell model(s)	Effect on ER expression	Level of regulation (mechanism)	Reference
17β-Estradiol	MCF-7	↓	Transcriptional Posttranscriptional Posttranslational	[26, 27, 29, 30, 34, 187]
Epidermal growth factor	MCF-7	↓	Transcriptional Posttranscriptional	[152, 188]
Insulin-like growth factor 1	MCF-7	↓	Transcriptional	[154]
Transforming growth factor β 1	MCF-7	↓	Transcriptional	[153]
Tumor necrosis factor ⍺	MCF-7, ZR-75-1	↓	Posttranslational	[189]
27-Hydroxycholesterol	MCF-7	↓	Posttranslational	[190]
Oncostatin M	MCF-7, T47D	↓	N/A	[191]
Prolactin	MCF-7	↑	N/A	[192]
Reactive oxygen species	MCF-7	↓	N/A	[193]
Heavy metals	MCF-7	↓	N/A	[194]
Cadmium	MCF-7	↓	Transcriptional	[195]
ERK/MAPK	MCF-7	↓	Transcriptional	[120, 121, 196]
p38 MAPK	MCF-7	↓	Posttranslational	[197]
PI3K/mTOR/Akt	MCF-7, T47D	↓	Transcriptional	[122]
Hypoxia	MCF-7, ZR-75-1	↓	Transcriptional Posttranslational	[158, 198]
IL-1β	MCF-7	↓	Transcriptional	[199]

Table 3.

Regulation of ER expression via co-culture with different cell types

Cell type	Breast cancer cell line(s)	Co-culture method	Δ ESR1 mRNA	Δ ER protein	Reference
Macrophages	MCF-7	CM	↓	↓	[156]
Primary human mammary fibroblasts (normal)	MCF-7, HMEC	Indirect (3D), direct (3D), CM	↓	N/A	[200]
Adipocytes	MCF-7	Indirect (transwell)	↓	N/A	[157]
Human mesenchymal stem cells	MDA-MB-231, T47D	Indirect (transwell), direct (in vivo mouse study)	Potential ↓(miR-221/222)	N/A	[201]
HS-5, HS-27a, primary human fibroblasts (CD146−/+ cancer-associated breast cancer tissues)	MCF-7	CM	↓	↓	[162]
HS-5, HS-27a, KM101, hFOB	MCF-7, T47D, BT-474	CM	↓	↓	[161]
HS-5, MDA-MB-231, MEF, MDA-MB-468, HS578T, HEK293, RMF/EG, MRC-5, THP-1, SK-BR-3, CV-1, MCF10A, BT-20	MCF-7, T47D	Direct	↓(HS-5)	↓(HS-5, RMF/EG, MRC-5, MDA-MB-231, CV-1, MCF10A, MDA-MB-468, HS-578T, THP-1)	[155]
HMF cell line	MCF-7	Indirect (3D)	↓	↓	[202]
Primary human fibroblasts (cancer-associated breast cancer tissues)	MCF-7	CM	↓(CAF derived from basal tumor)	↓(CAF derived from basal tumor)	[203]
Primary bone marrow stroma (normal donors)	T47D, MDA-MB-231	Direct	Potential ↓(miR-221/222)	N/A	[204]
Primary bone marrow stroma (normal and cancer-associated), HS-5	MCF-7, T47D	CM	↓	↓	[163]

To affect ESR1 expression, soluble factors, conditions promoting cell stress, and different cell types can transcriptionally, posttranscriptionally, or posttranslationally regulate ER expression. Of these mechanisms and given the complexity of ESR1 transcriptional regulation, the effects of the TME on ESR1 transcriptional regulation appears to be the least well-characterized. Growth factors are one of the most-studied factors in relation to breast cancer since these factors can alter sensitivity to hormone therapy and activate steroid receptors [143–151]. For example, Stoica et al. investigated the effects of three growth factors on ER expression: transforming growth factor β (TGF-β), epidermal growth factor (EGF), and insulin-like growth factor (IGF-1) [152–154]. MCF-7 cells were treated with these growth factors and each growth factor caused a decrease in both ESR1 mRNA and ER protein. Measuring ER protein half-life and ESR1 mRNA stability neither appeared to decrease following treatment with these growth factors. Using nuclear run-on assays, transcription of ESR1 decreased significantly with TGF-β and EGF treatment [152, 153]. Since neither ER protein nor ESR1 mRNA stability changed with IGF-1 treatment, it is likely a transcriptional mechanism is involved in regulating ER expression with IGF-1 as well [154]. While these three growth factors caused decreased expression of ESR1 mRNA, several additional findings were observed for each growth factor. With TGF-β, the activity of the ESR1 promoter (measured via chloramphenicol acetyltransferase reporter activity) decreased, suggesting transcriptional repression by TGF-β may be mediated via the ESR1 promoter [153]. With EGF, serum was required in the cell culture media to induce transcriptional repression of ESR1, implicating that additional factors are required in order to coordinate with the effects of EGF and repress ESR1 [152]. With IGF-1, the PI3K signaling pathway and protein kinase C were required to repress ESR1 [154]. Despite the differences for how each growth factor regulates ESR1, transcriptional repression of ESR1 appears to be a general mechanism by which the TME represses ER expression.

The cumulative effects of the TME has been primarily studied through co-culture with different cells in order to assess changes in tumor cell behavior. Given that multiple factors and cell types appear to regulate ER expression, our lab first addressed whether there is a degree of specificity in the ability of different cells to regulate ER expression. We directly co-cultured MCF-7 cells with a panel of cell lines and measured changes in ER protein expression [155]. We observed that only some of these cell lines were able to downregulate ER expression, which suggests that there is a degree of specificity in the regulation of ER expression that is not limited by cell type nor site of cell origin.

Decreases in ER expression by other cells have been shown to occur on multiple levels of regulation. Macrophages, adipocytes, and stromal cells mediate transcriptional repression of the ESR1 gene via different mechanisms. Stossi et al. co-cultured pro-inflammatory macrophages with MCF-7 cells and observed downregulation of ER protein and ESR1 mRNA [156]. Using an inhibitor screen of various signaling pathways, only inhibitors of the ERK/MAPK signaling pathway prevented the decrease in ER expression. To mediate repression of ESR1 by macrophages, cJun was essential for the recruitment of ERK to multiple binding sites on the gene. While a direct mechanism of ESR1 regulation by cJun or ERK was not described, Stossi et al. posit that macrophages regulate ER expression via direct ESR1 transcriptional repression. Yao-Borengasser et al. co-cultured human adipocytes with MCF-7 cells, which induced repression of ESR1 [157]. Through gene expression analysis of the co-cultured adipocytes, hypoxia inducible factor 1 subunit α (HIF1α) gene expression was significantly elevated relative to monoculture. Knockdown of HIF1α in the co-cultured MCF-7 cells relieved repression on ESR1. Given the role of HIF1α in response to hypoxic conditions, Yao-Borengasser et al. proposed that adipocytes may generate hypoxic conditions to repress ESR1, but it should be noted that signs of hypoxia such as low oxygen and increased ROS production were not measured in this study [157]. However, their hypothesis was supported in a study by Ryu et al., which showed MCF-7 cells cultured in hypoxic conditions resulted in transcriptional repression of ESR1 via HIF1α [158]. Similar findings were also described in primary breast tumors samples [159].

Recently, the effects of the bone microenvironment on breast cancer cell behavior has been of significant interest. The bone is the most common site of breast cancer metastasis, and despite the majority of breast cancer metastases maintaining their ER-positive status, most bone metastases have reduced responsiveness to endocrine therapies [160]. To explain the disparity in responsiveness, Priedigkeit et al. showed 7 out of 11 bone metastases exhibited repressed ESR1 mRNA expression compared to that in patient-matched primary tumor samples [142]. This suggests that the decrease in responsiveness to endocrine therapies may be attributed (at least in part) to a concomitant decrease in ER expression.

Our lab first described that HS5 cells, a bone marrow-derived cell line, regulated both ER protein and ESR1 mRNA expression in both direct and indirect co-culture systems [155]. Our results regarding downregulation of ESR1 expression by HS5 cells were corroborated by two additional publications, which reported that HS5 cells caused a decrease in both ER protein and ESR1 mRNA expression in ER-positive breast cancer cells [161, 162]. After assessing ESR1 mRNA and ER protein stability, we determined that transcriptional regulation of ESR1 was primarily responsible for ER downregulation by bone marrow stromal cells (BMSCs) [163]. Specifically, conditioned media (CM) from both cancer-associated BMSCs and HS5 cells caused a decrease in nascent ESR1 transcripts in both MCF-7 and T47D cells. HS5 CM treatment decreased Pol II occupancy on the A promoter and ENH1 while H3K27Ac levels on ENH1 decreased in MCF-7 cells following HS5 CM treatment. The decrease in H3K27Ac levels suggest that ENH1 may play a major role in regulating ESR1 by BMSCs. Interestingly, neighboring genes of ESR1 were also co-repressed in the presence of BMSC CM, similar to the pattern of co-regulation described by Bailey et al. and Dunning et al. As of this review, this was the first observation that suggested long-range chromatin interactions may be involved in ESR1 regulation by the TME. Overall, the complexity of ESR1 transcriptional repression is evident as ESR1 appears to be regulated via different mechanisms in each study. However, further investigation is needed to determine how the chromatin environment is affected by the cumulative effects of the TME and, ultimately, general mechanisms behind ESR1 repression.

Conclusions

Over the past 30 years, the field has made significant strides in piecing together the components of ESR1 regulation, identifying a number of transcription factors, chromatin states, signaling pathways, and chromosomal arrangements that all play a role in regulating the gene. This regulation likely involves an intricate coordination of these elements, and new discoveries of additional transcription factors, regulatory elements, and long-range chromatin interactions as a result of large sequencing datasets will further our understanding of how ESR1 is regulated. Considering the number of mechanisms affecting ESR1 expression, it is clear that modulating ER expression levels and ultimately hormone responsiveness is of significant biological importance. Dysregulation of ER expression has been implicated in the development of endocrine therapy resistance since ER expression levels directly correlates with endocrine therapy responsiveness. Continuing to investigate the mechanisms involved in dysregulation of ESR1 will ultimately provide insight on maintaining or even restoring the effectiveness of endocrine therapies in ER-positive breast cancer.

Funding Information

This work was supported by the National Institute of Health Grant T32 CA009135 (to D.K.L.) and McArdle Laboratory for Cancer Research.

Compliance with Ethical Standards

Conflict of Interest

The authors declare they have no conflicts of interest.

Ethics Approval

Not applicable

Consent to Participate

Not applicable

Consent for Publication

Not applicable