miRNAs regulated by estrogens, tamoxifen, and endocrine disruptors and their downstream gene targets.

MicroRNAs (miRNAs) are short (22 nucleotides), single-stranded, non-coding RNAs that form complimentary base-pairs with the 3' untranslated region of target mRNAs within the RNA-induced silencing complex (RISC) and block translation and/or stimulate mRNA transcript degradation. The non-coding miRBase (release 21, June 2014) reports that human genome contains ∼ 2588 mature miRNAs which regulate ∼ 60% of human protein-coding mRNAs. Dysregulation of miRNA expression has been implicated in estrogen-related diseases including breast cancer and endometrial cancer. The mechanism for estrogen regulation of miRNA expression and the role of estrogen-regulated miRNAs in normal homeostasis, reproduction, lactation, and in cancer is an area of great research and clinical interest. Estrogens regulate miRNA transcription through estrogen receptors α and β in a tissue-specific and cell-dependent manner. This review focuses primarily on the regulation of miRNA expression by ligand-activated ERs and their bona fide gene targets and includes miRNA regulation by tamoxifen and endocrine disrupting chemicals (EDCs) in breast cancer and cell lines.

1. Introduction

The three primary estrogenic steroid hormones: estradiol, estrone, and estriol regulate fertility, development, and homeostasis in various tissues including the brain, breast, cardiovascular system, colon, skin, brain, lung, and reproductive tract in both women and men. The word estrogen is often used in studies when referring to the use of estradiol (E2), the primary circulating estrogen in premenopausal women which is synthesized from cholesterol in the granulosa cells in the ovary in response to luteinizing hormone (LH). Estrone (E1) is the primary estrogen in postmenopausal women, synthesized primarily in adipose from adrenal androgens. E2 and E1 can also be formed locally, e.g., in breast (1) and lung (2).

Lifetime estrogen exposure is widely accepted as a major risk factor for the development of breast cancer (3). Because estrogens have a clear role in the majority of breast cancers and since estrogen receptor α (ERα)is the best prognostic indicator for breast cancer patients and is considered to be the most successful molecular target in the history of cancer drug discovery (4), much is known about the molecular mechanisms of estrogen regulation of transcription.

Data from ENCODE (Encyclopedia of DNA Elements, http://www.nature.com/encode/) revealed that ~ 75 % of the human genome is transcribed while only ~ 1% is protein-coding mRNA, suggesting that other RNA transcripts, including long non-coding RNAs (lncRNAs) and small RNAs (85% of which correspond to four major classes: small nuclear (sn)RNAs, small nucleolar (sno)RNAs, micro (mi)RNAs and transfer (t)RNAs), have regulatory functions (5). Next-generation sequencing (NGS) by RNA sequencing (RNA seq), also called ‘whole transcriptome shotgun sequencing’, is used to identify the transcriptome (6). The transcriptome includes all the RNAs in that source: mRNA, rRNA, and tRNA; and the non-coding RNAs (ncRNAs): miRNAs, enhancer RNAs (eRNAs), endogenous small-interfering RNAs (siRNAs), Piwi-interacting RNAs (piRNAs), and lncRNAs ranging from 1,000 to> 90,000 bases (7). Like miRNAs, siRNAs and piRNAs bind Argonaute family members and base pair with target RNA to cause RNA degradation and/or translation repression (8). LncRNAs are involved in assembly of active e.g., Neat1, or repressed, e.g., Xist, nuclear domains for transcription in a cell-dependent manner (9). This review focuses on estrogen regulation of miRNAs.

miRNAs, first described in 1993, are small (22 nucleotides), single-stranded non-coding, evolutionarily conserved RNA molecules that are related to, but distinct from, small interfering RNAs (siRNAs) which regulate mRNA translation or stability (10–12). Comparative genomics analyses have revealed > 45,000 miRNA binding sites within human 3'UTRs that are conserved, indicating that > 60% of human protein-coding genes have been under selective pressure to maintain pairing to miRNAs (13). Compared to transcriptome or microarray analyses identifying miRNA expression patterns in different human cells, tissues, or with various treatments, there are far fewer published reports of estrogen or tamoxifen regulation of miRNAs expression in human cells or tissues (Figure 1). The pace of publication on miRNAs in humans has slowed since 2013 and publication rate on estrogen and human miRNA peaked in 2012 and is in decline. Given the role of estrogens in stimulating breast cancer, it is not surprising that most studies have examined changes in miRNA expression and their correlation with diagnostic markers used in breast cancer therapies, e.g., ERα and tumor grade (14–24). Estrogens regulate miRNA expression by both genomic (transcriptional) and non-genomic/membrane-initiated mechanisms of action. Identification and characterization of estrogen-regulated miRNAs and their targets may provide new biomarkers and therapeutic targets in in diseases including breast cancer. There are many online resources about miRNA-mRNA targets recently compiled in http://multimir.ucdenver.edu/ and reviewed in (25).

Figure 1

History of PubMed citations on human miRNA, estrogen AND miRNA, and tamoxifen AND miRNA

The search terms used were human AND miRNA (black closed circles) and human AND miRNA AND estrogen. Each point is the number of publications in the calendar year indicated. The number of citations was taken directly from an advanced search of PubMed and was not hand-curated to remove non-relevant citations.

2. Genomic ER activities

Transcription is initiated through a complex series of activities occurring through the cooperative interaction of multiple factors at the target gene promoter in association with interactions with other chromatin regions at great distances from the transcription start site- and even on different chromosomes (26). I will use the term ER to refer to either ERα or ERβ or to both subtypes. I will refer to each subtype individually when appropriate to differentiate their established differences. Estrogens bind the ligand binding domains (LBD) of ERα and ERβ which are members of the 48 member steroid/nuclear receptor (NR) superfamily of proteins (27). ERα and ERβ are highly conserved within the DNA binding domain (DBD, C domain), but differ in their N- and C- termini (28).

Crystal structure studies of the LBD of ERα, excluding the F domain, identified 12 alpha helices and found that E2 binding repositions helix 12 that acts as a “switch” controlling accessibility of coregulator interaction site: the ‘coactivator binding groove’ (29).

Chromatin forms a barrier for transcription factor binding. FoxA1, PBX, TLE1, AP2g, and GATA3 act as “pioneer factors” that remodel condensed chromatin to facilitate ERα binding (reviewed in (30)). ERα interacts directly with high affinity to a specific DNA sequence called the estrogen response element (ERE = 5’-AGGTCAnnnTGACCT-3’) (28). ER-ERE binding enhances the recruitment of coactivator/chromatin remodeling complexes resulting in histone modifications, nucleosomal repositioning, increased accessibility to the DNA template for RNA polymerase II interaction, and increased target gene transcription (reviewed in (31, 32)). Chromatin immunoprecipitation (ChIP) of ERα in cell lines, most notably MCF-7 human breast cancer cells, followed by sequencing of the bound DNA(ChIP seq) has established that EREs are located in gene promoters and at great distances from the transcription start site, including in the 3’ flanking regions of regulated genes (33–40). Cell-specific ERα cistromes have been identified in ERα-transfected U2OS cells (41), MDA-MB-231 breast cancer cells(40)., and HeLa cells (42). In another example, ERα overexpression in ERα- HeLa cells identified only 9% of common promoter binding sites with MCF-7

In addition to direct ER-ERE binding, ER also activates transcription via a “tethering mechanism” whereby ER interacts directly with transcription factors, e.g. Sp1 (43) and AP-1 (44), bound to their response elements. ERβ binding sites appear enriched for AP-1 sites (45). ChIP-seq, ChIP-PET (ChIP for ERα followed by paired-end tag sequencing) and ChIP-chip experiments identified a number of transcription factor binding sites with which ERα interacts in MCF-7 cells including: AP-1, CEBP, FOXA1, PAX6, RORA, PITX2, and GATA2 (46).

3. Rapid, membrane-initiated, nongenomic estrogen action

In addition to its classical genomic/transcriptional effects mediated by ER-DNA interaction, described above, E2 has rapid “nongenomic, extra-nuclear, or membrane-initiated” effects that occur very rapidly, i.e., within seconds-minutes after E2 administration (reviewed in (47, 48)) These effects are independent and distinct from the genomic, i.e., ER-mediated transcription, activities reviewed in the preceding section. Rapid estrogen-stimulated intracellular activities are mediated by plasma membrane (PM)-associated ERα, ERβ, ERα splice variants: ERα46, ERα36, and/or by an ‘orphan’ G-protein coupled estrogen receptor GPR30/GPER (49–60). Palmitoylation of ERα46 helps it to localize to the PM (61–64). ERα36 is also recruited to the PM by palmitoylation (65). Evidence of the biological function of PM-associated ERs, including GPER, is supported by experiments in which cell-impermeable E2–bovine serum albumin (E2-BSA) or other E2-conjugates rapidly initiated intracellular kinase cascade activities including MAPK/ERK (p42/p44 MAPKs), endothelial nitric oxide synthase (eNOS), and PI3K/AKT (66–75). Increased E2 during pregnancy activates GPER which, with activation of glucagon-like peptide 1 (GLP1) receptor, increases cAMP-PKA and decreases miR-338-3p resulting in increased expression of proliferation and/or anti-apoptotic genes and β-cell proliferation (76). Overexpression of ERα46 stimulates E2-induced endogenous miR-21 transcription and reduced miR-21 targets PTEN and PDCD4 in MCF-7 cells (77). ERα36 and miR-210 expression were correlated in TNBC tumors (78), but to my knowledge, no mechanistic studies have been performed on ERα36 regulation of miRNA transcription.

4. miRNA processing and general activity

The human genome contains ~ 2,588 mature miRNAs (June 2014, http://www.mirbase.org/) (79). The term miRNome is defined as the full spectrum of miRNAs for a specific genome (80). About half of miRNAs are expressed from introns of protein-coding transcripts and miRNAs have 5' and 3' sequence features that form boundaries including transcription start sites, CpG islands, and transcription factor binding recognition elements (81). miRNAs may be differentially processed from the sense and antisense strands of the same hairpin RNA or transcripts from the same locus (82). miRNAs are produced by canonical miRNA processing or noncanonical pathways (83).

The canonical and noncanonical pathways of miRNA biogenesis and the regulation of components of this pathway by miRNAs, phosphorylation, and protein: protein interactions and E2are depicted in Figure 2. miRNAs are transcribed as primary-micro-RNAs (pri-miRNAs) by RNA polymerase II either as independent transcription units or are cotranscribed within introns of pre-mRNAs (84). Pri-miRs are caped and polyadenylated (85). The self-base-pairing stem-loop structure of the pri-miR is cleaved by the microprocessor complex with catalytic Drosha (RNASEN), an RNAse III family endonuclease, and its cofactor DGCR8 (DiGeorge syndrome critical region 8 gene) into shorter (60 to 70 nt) imperfect hairpin-containing precursor-miRNAs (pre-miRNAs) (86). DGCR8 functions as an anchor by binding the pri-miRNA to direct cleavage by Drosha 11 bp from the dsRNA-ssRNA junction (84). The Drosha microprocessor also binds and regulates other cellular RNAs (84) and includes other proteins and hnRNPs shown in Figure 2: EWSR1, FUS< Nucleolin, p68, p72 which interacts with YAP2.

Figure 2

Model of canonical miRNA biogenesis and function

Primary transcripts of microRNAs (pri-miRNAs) are transcribed by RNA polymerase II, processed by the RNAse III enzyme, Drosha and its cofactor DGCR8, to precursor microRNAs (pre-miRNAs) which are exported from the nucleus by Exportin/RAN-GTP (85). In the cytoplasm, pre-miRNAs are processed by the Microprocessor complex that includes Dicer, an RNAse III enzyme, to form mature ~22 nt transiently double-stranded miRNA duplexes that are transferred to Argonaute proteins (most notably AGO2 in the RNA-induced silencing complex (RISC), leading to unwinding of the duplexes to form single stranded miRNAs. The RISC complex binds either to the 3’ untranslated region (3’ UTR) or to the open reading frame (ORF) of its target mRNA. Binding of miRNA/RISC complex with the 3’UTR causes translational repression (18).

Exportin and Ran-GTP or CRM1 export pre-miRNAs from the nucleus. In the cytoplasm, pre-miRNAs are cleaved to the mature ~22 nt transiently double-stranded miRNA duplexes by the RNAse III enzyme Dicer. Dicer with its associated cofactors TRBP (TAR (transactivating response) RNA-binding protein) and PACT (protein activator of the interferon-induced protein kinase) transfers the miRNA to the RNA-induced silencing complex (RISC) containing the catalytic Argonaute proteins (AGO1, AGO2, AGO3, and AGO4 (87)) which unwind the duplexes to form single stranded miRNAs. One strand miRNA is preferentially selected to bind one of the AGO proteins and by base-pairing directs translational inhibition and/or mRNA degradation by binding either to the 3’ untranslated region (3’ UTR) or to the open reading frame (ORF) of its target mRNA (88–91). AGO2 is the catalytic component of RISC. Dicer binds not only miRNAs, but also tRNAs, snoRNAs, mRNA and promoter RNAs (92). The widespread reduction of miRNAs in cancers is considered to be the result of defective miRNA processing as reflected in increased pri-miRNAs due to Hippo signaling regulation of p72 nuclear function by YAP sequestering p72 from the Microprocessor in a cell-density-dependent manner (93).

The non-canonical pathways of miRNA generation include the generation of mirtrons which are short hairpin pre-miRNAs directly produced by splicing, thus bypassing Drosha-mediated cleavage (94, 95). Some miRNAs function as bimodal miRNAs controlling different target gene sets depending on the region used for interaction. i.e., a canonical seed in positions 2–8 or positions nt 6–12, e.g., miR-4728-3p, encoded in intron 24 of HER2 gene (96) which downregulates ESR1 expression through an internal seed interaction (97).

Just like protein-coding genes, complexity of the miRNome has increased with further research. miRNAs are heterogeneous in length and sequence with isomiRs that are sequence variants of the canonical miRNA currently in the miRBase generated from a single miRNA locus by template and non-template variants (98). Templated isomiRs match the genomic sequence, but have different 5′-start and/or 3′-ends, resulting from imprecise Drosha or Dicer cleavage (99), whereas non-templated isomiRs diverge from the genomic sequence due to post-transcriptional enzymatic modification. The most common non-templated modification is adenylation, catalyzed by the adenosine deaminase (ADAR) family of enzymes (100). The expression isomiRNAs is dynamic, with differences between cell types and tissues. A tool called IsomiRage http://cru.genomics.iit.it/Isomirage/ is available for profiling the miRNAs/isomiRs and corresponding differential expression patterns using Illumina next-generation sequencing datasets of small RNA (99). When applied to primary breast normal and cancer cells the IsomiRage increased the number of detected miRNA species by ~40%, thus revealing additional information “hidden” in sequencing datasets (99). These isomiRNAs are effectively loaded on AGO/RISC complexes and thus are thought to function as canonical miRNAs, thus increasing the repertoire of mRNA targets.

Not only are miRNAs active in the cells in which they are transcribed, but miRNAs circulate in exosomes: 40–100 nM membrane-bound vesicles composed of different growth factors, cytokines, lipids, cytoplasmic proteins, and nucleic acids, including miRNAs, which circulate in the blood and lymph and deliver molecules between tissues (101). The exosomal content is tightly regulated by endosomal sorting complexes required for transport (ESCRT) (102). Specific cell surface markers allow cellular uptake of exosomes with high specificity. The physiological role of exosomes is controversial. Exosomes can facilitate tumor progression by supplying tumor niches with factors that favor proliferation, invasion, drug resistance, and metastasis (101). Circulating miRNAs embedded in exosomes reprogram cellular mechanisms in recipient cells (103, 104). Whether exosomal miRNAs will be makers in cancer is currently speculative. A recent study appears to be the first comparison between cell-free and exosomal miRNAs in breast cancer patients and healthy women (105). The authors reported higher exosomal miR-372 and cell free (not exosomal) miR-373 in triple negative breast cancer compared to luminal breast cancer patients and higher cell free miR-101 in both groups (102).

5. miRNA-mRNA interaction

The critical, perfectly complementary basepairing between 7 to 8 nucleotides at the 5’ end of the miRNA and its target mRNA is referred to as the ‘seed sequence’. Base pairing of the miRNA-RISC complex within the ORF requires almost perfect complimentarity and the mRNA is either degraded or translation is blocked (85). RNA binding proteins (RBP), e.g. HuR, hnRNP E1, and hnRNP L, and miRNAs compete and collaborate to regulate mRNA stability and RBPs can recruit miRNA-containing RICSs to target lncRNAs (106). There is evidence that miRNA-mRNA gene silencing occurs in the rough endoplasmic reticulum (RER) by interaction of components of Dicer, TRBP and PACT with the RER (107).

Most commonly, because of imperfect base pairing between the miRNA and the 3’UTR, the RISC complex causes translational repression by interaction with eIF6 which prevents 80S ribosomal assembly (108) or by inhibition of translation (18). The exact mechanisms of translational inhibition versus mRNA degradation have not yet been fully elucidated (109). miRNAs initiate target mRNA degradation by recruiting mRNA decay pathway effectors such as de-adenylation and de-capping enzymes (110). The miRNA-containing ribonucleoprotein particle (miRNP)-silenced mRNA is directed to the P-bodies and the mRNA is either released from its inhibition upon a cellular signal and/or actively degraded (111). Some miRNAs may also increase translation of select mRNAs in a cell cycle-dependent manner (112).

miRNAs are considered highly stable, although this is cell-type, cell cycle, and miRNA- specific; further target regulation can promote miRNA’s 3’-end uridylation and degradation (106). This means that an increase in target mRNA leads to a decrease in its target miRNAs. miRNAs are regulated by competing endogenous RNAs (ceRNAs) (113) which contain miRNA target sties and thus act as miRNA ‘sponges’ and sequester miRNAs from interaction with target mRNAs. Circular RNAs (circRNAs) are ceRNAs that contain miRNA binding sites and are resistant to miRNA-mediated destabilization (reviewed in (114)). Multiple non-coding RNA species, including sncRNAs, pseudogenes, lncRNAs and circRNAs appear to possess ceRNA activity (114).

miRNAs have important roles in regulating cellular processes including replication, differentiation, and apoptosis. In cancer, miRNAs can either act as ‘oncosuppressor miRNAs’ which are often downregulated in cancer, e.g., the miR29b-1/a in acute myeloid leukemia resulting in upregulation of oncoprotein BCL-2 (115),or, as ‘oncomiRs’, by decreasing the levels of tumor suppressor proteins, e.g., miR-21 decreasing PDCD4 (116). MiRNAs are expressed in a tissue-specific manner (117). Each miRNA targets ~ 200 transcripts directly or indirectly (118), but the bone fide physiological targets of the vast majority of miRNAs remain to be experimentally verified.

6. HITS-CLIP to identify miRNA-mRNA interaction by Ago2 immunoprecipitation

High-throughput RNA-seq isolated by crosslinking immunoprecipitation (HITS-CLIP) of Argonaute 2 (Ago 2, catalytic component of the RISC complex (119)) is used to identify putative miRNA-mRNA ternary complexes (120, 121). HITS-CLIP of E2-treated MCF-7 cells revealed Ago 2 footprints throughout ESR1 mRNA, including peaks in the 3’UTR and within the coding region, and follow-up experiments identified miR-9-5p binding the 3’ UTR, directly downregulating ERα protein levels (122).

7. Nomenclature of miRNA

miRNAs are preceded a three lettered prefix indicating the species of origin e.g., hsa for homo sapiens and mmu for mouse (123). miRNAs originating from different genomic loci are assigned a numerical suffix, i.e, hsa-miR-29b-1 and hsa-miR-29b-2. If transcripts are equally expressed they are referred to as miR-21-5p (from the 5’ arm) and miR-21-3p (from the 3’ arm) arise from the same hairpin precursor. Alternatively, miR-21* indicates the less predominant species in RISC (124). miRNAs differing by a few bases are given a lettered suffix, e.g.,miR-125a and miR-125b. miRNA families arise from a common ancestor and have similar sequences, e.g. miR-221 and miR-222 family. 61% of mammalian miRNAs are expressed from polycistronic clusters. reflecting shared biological functions for unrelated miRNAs in the same primary transcript (125). miRNA clusters arise due to gene duplication, e.g., the miR-200 cluster of miRNAs are located in two chromosomes, i.e., miR-200a, miR-200b, and miR-429 are located on chromosome 1 and miR-200c and miR-141 are located on chromosome 12 (126). Each cluster is transcribed into a common precursor RNA.

8. Regulation of miRNA expression

Levels of mature miRNA are regulated transcriptionally and by processing of pri-miRNAs and pre-miRNAs. In the microprocessor complex the ratio of Drosha and DGCR8 are tightly regulated (127). DGCR8 stabilizes Drosha and Drosha cleaves and inactivates DGCR8; providing a tight feedback loop (128). ERα interacts directly with helicases p68 and p72 (which are established ERα coregulators (129)). ERα-p68 interaction was reported to inhibit Drosha complex formation (130), and thus repress pri-miRNA processing. Importantly, this work was recently retracted (131). However, another group of investigators also reported that Drosha and p68/DDX5 could be co-purified with ERα in MCF-7 cells, but not with ERβ in ERβ-stably transfected MCF-7 cells (132). This report has not been confirmed.

Dicer processes pre-miRNA to mature miRNA. Dicer activity is enhanced by MAPK-phosphorylation of TRBP (Figure 2) which promotes miRNA processing (133). The RNA coactivator SRA (steroid receptor RNA activator) binds Dicer complex components PACT, TRBP, and PKR in various cell lines and also binds NRs, including ERα (134). Dicer acts as a NR coactivator in MCF-7 cells and is recruited to the PSA gene promoter in DHT-treated LNCaP prostate cancer cells with androgen receptor (AR) (134). These findings suggest that pre-miR processing may be coupled with ERα and AR regulation of gene transcription.

AGO2 is the catalytic component of the RISC complex and serves as a platform to recruit additional regulators of mRNA stability (125). AGO2 is regulated at the transcriptional and post-transcriptional level. For example, in MCF-7 breast cancer cells, E2 inhibits AGO2 expression by activating epidermal growth factor (EGF)-MAPK signaling (135). Direct interaction of EGF receptor (EGFR) with AGO2 in the cytoplasm phosphorylates AGO2 at Tyr 393 which reduced AGO2 association with Dicer (Figure 2) and TRBP suppresses maturation of specific tumor suppressor miRNAs under hypoxic conditions (136).

Nucleolin is a multifunctional protein concentrated in the nucleolus, but located throughout the cell, including the plasma membrane, and has roles in transcription, ribosome biogenesis, DNA replication, chromatin remodeling, apoptosis, and macropinocytosis (137, 138). There are several examples of nucleolin functioning as a transcription factor or as a coregulator through its interactions with other proteins (reviewed in (139)). Nucleolin was reported to promote the maturation of specific miRNAs implicated in carcinogenesis in MCF-7 and HeLa cells: miR-21, miR-103, miR-221, and miR-222 (140).

9. Estrogen regulation of miRNA expression overview

Regulation of miRNA expression by estrogens in animals, fish, and humans has been reviewed by us (141, 142) and others (143). Since my previous review, a non-inclusive list of new studies of E2 regulation of miRNA expression in animals includes: female Fischer 344 rat brain, specifically in the ventral and dorsal hippocampus, central amygdala, and paraventricular nucleus and as a function of aging (144); in female ACI rats in an E2-induced mammary carcinogenesis model (145); mouse aorta (146); mouse liver and primary murine hepatocytes (147); rat cardiac fibroblasts (148). I will not review these studies, but will focus on human cell lines and tissues.

10. ERα and ERβ regulate miRNA expression in a ligand-independent manner

ChIP studies have shown that ‘unliganded’ ERα (149)and ERβ (150) bind DNA in cells grown in serum-free or charcoal-stripped serum medium. Overexpression of ERα in MCF-7 cells upregulated mIR-17 (151). Overexpression of ERβ in non-hormone treated MCF-7 and ZR-75.1 human breast cancer cell lines was reported to regulate the expression of > 450 miRNAs in next-gen RNA sequencing experiments (152). Here I will focus on updating reports on ER ligand-responsive regulation of miRNA expression in human cell lines and tissues.

11. E2 and other ER ligands regulate miRNA expression in human cell lines and tissues

The hope of current studies of E2 regulation of miRNA expression in breast cancer cell lines is that identification of E2-regulated miRNAs and their gene targets may offer insight into mechanisms of estrogen in breast carcinogenesis and progression and identify targets for therapeutic interference. By far and large, E2 regulation of the transcriptome, including miRNAs is best characterized in breast cancer cell lines with MCF-7 studies predominant. This will be apparent in Tables 1 and 2 which summarize the regulation of miRs and their bona fide targets by ER ligands including E2, tamoxifen, 4-OHT, and endocrine disruptors in human cell lines and tissues. It is worth noting that there are conflicting results of E2 and other ER ligand regulation of miRNAs within cell lines, e.g., MCF-7 and T47D, between reports from different investigators and even within the same lab group in different publications. There are many likely explanations for these differences including cell lines and variations in cell treatment conditions, circadian regulation of ERα expression (153), normalization of data (154), and control genes used for qPCR (155).

Table 1

miRNAs upregulated by estradiol (E2), tamoxifen (TAM), 4-hydroxytamoxifen (4-OHT), Fulvestrant (ICI 182,780), or endocrine-disrupting chemicals (EDC) in animal studies and human cell lines

The bona fide targets of the miRNAs are experimentally proven in the reference cited; however, this direct targeting is not necessary substantiated in E2 regulation in the cells indicated in column 3. DIANA-TarBase v7.0 (228) web site has a list of bona fide targets of miRNAs : http://diana.imis.athena-innovation.gr/DianaTools/

miRNA	Ligand	Human cell line/tissue	Comments	Bona fide targets
Let-7a,b,c,d,e,f, g, i	E2	MCF-7 cells stably expressing a biscistronic vector control (157). MCF-7 cells (141, 229).1 µM E2 in Ishikawa and ECC-1 ERα+ human endometrial cancer cells (158).Let-7a and let-7f-1* were increased at 6,12, and 72 h but decreased at 24 h with 10 nM E2 in MCF-7 cells (159).Let-7a* was increased in response to 10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).	Oncosuppressor miR- stimulate apoptosis (230)	DICER1 (231); let-7g:COL1A2 (232)
miR-7	E2	10 nM E2 MCF-7 cells (141, 233)	oncomiR	XRCC2 (234)KLF4 (235)
miR-10amiR-10b	E2	10 nM E2 24 h ERβ stably expressing SW480 colon cancer cells (236).	miR-10b is down-regulated in breast tumors and upregulated in sera (237).	BUB1, PLK1, CCNA2 (238)
miR-15a	E2	10 nM E2 MCF-7 cells (141).	Upregulated by E2F1 (239).	CCNE1 = CyclinE (239)
miR-16-1*	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-16-2*	E2	10 nM E2 for 24 h in T47D cells (154).
miR-17*	E2	10 nM E2 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).
miR-17-3p	E2	MCF-7 stably transfected to overexpress the aromatase gene (MCF-7aro) (240).
miR-17-92	E2	MCF-7 cells (233, 241, 242).	miR-17-92 cluser encodes miR-17, 18, 19, 20, 19b-1, 92-1	miR-19a and miR-92a: PTEN (243)
miR-18a	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-18a*	E2	10 nM E2 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).	miR-18a is higher in ERα-breast tumors (244)	ERα (241)
miR-18b	E2	10 nM E2 6, 12, 24, and 72 h in MCF-7 cells (159).10 nM E2 for 6, 12 h in MCF-7 cells stably overexpressing inducible ERβ or ERα-downregulated at 24 and 72 h (132).
miR-18b*	BPA	10 µM BPA for 18 h in MCF-7 cells (161)
miR-19a, 19b	E2	10 nM E2 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159)miR-19a and 19a* were increased by 10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-19b-1	E2	10 nM E2 6, 12, 24, and 72 h in MCF-7 cells (159).
miR-19b	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-20a*	E2	10 nM E2 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-21	Fludioxonil fenhexamid4-OHT	MCF-7 cells (176)MCF-7 cells (156)	oncomiRFludioxonil and fenhexamid are endocrine disruptors	NFIB (245); PTEN, PDCD4 (156); RASA1 and RASA2 (148)
miR-22	E2EDC	1 nM E2, 1 µM triclosan or 1 µM triclocarban for 18 h in MCF-7 cells (178).	EDC
miR-23b*	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ but not ERα (132).
miR-24	E2	1 nM E2 for 18 h in MCF-7 cells (161).
miR-24-1*	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ but not ERα (132).
miR-25	E2	MCF-7 cells (141, 233).	miR-106b-25 cluster encodes miR-106b, miR-93, and miR-25in the 13th intron of the MCM7 gene (246)	BIM (247); DR4 (248); MCU (249); Smad7 (250); LATS2 (251); RECK (252)
miR-25*	E2	10 nM E2 12 and 24 h in MCF-7 and ZR-75-1 cells (159).
miR-26a	E2 and fulvestrant	Primary human myometrial smooth muscle cells (MSMC) (253)	Oncosuppressor miR	ESR1 (254)CHD1, GREB1, and KPNA2 (255)
miR-27a	E2	1 µM E2 in Ishikawa and ECC-1 ERα+ human endometrial cancer cells (158).	OncomiR	EGFR (256)
miR-27b	E2	MCF-7 cells (233).	Oncosuppressor miR	Sp1 (257); LIMK1 (258); PPARγ (259)
miR-29a	E2	MCF-7 cells (233).	OncomiR: stimulates migration and invasion; Repressed by c-myc, YYI, NFκB, CEBPA and stimulated by p53 (260)	BCL2, CDC42, CDK6, DNMT, MCL1, Osteonectin, TGFβ3m, TTP, TGF-β1, TGF-β2, TTP (260)
miR-29b-2*	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ but inhibited by ERα (132).
miR-29c	E2	10 nM E2 for 24 h in T47D cells (154).
miR-30b	E2	MCF-7 cells (141)	Oncosuppressor miR	CCNE2(261); KRAS, PIK3CD and BCL2(262)
miR-30d	E2	1 µM E2 in Ishikawa ERα+ human endometrial cancer cells (158).10 µM BPA for 18 h in MCF-7 cells (161).
miR-32	E2	10 nM E2 72 h in MCF-7 cells stably overexpressing inducible ERβ (132).
miR-33a	E2	10 nM E2 6, 12, 24, and 72 h in MCF-7 cells (159).
miR-92	E2	10 nM E2 24 and 72 h in MCF-7 cells (159)
miR-92a	E2	1 µM E2 in ECC-1 ERα+ human endometrial cancer cells (158).1 nM E2 for 18 h in MCF-7 cells (161).
miR-92a-1*	E2	10 nM E2 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159)
miR-92b	E2	1 nM E2 for 18 h in MCF-7 cells (161).
miR-93	E2	10 nM E2 24 h in MCF-10A and T47D cells (263). 1 nM E2 for 18 h in MCF-7 cells (161).
miR-98	E2BPA	MCF-7 cells (141).10 µM BPA for 18 h in MCF-7 cells.
miR-99b	E2	1 nM E2 for 18 h in MCF-7 cells (161).
miR-101	E2	10 nM E2 24 h in MCF-7 cells (264).
miR-101*	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ (72 h) but not ERα (132).
miR-103	E2	1 nM E2 for 18 h in MCF-7 cells (161).
miR-122	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-124	E2	MCF-7 cells (233).	Oncosuppressor miR	Ets1 (265)miR-124-5p: LAMB1 (266)ROCK1 (267)FLOT1(268)SphK1 (269)CD151 (270)iASPP (271)Slug (272)
miR-130b	E2	MCF-7 cells (242).		TP53INP1 (273); DICER1 (274)
miR-135a	E2	10 nM E2 6 h in MCF-7 cells (264)10 nM E2 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159)
miR-135b	E2	10 nM E2 for 6 and 72 h in ZR-75-1 cells, but no change at 12 or 24 h (159).
miR-142-3p	E2	10 nM E2 6, 12, 24, and 72 h in MCF-7 cells (159).
miR-148	E2	MCF-7 cells (233).	miRNA-148/152 family include miR-148a, miR-148b, miR-152 (275)	PXR, DNMT1, CAND1, BCL2, p27, ACVR1, PETN, WNT10B, MSK1, CDC25B, ROCK1, CCKBR, CCK2R, IGF-1R, IRS1 (275)
miR-149	E2	MCF-7 cells (233).		GSK3α (276)GIT1 (277)AKT and E2F1 (278)
miR-151-5p	E2	1 nM E2 for 18 h in MCF-7 cells (161).
miR-155	E2	100 nM E2 for 48 h in MCF-7 cells (279).Higher levels circulating in the serum of breast cancer patients than healthy women (280).	oncomiR	TRF1 (281).TP53INP1(282)
miR-181a	E2	1 µM E2 in Ishikawa ERα+ human endometrial cancer cells (158).
miR-181d	E2	MCF-7 cells (233)		CCND1 (245)
miR-186	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159).
miR-190	E2	10 nM E2 for 6, 12, and 72 h in ZR-75-1 cells, but not 24 h (159).
miR-190a	E2	100 nM E2 in MCF-7 cells increased ERα recruitment to the miR-190a promoter containing a half-site ERE (283).		PAR-1 (283)
miR-190b	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159).
miR-191	E2	10 nM E2 for 6 h in MCF-7 cells (284).10 nM E2 (24 h) stimulation was inhibited by 100 nM tamoxifen and by siERα and siERβ in MCF-7 cells (285). ERα and ERβ ChIPped to the miR-191 promoter in MCF-7 cells (285). 1 nM E2 for 18 h in MCF-7 cells (161).		EGR1 (284)CDK6, BDNF, and SATB1 (285)
miR-193a-5p	E2	1 nM E2 for 18 h in MCF-7 cells (161).
miR-193b	E2EDC	MCF-7 cells (242)1 nM E2, 1 µM triclosan or 1 µM triclocarban for 18 h in MCF-7 cells (178).		uPA (286); YWHAZ, SHMT2, AKR1C2 (287); miR-193-3p: MYB (288)
miR-194	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159).
miR-195	E2	MCF-7 cells (141)		CCND1 (245)
miR-195*	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells- highest at 6 h (159).		ASF1B, BIM, BCL2L2, CCL5, CADM1, EZH2, FGF$1, HDGF, LTF, MAP2K3, NRAS, PTEN, TP53, TWIST1, XBP1 (and others) (289)
miR-196a2*	E2	10 nM E2 6 h in MCF-7 cells (264)	Mediated by ERα and the protein kinase ERK2 (264). By ChIP assay, both ERα and ERK2 were recruited to chromatin with 45 min 10 nM E2 alone with increased pSer5 RNA pol II recruitment (264).	TP63 (264)
miR-198	E2	10 nM E2 for 24 h in T47D cells (154).
miR-199a/b-3p	E2	10 nM E2 for 12, 24, and 72 h in ZR-75-1 cells, but not at 6h (159).
miR-199a-5p	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-200a	E2	MCF-7 cells (141)		BAP1, PTPRD, KLF11, SEPT7, HOX5B, ERBB2IP, RASSF2, ELMO2, SHC1, VAC14 (DIANA)
miR-200c	none	Endogenous ERα in MCF-10A cells ChIPed to the miR-200c promoter and Overexpression of ERα in MCF-10A cells increased miR-200c expression (290).1 nM E2 for 18 h in MCF-7 cells (161).
miR-203	E2	MCF-7 cells (141)
miR-205	E2	10 nM E2 24 h ERβ stably expressing SW480 colon cancer cells (236).
miR-206	DPNE2EDC	ERβ-selective agonist in MCF-7 cells (181).1 nM E2, 1 µM triclosan or 1 µM triclocarban for 18 h in MCF-7 cells (178)	Oncosuppressor miR
miR-210	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-216a	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-219-5p	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).
miR-222	E2BPA	1 nM E2 or 10 µM BPA for 18 h in MCF-7 cells (161).		KIT (291); PPP2R2A (292); CDKN1C (293); CDK1B (294); DICER1 (229)
miR-223	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).
miR-301b	E2	MCF-7 cells (242)
miR-320	E2	1 µM E2 in Ishikawa and ECC-1 ERα+ human endometrial cancer cells (158).
miR-320a	E2	1 nM E2 or 10 µM BPA for 18 h in MCF-7 cells (161).
miR-320c	E2BPA	1 nM E2 or 10 µM BPA for 18 h in MCF-7 cells (161).
miR-330-5p	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ not ERα (132).
miR-335	E2	10 nM E2 for 6, 12 and 72 h in MCF-7 and ZR-75-1 cells, but not at 24 h (159).
miR-342	E2; Not blocked by 1 µM 4-OHT	MCF-7-HER2 cells, MCF-7 cells stably overexpressing HER2, but still tamoxifen-sensitive (220)
miR-363	E2	10 nM E2 for 12 and 24 h in ZR-75-1 cells, but not 6 or 72 h (159).
miR-365	E2	MCF-7 cells (141)
miR-374a*	E2	10 nM E2 for 6, 12 and 72 h in MCF-7 and ZR-75-1 cells, but repressed > 1.5 fold at 24 h (159).
miR-375	E2	10 nM E2 for 24 and 72 h in ZR-75-1 cells, but not 6 or 12 h (159).
miR-376b	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells- highest at 6 h (159).
miR-423-5p	E2	1 nM E2 for 18 h in MCF-7 cells (161).
miR-424	E2	MCF-7 cells (165)10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159).
miR-424*	E2	10 nM E2 for 6, 12 and 72 h in MCF-7 and ZR-75-1 cells, but not at 24 h (159).
miR-425	E2	1 µM E2 in Ishikawa and ECC-1 ERα+ human endometrial cancer cells (158).10 nM E2 for 6 h in MCF-7 cells (284).		EGR1 (284)
miR-449a	E2	10 nM E2 for 6, 12and 24 h in ZR-75-1 cells, but not 72 h (159)
miR-450b-3p,5p	E2	10 nM E2 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells- highest at 72 h (159)
miR-455-5p, 455-3p	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159)
miR-484	E2	1 nM E2 for 18 h in MCF-7 cells (161).
miR-489	E2	10 nM E2 12, 24, and 72 h in MCF-7 and ZR-75-1 cells, but not at 6 h(159)
miR-491-3p	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-499-5p	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-515-5p	tamoxifen	100 nM tamoxifen for 48 h ~ 25% decrease in MCF-7 cells (295).		SK1 (295)
miR-520d	E2	MCF-7 cells stably expressing a constitutively active AKT (157)
miR-542-5p	E2	10 nM E2 for 72 h in MCF-7 cells (159)
miR-542-3p	E2	10 nM E2 for 72 h in MCF-7 and ZR-75-1 cells (159)
miR-548d-3p	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ, but not ERα (132).
miR-548e	E2	10 nM E2 for 6, 24, and 72 h in ZR-75-1 cells (159).
miR-550	E2	10 nM E2 for 72 h in MCF-7 cells (159).
miR-556-5p	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells, but not at 24 h (159).10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-560:9.1	E2	10 nM E2 6, 12, 24, and 72 h in MCF-7 cells (159).
miR-564	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-574-5p	E2PhIP	1 µM E2 in Ishikawa ERα+ human endometrial cancer cells (158).10 nM E2 or 100 nM PhIP for 4, 8, 12, or 24 h in MCF-7 cells (160).
miR-574-3p	E2 or PhIP	10 nM E2 or 100 nM PhIP for 4, 8, 12, or 24 h in MCF-7 cells (160).
miR-579	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159)
miR-590-3p	E2	10 nM E2 highest stimulation at 6, 12 and 72 h in ZR-75-1 cells with no change detected at 24 h (159)
miR-594:9.1	E2	10 nM E2 6, 12, 24, and 72 h in MCF-7 cells (159)
miR-615-3p	E2	10 nM E2 6 h in MCF-7 cells (264)
miR-628-5p	E2	10 nM E2 for 12, 24, and 72 h in ZR-75-1 cells (159).
miR-638	E2	1 nM E2 for 18 h in MCF-7 cells (161).
miR-643	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-651	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-652	E2	10 nM E2 for 24 and 72 h in ZR-75-1 cells, but not at 6 or 12 h (159).
miR-653	E2	10 nM E2 for 72 h in MCF-7 and ZR-75-1 cells(159).
miR-653:9.1	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).
miR-660	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-663	E2	1 nM E2 for 18 h in MCF-7 cells (161).
miR-663b	E2	10 nM E2 for 6 and 24 h in ZR-75-1 cells (159).10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-708	E2	10 nM E2 for 12, 24, and 72h in ZR-75-1 cells, but not at 6 h (159)
miR-720	E2	1 nM E2 for 18 h in MCF-7 cells (161).
miR-760	E2	24 h and 3d in MCF-7 cells (165).10 nM E2 for 24 and 72 h in MCF-7 and ZR-75-1 cells (159).
miR-886-3p	E2	10 nM E2 for 24 h in MCF-7 and ZR-75-1 cells, but not at 6, 12, or 72 h (159).
miR-938	E2	10 nM E2 for 6 h in MCF-7 cells (66).
miR-939	E2	10 nM E2 for 72 h in MCF-7 cells (159)
miR-940	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159).
miR-942	E2	10 nM E2 for 72 h in MCF-7 and ZR-75-1 cells, but not 6, 12, or 24 h (159).
miR-944	E2	10 nM E2 for 6 h in MCF-7 cells (66)
miR-1206	E2	10 nM E2 for 72 h in MCF-7 cells (159)
miR-122	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159).
miR-1248	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159).
miR-1268	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159),
miR-1275	E2	1 nM E2 for 18 h in MCF-7 cells (161).
miR-1305	E2	10 nM E2 for 12 and 72 h in MCF-7 and ZR-75-1 cells (159)
miR-1323	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-1826	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-1915	E2BPA	1 nM E2 for 10 µM BPA for 18 h in MCF-7 cells (161).

Table 2

Estradiol- and tamoxifen- inhibited miRNAs

This table lists miRNAs whose expression is decreased by E2, tamoxifen, or 4-OHT. MCF-7, T47D, ZR-75-1, BT-474, and BG1 are ERα positive breast cancer cells.

miRNA	Ligand	Species/tissue/cell line	Comments	Bona fidetargets
Let-7g, -7f, -7a, -7c	E2	10 nM E2 48 h in MCF-7 cells; Also repressed in T47D, ZR-75-1, BT-474, and BG1, but not SKBR3 breast cancer cells (296).10 nM E2 6 h in MCF-7 cells (141).Let-7g in MCF-7 cells (297).10 nM letrozole stimulated Let-7 expression in MCF-7 cells co-cultured with primary human stromal cells (167).1 nM E2 for 18 h in MCF-7 cells (161).	Blocked by fulvestrant	GAB2; FN1 (297)
Let-7b	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).
Let-7f	4-OHT	1 µM 4-OHT for one month in MCF-7 cells (298)
Let-7i	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-7-1	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159).
miR-9, miR-9-d	E2	10 nM E2 for 24 h in ER β stably expressing SW480 colon cancer cells (236).
miR-15a*	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-16	E2	10 nM E2 for 6, 24, and 48 h in MCF-7 cells; blocked by pretreatment with 1 uM ICI 182,780 (299).1 nM E2 for 18 h in MCF-7 cells (161).
miR-16-1*	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-17	E2	10 nM E2 for 24 h ER β stably expressing SW480 colon cancer cells (236).	Oncosuppressor miR206
miR-17*	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ but increased by ERα (132).
miR-18a, miR-18b	E2	10 nM E2 24 h ER β stably expressing SW480 colon cancer cells (236)
miR-19a, 19b	E2	10 nM E2 24 h ER β stably expressing SW480 colon cancer cells (236).
miR-20a	E2	24 h 10 nM E2 in isolated human endometrial glandular epithelial cell; blocked by ICI 182,780 (300).10 nM E2 for 24 h ERβ stably expressing SW480 colon cancer cells (236).
miR-21	E2	24 h 10 nM E2 in isolated human endometrial glandular epithelial cells and in Primary human leiomyoma smooth muscle cells (LSMC) (253)10 nM E2 for 48 h in MCF-7 cells (296) (181). 10 nM E2 6 h: ~ 60% reduction in miR-21 in MCF-7 cells (156)10 nM E2 for 12 or 24 h in MCF-7 cells (264). 10 µM E2 for 24 h in MCF-7 cells, no effect in MDA-MB-231 cells (301). 10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159). 10 nM E2 or 100 nM PhIP for 24 h in MCF-7 cells (160). 1 nM E2 for 18 h in MCF-7 cells (161).	blocked by ICI 182,780 isolated human endometrial glandular epithelial cellsERα or ERK2 knock-down reduced E2-downregulation of miR-21 expression(264)	PTEN, PDCD4 (156)JAG1 (301)
miR-22, 22*	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159).
miR-23a.23b	E2	10 nM E2 48 h in MCF-7 cells; Also repressed in T47D, ZR-75-1, BT-474, and BG1, but not SKBR3 breast cancer cells (296).miR-23a: 10 nM 3 h in MCF-7 cells (302) and 10 nM E2 for 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).
miR-24	E2	10 nM E2 48 h in MCF-7 cells; Also repressed in T47D, ZR-75-1, BT-474, and BG1, but not SKBR3 breast cancer cells (296)
miR-25	E2	10 nM E2 for 24 h ERβ stably expressing SW480 colon cancer cells (236).
miR-26a	E2	24 h 10 nM E2 LSMC(253).1 nM E2 for 18 h in MCF-7 cells (161).
miR-26a-2*	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).
miR-26b	E2	10 nM E2 48 h in MCF-7 cells; Also repressed in T47D, ZR-75-1, BT-474, and BG1, but not SKBR3 breast cancer cells (296).10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).1 nM E2 for 18 h in MCF-7 cells (161).
miR-27a*	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-27b	E2	10 nM E2 48 h in MCF-7 cells; Also repressed in T47D, ZR-75-1, BT-474, and BG1, but not SKBR3 breast cancer cells (296).10 nM E2 for 72 h in MCF-7 cells (159)	Oncosuppressor miR
miR-29a	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-29a*	E2	10 nM E2 6, 12, 24, and 72 h in MCF- 7 cells (159)
miR-29b-1*, 29b-2*	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159)
miR-30a	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ (132).	ERβ ChIPed to the promoter (132).
miR-30c-2*	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-30d	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159).10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132). ERα was more inhibitory than ERβ.
miR-34a	E2	10 nM E2 for 24 h MCF-7 cells (303)10 nM E2 for 6 h in HUVEC, LNCaP, C38IM, and C27IM human prostate cancer cells (304).Higher levels circulating in the serum of breast cancer patients than healthy women (280).10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159)	Oncosuppressor miR- stimulate apoptosis (230)	LMTK3 (303)SIRT1 (305)
miR-92a	E2	10 nM E2 24 h ERβ stably expressing SW480 colon cancer cells (236)
miR-99a	E2	10 nM E2 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159)
miR-99b	E2	10 nM E2 for 6,12, 24 and 72 h in ZR-75-1 cells, most repressed at 72 h (159).10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-105-2	4-OHT	1 µM 4-OHT for one month in MCF-7 cells (298)
miR-106	E2	10 nM E2 24 h ERβ stably expressing SW480 colon cancer cells (236)
miR-106b	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159).1 nM E2 for 18 h in MCF-7 cells (161).
miR-107	E2	10 nM E2, for 6, 12, 24 h and 3 d in MCF-7 cells (306).10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159).
miR-125a-3p	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).	Oncosuppressor miR
miR-125a	4-OHT	1 µM 4-OHT for one month in MCF-7 cells (298)
miR-125b-2*	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159).	Oncosuppressor miR	BAK1, BCL2, DICER1, ERBB2, ERBB3, ETS1, FGFR2, IL6R, JUN, LIN28A, LIN28B, MCL1, MUC1, NCOR2, SIRT7, STAT3, TNF, TP53 (and others)(289)
miR-128a:9.1	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159).	oncomiR
miR-130b*	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).
miR-132*	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159).
miR-135a	E2	10 nM E2 for 24 h ERβ stably expressing SW480 colon cancer cells (236).10 nM E2 24 h in MCF-7 cells (66).
miR-139-5p	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-140	E2	10 nM E2 for 24 h in ERα-stably transfected MCF-10A cells (307).ERα binds the miR-140 promoter in E2 or BPA-treated MCF_7 cells.		SOX2 (307)
miR-140-5p	E2	10 nM E2 for 24 h ERβ stably expressing SW480 colon cancer cells (236)
miR-141	E2 or PhIP	10 nM E2 or 100 nM PhIP for 24 h in MCF-7 cells (160).
miR-142-3p	E2	10 nM E2 24 h ERβ stably expressing SW480 colon cancer cells (236)
miR-143	E2	10 nM E2 for 6, 24, and 48 h in MCF-7 cells; blocked by pretreatment with 1 uM ICI 182,780 (299).
miR-148b*	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-149	E2	10 nM E2 6 h in MCF-7 cells (141)
miR-142-3p	E2	10 nM E2 24 h ERβ stably expressing SW480 colon cancer cells (236)
miR-146b-5p	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159).
miR-181a, 181b, 181d	E2	10 nM E2 48 h in MCF-7 cells; Also repressed in T47D, ZR-75-1, BT-474, and BG1, but not SKBR3 breast cancer cells (296).miR-181a and 181b inhibited by 100 nM E2 in MCF-7 cells (163).
miR-181	4-OHT	100 nM 4-OHT for 6 h in MCF-7 cells (155).
miR-181a*, 181c*	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).miR-181c* 10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-181c	E2	1 nM E2 for 18 h in MCF-7 cells (161).
miR-183	E2	10 nM E2 24 h ERβ stably expressing SW480 colon cancer cells (236).1 nM E2 for 18 h in MCF-7 cells (161).
miR-185*	E2	10 nM E2 for 12 and 72 h in ZR-75-1 cells, but not 6 or 24 h (159).
miR-186	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ, but increased by ERα (132).
miR-192	E2	10 nM E2 24 h ERβ stably expressing SW480 colon cancer cells (236)
miR-193a	E2	10 nM E2 48 h in MCF-7 cells; Also repressed in T47D, ZR-75-1, BT-474, and BG1, but not SKBR3 breast cancer cells (296).
miR-193a-3p	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159)
miR-193b*	E2	10 nM E2 for 72 h in ZR-75-1 cells (159).
miR-194	E2	10 nM E2 24 h ERβ stably expressing SW480 colon cancer cells (236).
miR-194b*	E2	10 nM E2 for 72 h in ZR-75-1 cells (159).
miR-196a	E2	10 nM E2 24 h ERβ stably expressing SW480 colon cancer cells (236).
miR-196b	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-199a/b-3p	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ but increased by ERα (132).
miR-199b-5p	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα-except that 24 h of E2 increased miR-199b-5p in ERα-MCF-7 cells (132).
miR-200a	E2	10 nM E2 24 h ERβ stably expressing SW480 colon cancer cells (236)10 nM E2 6 h MCF-7, LCC1, and LCC2 breast cancer cells (214).
miR-200b	E24-OHT	10 nM E2 24 h ERβ stably expressing SW480 colon cancer cells (236).10 nM E2 6 h MCF-7, LCC1, LCC2, and LCC9 breast cancer cells (214).500 nM 4-OHT for h in ECC-1 and Ishikawa endometrial cancer cells (308).	4-OHT induced c-Myc that inhibited miR-200a, miR-200b, and miR-429 transcription (308).miR-200b promoter P2 is hypermethylated in primary breast tumors and ERα-negative cell lines (309).	ZEB2 (308)
miR-200c	E24-OHT	10 nM E2 for 6 h in MCF-7 cells (141)10 nM E2 for 6 h MCF-7, LCC1, LCC2, and LCC9 breast cancer cells (214). 500 nM 4-OHT for h in ECC-1 and Ishikawa endometrial cancer cells (308).		ZEB2 (308)
miR-203	E2	10 nM E2 for 6, 24, and 48 h in MCF-7 cells; blocked by pretreatment with 1 uM ICI 182,780 (299).1 nM E2 for 18 h in MCF-7 cells (161).
miR-204	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-205	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).	Oncosuppressor miR
miR-206	1 nM E2 or 10 nM PPT (an ERα-selective agonist)	MCF-7 cells (181).	80% reduction in expression with 24 h treatment
miR-218	E2	10 nM E2 for 24 and 72 h in MCF-7 cells (159).
miR-220c	E2	10 nM E2 for 24 h in T47D cells (154).
miR-221	E2	10 nM E2 for 24 h ~ 80% reduction in MCF-7 and T47D cells (294).Repressed by ERα knockdown10 nM E2 48 h in MCF-7 cells (202).10 nM E2 24 h ERβ stably expressing SW480 colon cancer cells (236).1 nM E2 for 18 h in MCF-7 cells (161).	pro-metastatic/pro-proliferative	ESR1 = ERα (reviewed in (310))
miR-221*	E2	10 nM E2 for 12, 24, and 72 h in ZR-75-1 cells (159).
miR-222	E2BPA	10 nM E2 for 24 h ~ 80% reduction in MCF-7 and T47D cells (294).Repressed by ERα knockdown10 nM E2 for 48 h in MCF-7 cells (202).
miR-223	E2	10 nM E2 for 3 h in MCF-7 cells (302)
miR-301a	E2	10 nM E2 24 h ERβ stably expressing SW480 colon cancer cells (236)
miR-320bmiR-320d	E2	1 nM E2 for 18 h in MCF-7 cells (161).
miR-328	E2	10 nM E2 6 h in MCF-7 cells (141).10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-330-5p	E2PhIP	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).10 nM E2 or 100 nM PhIP for 4, 8, 12, or 24 h in MCF-7 cells (160).
miR-338-3p	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159). 10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-342	E2	10 nM E2 for 6 h in MCF-7 cells (141).
miR-345	E2	10 nM E2 for 72 h in ZR-75-1 cells (159).
miR-362-5p	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-365	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-374b*	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-375	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-376a	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-377	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-379	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-429	4-OHT	500 nM 4-OHT for h in ECC-1 and Ishikawa endometrial cancer cells (308).
miR-451	tamoxifen	1 µM tamoxifen repressed by 4 h and 90% at 24 h (311).	Expression ~ 2-fold lower in tamoxifen-resistant MCF-7 cells (311)
miR-487b	E2	10 nM E2 for 6,12, and 72 h in ZR-75-1 cells, but no significant expression at 24 h (159).
miR-499	E2	10 nM E2 for 48 h in MCF-7 cells; Also repressed in T47D, ZR-75-1, BT-474, and BG1, but not SKBR3 breast cancer cells (296).
miR-504	E2	10 nM E2 for72 h in MCF-7 and ZR-75-1 cells (159).
miR-	E2	10 nM E2 for 24 h in MCF-7 cells (159)
miR-515-5p	E2	10 nM E2 48 h in MCF-7 cells mediated by ERα binding (295).		SK1 (295)
miR-518c*	E2 or PhIP	10 nM E2 or 100 nM PhIP for 4, 8, 12, or 24 h in MCF-7 cells (160).
miR-520d	E2	10 nM E2 48 h in MCF-7 cells; Also repressed in T47D, ZR-75-1, BT-474, and BG1, but not SKBR3 breast cancer cells (296)
miR-548g	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159)
miR-570	E2	10 nM E2 for 6, 12, 24 h and 3 d in MCF-7 cells (306).10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159).
miR-574-3p	4-OHT	1 µM 4-OHT for one month in MCF-7 cells (298)		Clathrin heavy chain (CLTC) (298)
miR-579	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-582-3p	E2	10 nM E2 for 12, 24, and 72 h in ZR-75-1 cells (159).
miR-583-5p	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).
miR-584	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).
miR-589	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-590-5p	E2	10 nM E2 24 h ERβ stably expressing SW480 colon cancer cells (236)
miR-610	E2	10 nM E2 for 6,12, 24 and 72 h in ZR-75-1 cells, most repressed at 72 h (159).
miR-615-5p	E2 or PhIP	10 nM E2 or 100 nM PhIP for 24 h in MCF-7 cells (160).
miR-616	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-618	E2	10 nM E2 for 6, 12, 24 h and 3 d in MCF-7 cells (306). 10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).
miR-632	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-638	E2 or PhIP	10 nM E2 or 100 nM PhIP for 4, 8, 12, or 24 h in MCF-7 cells (160).
miR-646	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159)
miR-650	E2	10 nM E2 for 24 h in T47D cells (154).
miR-663	E2 or PhIP	10 nM E2 or 100 nM PhIP for 4, 8, 12, or 24 h in MCF-7 cells (160).
miR-671:9-1, 671-3p	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).
miR-708*	E2	10 nM E2 for 6, 24, and 72 h in ZR-75-1 cells, but not 12 h (159).10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-874	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).
miR-877	4-OHT	1 µM 4-OHT for one month in MCF-7 cells (298)
miR-935	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-938	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159)
miR-1225	E2	10 nM E2 for 12, 24, and 72 h in ZR-75-1 cells (159).
miR-1228	E2	10 nM E2 for 24 h in T47D cells (154).
miR-1229	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-1234	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159).
miR-1238	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 and ZR-75-1 cells (159) =.
miR-1257	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-1267	E2	10 nM E2 in MCF-7 cells stably overexpressing inducible ERβ or ERα (132).
miR-1301	E2	10 nM E2 for 6, 12, 24, and 72 h in ZR-75-1 cells (159).
miR-1303	E2	10 nM E2 for 12, 24, and 72 h in ZR-75-1 cells (159).
miR-1468	E2	10 nM E2 for 6, 12, 24, and 72 h in MCF-7 cells (159).

Identification of E2- and 4-OHT- regulated miRNAs was originally performed by microarray by us (155, 156) and others (132, 157–161). These reports are summarized in Tables 1 and 2. An Illumina human MicroRNA Expression Profiling Beadchip was used to identify E2-regulated miRNAs in MCF-7 and ZR-75.1 cells after 6, 12, 24, and 72 h of treatment following an initial 4 days of ‘hormone deprivation’ in medium containing 5% dextran-coated charcoal stripped FBS (159). The authors reported 230 significant miRNA changes (up- and down- regulation) that are summarized in Tables 1 and 2. The authors correlated miRNA expression with ERα in vivo binding in published data sets and found ERα binding within 10 kB of miR-125a-2, miR-181c, miR-23a, miR-27a, miR-24-2, and mIR-26 and ERα binding sites within 50kB of genes in which miRs are embedded: miR-25 in MCM2; miR-26a in CTDSP2, miR-424 in GBC16121, miR-618 in LIN7A, miR-760 in BCAR3, and miR-942 in TTF2(159). The authors noted that they found more of miR * strands regulated by E2 and suggested a possible role of ER in strand selection. Since the * strands are now known to be functional in Ago2-RISC complexes (162), these findings appear to reflect the wide range of miRNAs functionally regulating estrogen action in vivo.

GRO-seq (global nuclear run-on and sequencing) identified all RNA transcripts in E2-treated MCF-7 cells (163). The authors identified 119 miRNA transcripts as regulated by E2 at minimally one of the time points (10 and 40 min) examined with half of the miRNAs upregulated and half downregulated, the same as protein-coding transcripts. However, GRO-seq is unable to detect miRNAs that are co-transcribed as a part of their host gene within which they are embedded (164). Another genome wide analysis of E2-regulated miRNA expression was performed in MCF-7 and ZR-75-1 luminal-like breast cancer cells (165). In that study, E2 increased miR-760 and miR-424 and decreased miR-618, miR-570, and miR-107 expression. It will be of interest to correlate binding events, transcriptional regulation, and functional outcome in these large-scale studies.

Aromatase inhibitors are used to inhibit the endogenous synthesis of estrogens in postmenopausal breast cancer patients (166). The aromatase inhibitor letrozole (10 nM) stimulated the expression of let-7f, miR-146a, miR-150, miR-27a, miR-263, miR-19a, miR-372, miR-23b, miR-203, miR-10b, miR-128a, miR-9, and miR-126 and inhibited miR-134, miR-142-5p, miR-96, miR-148b, and miR-222 expression in MCF-7 cells co-cultured with primary human stromal cells (167). If these are E2-regulated miRNAs in MCF-7 cells, then we would expect E2 to increase miR-134, miR-142-5p, miR-96, miR-148b, and miR-222 and inhibit let-7f, miR-146a, miR-150, miR-27a, miR-263, miR-19a, miR-372, miR-23b, miR-203, miR-10b, miR-128a, miR-9, and miR126. We compared these expected results with published data summarized in Tables 1 and 2. E2 has not been reported to increase miR-134, miR-148b, or miR-96; however, in agreement with the expected results, E2 increased miR-142-3p and miR-222 in MCF-7 cells (Table 1). E2 has not been reported to inhibit miR-146a, miR-150, miR-263, miR-372, miR-10b, miR-9, or miR-126; however, E2 reduced let-7f, miR-27a, miR-19a, miR-23b, miR-203, miR-128a:9.1, in MCF-7 cells (Table 2).

12. Endocrine disrupting chemicals regulating miRNA expression

Endocrine disrupting chemicals (EDC) are environmental chemicals that mimic or block transcriptional activation elicited by naturally circulating steroid hormones by binding to steroid hormone receptors and either acting as agonists or antagonists of that receptor (168, 169). EDC may also affect the levels or activities enzymes involved in steroid hormone synthesis or metabolism, alter the expression or activities of transcriptional coregulators, and cause epigenetic changes(170) (168). The role of EDC in breast cancer is suspected, but not proven (171). Based on their widespread use, environmental persistence, the possible role of EDC in hormone-related cancers is of keen interest (168, 171, 172).

There are few reports examining how EDC affect miRNA expression in fish, animals or animal cell lines (173). Treatment of mouse TM4 Sertoli cells with 10 µ g/mL nonylphenol (NP) increased the expression of 47 miRNAs and down-regulated the expression of 100 miRNAs with 24 h of treatment (174). Only 10 miRNAs were increased > 1.5-fold with mmu-miR-135* being increased ~ 4-fold. The authors correlated the increase in miR-135* with decreased expression of 18 mRNAs in NP-treated cells, but did not confirm changes at the protein level or whether these are bona fide mRNA targets of mmu-miR-135a* (174). Neonatal exposure to the estrogenic analog estradiol benzoate (EB) from postnatal days (PND)1–5 with doses of 0, 0.75, 1.25, 2.5, or 25 µ g/d given sc, increased miR-29 (a,b, and c) in adult (PND90) rat testicular tissue with a concordant decrease in miR-29 target Mcl-1 protein (175).

To my knowledge, based on searching PubMed, there are only four studies of the effect of EDC on miRNA expression in human cell lines. One study showed that, like E2 (156), 10 µ M o,p-dichlorodiphenyltrichloroethane (DDT) and 10 µ M bisphenol A (BPA) activate ERα in MCF-7 cells and downregulated miR-21 (161). In addition, the authors reported that treatment of MCF-7 cells with 1 nM E2, 10 µ M BPA, or 10 µ M DDT reduced the expression of let-7a, b, c, d, e, and f, miR-15b, and miR-28b and upregulated miR-638, miR-663, and miR-1915. We reported that the anti-fungal agents fenhexamid and fludioxonil increased miR-21 expression in MCF-7, T47D, and MDA-MB-231 human breast cancer cells and reduced the expression of miR-125b and miR-181a (176). In MCF-7 cells, fenhexamid and fludioxonil induction of miR-21 was inhibited by fulvestrant; by AR antagonist, bicalutamide; by actinomycin D and cycloheximide, and by inhibitors of the mitogen-activated protein kinases (MAPK) and phosphoinositide 3-kinase (PI3K) pathways. Fenhexamid activation was inhibited by the arylhydrocarbon receptor antagonist α-napthoflavone.

The cooking of meat, particularly at high temperature with browning, e.g. grilling on a charcoal grill, results in the formation of heterocyclic amines (HCA), including the most abundant: 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) which is considered a mammary carcinogen (177). Treatment of MCF-7 cells with 100 nM PhIP decreased miR-21, miR-1, and miR-106b expression and increased miR-923, miR-574-3p, miR-574-5p, and miR-494 (160). Other miRNAs regulated by PhIP are listed in Tables 1 and 2.

The antimicrobial agents triclosan (TCS) and triclocarban (TCC) are widely used in many consumer products including soaps, skin creams, toothpastes and deodorants and are present in the aquatic and terrestrial environment (178). TCS and TCC are established EDS that compete with E2 for ERα and ERβ binding, albeit with lower affinity (179). TCS and TCC (each at 1 µ M) increased the expression of miR-22, miR-206, and miR-193b (2–3-fold) in MCF-7 cells, similar to the stimulation with 1 nM E2 (178).

13. miRNAs regulating ER expression

miRNAs can influence estrogen-regulated gene expression by directly reducing ERα mRNA stability or translation. Nine miRNAs have been reported to reduce ERα protein levels: miR-18a, miR 18b, miR-193b, miR-302c, miR-22 (180), miR-201, miR-221, and miR-222 (142), miR-206 (181), miR-222-3p (182), miR-4728-3p (97), miR-373 (105); miR-9-5p (122). let-7a, let-7b, and let-7i (183) (Figure 3). MiR-206 is inversely correlated with ERα expression, but not ERβ, in human breast tumors (184). miR-221/222 is higher in ERα negative than ERα positive breast cancer cell lines and human breast tumors (185, 186). Anti-miR-221 suppressed the growth of TAM-resistant breast cancer cells as xenografts in nude mice (187). Similarly, the expression of miR-22 was significantly lower in MCF-7, T-47D and BT474 ERα-positive versus ERα-negative MDA-MB-231 and SK-BR-3 breast cancer cells (188). A protein lysate microarray (LMA)-based strategy in which a library of pre-miRs was transiently transfected into MCF-7 and BT-474 cells in 384-well plates and ERα protein was subsequently analyzed in protein lysates that were printed on nitrocellulose-coated slides (189). miR-18a, miR-18b, miR -193, miR -206, and miR-302c reduced ERα by directly binding sites in the 3’UTR of ERα. Further, the authors reported an inverse correlation between the expression of miR-18a, -18b and ERα-negative breast tumor samples (189). ERα is upregulated during breast carcinogenesis and cancer stem cells (CSCs) isolated from MCF-7 and T47D cells had increased ERα and decreased let-7a, let-7b, let-7c, let-7d, let-7g levels (190). miR-873 was reported to inhibit E2-ERα-regulated gene transcription and cell proliferation by directly targeting CDK3, thus inhibiting ERα phosphorylation (Ser104, 106, and 118) and thus, ERα activity in MCF-7 cells (191). Stable overexpression of miR-873 in tamoxifen-resistant MCF-7 cells sensitized cells to tamoxifen (191).

Figure 3

Overview of miRNAs regulating ERα and ERβ expression and function

MiRNAs that inhibit ERα, ERβ, and coregulators involved in gene transcription are indicated as discussed in the text.

14. miRNAs that regulate ER coregulators

miRNAs may also affect estrogen-regulated gene expression by reducing the expression of ER-interacting coactivators. miR-17-5p inhibited translation of coactivator SRC-3/AIB1/NCOA3 and reduced E2-ERα-ERE-luciferase activity in transfected cells (192). miR-195 inhibited SRC-3 expression in HepG2 cells by direct interaction with the 3’UTR region (193). There are 3 reports on miRNA regulation of corepressors that target ERα. miR-10a and -10b repress SMRT/NCOR2 (194). miR-184 (195) and miR-16 (196) represses SMRT/NCOR2 translation, but how they affect ER activity is unknown. MTA1 (metastatic tumor antigen 1) repressed miR-661, but the effect on ERα transcription was not evaluated (197). miR-615-3p repressed LCoR expression (198), but whether this affects ERα was not studied. Clearly, little is known about regulation of ER coactivators and corepressors by miRNAs.

15. E2 regulation of AGO2 in human breast cancer cell lines

The expression of Argonaut-2 (Ago2), the catalytic subunit of the RISC complex that mediates miRNA-dependent cleavage/degradation in mammals is higher in ERα-negative, HER2-positive than ERα-positive/HER2 negative (luminal) human breast cancer cell lines and tumors (16). However, E2 and the ERα-agonist PPT, but not the ERβ-agonist DPN, increased AGO2 protein expression in MCF-7 cells (16). Further studies showed that EGF acts through the MAPK pathway to increase Ago2 protein stability, but there were no studies examining the mechanism by which E2 and PPT, presumably through ERα, increase Ago2 protein levels. Surprisingly, Ago2 overexpression in MCF-7 cells increased ERα protein levels by 3-fold, despite also increasing miR-206 that reduces ERα (16). The authors concluded that this “discordant” finding indicates that there is a greater concentration of miRNAs than target proteins involved in ERα suppression than those that target ERα itself” (16). Microarray profiling shows that the expression of Ago1 and Ago2 proteins is higher while Dicer and TRBP1 is lower in ERα-negative versus ERα-positive breast cancer cells (199).

16. MicroRNA and endocrine-resistant breast cancer

Altered miRNA expression is likely to play a role in endocrine-resistance in breast cancer. A PubMed search for ‘MicroRNA and endocrine resistance in breast cancer’ generated nine new publications since my previous review (200). A recent review of mechanisms of endocrine resistance includes a paragraph on the upregulation of miR-221, miR-222, and miR-181b and downregulation of miR-21, miR-342, and miR-489 in tamoxifen-resistant breast cells (201). miR-221/222 promoted TAM-resistance by targeting ERα and the cell cycle regulator p27 (also known as Kip1) (185). Overexpression of miR-221/222 also associates with Fulvestrant-resistance (202). miR-221/222 is also increased in CD44+CD24−/low human breast cancer stem cells, indicating a role for these stem cells in endocrine resistance (203). miRNAs in CSCs and their role in chemoresistance has been recently reviewed (204).

My laboratory identified miRNAs that are differentially regulated by TAM in endocrine-sensitive MCF-7 and endocrine-resistant LY2 human breast cancer cells (155). LY2 cells were derived from MCF-7 by serial passage in the antiestrogen LY 117018, a precursor to Raloxifene (RAL) (205), and express wild-type ERα mRNA levels similar to MCF-7 cells (206), but are resistant to TAM, RAL, and Fulvestrant (ICI 182,780) (207). We identified 97 miRNAs regulated in the opposite direction in MCF-7 and LY2 cells. Quantitative real-time PCR (qPCR) selectively confirmed higher miR-200a, miR-200b, and miR-200c in MCF-7 than LY2 cells and higher miR-10a, miR-22, miR-29a, miR-125b, and miR-222 in LY2 than in MCF-7 cells (155). Some of the mRNA targets include PDCD4, BCL2, CYP1B1, and ERBB3.

Members of the miR-200 family and miR-221/222 are implicated in epithelial-mesenchymal transition (EMT) and metastasis (208). Many studies have identified an inverse relationship between the expression of the miR-200 family and its targets ZEB1/2 in cells (209–213). ZEB1, a target of miR-200 family of miRNAs and a promoter of EMT, was found to be overexpressed in LY2 cells when compared to MCF-7 cells (155). We observed a progressive decrease in the expression of miR-200a, miR-200b, and miR-200c in an MCF-7-derived cell line model of TAM/endocrine resistance, i.e., decreasing from MCF-7, LCC1 (E2-independent, but TAM-sensitive; to the TAM-resistant LCC2, LCC9, and LY2 cell lines, respectively (214). Concurrently, we detected an increase in ZEB1 expression in LCC9 and LY2 cells. Overexpression of miR-200b and miR-200c enhanced the sensitivity of LY2 breast cancer cells to growth inhibition by antiestrogens 4-OHT and fulvestrant. These data are in agreement with other reports showing an inverse correlation between miR-200 family and ZEB1 expression in basal-like, triple negative breast cancer (TNBC) cells such as MDA-MB-231 and BT549 (210, 212, 213, 215). CpG island methylation of miR-200c/miR-141 promoter has been reported in breast and prostate cancer cells (216–218). Treatment of MDA-MB-231 and BT549 breast and PC3 prostate cancer cells with 5-aza-2’-deoxycytidine (5-aza-dC), a demethylating agent, increased miR-200c and miR-141 expression (216). Our study agrees with these reports of epigenetic silencing of the miR-200 family, because we demonstrated that treatment of LY2 cells with 5-aza-dC + histone deacetylase inhibitor trichostatin A (TSA) increased miR-200b and miR-200c expression (214). There was a concomitant decrease in the expression of ZEB1 mRNA and protein and the LY2 cells appeared more epithelial in in morphology and were sensitized to TAM and fulvestrant inhibition. Likewise, knockdown of ZEB1 increased antiestrogen sensitivity of LY2 cells resulting in inhibition of cell proliferation (214).

Global miRNA analysis of 153 ERα+ primary breast tumors from women who subsequently took tamoxifen as an adjuvant mono-therapy revealed that no single miRNA profile was predictive of patient outcome (219). Decreased expression of miR-190b, miR-339-5p, miR-520c-3-, miR-520g, miR-520h, miR-139-3p, miR-204, miR-502-5p, miR-365, and miR-363 in the primary tumors was associated with recurrence in tamoxifen-treated patients (219).

miR-342 was downregulated in two TAM-resistant cell lines derived from MCF-7 cells called LCC2 and TAMR1 (220). Overexpression of miR-342 conferred TAM-sensitivity and increased apoptosis. miR-451, an oncosuppressor miRNA, was downregulated in TAM-resistant breast cancer cells (221). miR-451 targets 14-3-3ζ an anti-apoptotic gene that is overexpressed in TAM-resistant tumors and is associated with lower survival (221). Increased expression of ERα36, a truncated form of the full length ERα66, that blocks ERα66 genomic activity while activating MAPK signaling, has been reported in TAM-resistant breast tumors (222). Let-7a targets ERα36 and loss of Let-7 family members conferred TAM-resistance by activating non-genomic estrogen signaling mediated by ERα36 (223).

miRNA microarray profiling identified 10 miRNAs downregulated in a TAM-resistant MCF-7 cell line compared with wt MCF-7 cells: miR-125a, miR-489, miR-375, miR-653, miR-135b, miR-556-3p, miR-190b, miR-556-5p, miR-561, and miR-548h; while 12 miRs were upregulated: miR-551b, miR-519a, miR-376a*, miR-31, miR-224, miR-521, miR-31*, miR-655, miR-205, miR-518f, miR-520h, miR-455-3p (224). Transfection of TAM-resistant MCF-7 cells with pre-miR-375 re-sensitized the cells to ~ 15% growth inhibition by 5 µ M TAM, reduced mRNA expression of EMT markers: FN1, ZEB1, and SNAI2, and reverted EMT-like invasive appearance of the cells (224). MTDH was identified as a direct target of miR-375 and siMTDH in TAM-resistant MCF-7 cells partially sensitized the cells to tamoxifen and higher TDFH was correlated with reduced disease-free survival in tamoxifen-treated breast cancer patients (224).

The miRNA cluster C19MC, encoding 59 miRNAs spanning ~ 100 kB(225), is the largest known cluster of miRNAs in the human genome (226). Many miRNAs of C19MC are oncomiRs when re-expressed in tissues (225). miRNA microarray profiling revealed that 18 miRNAs in the C19MC cluster were upregulated in in a TAM-resistant MCF-7 cell line compared with wt MCF-7 cells including miR- 520c-3p, miR-519d, miR-518b, miR-520h, miR-521, miR-518f, miR-520b, miR-518c, miR-512-5p, miR-512-3p, miR-518e*, miR-515-5p, miR-517c, miR-522, and miR-519a (227). Overexpression of a miR-519a mimic in MCF-7 cells resulted in TAM-resistance and transfection of TAM-resistant MCF-7 cells with a miR-519a inhibitor restored TAM-growth inhibition on the cells (227). The authors verified CDKN1A, RB1, and PTEN as bona fide targets of miR-519a and correlated increased miR-519a expression with poorer disease-free survival in ERα+ breast cancer patients (227).

CONCLUSION

Estrogens, most commonly E2, and other ER ligands including tamoxifen and endocrine disruptors regulate diverse physiological effects through genomic and nongenomic/membrane-initiated mechanisms that alter cellular expression of miRNAs. miRNAs are post-transcriptional regulators of mRNA translation and stability. Although miRNA changes in fish, mice, rats, and human breast cancer cells in response to E2 and tamoxifen have been reported, there are relatively few studies examining the detailed mechanisms for these responses and their downstream bona fide targets. The effect of E2 varies between and within cell lines depending on the ratio of ERs, including GPER, expressed, coregulators, chromatin structure, cell cycle, circadian rhythms, and numerous other physiological parameters. Future HITS-CLIP and global high-throughput studies are needed to elucidate the general principles while detailed biochemical/molecular studies are required to dissect the specific mechanisms involved in ER/miRNA interactions and their roles in human health and disease.