Content uploaded by Maria Ravo
Author content
All content in this area was uploaded by Maria Ravo
Content may be subject to copyright.
ORIGINAL ARTICLE
Direct regulation of microRNA biogenesis and expression by estrogen
receptor beta in hormone-responsive breast cancer
O Paris
1
, L Ferraro
1
, OMV Grober
1
, M Ravo
2
, MR De Filippo
1
,
2
, G Giurato
1
,
2
, G Nassa
2
,
R Tarallo
2
, C Cantarella
2
, F Rizzo
2
, A Di Benedetto
3
, M Mottolese
3
, V Benes
4
, C Ambrosino
5
,
ENola
1
and A Weisz
1
,
2
,
6
1
Department of General Pathology, Second University of Naples, Napoli, Italy;
2
Laboratory of Molecular Medicine and Genomics,
University of Salerno, Baronissi, Italy;
3
Department of Pathology, Regina Elena Cancer Institute, Rome, Italy;
4
Genomics Core
Facility, European Molecular Biology Laboratory, Heidelberg, Germany;
5
Department of Biological and Environmental Sciences,
University of Sannio, Benevento, Italy and
6
Division of Molecular Pathology and Medical Genomics, ‘SS. Giovanni di Dio e Ruggi
d’Aragona’ Hospital, University of Salerno, Salerno, Italy
Estrogen effects on mammary epithelial and breast
cancer (BC) cells are mediated by the nuclear receptors
ERaand ERb, transcription factors that display func-
tional antagonism with each other, with ERbacting as
oncosuppressor and interfering with the effects of ERaon
cell proliferation, tumor promotion and progression.
Indeed, hormone-responsive, ERaþBC cells often lack
ERb, which when present associates with a less aggressive
clinical phenotype of the disease. Recent evidences point
to a significant role of microRNAs (miRNAs) in BC,
where specific miRNA expression profiles associate with
distinct clinical and biological phenotypes of the lesion.
Considering the possibility that ERbmight influence BC
cell behavior via miRNAs, we compared miRNome
expression in ERbþvs ERbhormone-responsive BC
cells and found a widespread effect of this ER subtype on
the expression pattern of these non-coding RNAs. More
importantly, the expression pattern of 67 miRNAs,
including 10 regulated by ERbin BC cells, clearly
distinguishes ERbþ, node-negative, from ERb, meta-
static, mammary tumors. Molecular dissection of miRNA
biogenesis revealed multiple mechanisms for direct
regulation of this process by ERbþin BC cell nuclei.
In particular, ERbdownregulates miR-30a by binding to
two specific sites proximal to the gene and thereby
inhibiting pri-miR synthesis. On the other hand, the
receptor promotes miR-23b, -27b and 24-1 accumulation
in the cell by binding in close proximity of the
corresponding gene cluster and preventing in situ the
inhibitory effects of ERaon pri-miR maturation by
the p68/DDX5-Drosha microprocessor complex. These
results indicate that cell autonomous regulation of miRNA
expression is part of the mechanism of action of ERbin
BC cells and could contribute to establishment or
maintenance of a less aggressive tumor phenotype
mediated by this nuclear receptor.
Oncogene advance online publication, 9 January 2012;
doi:10.1038/onc.2011.583
Keywords: estrogen receptor beta; microRNA; breast
cancer; hormones; gene transcription
Introduction
Estrogens have a role in breast cancer (BC) pathogenesis
and progression by controlling mammary cell prolifera-
tion and key cellular functions via the estrogen receptors
(ERaand ERb: Heldring et al., 2007). ERs are members
of the nuclear receptors superfamily of ligand-dependent
transcription factors that both regulate gene expression
controlling the estrogen signal transduction cascade
with distinct and even antagonistic roles. In hormone-
responsive, ERa-positive BC cells ERbinhibits estro-
gen-mediated cell proliferation by increasing the expres-
sion of growth-inhibitory genes and by interfering
with activation of cell cycle and anti-apoptotic genes
by ERain response to 17b-estradiol (E2: Chang et al.,
2006; Grober et al., 2011). ERbis frequently lost in BC,
where its presence generally correlates with a better
prognosis of the disease (Sugiura et al., 2007), is a
biomarker of a less aggressive clinical phenotype
(Novelli et al., 2008; Shaaban et al., 2008) and its
downregulation has been postulated to represent a
critical stage in estrogen-dependent tumor progression
(Roger et al., 2001; Bardin et al., 2004). Despite the
direct relationships between estrogen and breast
carcinogenesis, the divergent roles of the two ER
subtypes in BC are not fully understood, mostly because
they are complex, involving genomic and non-genomic
actions, regulation of gene transcription and control of
mRNA stability and translation efficiency.
MicroRNAs (miRNAs) are small (20–25 nt) non-
coding RNAs that can regulate gene activity in a
posttranscriptional manner. These molecules, frequently
transcribed as polycistronic RNAs, are synthesized
in the nucleus by RNA polymerase II or III as
long primary transcripts (pri-miRNAs), that are then
Received 5 September 2011; revised 29 October 2011; accepted 3
November 2011
Correspondence: Dr A Weisz, Laboratorio di Medicina Molecolare e
Genomica, Universita
`degli Studi di Salerno, via Allende, Baronissi
SA 84081, Italy.
E-mail: aweisz@unisa.it
Oncogene (2012) 1–11
&
2012 Macmillan Publishers Limited
All rights reserved 0950-9232/12
www.nature.com/onc
processed by the class-2 RNase-III Drosha (Han et al.,
2004) in B70-nucleotide stem-loop RNAs (pre-miR-
NAs), that in turn are exported from nucleus to
cytoplasm by exportin 5 and Ran-GTP (Kim et al.,
2009) and cleaved by Dicer/TRBP endoribonuclease into
an imperfect miRNA/miRNA* duplex (Chendrimada
et al., 2005). Only one strand of the duplex is finally
selected to function as a mature miRNA, whereas the
other (passenger) strand is typically degraded (Okamura
et al., 2008; Newman and Hammond, 2010). Mature
miRNAs are then incorporated into an RNA-induced
silencing complex, which binds to target mRNAs,
determining gene silencing by either inhibition of
translation or mRNA degradation (Newman and
Hammond, 2010). miRNAs have been shown to
regulate a wide variety of cellular phenotypes, including
neoplastic transformation, cell proliferation, differentia-
tion and homeostasis (Garzon et al., 2009) and altered
expression of these small RNAs contributes to tumor-
igenesis, as some of them can function as either tumor
suppressors or oncogenes (Zhang et al., 2007; Croce,
2009). Interestingly, in solid tumors, such as prostate,
colon, stomach, pancreas, lung and breast, the spectrum
of miRNAs expressed (miRNome) is different from that
of the corresponding normal tissues (Volinia et al.,
2006), suggesting the involvement of miRNAs in
transformed cell biology. Differential expression of
miRNA genes was found associated with specific
pathological features of BC, where distinct miRNA
expression profiles in normal vs cancer tissue or between
different molecular and clinical tumor subtypes appears
to be the rule (Iorio et al., 2005; Lu et al., 2005;
Blenkiron et al., 2007; Tavazoie et al., 2008). There is
increasing evidence, in fact, that specific miRNAs may
be responsible at large for disease heterogeneity,
functioning as regulators of tumorigenicity, invasion
and metastasis (Tavazoie et al., 2008). Moreover, genetic
defects in key components of the miRNA biosynthetic
pathway have been described in tumors (Hill et al., 2009;
Melo et al., 2009, 2010), and several genes involved in
BC progression have been identified as targets of
miRNAs that, in turn, are found deregulated in BC
cells (Garzon et al., 2009).
Several evidences indicate that ERais among the
transcription factors regulating miRNA biogenesis in
hormone-responsive BC cells (Bhat-Nakshatri et al.,
2009; Castellano et al., 2009; Maillot et al., 2009;
Yamagata et al., 2009; Cicatiello et al., 2010; Ferraro
et al., 2010, 2011). More recently, global mapping of
ERbbinding to ERa-positive, hormone-responsive BC
cells chromatin in vivo showed ERbinteraction with
several miRNA genes, suggesting the possible involve-
ment of this receptor in hormonal control of small non-
coding RNA biogenesis in this cell type (Grober et al.,
2011). Starting from this observation, we investigated
here miRNA expression pattern in estrogen-responsive
BC cell lines engineered to express full-length ERband
in primary-tumor samples selected according to the
presence or absence of this nuclear receptor. Results
indicate a role of ERbin the control of miRNA
biogenesis and expression pattern in BC cells.
Results
ERbinduces widespread changes in miRNA expression in
hormone-responsive cells
In order to investigate the role of ERbin BC, we
generated MCF-7 cells stably expressing full-length
human ERb(ERb-1) fused at the N- (N-TAP-ERb)or
C- (C-TAP-ERb) terminus to a TAP tag in pTRE2pur-
HA expression vector (Puig et al., 2001). As shown in
Figure 1a, the expression levels of C-TAP-ERb(two
independent clones: lanes 2–3), N-TAP-ERb(lane 4) or
C-TAP-ERa(used as control: lane 5) are comparable to
those relative to endogenous ERa, as detected by WB
under comparable test conditions, to avoid toxic and
artifactual events consequent to overexpression of the
exogenous protein. The functional integrity of tagged
ERbwas assessed by measuring their ability to counter-
act induction of ERE-TK-luciferase reporter-gene
transcription by ligand-activated endogenous ERa.As
shown in Figure 1b, cell expressing TAP-ERbshow a
marked reduction in E2-mediated activation of reporter-
gene transcription compared with wt cells, a phenotype
that could be almost completely recovered by stimula-
tion with the ERa-selective ligand 4,40,400 -(4-propyl-
[1H]-pyrazole-1,3,5-triyl)trisphenol (PPT). TAP-ERb
effects on E2-induced MCF-7 cell proliferation and cell
cycle progression were also investigated and the results,
reported in Figures 1c and d, show that cells expressing
exogenous ERbgrow much slower in response to
estrogen than wt or C-TAP-ERacells, consequent to
reduced G1–S transition (Figure 1d). It is worth
mentioning that the cell cycle inhibitory effects of ERb
are well known (Heldring et al., 2007; Grober et al.,
2011, and references therein) and are more evident at
relatively higher concentrations of E2 (X10
10
), compa-
tible with the lower affinity of this ER subtype for the
hormone (compare, for each cell clone, the S þG2
fraction in hormone-stimulated vs -starved cells). The
efficiency of PPT in promoting cell cycle progression
(Figure 1d) relates to its ability to promote ERa-
mediated gene transcription (Figure 1b), confirming the
direct link between transcriptional activity of this
receptor subtype and the mitogenic effects of estrogen
(Cicatiello et al., 2010). Gene-expression profiling of
asynchronously growing cells showed no major differ-
ences between N- and C-TAP-ERbcells, whereas their
transcriptomes were significantly different from that of
C-TAP-ERacells (Supplementary Figure S1), confirm-
ing previous results obtained in E2-stimulated cells
(Grober et al., 2011). Based on these results, expression
of the TAP-ERbfusion proteins appears to significantly
affect ERa-mediated estrogen signal transduction to
target genes and the cell cycle, confirming previous
observations indicating that they are fully functional
in vivo (Grober et al., 2011; Nassa et al., 2011).
Multiple roles have been proposed for miRNAs in
hormone-responsive BC, where the presence of ERbhas
been shown to associate with less aggressive disease
forms. We decided to use our ERb-expressing cells to
investigate potential links between ERband miRNA
activity in hormone-responsive BC cells, as these
miRNA regulation by ERbin breast cancer
O Paris et al
2
Oncogene
represent a useful in vitro model to investigate the
molecular mechanisms underlying the biological effects
of this receptor subtype in hormone-responsive tumors.
To this aim, total RNA was extracted from wt MCF-7,
N-TAP-ERb, C-TAP-ERb(2 independent clones) and
C-TAP-ERacells. Global analysis of miRNome expres-
sion was peformed with microarrays detecting the vast
majority of known and characterized miRNAs (Illumina
MicroRNA Expression Beadchip, Illumina Italia, Mi-
lano, Italy) as described in Material and methods.
Results indicates that expression of ERbhas a deep
impact on BC cell miRNome, as 84 miRNAs were found
differentially expressed in three ERbþvs two ERb
cell lines, whereas no significant differences could be
detected among cells expressing the different tagged
forms of ERb, or between C-TAP-ERa,wt and MCF7-
TAP cells (not shown), that express only the TAP
peptide and show no differences in ERasignaling with
respect to wt cells (Ambrosino et al., 2010; Grober et al.,
2011). To validate this result, we performed miRNA
expression profiling with a different microarray plat-
form (Agilent Human microRNA Microarrays
18 15 K v3, Agilent Technologies Italia, Milano, Italy)
and compared the results obtained in the two experi-
mental settings. As expected, we observed some
differences between the two data sets, likely due to
technical differences between the two microarrays plat-
forms (in particular sensitivity and quality of the probes)
and the two probe sets (Supplementary Materials and
methods and data not shown). Nevertheless, 73 among
the differentially expressed miRNAs identified with the
Illumina platform were either fully confirmed with the
Agilent array or, in some instances, could not be
detected here due to a lower sensitivity of this platform.
For this reason, we performed a further validation of the
results obtained with the Illumina arrays analyzing by
real-time RT–PCR (reverse transcriptase–PCR) the
expression levels of 10 miRNAs selected according to
their relative expression level, ranging from very low to
high, and including also miRNA undetectable with
Agilent arrays or differentially expressed between cell
lines (except for miR-181c, that was not differentially
expressed and is included as negative control). Results
(reported in Supplementary Figure S2) show a very high
correlation between rtPCR and Illumina array data
(correlation coefficient: 0.76), indicating reliability of
this microarray platform. The 73 differentially expressed
miRNAs listed in Figure 2a and Table 1 were thus
considered validated. To gather insights on the mole-
cular mechanisms for ERbeffects on miRNAs, expres-
sion profiling was carried out in both cell types after
estrogen starvation. Under these conditions, no differ-
ences could be detected between ERbþand ERbcells
(left panel in Figure 2b, referring to average values
measured in ERbvs ERbþcell lines), indicating a key
role of the liganded in determining the observed
differences. For this reason, we next investigated
whether expression of the 73 miRNAs identified in the
Figure 1 Functional characterization of ERb-expressing MCF-7 cell clones. (a) Western blotting analysis of protein extracts from
control (wt, lane 1) and TAP-ERs (N-TAP-ERb: lanes 2–3, C-TAP-ERb: lane 4, C-TAP-ERa: lane 5) expressing cells. Asterisks mark
non-specific bands. (b) The ability of tagged ERbto interfere with ERaactivity was assessed by comparing estrogen effects in wt,
N-TAP-ERband C-TAP-ERbcells by ERE-TK-luciferase reporter-gene activation mediated by E2 or PPT (selective ERaagonist).
(c) Proliferation rate of wt,N-TAP-ERband C-TAP-ERbcells was measured in hormone-starved cells stimulated with 10
8
ME2, respect
to untreated cells. Cell counting was performed with a colorimetric assay at the indicated times. (d) Analysis of cell cycle progression after
estrogen stimulation of wt or TAP-ERb-expressing cells. The percent of S þG2 phase cells was determined by flow cytometry in estrogen-
starved cultures 27h after treatment with either vehicle alone (EtOH) or the indicated concentrations of E2 or PPT.
miRNA regulation by ERbin breast cancer
O Paris et al
3
Oncogene
previous experiment was affected by estrogen. To this
aim, N-TAP-ERb,C-TAP-ERband C-TAP-ERawere
E2-deprived and subsequently stimulated with 10
8
M
E2 for 6–72 hrs before miRNA analysis. Results
displayed in Figure 2b (right panel) show that all
investigated miRNAs respond to the hormone in a
time-dependent manner. Although kinetics and extent of
miRNA response to the stimulus were comparable
between the two ERbþcell lines, they were significantly
different in ERbþvs ERb(C-TAP-ERa) cells. Direct
comparison of the data from the two cell types indicates
that the differences in steady-state miRNA levels
consequent to ERbexpression are due to ERbantagon-
ism upon ERaactivity or to a specific effect of ligand-
activated ERb. As shown in Supplementary Figure S3,
for example, expression of hsa-miR30a* and, to a lesser
extent, hsa-miR30a shows a time-dependent decrease
following E2 stimulation only in ERbþcells, whereas it
is unaffected by the stimulus in the absence of ERb.On
the contrary, hsa-miR-23b and -23b*, hsa-miR-27b and -
27b* and hsa-miR-24 and -24-1* levels decrease in the
presence of E2 in ERbwhereas they increase in ERbþ
cells. The putative mRNA targets of the miRNAs
regulated by ERbwere searched with TargetScan and,
subsequently, analyzed for Gene Ontology term over-
representation, in order to identify biological processes
likely to be influenced by this ER subtype via miRNAs.
In this way, several cellular processes were found
downstream of ERb-responsive miRNAs, including those
known to be affected by ERb, such as response to
hormonal stimuli, regulation of transcription and cell
proliferation and others that represent key cellular
processes in malignant cells, including cell motility,
migration, adhesion, differentiation and fate determina-
tion, and are targeted by regulatory cascades in cancer
cells (Supplementary Figure S4).
The data described above were obtained in vitro in a
BC cell model that, although it has been shown to reflect
ER+ER-
ER-
ER+
72h
24h
12h
6h
C-TAP-ER
72h
24h
12h
6h
C-TAP-ER
72h
24h
12h
6h
N-TAP-ER
--
ER -
ER +
E2
MCF7-TAP
C-TAP-ER
C-TAP-ER_B
C-TAP-ER_A
N-TAP-ER
Figure 2 Correlations between ERband miRNome expression in hormone-responsive BC cells and primary breast carcinomas.
(a) Heatmap showing 73 miRNAs differentially expressed between ERbþand ERbcells maintained in standard culture conditions.
Data displayed represent the ratio between the fluorescence intensity value of each miRNA in a given array (cell line) vs the average of
the fluorescence intensity value of the same miRNA in all arrays. (b) Heatmaps showing relative expression of 73 ERb-responsive
miRNAs in ERbþand ERbcells treated with vehicle alone (EtOH; left panel) or with E2 for the indicated times (right panel). Data
displayed represent the ratio between the fluorescence intensity value of each miRNA at the indicated time after E2 stimulation ( þE2)
vs the same in hormone-starved cells (E2, control). (c) Top: Principal component analysis (PCA) relative to differential miRNA
expression in 17 ERbþand 19 ERbprimary BC samples. Bottom: Cluster analysis of 67 miRNAs discriminating between ERbþ
and ERbBC samples. Data displayed represent the ratio between the fluorescence intensity value of each miRNA in a given array
(tumor sample) vs the average of the fluorescence intensity value of the same miRNA in all arrays. miRNA marked in red were
differentially expressed both in ERbþcell lines and BC samples, whereas those marked in purple in C derive from the same pre-miR
of those differentially expressed in ERbþcell lines and tumors.
miRNA regulation by ERbin breast cancer
O Paris et al
4
Oncogene
only in part the complexity of the hormone-responsive
phenotype, in several cases provided molecular insights
that could be validated and find application in the
clinical setting. For this reason, we considered these
evidences as an indication that ERbmight indeed
influence miRNome activity also in primary breast
tumors. To this end, BC samples were selected, among
those originally included in the study reported by
Novelli et al. (2008), for presence or absence of
ERbexpression according to immuno-histochemistry
(Supplementary Figure S5). Tumors were divided in two
groups of 22 ERbþand 18 ERbtumors, respectively,
that did not show significant differences from each other
with respect to key clinical and molecular parameters,
summarized in Supplementary Table S1, with the
notable exception of the presence of lymphnodal
metastases and a worst tumor grading for ERb
tumors. RNA was extracted from formalin-fixed,
paraffin-embedded tissues and that from 17 ERbþ
and 19 ERbtumors was of quality and concentration
apt to perform miRNA expression profiling as described
(Ravo et al., 2008). This led to the identification of 67
miRNAs, whose expression level discriminates ERbþ
from the ERbbreast tumors, including 10 miRNAs
that were found differentially expressed also in ERbþ
vs ERbBC cells in vitro (Figure 2c and Supplementary
Table S2). These results confirm those obtained in cell
lines (Figures 2a and b), pointing to a role of ERbin the
control of BC miRNome and thereby indicating that
miRNAs are integral components of the gene regulation
cascade mediating the effects of this nuclear receptor in
tumor cells.
Direct regulation of miRNA biogenesis by
hormone-activated ERbin BC cells
Mature miRNA expression can be regulated through
control of either transcription or one of the key steps of
primary transcript (pri-miR) maturation. We analyzed
by chromatin immunoprecipitation sequencing (ChIP-
Seq) the entire ERaand ERbcistromes in the ERbþ
(Grober et al., 2011) and ERbcells (Cicatiello et al.,
2010) upon E2 stimulation. Aligning ER-binding sites
and miRNA gene positioning in the genome we
observed that several miRNA-encoding genes differ-
entially expressed in ERbþvs ERbcell lines
(Supplementary Table S3A) and/or mammary tumors
Table 1 Seventy-three miRNAs differentially expressed following ERbexpression in hormone-responsive human breast cancer cells
miRNA Fold-change
(ERbþ/ERb)
P-value miRNA Fold-change
(ERbþ/ERb)
P-value
HS_108.1 1.58 0.03780 hsa-miR-24-2* 1.73 0.00002
HS_131 1.46 0.00782 hsa-miR-23b* 2.07 0.00016
HS_166.1 2.00 0.00001 hsa-miR-27b* 1.60 0.00024
HS_266.1 1.62 0.01176 hsa-miR-24-1* 2.06 0.00000
HS_305_b 1.84 0.00000 hsa-miR-24-1*(miR-189:9.1) 2.51 0.00000
HS_99.1 1.39 0.00505 hsa-miR-29a* 1.50 0.00857
hsa-let-7a 1.50 0.00648 hsa-miR-29b-2* 1.41 0.00509
hsa-let-7a* 13.48 0.00000 hsa-miR-30a* 2.97 0.00012
hsa-let-7c 4.02 0.00000 hsa-miR-30c-2* 2.05 0.00003
hsa-let-7f 1.62 0.00510 hsa-miR-30d* 1.62 0.00046
hsa-miR-100 4.57 0.00000 hsa-miR-31 1.65 0.00435
hsa-miR-101* 1.53 0.00495 hsa-miR-32* 2.29 0.00000
hsa-miR-1257 1.46 0.00331 hsa-miR-330-5p 2.04 0.00003
hsa-miR-125b 2.37 0.00000 hsa-miR-335 4.61 0.00000
hsa-miR-1267 1.40 0.00514 hsa-miR-338-3p 1.94 0.00000
hsa-miR-1285 1.64 0.00090 hsa-miR-361-3p 1.44 0.00998
hsa-miR-1305 1.44 0.00254 hsa-miR-362-5p 1.67 0.00074
hsa-miR-148b* 1.41 0.00600 hsa-miR-365 2.92 0.00000
hsa-miR-15a* 2.05 0.00097 hsa-miR-374b* 1.53 0.00680
hsa-miR-16-1* 1.98 0.00004 hsa-miR-375 1.65 0.00274
hsa-miR-17* 1.51 0.00113 hsa-miR-450b-3p 1.69 0.00645
hsa-miR-181c* 1.70 0.00005 hsa-miR-452*:9.1 1.98 0.00024
hsa-miR-186 7.14 0.00000 hsa-miR-501-5p 1.79 0.00033
hsa-miR-18a* 1.40 0.01173 hsa-miR-542-3p 3.22 0.00000
hsa-miR-18b 1.78 0.00000 hsa-miR-548d-3p 1.97 0.00001
hsa-miR-196b 1.65 0.00101 hsa-miR-556-5p 4.07 0.00000
hsa-miR-199a-5p 1.78 0.00412 hsa-miR-579 2.31 0.00079
hsa-miR-199b-5p 1.52 0.00574 hsa-miR-616* 1.56 0.00366
hsa-miR-199a*:9.1 1.43 0.00283 hsa-miR-629* 1.67 0.00746
hsa-miR-199a-3p/199b-3p 1.72 0.00001 hsa-miR-642 1.79 0.00144
hsa-miR-19a 1.29 0.04889 hsa-miR-651 1.56 0.00148
hsa-miR-19a* 2.86 0.00000 hsa-miR-652 1.54 0.00179
hsa-miR-19b-1* 1.95 0.00800 hsa-miR-663b 1.37 0.02954
hsa-miR-20a* 1.46 0.00291 hsa-miR-708* 2.05 0.00478
hsa-miR-216a 2.47 0.02759 hsa-miR-935 1.53 0.00068
hsa-miR-23a* 1.93 0.00007 hsa-miR-99b* 1.56 0.00327
hsa-miR-27a* 2.10 0.00000
Bold entries denote miRNAs differentially expressed also in ERb-positive vs ERb-negative breast tumor biopsies.
miRNA regulation by ERbin breast cancer
O Paris et al
5
Oncogene
(Supplementary Table S3B) display ER-binding sites
within 10 kb of the transcription unit, including sites
where both ERs can be found together, likely associated
in heterodimers. This finding suggested us the possibility
that miRNA gene activity could be modulated in BC
cells by an interplay of the two ER subtypes bound to
chromatin, with ERbantagonizing ERa-mediated reg-
ulation of pri-miR biosynthesis and/or maturation rate.
To verify this possibility, we choose to investigate in
detail differences in miRNA precursor levels in ERbþ
vs ERbcells following stimulation with E2, focusing
on miR-30a gene and the miR-23b/27b/24-1 chromoso-
mal cluster. The first was selected as it encodes two
miRNAs (miR-30a and -30a*) that are downregulated
by estrogen in ERbþcells only (Supplementary Figure
S3) and it shows two binding sites for ERbin close
proximity —one upstream and one downstream— of the
transcription unit, but no ERasites (Figures 3a and b).
The second caught our attention, instead, as it shows
sites for both receptors (Figure 4a and Supplementary
Figure S6) and it encodes three distinct couples of
miRNAs, all accumulating in ERbþcells and decreas-
ing in ERbcells in response to the hormone
(Supplementary Figure S3). Interestingly, in both cases
the effect of the hormone was more evident on the ‘star’
strand that, for this reason, led us first to their
identification (Figure 2a) and was routinely used here
to monitor ERbeffects.
The results relative to the miR-30a locus are reported
in Figure 3 and show that ERbbinding results in a
significant reduction of pri-, pre- and mature miR-30a
levels following E2 stimulation, detectable already after
2 h (Figures 3c–e), to indicate that the predominant
effect of ligand-activated ERbis to trans-repress basal
gene transcription by direct binding to this transcription
unit. Noteworthy, activation of ERaalone (wt cells) did
not affect miR-30a biogenesis, in agreement with the
lack of binding of this receptor to the locus (Figure 3a).
When combined, these results indicate a specific and
direct role of ERbin repression of miR-30a expression
in BC cells, possibly mediated by promoter trans-
repression. This could be due to direct transcriptional
repression, via recruitment of a repressor complex to the
chromatin by ligand-activated ERb, or, alternatively, to
inhibition of gene trans-activation caused by tethering of
ERbto a transcription factor constitutively bound to
the locus, resulting in displacement or inhibition of an
activator complex. The latter possibility, that could
explain also lack of ERabinding to such regulatory site,
is worth investigating further, extending the analysis
Figure 3 Analysis of ERbregulation of miR-30a and precursor biogenesis. (a) Genome browser view of the two ERb-binding sites
within 10 kb upstream or downstream from hsa-miR-30a locus on chromosome 6. (b) Validation of ERb-binding site by ChIP and real-
time PCR in wt or ERbþcells before (-) and after stimulation with E2 for 45 min. (c,d) Real-time rtPCR analysis of pri-miR-30a (c)
and pre-miR-30a (d)inwt or ERbþcells before (-) and after stimulation with E2 for the indicated times. (e) Real-time rtPCR analysis
of mature hsa-miR-30a* in wt or ERbþcells (C- and N-TAP-ERbcell RNA combined) before (-) and after stimulation with E2 for
the indicated times.
miRNA regulation by ERbin breast cancer
O Paris et al
6
Oncogene
also to other genetic loci selectively regulated by ERbin
BC cells under the same conditions.
Our attention focused next on the miR-23b/27b/24-1
cluster on chromosome 9, whose organization is showed
in Figure 4a. In this case, both ERb- and ERa-binding
sites are detected. Noteworthy, the two ERa-binding
sites identified by ChIP-Seq were also found by ChIP-
on-chip in an independent study (Hurtado et al., 2008)
and binding of the two ERs to both sites identified here
in ERbþcells was confirmed by ChIP (Supplementary
Figure S6). The effects of ERbin regulation of the first
step in miRNA biogenesis were investigated by measur-
ing changes in pri-miR expression in control (wt),
N-TAP- and C-TAP-ERbcells before and after E2
stimulation. Results show that in the absence of ERb,
estrogen stimulation did not influence primary-tran-
script levels, assessed by both quantitative real-time
rtPCR and RNA-expression profiling (c9orf3 RNA;
Figure 4b and Cicatiello et al., 2010). On the other hand,
a slight but reproducible accumulation of pri-miR-23b/
-27b/-24-1 was detectable in ERbþcells already 2 h
after E2 (Figure 4b). We next measured the intracellular
concentration of the individual pre-miR deriving from
this primary transcript (pre-miR-23b, -27b and -24-1) in
both cell types and the results obtained were surprisingly
very different. Indeed, as shown in Figure 4c, whereas
stimulation with E2 of ERbcells caused a substantial
loss of pre-miR (ranging from 20 to 75%), the same
treatment caused instead accumulation of these pre-
miRs in ERbþcells. This was reflected in comparable
changes in expression of the corresponding mature
miRNAs for up to 72 h after E2 stimulation (Figure 4d
and Supplementary Figure S3). These results indicate
that the presence of ERbin ERa-expressing, estrogen-
responsive BC cells can modify substantially the
response of miRNA genes to hormonal stimulus. In
the case of the miR-23b/27b/24-1 gene cluster, this
results from changes in pri-miR maturation, rather than
synthesis, leading to increase in pre-miR biosynthesis in
the presence of chromatin-bound ERb.
ERbinterferes with ERa-mediated recruitment of Drosha
in inactive chromatin-bound complexes
Yamagata et al. (2009) reported ERa-mediated regula-
tion of miRNA maturation by direct interaction in the
nucleus of ERawith a protein complex comprising
Drosha and the DEAD box RNA helicase p68/DDX5,
resulting in inhibition of pri- to pre-miRNA conversion
by Drosha. We thus considered the possibility that the
enhancing effect of ERbon pri-miR-23b/-27b/-24-1
maturation shown in Figure 4 could result from
competition for binding of the ERa-p68-Drosha com-
plex to this locus by ERb, as recently described for other
target genes (Grober et al., 2011), thereby preventing the
inhibitory effect of ERaon nascent pri-miR maturation.
In both cases, we should expect inhibition of ERa-
mediated p68/DDX5-Drosha recruitment to miR-23b/-
27b/-24-1 chromatin by ERb. Indeed, this appears to be
the case, as E2-induced p68/DDX5 and Drosha binding
to ERb-G9242 and ERb-G9242 chromatin sites was
strongly reduced in C-TAP-ERbcompared with wt cells,
concomitant with a reduction of ERaand appearance
of ERb(upper panel of Figure 5a and data not shown).
It is worth mentioning that binding of ERato chromatin
in ERbþcells occurs mainly via heterodimerization
with ERb(Grober et al., 2011). ERb-mediated inhibi-
tion of p68/DDX5 binding could be observed also at
the ERb-G5984 site of the TFF1/pS2 gene promoter,
Figure 4 Analysis of ERbregulation of miR-23b/27b/24-1 and
precursor biogenesis. (a) Genome browser view of ERb- and
ERa-binding sites within 10 kb upstream or downstream from
miR-23b/27b/24-1 cluster within the c9orf3 locus on chromosome 9.
(b–d) Real-time rtPCR analysis of the 23b/27B724-1 pri-miR (b),
pre-miR (c) and mature miRNA* (d)inwt or ERbþcells (C- and
N-TAP-ERbcell RNA combined) before (-) and after stimulation
with E2 for the indicated times. The right columns in (b) show the
relative expression of c9orf3 RNA in ERbþcells following E2
stimulation, measured by mRNA expression profiling (Grober
et al., 2011).
miRNA regulation by ERbin breast cancer
O Paris et al
7
Oncogene
although Drosha could not be detected tethered to this
site under any condition (lower panel of Figure 5b),
suggesting that association of this enzyme to chromatin
may be promoted by ERaonly at sites of pri-miR
synthesis, where Drosha could be ‘locked’ in an inactive
complex comprising ERaand the hairpin structure of
the nascent pri-miR. Concerning the nature of the
physical interaction between ERaand Drosha, it was
suggested that this is mediated by p68/p72 RNA
helicases (Yamagata et al., 2009). Interestingly, a
systematic analysis of the ERbinteractome of MCF-7
cell nuclei (Nassa et al., 2011) failed to identify p68/
DDX5 binding to this receptor subtype as well as to
ERa/ERbheterodimers, suggesting that the presence of
ERbcould determine inhibition of p68/DDX5-mediated
sequestering of Drosha to the chromatin in an inhibitory
complex. This possibility would provide a rationale for
the ChIP results obtained in ERbþcells, where we
failed to detect these two proteins in the presence of
both ERs (Figure 5a). To verify this possibility, we
performed co-purification analysis of all these proteins
in nuclear extracts from wt, C-TAP-ERaor C-TAP-
ERbcells. The two ERs were adsorbed to Sepharose-
bound IgG via their TAP tag, as described (Ambrosino
et al., 2010; Nassa et al., 2011). As shown in Figure 5b,
Drosha and p68/DDX5 could be co-purified with ERa
but not with ERb, demonstrating that ERbis unable
to bind these proteins. It is worth mentioning that as
under these experimental conditions ERaco-purifies
with C-TAP-ERb(Nassa et al., 2011 and lower section
of Figure 5b), ERa/ERbheterodimers do not bind
Drosha and p68/DDX5.
Discussion
The results described here demonstrates that ERb
controls synthesis, maturation and steady-state levels
of a significant number of miRNAs in BC cells by
interfering with ERaactivity or acting autonomously, as
demonstrated here for the miR-23b/-27b/-24-1 cluster
and the miR-30a gene, respectively. This, in turn,
determines a profound effect on miRNome expression
and activity in tumors expressing ERb, which could help
explain their less aggressive clinical phenotype (Novelli
et al., 2008; Shaaban et al., 2008). Identification of the
intracellular targets of these ERb-regulated miRNAs,
and the effects they exert on key cellular functions of BC
cells, will now provide a new venue to understand the
pleiotropic role of this oncosuppressive factor in breast
carcinogenesis and tumor progression. Furthermore, it
is reasonable to conceive that proteins encoded by the
mRNAs targeted by these miRNA may represent
molecular markers exploitable for prognostic evaluation
of primary breast tumors or for prediction of the disease
responsiveness to hormonal therapy.
Materials and methods
Cell Culture, transient transfection and cell cycle analyses
Human hormone-responsive BC cells MCF-7 Tet-Off
(Clontech-Takara) expressing TAP (control cells), C-TAP-
ERa, C-TAP-ERbor N-TAP-ERbwere described previously
(Ambrosino et al., 2010; Nassa et al., 2011). They were
propagated, hormone starved and analyzed for estrogen
signaling, cell cycle progression and cell proliferation as
described earlier (Cicatiello et al., 2000; Grober et al., 2011).
RNA purification
Total RNA was extracted from hormone-starved ( þEtOH,
E2) or stimulated ( þE
2
) cell cultures as described previously
(Cicatiello et al., 2004). FFPE tumor samples were cut in
5-mm-thick sections on a microtome with a disposable blade.
RNA was extracted from three and eight sequential sections as
described (Ravo et al., 2008). RNA concentration in each
sample was determined with a NanoDrop-1000 spectrophot-
ometer (Thermo Fisher Scientific Italy, Cinisello Balsamo,
Italia) and quality assessed with the Agilent 2100 Bioanalyzer
and Agilent RNA 6000 cartridges (Agilent Technologies). For
microarray analysis, RNAs extracted from replicate samples of
the same tumor were pooled.
Figure 5 ER/Drosha interaction in MCF-7 cell nuclei. ChIP real-
time PCR results showing binding of ERa,ERb, p68 and Drosha
to miRNA 23b/27b/24-1 cluster and the TFF1/pS2 loci (a) in the
TFF1/pS2 loci (b)inwt and C-TAP-ERbcells; data are expressed
as % occupancy respect to input chromatin. (b) Western blot
analysis of whole nuclear extracts (lanes 1–3) and IgG-Sepharose-
affinity-purified nuclear extracts (lanes 4–6) from wt (lanes 1 and 4),
C-TAP-ERa(lanes 2 and 5) or C-TAP-ERb(lanes 3 and 6) cells,
probed with the indicated antibodies. Asterisks mark non-specific
bands.
miRNA regulation by ERbin breast cancer
O Paris et al
8
Oncogene
Microarray analyses
See Supplementary Materials and methods.
Protein-complex immunoprecipitations and analysis
Cells were hormone starved for 5 days and following
stimulation with 10
8
ME2 for 2 h, nuclear proteins were
extracted and incubated with IgG-Sepharose beads (GE
Healthcare, Milano, Italy) for 4 h at 4 1C, as described earlier
(Ambrosino et al., 2010). Affinity-purified complexes were
resuspended in SDS sample buffer (Invitrogen Life Technol-
ogies Italia, Milano, Italy) and analyzed by SDS–PAGE and
western blotting by using anti-TAP (CAB1001, Open Biosys-
tems, Euroclone Spa, Milano, Italy), anti-ERa(sc-543, Santa
Cruz Biotechnology), anti-Drosha (ab12286, Abcam, Cam-
bridge, UK) and anti-DDX5 (ab21696, Abcam) antibodies.
The primary antibodies were detected with a horseradish
peroxidase-conjugated anti-rabbit antibody (GE Healthcare)
and revealed by chemiluminescence and autoradiography.
Chromatin immunoprecipitation
Cells were hormone deprived for 4 days and chromatin was
extracted from replicate samples before (E2) or 45 min after
stimulation with E2 as described previously (Cicatiello et al.,
2010; Grober et al., 2011). Chromatin samples were incubated
at 4 1C overnight with Abs against the C- (HC-20, from Santa
Cruz Biotechnology, Europe) or N- (18–32, Sigma Aldrich
Italia, Milano, Italy) terminus of human ERa, anti-Drosha
(ab12286, Abcam, used as described by Nakamura et al.
(2007), anti-DDX5 (ab21696, Abcam) or, for TAP-ERb, with
IgG Sepharose 6 fast Flow (GE Healthcare) as described
earlier (Grober et al., 2011). As control, aliquots of the same
chromatin were processed in the same way but Abs were
omitted from the incubation mixtures ( þE2/-Abs) or, where
required, underivatized Sepahrose was used.
Quantitative real-time rtPCR
Total RNA was extracted from cell lines (as described before) after
stimulation for 2 h and 4 h with10
8
ME2. For miRNA analysis,
mature miRNA was reverse transcribed using a miRNA-specific
stem-loop reverse transcriptase and real-time PCR was performed
using Taqman microRNA assays (Assay ID: 2822, 416, 2439,
2445, 2441, 2126, 2174, 2440, 2333, 482; Applied Biosystems
Italia, Monza, Italy) according to the manufacturer’s instruction.
RNU49 was used as an internal control to normalize all data
using the Taqman RNU49 assay (Applied Biosystems Italia).
RNU49 was unaffected by hormone treatment. For pre-miRNA
and pri-miRNA analysis, RNA was reverse transcribed using
Quantitect Rev. Transcription kit (Qiagen Italy, Milano, Italia)
and real-time PCR was performed in triplicates in three
independent experiments using Power Syber Green PCR Master
Mix (Applied Biosystems Italia) and normalized to U6 snRNA.
All the real-time PCR were performed on a MJ Research PTC-
200 Opticon Instrument (MJ Research, Waltham, MA, USA).
Primers used are listed in Supplementary Table 4.
ChIP-Seq data analysis
For ER-binding-site mapping in genome, ChIP-Seq data
relative to ERb(Grober et al., 2011; accession number
E-MTAB-345) and ERa(Cicatiello et al., 2010; accession
number E-MTAB-131) were analyzed as follows. Enriched
ChIP-Seq peaks were identified using FindPeaks (Fejes et al.,
2008), with a subpeaks value of 0.5. To select only highly
relevant sites, the statistical cut-off of the first quartile was
applied. The binding sites supported by a number of tags lower
than 25% of the range of the values was discarded. This led to
re-mapping of ERb-binding sites (renumbered here from
ERb_G1 to ERb_G12430); for ERa-binding sites, numbering
was the same as previously described (Cicatiello et al., 2010).
miRNA target prediction and functional analysis of their
putative mRNA targets
For comprehensive prediction of miRNA-target genes, we
used TargetScan, release 5.1 (http://www.targetscan.org). To
identify statistically overrepresented ‘biological process’ gene
ontology terms among sets of selected mRNA target, we used
the Database for Annotation, Visualization and Integrated
Discovery (DAVID, http://david.abcc.ncifcrf.gov/) functional
annotation tool (Dennis et al., 2003; Huang et al., 2009). To
this aim, we used as background data coming from gene
expression-profiling experiments performed on the same cell
line and under the same experimental conditions used in this
study.
Immunohistochemistry
See: Supplementary Materials and methods.
BC samples clinical hallmarks
For the purpose of this study, 40 breast carcinomas were
selected from a series of 936 cases with a median follow up
(FU) of 50 months (min 1–max 108) subjected to breast surgery
at the Regina Elena Cancer Institute between 2001 and 2005
(Novelli et al., 2008). Of these, 22 were ERbþwithout any
recurrence, whereas 18 were ERband presented local or
distant metastasis. In these patients, ERbexpression was
routinely determined at the time of surgical treatment along
with other conventional biological factors namely ERaand
progesterone receptors (PgR), HER2 and Ki-67, before any
adjuvant therapy was planned. As showed in Supplementary
Table S1, the group included 37 (92.5%) invasive ductal
carcinomas and 3 (7.5%) invasive lobular carcinomas. Among
these, 28 (70%) were pT1, 9 (22.5%) pT2 and 3 (7.5%) pT3-4,
27 (67.5%) were node negative and 13 (32.5%) were node
positive, 29 (72.5%) G1-2 and 11 (27.5%) G3. ERawas
positive in 37 tumors (92.5%) and negative in 3 (7.5%), PgR
was positive in 31 tumors (77.5%) and negative in 9 (22.5),
HER2 was positive in 12 tumors (30%) and negative in 28
(70%) and Ki-67 was positive in 16 tumors (40%) and negative
in 24 (60%). Tumors were graded according to Bloom and
Richardson and staged according to the Unione Internationale
Contre le Cancer tumor-node-metastasis system criteria, and
histologically classified according to the World Health Orga-
nization (Tavassoli and Devilee, 2003). In the selected group,
ERbþwas significantly associated to negative lymphnodes
(Po0.0001) and low tumor grade (G1-2) (P¼0.03) whereas, as
already described on a large series of BC patients, (Novelli
et al., 2008) no significant correlation was observed between
ERbexpression and the other parameters analyzed. Follow-up
data were obtained from hospital charts and by corresponding
with the referring physicians.
Conflict of interest
The authors declare no conflict of interest.
Acknowledgements
We thank Manuela Ferracin and Massimo Negrini for help
with analysis of miRNA expression data from primary tumor
miRNA regulation by ERbin breast cancer
O Paris et al
9
Oncogene
samples and useful suggestions, and Luigi Cicatiello, Mar-
gherita Mutarelli and Maria Francesca Papa for technical
assistance. Work supported by: Italian Association for Cancer
Research (grants IG-8586 to AW and IG-8706 to MM),
European Union (CRESCENDO IP, contract number LSHM-
CT2005-018652), Italian Ministry for Education, University
and Research (grant PRIN 2008CJ4SYW_004), University of
Salerno (Fondi FARB 2011) and Fondazione con il Sud (grant
2009-PdP-22). CC, FR, GN, MR and RT are fellows of
Fondazione con il Sud.
References
Ambrosino C, Tarallo R, Bamundo A, Cuomo D, Franci G, Nassa G
et al. (2010). Identification of a hormone-regulated dynamic nuclear
actin network associated with estrogen receptor alpha in human
breast cancer cell nuclei. Mol Cell Proteomics 9: 1352–1367.
Bardin A, Boulle N, Lazennec G, Vignon F, Pujol P. (2004). Loss of
ERbeta expression as a common step in estrogen-dependent tumor
progression. Endocr Relat Cancer 11: 537–551.
Bhat-Nakshatri P, Wang G, Collins NR, Thomson MJ,
Geistlinger TR, Carroll JS et al. (2009). Estradiol-regulated
microRNAs control estradiol response in breast cancer cells. Nucleic
Acids Res 37: 4850–4861.
Blenkiron C, Goldstein LD, Thorne NP, Spiteri I, Chin SF, Dunning
MJ et al. (2007). MicroRNA expression profiling of human breast
cancer identifies new markers of tumor subtype. Genome Biol 8:
R214.
Castellano L, Giamas G, Jacob J, Coombes RC, Lucchesi W,
Thiruchelvam P et al. (2009). The estrogen receptor-alpha-induced
microRNA signature regulates itself and its transcriptional
response. Proc Natl Acad Sci USA 106: 15732–15737.
Chang EC, Frasor J, Komm B, Katzenellenbogen BS. (2006). Impact
of estrogen receptor beta on gene networks regulated by estrogen
receptor alpha in breast cancer cells. Endocrinology 147: 4831–4842.
Chendrimada TP, Gregory RI, Kumaraswamy E, Norman J, Cooch
N, Nishikura K et al. (2005). TRBP recruits the Dicer complex to
Ago2 for microRNA processing and gene silencing. Nature 436:
740–744.
Cicatiello L, Addeo R, Altucci L, Belsito Petrizzi V, Boccia V,
Cancemi M et al. (2000). The antiestrogen ICI 182 780 inhibits
proliferation of human breast cancer cells by interfering with
multiple, sequential estrogen-regulated processes required for cell
cycle completion. Mol Cell Endocrinol 165: 199–209.
Cicatiello L, Mutarelli M, Grober OM, Paris O, Ferraro L, Ravo M
et al. (2010). Estrogen receptor alpha controls a gene network in
luminal-like breast cancer cells comprising multiple transcription
factors and microRNAs. Am J Pathol 176: 2113–2130.
Cicatiello L, Scafoglio C, Altucci L, Cancemi M, Natoli G, Facchiano
Aet al. (2004). A genomic view of estrogen actions in human breast
cancer cells by expression profiling of the hormone-responsive
transcriptome. J Mol Endocrinol 32: 719–775.
Croce CM. (2009). Causes and consequences of microRNA dysregula-
tion in cancer. Nat Rev Genet 10: 704–714.
Dennis Jr G, Sherman BT, Hosack DA, Yang J, Gao W, Lane HC
et al. (2003). DAVID: Database for Annotation, Visualization, and
Integrated Discovery. Genome Biol 4: P3.
Fejes AP, Robertson G, Bilenky M, Varhol R, Bainbridge M,
Jones SJ. (2008). FindPeaks 3.1: a tool for identifying areas
of enrichment from massively parallel short-read sequencing
technology. Bioinformatics 24: 1729–1730.
Ferraro L, Ravo M, Nassa G, Tarallo R, De Filippo MR, Giurato G
et al. (2011). Effects of estrogen on microRNA expression in
hormone-responsive breast cancer cells. Horm Cancer (in press).
Ferraro L, Ravo M, Paris O, Giurato G, De Filippo MR, Stellato C
et al. (2010). Identification of a gene network controlled by estrogen
receptor ain human breast cancer cells by miRNA expression
profiling and ChIP-Seq. Am J Pathol 177: S34, Abs. STE 12.
Garzon R, Calin GA, Croce CM. (2009). MicroRNAs in cancer. Annu
Rev Med 60: 167–179.
Grober OMV, Mutarelli M, Giurato G, Ravo M, Cicatiello L, De
Filippo MR et al. (2011). Global analysis of estrogen receptor beta
binding to breast cancer cell genome reveals extensive interplay with
estrogen receptor alpha for target gene regulation. BMC Genomics
12: 36.
Han J, Lee Y, Yeom KH, Kim YK, Jin H, Kim VN. (2004). The
Drosha-DGCR8 complex in primary microRNA processing. Genes
Dev 18: 3016–3027.
Heldring N, Pike A, Andersson S, Matthews J, Cheng G, Hartman J
et al. (2007). Estrogen receptors: how do they signal and what are
their targets. Physiol Rev 87: 905–931.
Hill DA, Ivanovich J, Priest JR, Gurnett CA, Dehner LP, Desruisseau
Det al. (2009). DICER1 mutations in familial pleuropulmonary
blastoma. Science 325: 965.
Huang W, Sherman BT, Lempicki RA. (2009). Systematic and
integrative analysis of large gene lists using DAVID Bioinformatics
Resources. Nature Protoc 4: 44–57.
Hurtado A, Holmes KA, Geistlinger TR, Hutcheson IR,
Nicholson RI, Brown M et al. (2008). Regulation of ERBB2 by
oestrogen receptor-PAX2 determines response to tamoxifen. Nature
45: 663–666.
Iorio MV, Ferracin M, Liu CG, Veronese A, Spizzo R, Sabbioni S
et al. (2005). MicroRNA gene expression deregulation in human
breast cancer. Cancer Res 65: 7065–7070.
Kim VN, Han J, Siomi MC. (2009). Biogenesis of small RNAs in
animals. Nat Rev Mol Cell Biol 10: 126–139.
Lu J, Getz G, Miska EA, Alvarez-Saavedra E, Lamb J, Peck D et al.
(2005). MicroRNA expression profiles classify human cancers.
Nature 435: 834–838.
Maillot G, Lacroix-Triki M, Pierredon S, Gratadou L, Schmidt S,
Be
´ne
`sVet al. (2009). Widespread estrogen-dependent repression of
micrornas involved in breast tumor cell growth. Cancer Res 69:
8332–8340.
Melo SA, Moutinho C, Ropero S, Calin GA, Rossi S, Spizzo R et al.
(2010). A genetic defect in exportin-5 traps precursor microRNAs in
the nucleus of cancer cells. Cancer Cell 18: 303–315.
Melo SA, Ropero S, Moutinho C, Aaltonen LA, Yamamoto H,
Calin GA et al. (2009). A TARBP2 mutation in human cancer
impairs microRNA processing and DICER1 function. Nat Genet
41: 365–370.
Nakamura T, Canaani E, Croce CM. (2007). Oncogenic All1 fusion
proteins target Drosha-mediated microRNA processing. Proc Natl
Acad Sci USA 104: 10980–10985.
Nassa G, Tarallo R, Ambrosino C, Bamundo A, Ferraro L, Paris O
et al. (2011). A large set of estrogen receptor b-interacting proteins
identified by tandem affinity purification in hormone-responsive
human breast cancer cell nuclei. Proteomics 11: 159–165.
Newman MA, Hammond SM. (2010). Emerging paradigms of
regulated microRNA processing. Genes Dev 24: 1086–1092.
Novelli F, Milella M, Melucci E, Di Benedetto A, Sperduti I, Perrone-
Donnorso R et al. (2008). A divergent role for estrogen receptor-
beta in node-positive and node-negative breast cancer classified
according to molecular subtypes: an observational prospective
study. Breast Cancer Res 10: R74.
Okamura K, Phillips MD, Tyler DM, Duan H, Chou YT, Lai EC.
(2008). The regulatory activity of microRNA* species has sub-
stantial influence on microRNA and 30UTR evolution. Nat Struct
Mol Biol 15: 354–363.
Puig O, Caspary F, Rigaut G, Rutz B, Bouveret E, Bragado-Nilsson E
et al. (2001). The tandem affinity purification (TAP) method: a general
procedure of protein complex purification. Methods 24: 218–229.
Ravo M, Mutarelli M, Ferraro L, Grober OMV, Paris O, Tarallo R
et al. (2008). Quantitative expression profiling of highly degraded
miRNA regulation by ERbin breast cancer
O Paris et al
10
Oncogene
RNA from formalin-fixed, paraffin-embedded breast tumor biopsies
by oligonucleotide microarrays. Lab Invest 88: 430–440.
Roger P, Sahla ME, Ma
¨kela
¨S, Gustafsson JA, Baldet P, Rochefort H.
(2001). Decreased expression of estrogen receptor beta protein
in proliferative preinvasive mammary tumors. Cancer Res 61:
2537–2541.
Shaaban AM, Green AR, Karthik S, Alizadeh Y, Hughes TA, Harkins
Let al. (2008). Nuclear and cytoplasmic expression of ERbeta1,
ERbeta2, and ERbeta5 identifies distinct prognostic outcome for
breast cancer patients. Clin Cancer Res 14: 5228–5235.
Sugiura H, Toyama T, Hara Y, Zhang Z, Kobayashi S, Fujii Y et al.
(2007). Expression of estrogen receptor beta wild-type and its
variant ERbetacx/beta2 is correlated with better prognosis in breast
cancer. Jpn J Clin Oncol 37: 820–828.
Tavassoli FA, Devilee P. (2003). World Health Organization
Classification of Tumors. Pathology and Genetics of Tumours of the
Breast and Female Genital Organs. IARC Press: Lyon.
Tavazoie SF, Alarco
´n C, Oskarsson T, Padua D, Wang Q, Bos PD
et al. (2008). Endogenous human microRNAs that suppress breast
cancer metastasis. Nature 451: 147–152.
Volinia S, Calin GA, Liu CG, Ambs S, Cimmino A, Petrocca F et al.
(2006). A microRNA expression signature of human solid tumors
defines cancer gene targets. Proc Natl Acad Sci USA 103: 2257–2261.
Yamagata K, Fujiyama S, Ito S, Ueda T, Murata T, Naitou M et al.
(2009). Maturation of microRNA is hormonally regulated by a
nuclear receptor. Mol Cell 36: 340–347.
Zhang B, Pan X, Cobb GP, Anderson TA. (2007). microRNAs as
oncogenes and tumor suppressors. Dev Biol 302: 1–12.
Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)
miRNA regulation by ERbin breast cancer
O Paris et al
11
Oncogene