Retinoic acid-induced HOXA5 expression is co-regulated by HuR and miR-130a

Cellular Signalling 25 (2013) 1476–1485

Contents lists available at SciVerse ScienceDirect

Cellular Signalling journal homepage: www.elsevier.com/locate/cellsig

Retinoic acid-induced HOXA5 expression is co-regulated by HuR and miR-130a Fan Yang a, Lin Miao b, Yide Mei a,⁎, Mian Wu a,⁎ a b

Hefei National Laboratory for Physical Sciences at Microscale and School of Life Sciences, University of Science and Technology of China, Hefei, Anhui, 230027, China Scientific and Educational Department, The second hospital of Anhui Medical University, Hefei, Anhui, 230061, China

a r t i c l e

i n f o

Article history: Received 2 February 2013 Received in revised form 16 March 2013 Accepted 16 March 2013 Available online 23 March 2013 Keywords: HuR miR-130a HOXA5 Retinoic acid

a b s t r a c t Retinoic acid (RA) has been used as a chemopreventive agent for breast cancer. It has been shown that HOXA5 is a critical mediator of RA-induced cell growth inhibition. However, the molecular mechanisms underlying RA-induced HOXA5 expression remain largely unknown. Here we report that in addition to transcriptional regulation, post-transcriptional regulation also contributes to RA-induced HOXA5 expression. miR-130a, a c-Myc responsive miRNA, represses HOXA5 cellular levels under unstressed condition. Upon RA treatment, c-Myc is quickly degraded via the proteasome-dependent pathway. This in turn decreases miR-130a levels and de-represses the translation of HOXA5. We also show that the de-repression of HOXA5 translation is dependent on the RNA-binding protein Human antigen R (HuR), which binds to 3′UTR of HOXA5 mRNA and increases its stability in response to RA treatment. Collectively, these results demonstrate that HuR and miR-130a dynamically regulate HOXA5 gene expression via modulating HOXA5 mRNA turnover and translation, respectively, thereby contributing to RA-induced growth inhibition. © 2013 Elsevier Inc. All rights reserved.

1. Introduction Unlike prokaryotes, whose gene expression is mainly regulated at the transcription level, eukaryotes regulate gene expression not only through the transcriptional modification, but also via the post-transcriptional mechanisms. It has been well recognized that RNA-binding proteins (RBPs) and microRNAs (miRNAs) are two major factors involved in the post-transcriptional regulation of gene expression. miRNAs are 20- to 22-nt-long regulatory RNAs expressed in both plants and metazoan animals [1]. It has been estimated that approximately 30% of all human genes are regulated by miRNAs [2]. In principle, miRNAs base-pair with their target mRNAs, promote mRNA degradation and/or inhibit mRNA translation, thereby regulating gene expression post-transcriptionally. Over the past decade, the function of miRNAs has been intensely investigated. It has been shown that miRNAs play critical roles in the regulation of a variety of cellular functions as well as many disease processes [3,4]. Similar to miRNAs, RBPs are solidly established regulators of mRNA stability and translation in response to environmental

⁎ Corresponding authors. E-mail addresses: [email protected] (Y. Mei), [email protected] (M. Wu). 0898-6568/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.cellsig.2013.03.015

changes in metazoans. By binding to the specific RNA sequences in the target mRNAs, RBPs can modulate mRNA turnover and translation [5]. Given the critical roles of RBPs in the post-transcriptional regulation of gene expression, a great deal of work has been done with RBPs [6–8]. It is not surprising that RBPs exert profound effects on cellular adhesion and invasion during cancer progression. It is well accepted that both miRNAs and RBPs have the opportunity to regulate the common target mRNA, yet existence of the dynamic regulation of mRNA orchestrated by miRNAs and RBPs remains to be determined. Retinoic acid (RA), the natural or synthetic derivative of vitamin A, participates in the regulation of a vast spectrum of biological processes such as cell growth, differentiation, apoptosis, and morphogenesis [9,10]. RA has been increasingly used for the treatment of a variety of cancers, underscoring the important role of RA in the anti-cancer therapy. Because of this, understanding the molecular mechanisms by which RA induces cytotoxicity of tumor cells has attracted increasing attentions recently. HOXA5 has been well recognized as a critical mediator of RA-induced growth inhibition. It has been shown that HOXA5 is strongly up-regulated following RA treatment [11]. Here, we set out to investigate the molecular mechanisms underlying the RA-induced HOXA5 expression and the subsequent cell death. Our findings reveal that both miR-130a and HuR are involved in the up-regulation of HOXA5 in response to RA treatment. Also, miR-130a- and HuR-mediated HOXA5 regulation is functionally important for RA-induced cell death. Our data highlight the importance of post-transcriptional modulation in the

F. Yang et al. / Cellular Signalling 25 (2013) 1476–1485

regulation of RA-induced HOXA5 expression, and also implicate miR130a and HuR as potential therapeutic targets for cancer treatment.

1477

shRARβ: CTGGGTAAATACACCACGAAT,GCCACCTCTCATTCAAG AAAT.

2. Materials and methods

2.3. Plasmid construction

2.1. Reagents and antibodies

HuR and HOXA5 cDNAs were amplified by RT-PCR from total RNA purified from HeLa cells. The generated DNA fragments were cloned into p3XFLAG-myc-CMVTM-24 expression vector (Sigma). DNA fragments containing wild-type or mutant c-Myc responsive elements of miR-130a promoter were amplified by PCR, digested using XhoI and HindIII, and then ligated into a linearized pGL3Basic vector (Promega). The HOXA5 mRNA 3′UTR and the deletion mutants were generated and ligated into psiCHECK-2 vector (Promega). All PCR products were verified by DNA sequencing.

The following antibodies were used in this study: anti-HOXA5 (Sigma), anti-β-Actin, anti-RARβ and anti-HuR (Santa Cruz), anti-c-Myc and anti-Ago2 (Cell Signaling Technology), mouse IgG1 (M075-3, MBL), and rabbit IgG (Sigma). MG132 was purchased from Calbiochem (La Jolla, CA, USA). 9-Cis-RA, 13-cis-RA, all-trans-RA, α-amanitin and cycloheximide were obtained from Sigma-Aldrich Corporation. 2.2. Oligonucleotides Mimics and inhibitors of miR-130a were synthesized by Genepharma Company (Shanghai, China). Transfection of miR-130a mimics and inhibitors using lipofectamine 2000 (Invitrogen) was performed according to the manufacturer's instruction. The sequences of the oligonucleotides used in this study are as listed below: Primers used in qRT-PCR/RT-PCR assays HOXA5 Fw: 5′AGATCTACCCCTGGATGCGC3′ Rev: 5′CCTTCTCCAGCTCCAGGGTC3′ β-Actin Fw: 5′GACCTGACTGACTACCTCATGAAGAT3′ Rev: 5′ GTCACACTTCATGATGGAGTTGAAGG3′ U6(RT): 5′CGCTTCACGAATTTGCGTGTCAT3′ U6 Fw: 5′GCTTCGGCAGCACATATACTAAAAT3′ U6 Rev: 5′CGCTTCACGAATTTGCGTGTCAT3′ miR-130a (RT): 5′GTCGTATCCAGTGCGTGTCGTGGAGTCGGCAATT GCACTGGATACGACATGCCC3′ miR-130a Fw: 5′GTCAGTGCAATGTTAAAAGGGCAT3′ miR-130a Rev: 5′CAGTGCGTGTCGTGGAGT3′ pri-miR-130a Fw: 5′GGTGGTCTCTGTGCTGGGGGTCAGG3′ pri-miR-130a Rev: 5′ ATGCTGAGGAGGCAGCCAGCGCTGGGTAG3′ c-Myc Fw: 5′ CAGCTGCTTAGACGCTGGATT3′ c-Myc Rev: 5′ GTAGAAATACGGCTGCACCGA3′ GAPDH Fw: 5′ CCTGTTCGACAGTCAGCCG3′ GAPDH Rev: 5′ CGACCAAATCCGTTGACTCC3′. Synthetic Renilla luciferase gene (hRluc) hRluc Fw: 5′AGACAAGATCAAGGCCATCGTCCA3′ hRluc Rev: 5′TTTCTCGCCCTCTTCGCTCTTGAT3′. Synthetic Firefly luciferase gene (hluc) hluc Fw: CTGCTGAACAGCATGGGCATTTCT hluc Rev: ATGTGTACATGCTCTGGAAGCCCT miR-130a mimic sequence: sense 5′CAGUGCAAUGUUAAAAGGG CAU3′ and antisense 5′GCCCUUUUAACAUUGCACUGUU3′ miR-130a inhibitor sequence: 5′AUGCCCUUUUAACAUUGCACU G3′. Primers used in ChIP assays miR-130a c-Myc responsive element: Fw: 5′TCTGTGCTGGGGGTCAGGGGGTTG3′ Rev: 5′CCTCCTCACTCCTTCTCCCAGTCCC3′. Oligonucleotide sequence of shRNAs: siAgo2-1: CGGCAAGAAGAGATTAGCAAA siAgo2-2: CGTCCGTGAATTTGGAATCAT shc-Myc-1: CAGTTGAAACACAAACTTGAA shc-Myc-2: CCTGAGACAGATCAGCAACAA shHuR: GAGGCAATTACCAGTTTCA shHOXA5: GCCATTATAGCGCCTGTATAA, CCGCAGAAGGAGGATTG AAAT

2.4. Real-time RT-PCR and RT-PCR Total RNA was isolated using TRIzol (Invitrogen). 1 μg of RNA was used to synthesize cDNA using PrimeScript™ RT reagent kit (Takara) according to the manufacturer's instruction. Real-time PCR was performed using SYBR Green real-time PCR analysis (Takara). PCR results, recorded as threshold cycle numbers (Ct), were normalized against an internal control (β-actin). RT-PCR was carried out by using One-Step RNA PCR kit (AMV) (Takara, Tokyo, Japan) according to the supplier's protocol. For miR-130a and U6 RNA level examination, reverse transcription reaction was carried out with the indicated RT-primer (miR-130a-RT, U6-RT) before real-time PCR analysis. 2.5. Luciferase assay To investigate whether miR-130a is transcriptionally regulated by c-Myc, MCF7 cells expressing either control shRNA or c-Myc shRNA were transfected with the pGL3-based construct containing miR-130a promoter plus Renilla luciferase reporter plasmid. Twenty-four hours later, the reporter activity was measured by using a luciferase assay kit (Promega) and plotted after normalizing with respect to Renilla luciferase activity (mean ± SD). To determine the effect of HuR on HOXA5 mRNA stability, MCF7 cells expressing either control shRNA or HuR shRNA were transfected with the psiCHECK-2 based construct containing full length or the indicated fragment of HOXA5 3′UTR. Twenty-four hours later, the reporter activity was measured by a luciferase assay kit and plotted after normalizing with respect to Firefly luciferase activity. The data were shown as mean ± SD of three independent experiments. 2.6. RNA interference To generate lentiviruses expressing RARβ, HuR, c-Myc, HOXA5, or control shRNA, HEK293T cells grown on a 6-cm dish were transfected with 2 μg of RARβ shRNA, HuR shRNA, c-Myc shRNA, HOXA5 shRNA (cloned in PLKO.1) or control vector, 2 μg of pREV, 2 μg of pGag/Pol/ PRE, and 1 μg of pVSVG. 24 h after transfection, cells were cultured with DMEM medium containing 20% FBS for another 24 h. The culture medium containing lentivirus particles was cleaned by centrifugation to get rid of the cell debris at 12,000 ×g for 5 min, and used for the target cell infection. 2.7. ChIP The ChIP assay was performed as previously described [12]. 2.8. IP–RT-PCR Immunoprecipitation–Reverse transcription polymerase chain reaction (IP–RT-PCR) was performed as described previously [13]. Briefly,

1478

F. Yang et al. / Cellular Signalling 25 (2013) 1476–1485

A

B

C RA

All-trans-RA

1

HOXA5

10 100 (µM)

RARβ β-Actin

β-Actin

β-Actin 3

1

4

4 3 2 1 0

2

3

4

Relative RNA level of HOXA5

2

Relative RNA level of HOXA5

1

Relative RNA level of HOXA5

0

HOXA5

HOXA5

4 3 2 1

0 All-trans-RA

0

1

1

2

3

4

+ -

+ +

+ -

+ +

4 3 2 1

0 Ctrl shRNA RARβ shRNA RA

10 100 (µM)

E of HOXA5

D Relative RNA level

sh Ctrl sh RARβ + - +

RA 0 h RA 1 h RA 2 h RA 4 h

5 4 3 2 1 0

α-AM

DMSO RA

0 1

2

4

0 1

2 3

4

5

CHX 2 4

0 1

2 4 (h)

HOX A5 β-Actin 1 DMSO

α-AM

6 7

8

9 10 11 12

CHX

Fig. 1. RA-induced expression of HOXA5 occurs at both transcriptional and post-transcriptional levels. (A) Analysis of HOXA5 protein and mRNA expression in MCF7 cells upon treatment with the retinoic acid (RA) analogs. Cells were treated with the indicated RA at 100 μM for 2 h. Cell lysates and total RNA were subjected to Western blot and real-time RT-PCR analyses, respectively. (B) MCF7 cells were treated with the indicated concentration of all-trans-RA for 2 h. Cell lysates and total RNA were analyzed by Western blotting and real-time RT-PCR, respectively. (C) MCF7 cells were infected with lentiviruses expressing either control or RARβ shRNA. Twenty-four hours after infection, cells were treated with or without 100 μM all-trans-RA for 2h. Cell lysates and total RNA were analyzed by Western blotting and real-time RT-PCR, respectively. For convenience, all-trans-RA was hereafter referred to as RA. (D and E) MCF7 cells were pre-treated with either a-amanitin (a-AM) or cycloheximide (CHX) for 2h. Cells were then incubated with 100 μM RA for the indicated periods of time. D) Cell lysates were analyzed by Western blotting with the indicated antibodies. E) Total RNA was subjected to real-time RT-PCR analysis.

siAgo2-2

HOXA5 Ago2

Relative RNA level of HOXA5

B

RA

sictrl

siAgo2-2

sictrl

siAgo2-1

DMSO

siAgo2-1

A

β-Actin 1 2

4

HOXA5 mRNA

2

Ago2

1

HOXA5 mRNA

0

Ago2

RA

Relative RNA level of pri-miR-130a

0.5

0

RA

0

1

2

4 (h)

Anti-Ago2

12 3 4 5 67 8

E 1

IP

Anti-IgG

DMSO

***

Input

RA(h) 0 1 2 4 0 1 2 4

3

3 4 5 6

D Relative RNA level of miR-130a

C

sictrl siAgo2-1 siAgo2-2

F ** 1

0.5

DMSO

RA

Ctrl miR130a mimics

+ - +

- + - - - + -

miR130a inhibitor

-

+ -

-

- +

HOX A5 Ago2

0

RA

0

1

2

4 (h)

β-Actin 1 2 3 4 5 6

Fig. 2. miR-130a regulates RA-induced HOXA5 expression. (A and B) MCF-7 cells were transfected with siRNA against either control (Si Ctrl) or Ago2 (Si-Ago2-1 and Si-Ago2-2). Forty-eight hours after transfection, cells were treated with 100 μM RA for 2 h. A) Cell lysates were subjected to Western blot analysis with the indicated antibodies. B) Total RNA was subjected to real-time RT-PCR analysis to examine HOXA5 mRNA expression. (C) MCF7 cells were treated with 100 μM RA for the indicated periods of time. Cell lysates were immunoprecipitated with anti-Ago-2 antibody. HOXA5 mRNA present in the immunoprecipitates was examined by RT-PCR analysis. (D and E) MCF7 cells were treated with 100 μM RA for the indicated periods of time. Total RNA was subjected to real-time RT-PCR analysis to determine the expression levels of mature form (D) and primary transcript (E) of miR-130a. The data were represented as mean ± SD of three independent experiments. ** and *** indicate P b 0.01 and P b 0.001, respectively. (F) MCF7 cells were transfected with either mimics or inhibitors of miR-130a. Forty-eight hours after transfection, cells were treated with or without 100 μM RA for 2 h. Cell lysates were then analyzed by Western blotting with the indicated antibodies.

B

miR-130a

A -63--55

1479

C Input

IP

IgG c-Myc IgG c-Myc c-Myc consensus sequence:XAXCACGTGXX Binding site WB

c-Myc binding site (BS): GAGCCCGTGA BS mutant: GAGAAAAAAA

1

IgG Hc

4

0

1 0.5

0 Vector BS (wt) BS (mt)

F Dox 0 h

0.5

Dox 48 h Wash 48 h

Dox 24h

1 0.5

0 pri-miR-130a mature-miR-130a

pri-miR-130a mature-miR-130a

Wash Dox

shRNA Ctrl c-Myc-1 c-Myc-2 c-Myc

0

24

48

0% FBS 10% FBS

5 4 3 2 1 0

shRNA

+ 48 (h)

Ctrl

FBS(%)

3

H

G RA

0

1

2

4

(h)

c-Myc β-Actin 1

2

3

4

J

1

Relative RNA level of c-Myc

2

1

2

3

10

1

2

0

10

3

4

I

4

1

MG132 RA c-myc

0.5

K 0h 1h 2h 4h

0

β- Actin

β-Actin 1

c-Myc-2

c-Myc

c-Myc

β- Actin

Relative luciferase activity

3

Ctrl shRNA c-Myc shRNA-1 c-Myc shRNA-2

Relative RNA level of miR-130a

1

2

E

Ctrl shRNA c-Myc shRNA-1 c-Myc shRNA-2

Relative RNA level

Relative RNA level

D

c-Myc

Relative luciferase activity

F. Yang et al. / Cellular Signalling 25 (2013) 1476–1485

0

1

2

4

+ 4

1

2

3

4

5

(h)

β-Actin

0 RA

0

1

2

4 (h)

miR-130a mimics - sh c-Myc - sh Ctrl RA

+ + - +

- + +

+ + + +

- - +

- - +

HOXA5

0.5

c-Myc β- Actin

0 Vector

BS (wt

BS (mt)

1 2 3 4 5

6

Fig. 3. c-Myc transcriptionally regulates miR-130a expression. (A) Schematic illustration of consensus c-Myc binding sites A and B in miR-130a gene promoter. (B) Lysates from MCF7 cells were subjected to ChIP assay. ChIP products were amplified by PCR reaction. β-Actin was used as a negative control. The successful pulldown of c-Myc by anti-c-Myc antibody was also confirmed by Western blot analysis. (C) MCF7 cells expressing shRNA against either control (Ctrl shRNA) or c-Myc (c-Myc-1 and c-Myc-2 shRNA) were co-transfected with the indicated reporter constructs and Renilla luciferase plasmid. Twenty-four hours later, reporter activity was measured by using luciferase assays and plotted after normalizing with respect to Renilla luciferase activity (mean ± SD). (D) Lysates and total RNA from MCF7 cells expressing shRNA against either control (Ctrl shRNA) or c-Myc (c-Myc-1 shRNA and c-Myc-2 shRNA) were analyzed by Western blotting and real-time RT-PCR, respectively. (E) P493-6 cells were treated with doxycycline (Dox) for the indicated different lengths of time to inhibit c-Myc expression or were treated first with doxycycline for 48 h and then washed (Wash) to remove doxycycline for recovering c-Myc expression by waiting for another 48 h. Cells were then harvested for western blot and real-time RT-PCR analyses. (F) MCF7 cells expressing shRNA against either control (Ctrl shRNA) or c-Myc (c-Myc-1 and c-Myc-2 shRNA) were serum-starved for 24 h before they were cultured in DMEM culture medium containing 10% FBS for another 2 h. Cell lysates and total RNA were analyzed by Western blotting and real-time RT-PCR, respectively. (G and H) MCF7 cells were treated with RA for the indicated periods of time. Cell lysates and total RNA were analyzed by Western blotting (G) and real-time RT-PCR (H), respectively. (I) MCF7 cells were treated with RA in the presence or absence of MG132 as indicated. Cell lysates were analyzed by Western blotting with the indicated antibodies. (J) MCF7 cells were co-transfected with the indicated reporter constructs and Renilla luciferase plasmid. Twenty-four hours after transfection, cells were treated with RA for the indicated periods of time. Reporter activity was then measured by using luciferase assays and plotted after normalizing with respect to Renilla luciferase activity (mean ± SD). (K) MCF7 cells expressing either control shRNA or c-Myc shRNA were transfected with or without miR-130a mimics, followed by treatment with or without RA for 2 h before they were harvested and subjected to Western blot analysis.

1 × 107 cells were lysed in hypotonic buffer supplemented with RNase A inhibitor and DNAse I before centrifugation. Cell lysates were pre-cleared with protein A/G beads (Pierce) before they were incubated with protein A/G beads coated with the indicated antibodies at 4 °C for 3 h. After extensive washing, the bead-bound immunocomplexes were eluted using elution buffer [50 mM Tris (pH 8.0), 1% SDS, and 10 mM EDTA] at 65 °C for 10 min. To isolate protein-associated RNA from the eluted immunocomplexes, samples were treated with proteins K,

and RNA was extracted by phenol/chloroform. Purified RNA was then subjected to RT-PCR analysis.

2.9. In vitro caspase activity assay In vitro caspase activity assay was performed as described previously [14].

F. Yang et al. / Cellular Signalling 25 (2013) 1476–1485

B

sh Ctrl

RARA+

1

RA 0 1 2 4 0 1 2 4 (h) HOX A5

0.8

HuR

0.6 0.4

β-Actin

0.2

HOX A5

0

α-AM

0

D

3

6

sh Ctrl

miR-130a inhibitor - - + + RA - + - +

9

12

15

sh HuR - - ++ - + - +

(h)

β-Actin

E

HuR β-Actin 1 2 3 4 5 6 7 8

sh Ctrl RA +

sh HuR +

Ago2

H

Ctrl + miR130a mimics miR130a inhibitor HuR

β-Actin

β-Actin 2

3

β-Actin 1 2 3 4 5 6 7 8 9 1011 12

1.5

4

1

+ -

+

2

3

RARA+

1 0.5 0

1 2 3 4 5 6 7 8

HuR

1

3Xflag-HuR HuR(endo)

F

sh Ctrl sh HuR miR-130a mimics - - + + - - + + RA - + - + - + - + HOXA5

HuR β-Actin

RA 0 1 2 4 0 1 2 4 0 1 2 4 (h) HOXA5

1 23 45 67 8

HOX A5

G

C

sh HuR

WB

Relative mRNA level of HOXA5

1.2

RT-PCR

A

Relative RNA level of miR-130a

1480

sh Ctrl sh HuR

Fig. 4. HuR is essential for HOXA5 induction upon RA treatment. (A) mRNA stability assay of HOXA5 in MCF7 cells treated with or without RA. MCF7 cells were treated with or without RA for 4 h before they were incubated with α-amanitin (α-Am) for the indicated periods of time. HOXA5 mRNA levels were analyzed by real-time RT-PCR. (B) HuR knockdown blocks RA-induced HOXA5 expression. MCF7 cells expressing either control or HuR shRNA were treated with RA for the indicated periods of time, followed by Western blot and RT-PCR analyses. (C) MCF-7 cells expressing either control or HuR shRNA were transfected with or without shRNA-resistant 3× Flag-HuR construct as indicated. Twenty-four hours after transfection, cells were treated with RA for the indicated periods of time. Cell lysates were then analyzed by Western blotting. (D) MCF7 cells expressing either control or HuR shRNA were transfected with or without miR-130a inhibitors. Forty-eight hours after transfection, cells were treated with or without RA for 2 h. Cell lysates were analyzed by Western blotting. (E) MCF7 cells expressing either control or HuR shRNA were transfected with or without miR-130a mimics. Forty-eight hours after transfection, cells were treated with or without RA for 2 h, followed by Western blot analysis. (F) MCF7 cells expressing either control or HuR shRNA were treated with or without RA. Total RNA was subjected to real-time RT-PCR analysis to examine miR-130a expression levels. (G) MCF-7 cells expressing either control or HuR shRNA were treated with or without RA before they were harvested and subjected to Western blot analysis. (H) MCF-7 cells were transfected with either mimics or inhibitors of miR-130a. Forty-eight hours after transfection, cell lysates were analyzed by Western blotting.

3. Results 3.1. RA-induced HOXA5 expression involves the post-transcriptional events To investigate the molecular mechanisms underlying RA-induced HOXA5 expression, we treated human breast cancer MCF7 cells with different RA analogs. Among the examined RA analogs, all-trans-RA exhibited the most potent ability to induce protein expression, as well as mRNA expression of HOXA5 (Fig. 1A). In addition, all-trans-RA induced protein and mRNA expressions of HOXA5 in a dose-dependent manner (Fig. 1B). Because of that, only all-trans-RA was used in the following experiments. For convenience, all-trans-RA hereafter was referred to as RA. It has been shown that the retinoic acid receptor β (RARβ) is involved in RA-induced transcriptional up-regulation of HOXA5 [11]. In agreement with this finding, when RARβ was knocked down in MCF7 cells, RA was no longer able to up-regulate mRNA levels of HOXA5 (Fig. 1C). However, protein levels of HOXA5 were still elevated in RARβ knockdown MCF7 cells following RA treatment, although to a lesser extent compared with those in RA-treated control cells (Fig. 1C). These data indicate the existence of the post-transcriptional regulation of HOXA5

expression in response to RA treatment. To further confirm this, we used α-amanitin (α-Am) and cycloheximide (CHX), which are specific inhibitors of transcription and translation, respectively. Although RAinduced HOXA5 mRNA expression was efficiently blocked by α-Am (Fig. 1D), HOXA5 protein expression was still elevated upon RA treatment in the presence of α-Am (Fig. 1E). In contrast, RA-induced HOXA5 mRNA expression was not affected by CHX (Fig. 1D), while RA-induced HOXA5 protein expression was completely inhibited in the presence of CHX (Fig. 1E), demonstrating that the de novo protein synthesis is required for RA-induced HOXA5 up-regulation. However, we currently cannot exclude the possibility that HOXA5 undergoes rapid degradation in the presence of CHX. Taken together, these data strongly suggest that in addition to transcriptional regulation, posttranscriptional regulation also plays an important role in RA-induced up-regulation of HOXA5 expression. 3.2. miR-130a is involved in RA-induced HOXA5 expression miRNA has been shown to play a critical role in the posttranscriptional regulation of gene expression, where miRNA functions in association with proteins to form the micro-ribonucleoprotein complexes (miRNPs) [15]. Argonaute (Ago) proteins, especially Ago2

F. Yang et al. / Cellular Signalling 25 (2013) 1476–1485

B

C Input

874

RT-PCR

HOXA5

3'UTR

1 62

IP

IgG HuR IgG HuR

HOXA5 mRNA(NM_019102.3) 5‘UTR CDS

Relative RNA level of HOXA5

A

1481

GAPDH

1657

IgG Hc (1149- 1192) uaacuauaacaa aaaaagaaaag agaaaaaaaaac acacaaaaa

(1572-1619) uuuuuuuuuuuuuu uuuuuuuuuuguau uuguuuuuaaggaa aaaaua

β-Actin IgG HuR

1

15 10

5 0 Vector

G

FL

A

B

sh Ctrl sh HuR

1

0.5

0 Vector FL

C

HuR 0 1 2 4 (h)

WB

HuR IgG Lc

GAPDH HuR

WB RT-PCR

Input HOXA5

C

1 2 3 4 5 6 7 8 9 10

H

I sh Ctrl sh Ctrl+RA sh HuR sh HuR+RA

1

0.5

0 α-AM 0

3

6

9

12

15 (h)

Relative luciferase activity

GAPDH

B

β-Actin

Relative RNA level of HOXA5

RT-PCR

RA 0 1 2 4 HOXA5

A

sh HuR - + - + - + - + - + sh Ctrl + - + - + - + - + HuR

IP IgG

2

F

Relative luciferase activity

hRluc 875 1657 hRluc FL 3‘UTR 875 1220 hRluc A 1430 1657 hRluc B 1221 1429 hRluc C

0

HuR

IgG Lc

Relative HOXA5 mRNA level in IP

hluc hluc hluc hluc hluc

0.5

sh Ctrl sh HuR

HuR

E

D

1

4

0h 1h 2h 4h

3 2 1 0 sh Ctrl

sh HuR

Fig. 5. HOXA5 mRNA is stabilized by HuR. (A) Schematic illustration of HOXA5 mRNA. Two AU-rich elements present in the 3′UTR of HOXA5 mRNA were also indicated. (B) Lysates from MCF7 cells were immunoprecipitated with anti-HuR antibody or an isotype-matched IgG. HOXA5 mRNA present in the immunoprecipitates was examined by RT-PCR. The protein levels of HuR in the immunoprecipitates were also examined by Western blot analysis. (C) Real-time RT-PCR analysis of HOXA5 mRNA expression in MCF7 cells expressing either control or HuR shRNA. (D) Schematic description of the pSICHECK2-based luciferase reporter constructs used in this study. These reporter plasmids containing full length or fragment of HOXA5 3′UTR were generated using standard PCR approaches. (E) MCF7 cells were transfected with the indicated luciferase reporter constructs. Twenty-four hours after transfection, cells were harvested and subjected to immunoprecipitation with anti-HuR antibody or an isotype-matched IgG. hRluc mRNA present in the immunoprecipitates was examined by real-time RT-PCR. (F) MCF7 cells expressing either control or HuR shRNA were transfected with the indicated luciferase reporter constructs. Twenty-four hours after transfection, cells were harvested. Reporter activity was then measured by using luciferase assays and plotted after normalizing with respect to Firefly luciferase activity (mean ± SD). The successful knockdown of HuR was confirmed by Western blot analysis. (G) MCF7 cells were treated with RA for the indicated periods of time. Cell lysates were then immunoprecipitated with anti-HuR antibody or an isotype-matched IgG. HOXA5 mRNA present in the input and immunoprecipitates was determined by RT-PCR analysis. The protein levels of HuR in the immunoprecipitates were also examined by Western blot analysis. (H) MCF7 cells expressing either control or HuR shRNA were treated with or without RA for 4 h before they were incubated with α-amanitin (α-Am) for the indicated periods of time. HOXA5 mRNA levels in these cells were examined by real-time RT-PCR analysis. (I) MCF-7 cells expressing either control or HuR shRNA were transfected with the pSICHECK2-based luciferase reporter plasmid containing full length of HOXA5 3′UTR. Twenty-four hours after transfection, cells were treated with or without RA for the indicated periods of time. Reporter activity was then measured by using luciferase assays and plotted after normalizing with respect to Firefly luciferase activity (mean ± SD).

protein, are the essential and best characterized components of miRNPs [16,17]. To investigate whether miRNA regulates RA-induced HOXA5 expression, we knocked down Ago2 using two different sets of siRNAs in MCF7 cells. Knockdown of Ago2 led to a marked increase in HOXA5 protein levels in RA-untreated cells, but minimally affected HOXA5 protein levels in RA-treated cells (Fig. 2A). However, Ago2 knockdown did not show significant effect on HOXA5 mRNA levels under both RA-treated and -untreated conditions (Fig. 2B). These combined data indicate that miRNA post-transcriptionally represses HOXA5 expression under unstressed condition. They also raise an intriguing possibility that the miRNA-mediated inhibitory effect on HOXA5 expression may be relieved upon RA treatment. In support of this, treatment of MCF7

cells with RA indeed inhibited the binding of Ago2 to HOXA5 mRNA in a time-dependent manner (Fig. 2C). We next sought to identify the miRNA(s) that directs Ago2 to HOXA5 mRNA and inhibits HOXA5 protein expression. It has been recently reported that miR-130a is an important miRNA that functions in angiogenesis through targeting HOXA5 [18]. We therefore examined whether miR-130a is involved in RA-induced HOXA5 expression. Intriguingly, levels of both mature form and primary transcript of miR-130a were down-regulated after treatment with RA (Fig. 2D and E). In addition, knockdown of miR-130a by its inhibitors increased protein expression of HOXA5 in RA-untreated cells (Fig. 2F, lane 1 vs. 3), whereas induction of miR-130a by its mimics decreased HOXA5 protein

1482

F. Yang et al. / Cellular Signalling 25 (2013) 1476–1485

expression in RA-treated cells (Fig. 2F, lane 4 vs. 5). These data indicate that high basal levels of miR-130a are important for maintaining the low expression of HOXA5 in unstressed cells, and elimination of miR-130a contributes to RA-induced HOXA5 expression. Taken together, these results suggest that miR-130a is involved in regulating HOXA5 expression in response to RA treatment. 3.3. c-Myc controls miR-130a expression in response to RA treatment To investigate the underlying mechanism(s) whereby miR-130a is down-regulated in response to RA treatment, we inspected the genomic sequence upstream of the mature miR-130a by using the genomatix suite of sequence analysis tools (MatInspector). One putative c-Mycbinding site (c-Myc-BS) was found in the upstream region of the mature miR-130a (Fig. 3A). Given that c-Myc has the ability to directly regulate the expression of a diverse set of miRNAs [19–21], we examined whether miR-130a is regulated by c-Myc. The Chromatin Immunoprecipitation (ChIP) analysis showed that the chromatin fragments, corresponding to the putative c-Myc-BS, were specifically present in anti-c-Myc immunoprecipitates (Fig. 3B). Knockdown of c-Myc decreased the activity of the pGL3 luciferase reporter plasmid containing the putative c-Myc-BS, but left the activity of the mutant plasmid unaffected (Fig. 3C). Furthermore, c-Myc knockdown resulted in the substantially lower levels of miR-130a expression (Fig. 3D). These data indicate that miR-130a is transcriptionally regulated by c-Myc. To further confirm the transcriptional regulation of miR-130a by c-Myc, we used the human P-493 B cells that bear a tetracyclinerepressible c-Myc construct. In the presence of doxycycline (Dox), c-Myc protein expression was inhibited, which was accompanied with the decreased levels of pri-miR-130a (Fig. 3E). However, when c-Myc expression was re-induced by removing Dox, levels of pri-miR-130a were also recovered (Fig. 3E). Since previous studies have suggested that both c-Myc and miR-130a could be induced by serum stimulation, we sought to determine whether serum stimulation induces miR-130a expression via c-Myc. We found that along with the induction of c-Myc, miR-130a was indeed strongly up-regulated in MCF7 cells upon serum stimulation. However, when c-Myc was knocked down, this effect was greatly reversed (Fig. 3F), suggesting that c-Myc is important for serum stimulation-induced miR-130a expression. We next examined the effect of RA treatment on c-Myc expression and pGL3-c-Myc-BS activity. Intriguingly, treatment of MCF7 cells with RA strongly decreased protein expression (Fig. 3G), but not mRNA expression of c-Myc in a time-dependent fashion (Fig. 3H). Also, this RA-induced down-regulation of c-Myc protein expression can be blocked by the proteasome inhibitor MG132 (Fig. 3I), indicating that c-Myc is quickly degraded though the proteasome pathway following RA treatment. Moreover, correlating with the decline in c-Myc protein levels, the activity of pGL3-c-Myc-BS was also decreased in response to RA treatment (Fig. 3J), further indicating the importance of c-Myc in regulating miR-130a expression in response to RA treatment.

Since c-Myc increases miR-130 expression and the latter in turn down-regulates HOXA5 cellular levels, we thus determined the effect of c-Myc on HOXA5 protein expression. Knockdown of c-Myc resulted in an increase in HOXA5 levels in RA-untreated cells (Fig. 3K, lane 1 vs 3), partially recapitulating the effect of RA treatment on HOXA5 expression (Fig. 3K, lane 2 vs 3), indicating that c-Myc regulates HOXA5 expression via miR-130a. In support of this notion, induction of miR-130a completely blocked RA-induced HOXA5 induction in c-Myc knockdown MCF7 cells (Fig. 3K, lane 4 vs. 6). Together, these results suggest that the c-Myc-miR-130a axis is important for controlling HOXA5 expression in response to RA treatment.

3.4. HuR is indispensable for RA-induced HOXA5 expression As shown above, knockdown of either miR-130a or c-Myc was not able to fully recapitulate the effect of RA treatment on HOXA5 expression (Fig. 2F, lane 3 vs. 4 and Fig. 3K, lane 2 vs. 3), suggesting that miRNA may not be the sole determinant of RA-induced up-regulation of HOXA5 expression. We therefore tested whether HOXA5 mRNA stability could be regulated by treatment with RA. We found that HOXA5 mRNA was more stable in RA-treated cells than in RA-untreated cells (Fig. 4A). It has been well known that mRNA stability is controlled by RNA binding proteins, and among which HuR is an important one and has been implicated in the different aspects of post-transcriptional regulation, such as modulating translation and increasing the stability of different target mRNAs [22]. We then examined whether HuR is involved in the induction of HOXA5 expression upon RA treatment. Knockdown of HuR by its specific shRNA abolished the induction of both protein and mRNA expression of HOXA5 following RA treatment (Fig. 4B), however, which was reversed by re-introduction of shRNA-resistant HuR into HuR-knockdown cells (Fig. 4C), indicating the specific effect of HuR on RA-induced HOXA5 expression. Combined with previously mentioned evidence, these findings suggest that induction of HOXA5 following RA treatment is co-regulated by HuR and miR-130a. To investigate the mechanisms underlying this regulation, MCF7 cells expressing either control shRNA or HuR shRNA were transfected with miR-130a mimics or inhibitors, followed by the treatment with or without RA. As was expected, knockdown of miR-130a by its inhibitors led to an increase in HOXA5 protein levels (Fig. 4D, lane 1 vs. 3). By contrast, induction of miR-130a by its mimics resulted in a decrease in HOXA5 protein levels after RA treatment (Fig. 4E, lane 2 vs. 4). However, when HuR was knocked down, all these effects were completely reversed (Fig. 4D and E). These data indicate that HuR is required for the function of miR-130a in regulating RA-induced HOXA5 expression. We next determined whether HuR and miR-130a mutually influences each other. As shown in Fig. 4F and G, HuR knock-down did not significantly change the expression levels of miR-130a and Ago2. Additionally, neither mimics nor inhibitors of miR-130a affected HuR protein expression (Fig. 4H), indicating that HuR and miR-130a do not

Fig. 6. HuR and miR-130a regulate RA-induced cell death via HOXA5. (A and B) MCF7 cells were transfected with miR-130a mimics alone or miR-130a mimics plus Flag-HOXA5. Twenty-four hours after transfection, cells were treated with or without RA for another 4 h. (A) Cell viability was measured using the MTT assay. The data were represented as mean ± SD of three independent experiments. (B) Cell lysates were analyzed by Western blotting. (C and D) MCF7 cells expressing either control shRNA or HOXA5 shRNA were transfected with miR-130a inhibitors. Twenty-four hours after transfection, cells were treated with or without RA for another 4 h. (C) Cell viability was measured using the MTT assay. The data were represented as mean ± SD of three independent experiments. (D) Cell lysates were analyzed by Western blotting. (E and F) MCF7 cells expressing either control shRNA or HuR shRNA were transfected with Flag-HOXA5 as indicated. Twenty-four hours after transfection, cells were treated with or without RA for another 4 h. (E) Cell viability was measured using the MTT assay. The data were represented as mean ± SD of three independent experiments. (F) Cell lysates were analyzed by Western blotting. (G) MCF7 cells expressing either control shRNA or HuR shRNA were transfected with Flag-HOXA5 as indicated. Twenty-four hours after transfection, cells were treated with or without RA for another 4 h. Cell lysates were then assayed for caspase-2 and caspase-8 activities. The data were represented as mean ± SD of three independent experiments. * and ** indicate P b 0.05 and P b 0.01, respectively. (H) MCF7 cells were transfected with miR-130a mimics alone or miR-130a mimics plus Flag-HOXA5. Twenty-four hours after transfection, cells were treated with or without RA for another 4 h. Cell lysates were then assayed for caspase-2 and caspase-8 activities. The data were represented as mean ± SD of three independent experiments. * and ** indicate P b 0.05 and P b 0.01, respectively. (I) MCF7 cells expressing either control shRNA or HOXA5 shRNA were transfected with miR-130a inhibitors. Twenty-four hours after transfection, cells were treated with or without RA for another 4 h. Cell lysates were then assayed for caspase-2 and caspase-8 activities. The data were represented as mean ± SD of three independent experiments. * and *** indicate P b 0.05 and P b 0.001, respectively. (J) A hypothetical model of HuR and miR-130a in the regulation of RA-induced HOXA5 expression. Under unstressed condition, miR-130a is highly expressed due to high levels of c-Myc protein in cancer cells. miR-130a post-transcriptionally inhibits HOXA5 expression and renders the cells resistant to apoptosis. However, in response to RA treatment, c-Myc is quickly degraded via the proteasome pathway. This results in the down-regulation of miR-130a expression and de-repression of HOXA5 expression. Additionally, RA treatment increases the binding of HuR to HOXA5 mRNA, which leads to the stabilization of HOXA5 mRNA. The up-regulated HOXA5 ultimately mediates the cytotoxic effect of RA.

F. Yang et al. / Cellular Signalling 25 (2013) 1476–1485

RT-PCR analysis showed that HOXA5 mRNA was specifically enriched in anti-HuR immunoprecipitates, but not in control IgG immunoprecipitates (Fig. 5B), indicating the specific association of HuR with HOXA5 mRNA. Furthermore, HuR knockdown led to a dramatic decrease in HOXA5 mRNA levels (Fig. 5C). These combined data suggest that HuR could bind to and stabilize HuR mRNA. To further confirm this notion, we generated a series of reporter constructs where the different fragments of HOXA5 mRNA 3′UTR were individually cloned into the down-

affect each other. Taken together, our data suggest that HuR plays a critical role in RA-induced HOXA5 expression. 3.5. HuR mediates HOXA5 mRNA stabilization upon RA treatment A detailed analysis revealed that the 3′UTR of HOXA5 mRNA contained two typical adenosine and uridine-rich elements (AREs), which were potential HuR binding sites (Fig. 5A). The subsequent IP–

C

0.5

0 Ctrl miR-130a mimics 3xflag-HOXA5

+ -

+ -

+ +

1

0.5

0 Ctrl miR-130a inhibitor

+ -

+

+

sh HOXA5

-

-

+

B

3xflag-HOXA5 HOXA5

1 0.5

+ -

+ -

+ +

F 3xflag-HOXA5 sh HuR RA -

sh HOXA5 - - - - + + miR-130a inhibitor - - + + + + RA - + - + - + HOXA5

HOXA5 HuR

1 2

1 2 3 4 5 6

- - - + + - + + + + + - + - +

3xflag-HOXA5

β-Actin

β-Actin

RARA+

0 Ctrl sh HuR 3xflag-HOXA5

D

3xflag-HOXA5 - - - - + + miR-130a mimics - - + + + + RA - + - + - +

1.5 (% control)

1

RARA+

Relative viable cells

RARA+

(% control)

1.5

E 1.5

Relative viable cells

Relative viable cells (% control)

A

1483

3 4 5

β-Actin

6

1 2 3 4 5 6

H 2

**

*

*

**

**

**

1.5 1 0.5

2

**

*

*

*

*

*

1.5 1 0.5

Casp 2 + + - - - - - ++ ++ - - - - ++

Casp 8 + + - - - Ctrl mimics - - ++ + + - - - - + + 3flag-HOXA5

RARA+ 2.5

*** *

2

***

Casp 2 + + - - - - - + + + + - - - - + +

c-Myc

1 0.5

Casp 8 + + - - - - - + + + + - - - - + +

Ctrl inhibitor sh HOXA5

Casp 2 + + - - - -

- + + + + - - - + +

RA treatment

HuR

c-Myc

High c-Myc expression

High miR-130a expression

ub ubub ub

Low HOXA5 expression

Cell survival

HuR HuR

c-Myc Low c-Myc expression

***

1.5

HuR

J c-Myc

*** *

0

0

0

Ctrl sh HuR 3flag-HOXA5

I RARA+

Relative caspase activity

RARA+

Relative caspase activity

Relative caspase activity

G

Low miR-130a expression

High HOXA5 expression

Cell death

Casp 8 + + - - - - + + + ++ - - - - ++

1484

F. Yang et al. / Cellular Signalling 25 (2013) 1476–1485

stream of luciferase gene (Fig. 5D). These reporter plasmids were then transfected into MCF7 cells, followed by IP–RT-PCR analysis to determine the association of HuR with the indicated chimeric mRNA. The results revealed that similar to full length 3′UTR of HOXA5 mRNA, both fragments A and B of HOXA5 mRNA 3′UTR containing each of the two AREs can interact with HuR (Fig. 5E). Conversely, the fragment C of HOXA5 mRNA 3′UTR containing no potential AREs did not bind to HuR (Fig. 5E). The subsequent luciferase assay showed that the activities of the reporter constructs containing at least one ARE (FL, A and B) were decreased when HuR was knocked down (Fig. 5F), suggesting that the ARE-containing chimeric mRNAs are unstable after knockdown of HuR. Therefore, these results further indicate that HuR binds to both fragments A and B of HOXA5 mRNA 3′UTR and increases HOXA5 mRNA stability. We next examined whether the association of HuR and HOXA5 mRNA is regulated by RA treatment. As shown in Fig. 5G, treatment with RA increased the HuR–HOXA5 mRNA interaction in a timedependent manner. Moreover, correlated with this enhanced interaction between HuR and HOXA5 mRNA, the stability of HOXA5 was enhanced after treatment with RA (Fig. 5H, blue line vs. pink line). However, when HuR was knocked down, treatment with RA was no longer able to enhance HOXA5 mRNA stability (Fig. 5H, green line vs. yellow line). These findings strongly indicate that treatment with RA enhances the HuR–HOXA5 mRNA binding and thereafter increases HOXA5 mRNA stability. In support of this notion, treatment with RA increased the activity of the luciferase reporter construct containing full length 3′UTR of HOXA5 mRNA in a time-dependent manner in control MCF7 cells, but not in HuR knockdown MCF7 cells (Fig. 5I). Taken together, our data suggest that HuR-mediated HOXA5 mRNA stabilization contributes to RA-induced up-regulation of HOXA5 expression. 3.6. HuR- and miR-130a-mediated HOXA5 regulation contribute to RA-induced cell death HOXA5 has been shown to mediate cytotoxic effect of RA in breast cancer cells [11]. To determine the physiological significance of HuRand miR-130a-mediated HOXA5 regulation, we evaluated the effects of HuR and miR-130a on RA-induced cell death. In consistent with the previous report [11], treatment of MCF7 cells with RA decreased cell viability, which was accompanied by the induction of HOXA5 (Fig. 6A and B). However, when HOXA5 expression was inhibited by miR-130a mimics, RA-induced cell viability inhibition was greatly reversed (Fig. 6A and B). Also, when Flag-HOXA5 was exogenously expressed, miR-130a mimics were no longer able to regulate RA-induced cell viability inhibition (Fig. 6A and B). Furthermore, treatment with miR-130a inhibitors plus RA expectedly decreased cell viability of control MCF7 cells, but failed to show any effect on cell viability of HOXA5 knockdown MCF7 cells (Fig. 6C and D). These data indicate that the miR-130a-HOXA5 axis is important for RA-induced cell death. We next determined the effect of HuR on RA-induced cell death. Compared with control cells, HuR knockdown MCF7 cells were resistant to RA-induced cell viability inhibition (Fig. 6E and F), indicating the important function of HuR in mediating RA-induced cell death. However, after re-introduction of exogenous HOXA5 into HuR knockdown MCF7 cells, these cells became sensitive to RA-induced cell viability inhibition. Collectively, these data suggest that HuR-mediated HOXA5 regulation plays an important role in RA-induced cell death. It has been shown that both caspase-2 and caspase-8 play pivotal roles in mediating RA-induced cell death [14]. We therefore examined the effect of HuR and miR-130a on RA-induced caspase-2 and caspase-8 activation. As expected, treatment with RA resulted in an increase in both caspase 2 and caspase 8 activities in MCF7 cells. Knockdown of HuR effectively inhibited RA-induced caspase-2 and caspase-8 activation (Fig. 6G). Similarly, the induction of miR-130a by its mimics compromised the activation of caspase 2 and caspase 8 induced by RA (Fig. 6H). However,

when HOXA5 was exogenously expressed, these effects of miR-130a and HuR on RA-induced caspase-2 and caspase-8 activation were reversed (Fig. 6G and H). Moreover, co-treatment with RA and miR-130a inhibitors substantially activates caspase-2 and caspase-8 in control MCF7 cells, but not in HOXA5 knockdown MCF7 cells (Fig. 6I). These data indicate that HuR and miR-130a could regulate RA-induced caspase activation via HOXA5. Taken together, these results suggest that both HuR- and miR-130a-mediated HOXA5 regulation contributes to RA-induced cell death. 4. Discussion It has been widely accepted that RA is not only a key morphogen in vertebrate development but is also a potent therapeutic agent for cancer [9,10,23]. The anti-cancer effect of RA is mainly mediated by HOXA5. It has been shown that HOXA5 is up-regulated following RA treatment, which in turn induces either cell growth arrest or cell death. However, the detailed mechanisms underlying RA-induced HOXA5 expression remain largely unknown. Here we provide evidence that both miR-130a and HuR are involved in RA-induced HOXA5 up-regulation, thus contributing to RA-induced cell death. It has previously been reported that HOXA5 expression is lost in more than 60% of primary breast carcinomas [24]. This finding suggests that HOXA5 plays an important role in the initiation and development of breast cancer, and also implicates HOXA5 as a potential therapeutic target for the treatment of cancer. However, since HOXA5 itself is an important and multifunctional transcription factor, as a practical matter, therapeutically targeting HOXA5 is likely to be difficult. It is plausible that targeting the upstream regulatory factors of HOXA5 could represent a more practical strategy for effective and radical anti-cancer therapy. Therefore, our data not only uncover the importance of posttranscriptional modulation in the regulation of RA-induced HOXA5 expression, but also suggest that miR-130a and HuR may represent novel therapeutic targets for cancer treatment. Previous studies have shown that HOXA5 is transcriptionally up-regulated by RARβ in response to RA treatment. However, in some breast cancer cell lines with normal RARβ expression, HOXA5 expression cannot be induced following treatment with RA [11], suggesting the existence of other mechanisms involved in RA-induced HOXA5 expression. In consistent with this notion, we show that miR-130a posttranscriptionally regulates the cellular levels of HOXA5. Intriguingly, miR-130a itself is controlled by the oncoprotein c-Myc. Since c-Myc is amplified in a number of different human cancers, it is conceivable that high levels of c-Myc protein in cancer cells maintain high expression of miR-130a, which in turn inhibits HOXA5 expression and renders these cells resistant to apoptosis. Therefore, our data suggest that the miR-130a-HOXA5 axis may contribute to the oncogenic activity of c-Myc protein by preventing apoptosis. We also show that, in response to RA treatment, c-Myc is quickly degraded by the proteasome. This leads to the down-regulation of miR-130 expression and the subsequent de-repression of HOXA5 expression. These combined data suggest that the c-Myc-miR-130a axis is important for controlling RAinduced HOXA5 expression. Our results also highlight the important role of HuR in RA-induced HOXA5 expression. HuR is a ubiquitously expressed RNA-binding protein involved in the post-transcriptional regulation of gene expression by modulating translation and/or stability of the target mRNA. HuR binds to the target mRNA containing AU-rich element (ARE) [25–27]. Our findings demonstrate that there are two potential AREs in HOXA5 mRNA 3′ UTR. HuR associates with these two sites to increase HOXA5 mRNA stabilization (Fig. 5D–F). More interestingly, the HuR–HOXA5 mRNA interaction becomes much stronger after RA treatment (Fig. 5G), reinforcing the importance of HuR in mediating RA-induced HOXA5 expression. However, the underlying mechanism of how the HuR–HOXA5 binding is regulated upon RA treatment awaits further elucidation. In addition, we also demonstrate that HuR is required for the function of miR-130a

F. Yang et al. / Cellular Signalling 25 (2013) 1476–1485

in regulating RA-induced HOXA5 up-regulation, which is revealed by the observation that the HuR knockdown blocks RA-induced HOXA5 expression regardless of miR-130a expression levels (Fig. 4D and E). One possible explanation could be that HOXA5 mRNA becomes extremely unstable in the absence of HuR expression, and in this situation, miR-130a cannot post-transcriptionally regulate HOXA5 expression. In summary, the findings presented here suggest the existence of the dynamic regulation of HOXA5 expression by HuR and miR-130a in response to RA treatment, and implicate the importance of HuR and miR-130a in the regulation of RA-induced cell death. We here propose a hypothetical model of HuR and miR-130a in the regulation of RA-induced HOXA5 expression and the subsequent cell death as shown in Fig. 6J. Acknowledgments We thank Dr. Ping Gao for kindly providing us with P493-6 cells. This work was supported by Grants (2011CB966302 and 2010CB912804) from the Ministry of Science and Technology of China and Grants (31030046 and 81101525) from the National Natural Science Foundation of China. References [1] V. Ambros, Nature 431 (2004) 350–355. [2] S.N. Bhattacharyya, R. Habermacher, U. Martine, E.I. Closs, W. Filipowicz, Cell 125 (2006) 1111–1124. [3] D.T. Humphreys, B.J. Westman, D.I. Martin, T. Preiss, Proceedings of the National Academy of Sciences of the United States of America 102 (2005) 16961–16966. [4] N. Liu, M. Landreh, K. Cao, M. Abe, G.J. Hendriks, et al., Nature 482 (2012) 519–523. [5] K. Abdelmohsen, S. Srikantan, X. Yang, A. Lal, H.H. Kim, et al., EMBO Journal 28 (2009) 1271–1282.

1485

[6] Y. Kuwano, H.H. Kim, K. Abdelmohsen, R. Pullmann Jr., J.L. Martindale, et al., Molecular and Cellular Biology 28 (2008) 4562–4575. [7] S. Yamasaki, G. Stoecklin, N. Kedersha, M. Simarro, P. Anderson, The Journal of Biological Chemistry 282 (2007) 30070–30077. [8] H. Yu, Y. Sun, C. Haycraft, V. Palanisamy, K.L. Kirkwood, Cytokine 56 (2011) 245–255. [9] A. Decensi, D. Serrano, B. Bonanni, M. Cazzaniga, A. Guerrieri-Gonzaga, Journal of Mammary Gland Biology and Neoplasia 8 (2003) 19–30. [10] T.R. Evans, S.B. Kaye, British Journal of Cancer 80 (1999) 1–8. [11] H. Chen, H. Zhang, J. Lee, X. Liang, X. Wu, et al., Cancer Research 67 (2007) 8007–8013. [12] L. Jin, W.L. Hu, C.C. Jiang, J.X. Wang, C.C. Han, et al., Proceedings of the National Academy of Sciences of the United States of America 108 (2011) 15840–15845. [13] Z. Yao, S. Duan, D. Hou, W. Wang, G. Wang, et al., Oncogene 29 (2010) 1821–1834. [14] H. Chen, S. Chung, S. Sukumar, Molecular and Cellular Biology 24 (2004) 924–935. [15] W. Hu, J. Coller, Cell Research 22 (2012) 1322–1324. [16] G. Meister, M. Landthaler, A. Patkaniowska, Y. Dorsett, G. Teng, et al., Molecular Cell 15 (2004) 185–197. [17] J. Liu, M.A. Carmell, F.V. Rivas, C.G. Marsden, J.M. Thomson, et al., Science 305 (2004) 1437–1441. [18] Y. Chen, D.H. Gorski, Blood 111 (2008) 1217–1226. [19] L. Ma, J. Young, H. Prabhala, E. Pan, P. Mestdagh, et al., Nature Cell Biology 12 (2010) 247–256. [20] K.A. O'Donnell, E.A. Wentzel, K.I. Zeller, C.V. Dang, J.T. Mendell, Nature 435 (2005) 839–843. [21] J.H. Schulte, S. Horn, T. Otto, B. Samans, L.C. Heukamp, et al., International Journal of Cancer 122 (2008) 699–704. [22] K. Mazan-Mamczarz, S. Galban, I. Lopez de Silanes, J.L. Martindale, U. Atasoy, et al., Proceedings of the National Academy of Sciences of the United States of America 100 (2003) 8354–8359. [23] R.M. Niles, Nutrition 16 (2000) 1084–1089. [24] V. Raman, S.A. Martensen, D. Reisman, E. Evron, W.F. Odenwald, et al., Nature 405 (2000) 974–978. [25] K. Abdelmohsen, R. Pullmann Jr., A. Lal, H.H. Kim, S. Galban, et al., Molecular Cell 25 (2007) 543–557. [26] V.E. Myer, X.C. Fan, J.A. Steitz, EMBO Journal 16 (1997) 2130–2139. [27] S.S. Peng, C.Y. Chen, N. Xu, A.B. Shyu, EMBO Journal 17 (1998) 3461–3470.