Sustained activation of ERK1/2 by NGF induces microRNA-221 and 222 in PC12 cells

Authors


G. Tsujimoto, 45-29 Yoshida-Shimo-Adachi-cho, Sakyo-ku, Kyoto 606-8501, Japan
Fax: +81 75 753 4523
Tel: +81 75 753 4544
E-mail: gtsuji@pharm.kyoto-u.ac.jp

Abstract

MicroRNAs (miRNAs) are small non-coding RNAs that regulate gene expression by inhibiting translation and/or inducing degradation of target mRNAs, and they play important roles in a wide variety of biological functions including cell differentiation, tumorigenesis, apoptosis and metabolism. However, there is a paucity of information concerning the regulatory mechanism of miRNA expression. Here we report identification of growth factor-regulated miRNAs using the PC12 cell line, an established model of neuronal growth and differentiation. We found that expression of miR-221 and miR-222 expression were induced by nerve growth factor (NGF) stimulation in PC12 cells, and that this induction was dependent on sustained activation of the extracellular signal-regulated kinase 1 and 2 (ERK1/2) pathway. Using a target prediction program, we also identified a pro-apototic factor, the BH3-only protein Bim, as a potential target of miR-221/222. Overexpression of miR-221 or miR-222 suppressed the activity of a luciferase reporter activity fused to the 3′ UTR of Bim mRNA. Furthermore, overexpression of miR-221/222 decreased endogenous Bim mRNA expression. These results reveal that the ERK signal regulates miR-221/222 expression, and that these miRNAs might contribute to NGF-dependent cell survival in PC12 cells.

Abbreviations
EGF

epidermal growth factor

ERK

extracellular signal-regulated kinase

MAPK

mitogen-activated protein kinase

MEK

MAPK/ERK kinase

miRNA

microRNA

NGF

nerve growth factor

MicroRNAs (miRNAs) are evolutionally conserved small non-coding RNAs that regulate gene expression at the post-transcriptional level and play important roles in a wide variety of biological functions, including cell differentiation, tumorigenesis, apoptosis and metabolism [1–3]. Approximately 30% of human protein-coding genes are predicted to be targets of miRNA [4,5]. Biogenesis of miRNA and the mechanism for regulation of target gene expression by miRNA are relatively well characterized. miRNA genes are initially transcribed mainly by RNA polymerase II as long primary transcripts (pri-miRNAs), processed by the nuclear RNase Drosha to produce precursor miRNAs (pre-miRNAs), and then exported to the cytoplasm. Pre-miRNAs are further processed into mature miRNAs by the cytoplasmic RNase Dicer [6]. miRNAs recognize and bind to partially complementary sites in the 3′ UTRs of target mRNAs, resulting in either translational repression or target degradation [7]. To further understand the functional significance of miRNA, the regulatory mechanism of miRNA expression needs to be better understood.

The mitogen-activated protein kinase (MAPK) cascades play an essential role in transducing extracellular signals to cytoplasmic and nuclear effectors, and regulate a wide variety of cellular functions, including cell proliferation, differentiation and stress responses [8,9]. Extracellular signal-regulated kinases 1 and 2 (ERK1/2) were the first recognized members of the MAPK family of proteins. These kinases are primarily activated by mitogenic factors, differentiation stimuli and cytokines such as nerve growth factor (NGF) and epidermal growth factor (EGF) [10–12]. Because both ERK signaling and miRNA function are involved in a variety of important biological responses, the significance of ERK signaling in terms of regulating miRNA expression is of great interest.

To study the role of the ERK1/2 pathway in the regulation of miRNA expression, we first determined the expression profile of miRNAs by using the growth factor-induced neural differentiation process of PC12 as a model. It is well known that NGF, but not EGF, induces neural differentiation in PC12 cells, although both growth factors potently activate ERK1/2 [13–16]. We identified miR221 and 222 as differentially regulated miRNAs. Although the expression of these miRNAs was found to be ERK-dependent, the effect of NGF and EGF on their expression was different; thus, only NGF, but not EGF, can induce their expression. Further, our study showed that the sustained activation of ERK1/2 by NGF, but not the transient activation of ERK, could effectively induce miR221 and 222 in PC12 cells. Finally we identified the BH3-only protein Bim, which is involved in NGF-dependent neuronal survival [17–19], as a potential target of miR-221 and 222.

Results

NGF stimulation induces miR-221/222 in PC12 cells

Treatment of PC12 cells with NGF for 48 h induced readily detectable neurite outgrowth (Fig. 1A), so we selected two points at 0 and 48 h after stimulation to compare the expression of miRNAs. We used a TaqMan miRNA assay, featuring reverse transcription using stem-loop RT-PCR primers followed by real-time PCR using TaqMan probes [20], to examine the expression of 156 rat miRNAs. We identified only two miRNAs, miR-221 and 222, as drastically up-regulated 48 h after NGF stimulation (data not shown). These miRNAs are encoded in tandem on the chromosome X (Fig. 1B).

Figure 1.

 NGF induces expression of the microRNAs miR-221 and 222 in PC12 cells. (A) NGF-induced differentiation of PC12 cells. (B) Schematic representation of the genomic structure of miR-221 and 222 and their corresponding sequences. (C, F) PC12 cells were treated with 100 ng·mL−1 NGF or 30 nm EGF for the indicated times. Cells were harvested and total RNA was prepared. The RNA was subjected to quantitative RT-PCR to assess the levels of mature miR-221/222. The data represent means of the Ct values (± SD, = 3). In (C), *< 0.01 for miR-221 versus 0 h point, and < 0.01 for miR-222 versus 0 h point. In (F), < 0.01 for comparisons indicated by asterisks (Tukey’s test). (D, E) Cell extracts were subjected to immunoblotting with α-phospho-ERK1/2 and α-ERK1/2 IgGs.

Next, we examined the time course of expression of miR-221 and 222. Quantitative RT-PCR analysis showed that these miRNAs had a very similar profile (Fig. 1C). An alternative RT-PCR analysis, using a set of primers that amplify a fragment located between these two miRNAs, demonstrated transcriptional induction of this region upon NGF stimulation (data not shown). These data support the notion that miR-221 and 222 derive from the same pri-miRNA [21]. Following NGF stimulation, the expression level of both miRNAs rapidly increased and reached a maximum at 3–6 h, which was then sustained to the last time point assayed at 48 h. Over this time course, expression of an internal control U6 remained unchanged (Fig. 1C). Based on the threshold cycle (Ct) changes, we estimate that the expression of miR-221 and 222 had increased by approximately 26-fold. NGF induced sustained activation of ERK1/2 over this time course (Fig. 1D). We further examined whether EGF stimulation also induces miR-221/222 in PC12 cells. Our analysis confirmed that EGF transiently activated ERK1/2, unlike NGF [13–16] (Fig. 1E). Because NGF-mediated expression of miR-221/222 peaked approximately 3 h after stimulation, we examined the expression of miR-221/ 222 at this time point in all further experiments. In contrast to NGF stimulation, EGF did not induce any detectable up-regulation of miR-221 and 222 at 3 h (Fig. 1F). Moreover, no such stimulation was found even after monitoring for up to 48 h (data not shown).

Sustained activation of ERK1/2 is necessary for induction of miR-221/222

We next studied whether the NGF-induced expression of miR-221/222 depends on ERK1/2 activation. We first examined the effect of a specific inhibitor (U0126) for MAPK/ERK kinase (MEK) 1/2, which is a direct activator of ERK1/2 [22]. As shown in Fig. 2A, pre-treatment of U0126 potently inhibited NGF-induced ERK1/2 activation. The same pre-treatment with U0126 completely blocked induction of miR-221 and 222 (Fig. 2B). Moreover, we found that expression of miR-221/222 dramatically increased when constitutively active MEK1 (MEK1SDSE) [23] was transiently expressed in PC12 cells (Fig. 2C,D).

Figure 2.

 Activation of ERK1/2 pathway is involved in miR-221/222 induction. (A, B) PC12 cells were pre-treated with 10 μm U0126 for 10 min before treatment with 100 ng·mL−1 NGF for 3 h. Inhibition of ERK1/2 activity by U0126 was confirmed by immunoblotting (A). The expression levels of miR-221/222 were examined as described in Fig. 1C (*< 0.05, Tukey’s test) (B). (C, D) PC12 cells were transfected either with empty vector or HA-tagged MEK1SDSE. The expression levels of miR-221/222 were examined 24 h after transfection as described in Fig. 1C (*< 0.05 versus empty vector, Student’s t test) (C). Expression of HA-tagged MEK1SDSE was confirmed by immunoblotting (D). DMSO, dimethylsulfoxide.

Taken together, these results indicate that induction of miR-221/222 depends on the activation of ERK1/2. However, transient activation of ERK/2 upon EGF stimulation did not induce miR-221/222 expression. This observation prompted us to hypothesize that induction of these miRNAs requires sustained activation of ERK1/2. To verify this hypothesis, we blocked NGF-induced sustained ERK1/2 activation by adding U0126 10 min after NGF treatment (Fig. 3A). As shown in Fig. 3B, addition of U0126 completely inhibited the sustained activation of ERK1/2. In this situation, the induction of miR221/222 was also completely suppressed (Fig. 3C). These results clearly demonstrate that sustained activation of ERK1/2 is required for induction of miR-221 and 222.

Figure 3.

 Sustained activation of ERK1/2 is required for NGF-induced miR-221/222 expression. (A) Schematic diagram of the experimental design. (B, C) PC12 cells were pre-treated with 10 μm U0126 for 10 min after treatment with 100 ng·mL−1 NGF for 3 h. Inhibition of sustained ERK1/2 activation by U0126 was confirmed by immunoblotting (B). The expression levels of miR-221/222 were examined as described in Fig. 1C (*< 0.05, Tukey’s test) (C). DMSO, dimethylsulfoxide.

However, the apparent induction of these miRNA molecules was only observed approximately 3 h after NGF stimulation (Fig. 1C), which implies that miR-221 and 222 are not the primary target genes regulated by NGF. We reasoned that this induction requires de novo protein synthesis. As shown in Fig. 4A, treatment with cycloheximide completely inhibited induction of these miRNAs, but had no significant effect on ERK1/2 activation (Fig. 4B). We also confirmed inhibition of protein synthesis by monitoring expression of c-Fos protein, which is a well-known NGF-induced immediate early gene [24]. Cycloheximide treatment completely blocked c-Fos protein synthesis (Fig. 4C, compare with control). These data indicate that de novo protein synthesis is required for the induction of miR-221/222.

Figure 4.

 Protein synthesis is required for NGF-induced miR-221/222 expression. PC12 cells were pre-treated with 10 μg·mL−1 cycloheximide for 30 min before treatment with 100 ng·mL−1 NGF for 3 h. The expression levels of miR-221/222 were examined as described in Fig. 1C (*< 0.05, Tukey’s test) (A). ERK1/2 activation in the presence of cycloheximide (B) and inhibition of protein synthesis by cycloheximide (C) were confirmed by immunoblotting. α-Tubulin was used as a loading control. DMSO, dimethylsulfoxide.

Pro-apototic Bim is a plausible target of miR-221/222

We used TargetScan [4] to identify the likely target genes of miR-221/222. Specifically, we focused on the pro-apototic Bim gene because Bim has been reported to be involved in NGF-dependent survival of PC12 cells [19]. The predicted target site for these miRNAs is conserved in human, mouse, rat, dog and chicken (Fig. 5A). The rat Bim gene had no annotated 3′ UTR, and so in the TargetScan program this is computationally determined based on the human Bim 3′ UTR sequence. Initially, we used RACE to verify whether the predicted 3′ UTR region is transcribed. 3′ RACE analysis detected products containing the terminal portion of the predicted Bim 3′ UTR. Moreover, RT-PCR analysis showed that a fragment containing the predicted target site was amplified (data not shown). To examine whether these miRNAs can target the Bim gene, we generated a luciferase construct harboring a fragment of the Bim 3′ UTR containing the target sequence (Fig. 5A). Co-expression of either miR-221 or miR-222 significantly (< 0.05) suppressed the reporter activity compared to the control (Fig. 5C, wt). Mutations introduced into the predicted binding site almost eliminated this suppression. These results suggest a direct interaction of these miRNAs with the predicted target site of the Bim 3′ UTR (Fig. 5B,C, mt). Furthermore, we investigated the effect of expression of miR-221/222 on endogenous Bim mRNA expression. Because the Bim gene has three alternative splice variants [25], we designed a set of primers to detect all three products. Overexpression of either miR-221 or miR-222 resulted in significant (< 0.05) down-regulation of Bim mRNA (Fig. 5D). To show that this down-regulation is a specific effect for Bim mRNA, we examined the mRNA level of an apoptosis-related gene, Bax, because the 3′ UTR of Bax mRNA has no predicted target site for miR-221/222. We found that overexpression of either miR-221 or miR-222 had no significant effect on the Bax mRNA level (data not shown). Taken together, our data suggest that miR-221 and 222 can target the Bim gene.

Figure 5.

 The Bim 3′ UTR is a target of miR-221/222. (A) Schematic representation of the reporter construct and conservation of the target site of miR-221/222 in the Bim 3′ UTR in vertebrates. The boxed region indicates the sites complementary to the seed sequence of miR-221/222. (B) Predicted base pairing between miR-221 and 222 and their target sites in the Bim 3′ UTR. Underlining indicate the seed sequences. The calculated free energy of hybridization of each miRNA with the target site is also indicated. These data were prepared using RNAhybrid. Lower-case letters indicate the sites introduced to the Bim 3′ UTR by mutation. The indicated bases were substituted with complementary nucleotides in the mutation construct. (C) The indicated RNA oligonucleotides (10 pmol per well) and reporter plasmid (200 ng per well) were co-transfected with the internal control plasmid (20 ng per well) into PC12 cells. After 24 h, the cells were harvested, and the lysates were subjected to a luciferase assay. The results are the ratio of firefly to renilla luciferase activity (mean ± SD, = 3). The luciferase activity ratio for control RNA transfection (mock) for each reporter was set at 1 (*< 0.01 versus mock; Student’s t test). (D) PC12 cells were transfected with the indicated RNA oligonucleotides. After 24 h, the cells were harvested and total RNA was prepared. The RNA was subjected to quantitative RT-PCR to assess the levels of Bim mRNA. The Bim mRNA expression was normalized against GAPDH mRNA expression (mean ± SD, = 3). The normalized Bim expression of control RNA transfection (mock) was set at 1 (*< 0.05 versus mock; Student’s t test).

Discussion

The present study has demonstrated that, in PC12 cells, miR-221 and 222 are transcriptionaly induced upon stimulation by NGF, and that this induction requires sustained ERK1/2 activation and de novo protein synthesis. Presumably, sustained ERK1/2 activation is required for the induction of transcriptional regulatory protein(s) that regulates miR-221/222 expression. Recently, miR-155 expression has been shown to be regulated by two MAPK pathways, the ERK1/2 and c-Jun N-terminal kinase pathways, through AP-1 signaling [26]. AP-1 family proteins are good candidates for mediating NGF-induced miR-221/222 expression in PC12 cells.

A previous study showed that miR-221 and 222 are up-regulated in quiescent cells that have been stimulated to proliferate by serum stimulation [27]. ERK1/2 is known to play a critical role in growth factor-stimulated cell-cycle progression from G0/G1 to S phase, and sustained activation is required for this progression [28]. Our finding that sustained ERK1/2 activation induces miR-221/222 expression is entirely consistent with this observation. miR-221 and 222 have also been reported to be up-regulated in some cancer cell lines and to function as oncogenic miRNAs by targeting the cyclin-dependent kinase inhibitor p27Kip1 [29–32]. These studies strongly suggest that miR-221 and 222 are involved in the regulation of cell growth. In PC12 cells, NGF stimulation of starved cells promotes cell survival and differentiation, and inhibits cell-cycle progression [16]. However, NGF-induced miR-221 and miR-222 expression in PC12 cells is probably not involved in cell-cycle progression. These apparently contradictory effects might be attributed to cell type-dependent differences.

We could not observe any detectable effect of NGF-induced neurite outgrowth when miR-221 and/or 222 were overexpressed in PC12 cells (data not shown). However, we show that the pro-apototic protein Bim is a plausible target of miR-221/222. Bim is known to be regulated at both the transcriptional and post-transcriptional level [33]. Here, we have confirmed that Bim is regulated at the translational level. In PC12 cells, the Bim gene is induced upon withdrawal of serum. When NGF stimulation occurs, activation of ERK1/2 causes phosphorylation of Bim proteins and inhibits their function, resulting in cell survival. This translational regulation of Bim ensures the down-regulation of Bim function and hence cell survival. These data indicate that NGF-induced miR-221/222 expression is involved in NGF-dependent cell survival. Interestingly, recent reports have shown that Bim is regulated by miR-32 and the miR-17–92 cluster of miRNAs, which are known to be up-regulated in several cancers [34–36]. Up-regulation of these miRNAs, which results in down-regulation of pro-apoptotic Bim mRNA, exerts an anti-apoptotic effect in some cancers [35,36]. The miR-17–92 cluster has also been shown to be involved in B-cell development and to target Bim mRNA [37,38]. In PC12 cells, the expression level of these miRNAs is moderate and remains unchanged upon NGF stimulation (data not shown). Thus, induced miR-221 and 222 might work cooperatively with these miRNAs.

miR-221 and 222 are highly conserved in vertebrates and have the same seed sequence. Moreover, they are encoded in tandem on the same chromosome in human, mouse and rat, indicating that they have important roles in biological processes. In this study, we found that miR-221/222 expression is regulated by the ERK1/2 pathway in PC12 cells. This regulation might also function in different biological processes. Our findings provide new insights into the MAPK signaling pathway.

Experimental procedures

Cell culture and transfection

PC12 cells were maintained in Dulbecco’s modified Eagle’s medium plus 10% fetal bovine serum, 5% donor horse serum and antibiotics at 37 °C in 5% CO2. The cells were seeded onto poly-l-lysine-coated 60 mm dishes (AGC Techno Glass Co. Ltd, Chiba, Japan) and incubated in a low concentration of serum (1% horse serum) for 24 h prior to treatment with 100 ng·mL−1 NGF (Alomone Labs Ltd, Jerusalem, Israel) or 30 nm EGF (PeproTech EC Ltd, London, UK). Transfections were performed according to the manufacturer’s instructions using LipofectAMINE 2000 (Invitrogen, Carlsbad, CA, USA). The miRNA precursors miR-221 and 222 and control RNA were purchased from Ambion (Austin, TX, USA).

RNA isolation and RT-PCR analysis

Total RNA was isolated using ISOGEN reagent (Nippon Gene Co. Ltd, Tokyo, Japan). miRNA expression was measured using TaqMan MicroRNA Assays (Applied Biosystems, Lincoln, CA, USA) according to the manufacturer’s protocol, except that all reactions were carried out at half scale. The rat miRNAs assayed in this study are listed in the microrna assay index file version 1 (Applied Biosystems). U6 snRNA was used as an internal control. For detection of Bim mRNA, reverse transcription was performed using a QuantiTect reverse transcription kit (Qiagen, Hilden, Germany). Prepared cDNAs were then subjected to quantitative PCR analysis using Power SYBR Green PCR Master Mix (Applied Biosystems). The primers for the PCR analysis were 5′-CTTCCATAAGGCAGTCTCAG-3′ and 5′-CGGAAGATGAATCGTAACAG-3′ for Bim, and 5′-TTGCTGACAATCTTGAGGGAG-3′ and 5′-GAGTATGTCGTGGAGTCTACTG-3′ for glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The data were obtained and analyzed using an ABI 7300 real-time PCR system (Applied Biosystems).

Plasmid construction

The plasmid pcDNA3-HA encoding Xenopus MEK1SDSE was supplied by E. Nishida (Graduate School of Biostudies, Kyoto University, Japan). A partial fragment of the 3′ UTR of rat Bim was amplified by PCR using the primers 5′-CCTGCCTCTTGAGGTACTGC-3′ and 5′-AGCTAGTCGCAAGTTTTA-3′ following reverse transcription from total RNA isolated from PC12 cells, and cloned into the pCR-Blunt vector (Invitrogen). Mutagenesis of the Bim 3′ UTR was performed using a QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA). An EcoRI site was introduced into the XbaI site of the luciferase reporter vector pGL4.23 (Promega, Madison, WI, USA) by ligation with the oligonucleotides 5′-CTAGACTGAATTC-3′ and 5′-CTAGGAATTCAGT-3′, yielding the pGL4.23EcoRI vector. EcoRI fragments of the wild-type (wt) and mutated (mt) Bim 3′ UTR forms of the pCR-Blunt vector were ligated into the EcoRI site of the pGL4.23EcoRI vector. The identity of all constructs was confirmed by DNA sequencing.

Immunoblotting

Cells were harvested by scraping from culture dishes in hot 1× SDS sample buffer, and the lysates were separated by SDS–PAGE and analyzed by immunoblotting. Anti-HA (3F10) rat monoclonal IgG was purchased from Roche (Basel, Switzerland). Anti-p44/42 MAP kinase, anti-phospho-p44/42 MAP kinase IgGs (numbers 9101 and 9102, respectively) and anti-c-Fos IgG (number 4384) were obtained from Cell Signaling (Danvers, MA, USA). Anti-α-Tubulin (B-5-1-2) mouse monoclonal IgG was purchased from Sigma (St Louis, MO, USA). Peroxidase-linked secondary antibodies were purchased from GE Healthcare (Chalfont St Giles, UK). An LAS3000 CCD imaging system (Fujifilm, Tokyo, Japan) was used for detection.

Reporter assay

Cells grown in 24-well plates (1.0 × 105 cells per well) were harvested for assays 24 h after transfection. The luciferase activity was measured using a dual-luciferase reporter assay system (Promega) with a Lumat LB9507 luminometer (Berthold Technologies, Bad Wildbad, Germany). As an internal control, a renilla luciferase vector pGL4.70 (Promega) was used. The data represent means and standard deviations of three independent experiments.

Statistical analysis

The data were analyzed using Student’s t test or ANOVA followed by Tukey’s test as indicated.

Acknowledgements

This work was supported in part by a grant from the Ministry of Education, Culture, Sports, Science and Technology of Japan (to K.T., F.S., K.S. and G.T.), the New Energy and Industrial Technology Development Organization (to K.T. and G.T.), and the Uehara Memorial Foundation (to G.T.).

Ancillary