miRNA-196b inhibits cell proliferation and induces apoptosis in HepG2 cells by targeting IGF2BP1

In order to explore the role of miRNAs in chemoresistance induced by hypoxia, the expression level of 365 miRNAs was compared in HepG2 cells exposed to normoxia or hypoxia with or without etoposide during 16 h. We identified 11 miRNAs for which expression was dysregulated with a fold change higher than 1.5 in cells incubated under normoxia with etoposide when compared to cells incubated under hypoxia with etoposide. The expression of 4 miRNAs was significantly up-regulated in cells incubated under hypoxia with etoposide compared to cells incubated under normoxia with etoposide (hsa-miR-220, −219, −210, −193a) while the expression of 7 miRNAs was significantly down-regulated in cells incubated under hypoxia compared to cells incubated under normoxia in the presence of etoposide (fold change > −1.5) (Figure 1A). To confirm the results of the TaqMan low-density array (TLDA) analysis, RT-qPCR assays were performed to evaluate miRNAs differentially expressed (Figure 1B). Among the 11 differentially expressed miRNAs identified by TLDA, changes in expression for only 3 miRNAs were validated. Etoposide significantly increased the miR-565 and miR-196b expression and hypoxia significantly decreased the expression of these two miRNAs induced by etoposide. On the contrary, hypoxia exposure with etoposide significantly increased miR-210 expression compared to cells incubated under normoxia with etoposide (Figure 1B). These results were consistent with the results obtained with the TLDA analysis.

However, miRBase.org indicated that the sequence of miR-565 is more likely a tRNA fragment and not a miRNA [21]. Thus, we decided to focus our study on the involvement of miR-210 and miR-196b in the chemoresistance induced by hypoxia.

We previously evidenced that hypoxia protects HepG2 cells from etoposide-induced apoptosis. In order to investigate the potential role of miR-196b and miR-210 in the hypoxia-induced resistance to etoposide, we performed western blot analysis for apoptotic markers in cells in which miR-210 or miR-196b was overexpressed by transfection with miRNA precursor (pre-miR). As shown in Figure 2, hypoxia decreased PARP and caspase 3 cleavage induced by etoposide both after 24 and 48 h incubation in the presence of etoposide in untransfected cells.

miR-210 overexpression decreased PARP and caspase 3 cleavage in cells exposed to normoxia with etoposide for 48 h, reaching level similar than the one observed in cells exposed to hypoxia with etoposide, in which miR-210 expression level is high (Figure 2A). These results were validated by a caspase 3/7 activity assay. Results showed that miR-210 overexpression significantly decreased the caspase 3/7 activity in cells incubated under normoxia with etoposide (Figure 2B). It must be noted that overexpression of miR-210 by pre-miR-210 transfection did not have any effect under hypoxia. Indeed, miR-210 is overexpressed under hypoxia and there is a strong protective effect of hypoxia on the etoposide-induced apoptosis. There is no difference in caspase 3 activity level, nor in caspase 3 and PARP cleavage in cells incubated with etoposide under hypoxia in comparison to control normoxic cells, so it is very unlikely that increasing miR-210 level could exert its anti-apoptotic function in these conditions. As our goal was to identify a miRNA that exerts a modulatory effect of the etoposide-induced apoptosis under hypoxia, we next focused our interest on the role of miR-196.

miR-196b overexpression strongly increased PARP and caspase 3 cleavage in cells incubated with etoposide under normoxia or hypoxia for 24 or 48 h (Figure 2C). As shown in Figure 2D, miR-196b overexpression induced a significant increase in caspase 3/7 activity. This activity reached comparable levels in cells incubated with etoposide under normoxia or hypoxia. Thus, miR-196b overexpression not only enhanced the apoptosis induced by etoposide under normoxia but also reversed the hypoxia-induced chemoresistance.

To further confirm the effect of miR-196b in hypoxia-induced chemoresistance, HepG2 cells were transfected with LNA-miR-196b (Locked Nucleic Acid) to inhibit the miR-196b expression. However, by using several LNA negative controls (LNA CTL- #1, LNA CTL- #2 or a mix of LNA CTL- #1 + #2), we observed that LNA negative controls had a strong impact on apoptotic markers studied (data not shown). It was thus impossible to interpret the results obtained with LNA-miR-196b.

Taken together, these results showed that miR-196b displays a pro-apoptotic function.

To unravel the potential role of miR-196b in HepG2 cells and to identify targets of miR-196b, we used a proteomic approach. HepG2 cells were transfected with pre-miR-196b or pre-miR negative control for 24 h. Cells were then harvested, lysed and subjected to subcellular fractionation to recover the cytosolic fraction. Relative protein abundance differences between negative control cells and cells that overexpress pre-miR-196b were then analyzed by 2D-DIGE gels in four independent replicates (Figure 3). Figure 3A shows a pair of representative 2-DE gel images of HepG2 cells transfected to pre-miR negative control compared to pre-miR-196b. Proteins from spots with significant changes in intensities (p < 0.05 up- or down-regulated) were cut out for identification by mass spectrometry. Thirty four spots differentially abundant were detected, of which 13 spots were up-regulated and 21 spots were down-regulated. The complete list of the 172 proteins identified by mass spectrometry is detailed in Additional file 1. We next compared the list of differentially expressed proteins from our proteomic study to the respective list of computationally predicted targets by several softwares (TargetScan, MicroCosm and miRDB) (Figure 3B). Since the number of differentially expressed proteins is typically much lower than the total number of predicted targets, we looked for proteins detected in our proteomic experiment that were also predicted as targets by the bioinformatics studies. We identified 8 differentially expressed proteins that were also predicted by at least one of the 3 algorithms as targets of miR-196b (Figure 3B). Among the 8 potential targets for miR-196b, three of them were involved in DNA (MCM4 cDNA and CTPS) or in protein (EEF2K) synthesis and the link between these processes and apoptosis is not clear. Similarly, AKR1B10 is an aldo-keto reductase enzyme that acts as all-trans-retinaldehyde reductase while LTA4H is an isoform of leukotriene A4 hydrolase. These metabolic pathways are most likely not involved either in regulating apoptosis. The last two proteins (retrotransposon-derived protein PEG10 and C22orf28) are of unknown function. We thus chose IGF2BP1 to investigate its role in the hypoxia-induced chemoresistance. IGF2BP1 was down-regulated by miR-196b overexpression as revealed by the zoom of spot 952 (Figure 3C) in the proteomic analysis as well as predicted to be a target of miR-196b by the TargetScan software.

TargetScan 6.2 revealed 3 putative miR-196b seed match sites present in the 3′UTR of IGF2BP1 mRNA (Figure 4A). IGF2BP1 is a RNA-binding protein that controls the localization, translation and/or turnover of specific target transcripts [22]. To experimentally validate that IGF2BP1 is an endogenous target of miR-196b, HepG2 cells were transfected with pre-miR-196b or pre-miR negative control and the IGF2BP1 mRNA and protein levels were measured by RT-qPCR and western blot, respectively (Figure 4B-4C). As shown in Figure 4B, IGF2BP1 mRNA abundance was significantly reduced by miR-196b overexpression compared to cells transfected with negative control 24 and 48 h after transfection. Similarly, the abundance of IGF2BP1 protein was decreased by miR-196b overexpression when compared to pre-miR negative control 24, 48 and 72 h after the transfection (Figure 4C). Even if miRNAs can act mainly through translational repression rather than mRNA degradation, our data showed that miR-196b regulates both IGF2BP1 mRNA and protein expression levels. These results suggest that IGF2BP1 could be a direct target of miR-196b.

To demonstrate that the negative regulatory effects of miR-196b exerted on IGF2BP1 expression were mediated through the binding of miR-196b to the predicted sites in the 3′UTR of IGF2BP1 mRNA, a reporter plasmid (pmiRGLO IGF2BP1 3′UTR) containing a part of IGF2BP1 3′UTR which includes 2 predicted binding site (out of 3 sites), downstream of the firefly luciferase reporter plasmid, was used (Figure 4D). The reporter plasmid and pre-miR negative control (or pre-miR-196b) were co-transfected in HepG2 cells. As expected, miR-196b overexpression resulted in a significant decrease in the luciferase reporter activity compared to cells transfected with pre-miR negative control (Figure 4E). Furthermore, a mutated reporter plasmid containing 3 nucleotide mutations in the miR-196b seed match sites in the IGF2BP1 mRNA 3′UTR was used (Figure 4D). In contrast to the wild-type reporter plasmid, miR-196b had no significant effect on the reporter luciferase activity of the mutated plasmid, indicating that miR-196b interacts directly with 3′UTR of IGF2BP1 (Figure 4E). These results demonstrated that miR-196b directly targets the 3′UTR of IGF2BP1 mRNA leading to the down-regulation of its expression.

Taken together, proteomic analysis, western blot, RT-qPCR and luciferase activity data provide strong evidence that IGF2BP1 mRNA is a direct target of miR-196b.

To study effects of IGF2BP1 down-regulation induced by miR-196b, HepG2 cells were transfected with either pre-miR-196b or with IGF2BP1 siRNA and cell viability, proliferation and apoptotic profile were evaluated in normal culture conditions.

We assessed the cell morphology by phase contrast microscopy, 72 h after pre-miR-196b or IGF2BP1 siRNA transfection. miR-196b overexpression reduced the cell number and floating cells were observed. On the other hand, IGF2BP1 silencing seems to change the cell morphology since an increase in cell size was observed (Figure 5A).

The number of viable cells was then assessed using a MTT assay. Results showed that the number of untransfected cells or cells transfected with negative controls (pre-miR negative control or RF) increased with time, suggesting cell proliferation. miR-196b overexpression decreased significantly the number of viable cells by 47 and 43.7%, 72 and 96 h post-transfection, respectively. IGF2BP1 silencing also decreased the number of viable cells by 51.9 and 65.2%, 72 and 96 h post-transfection, respectively (Figure 5B).

Proliferation was then directly measured by using BrdU incorporation 72 h after transfection. Results showed that the proliferation rate was significantly lower when cells overexpressed miR-196b compared to cells transfected with pre-miR negative control (33.8% inhibition of cell proliferation). A significant decrease was also observed when cells were transfected with IGF2BP1 siRNA compared to RISC Free 72 h post-transfection (59.2% inhibition of cell proliferation) (Figure 5C).

We then assessed apoptosis by studying PARP and caspase 3 cleavage and measuring caspase 3/7 activity assay. As already shown in Figure 4C, miR-196b overexpression decreased IGF2BP1 protein abundance. In addition, miR-196b overexpression induced an increase in PARP cleavage 48 and 72 h post-transfection (Figure 6A, left panel). However, even if we did not detect an increase in caspase 3 cleavage by western blot, miR-196b overexpression induced a significant increase in caspase 3/7 activity by 54%, 48 h post- transfection (Figure 6B). To evaluate if IGF2BP1 is involved in the pro-apoptotic effects of miR-196b, IGF2BP1 expression was silenced using siRNA. As shown in Figure 6A (right panel), IGF2BP1 silencing strongly decreased IGF2BP1 protein abundance 24, 48 and 72 h after transfection (56, 72 and 80% of inhibition, respectively). The inhibition induced an increase in PARP cleavage but not in caspase 3 cleavage. In parallel, IGF2BP1 silencing induced a significant increase in caspase 3/7 activity by 13% but to lower extent than the one induced by miR-196b overexpression (Figure 6B, right panel). These results suggest that miR-196b promoted apoptosis and that IGF2BP1 contributes at least in part to the pro-apoptotic action of miR-196b.

Taken together, these results showed that IGF2BP1 silencing as well as miR-196b overexpression had similar effects resulting in cell proliferation inhibition and in apoptosis initiation in HepG2 cells.

A major function of IGF2BP1 is the stabilization of numerous mRNAs. Among the known IGF2BP1 target mRNAs are actin and c-myc [23,24]. Myc represents an interesting target because of its role as a proto-oncogene. Furthermore, it was shown to be a target of miR-196b in endometriotic stromal cells [25]. Finally, several studies showed that IGF2BP1 promoted cell proliferation by enhancing c-myc expression [26]. Thus, the effects of IGF2BP1 silencing and of miR-196b overexpression on myc mRNA and protein levels were assessed in HepG2 cells. IGF2BP1 silencing decreased IGF2BP1 mRNA level by 70 and 66%, 24 and 48 h post-transfection respectively (Figure 7A). The silencing induced an inhibition by 48% of myc mRNA expression that was only detected 48 h after transfection (Figure 7B). Similarly, the abundance of myc protein was decreased by IGF2BP1 silencing compared to RF negative control 48 and 72 h after transfection (Figure 7B). However, miR-196b overexpression did not decrease the myc mRNA and protein levels (Figure 7C).

Taken together, results showed that miR-196b mediates its anti-proliferative and pro-apoptotic effects through a c-myc-independent pathway, while IGF2BP1 silencing might exert its anti-proliferative and pro-apoptotic effects at least in part through a c-myc-dependent pathway.

Since we demonstrated that IGF2BP1 is a target of miR-196b and because miR-196b was shown to modulate hypoxia-induced chemoresistance, we sought whether the chemoresistance induced by hypoxia is due to IGF2BP1 expression inhibition. For that, HepG2 cells were transfected with IGF2BP1 siRNA for 48 h and then incubated under normoxia or hypoxia with or without etoposide, and apoptotic markers were studied by western blot (Figure 8A). IGF2BP1 silencing decreased IGF2BP1 protein abundance in cells incubated under normoxia or hypoxia with or without etoposide 48 h post-transfection. However, IGF2BP1 silencing did not modify the chemoresistance induced by hypoxia. Indeed, no increase in PARP and caspase 3 cleavage was observed. These data were confirmed using a caspase 3/7 activity assay that showed no modification of caspase 3/7 activity when cells were transfected with IGF2BP1 siRNA (Figure 8B).

These results indicated that IGF2BP1 silencing alone is not sufficient to reverse the hypoxia-induced chemoresistance probably because other miR-196b targets could be involved.

Human hepatoma HepG2 cells were maintained in culture in 75-cm² polystyrene flasks (Costar, Lowell, MA, USA) with D-MEM low glucose (Invitrogen, Carlsbad, CA, USA) containing 10% of fetal calf serum (FCS) (Invitrogen) and incubated under an atmosphere containing 5% CO₂. For hypoxia experiments (1% O₂), cells were incubated in serum-free CO₂-independent medium (Invitrogen, cat n°18045) supplemented with 0.5 mM L-glutamine (Sigma, Saint Louis, USA) with or without 50 μM etoposide (Sigma). Normoxic control cells were incubated in the same conditions but kept in normal atmosphere (21% O₂).

At the end of 16 h incubation under normoxia or hypoxia with or without etoposide, total RNA was extracted from cells using the Total RNAgent extraction kit (Promega, Madison, WI, USA). mRNA contained in 2 μg total and miRNA was reverse transcribed using the “High Capacity cDNA Archive” kit (Applied Biosystems, Carlsbad, CA, USA) according to the manufacturer’s instructions. An amount corresponding to 100 ng of retrotranscribed total RNA in 50 μl was then mixed with 50 μl of the “TaqMan Universal PCR master Mix” (Applied Biosystems) and loaded into one of the 8 fill ports the microfluidic array. “TaqMan Array Human MicroRNA panel v1.0” (Applied Biosystems) is microfluidic card that contains assays for the detection of 365 miRNAs in addition to 2 endogenous controls (RNU44 and RNU48). miRNA expression level was quantified using the threshold cycle method [45].

miRNA were extracted from samples containing total RNA with TRI Reagent Solution (Invitrogen) and the miRNeasy mini kit (Qiagen, Hilden, Germany) following manufacturer’s instructions. Total RNA without miRNA were extracted with RNeasy mini kit (Qiagen) following manufacturer’s instructions. RNA concentration was determined by using Nanodrop.

miRNA expression was evaluated using TaqMan MicroRNA Assay (Applied Biosystems) as specified in manufacturer’s instructions. Briefly, single‑stranded cDNA was synthesized by the TaqMan MicroRNA Reverse Transcription Kit from 200 ng of total RNA in a 15 μl reaction. The reactions were incubated sequentially at 16°C for 30 min, at 42°C for 30 min and then at 85°C for 5 min. cDNAs were then amplified by quantitative PCR in a 7900HT Sequence Detection System using the TaqMan microRNA Assays Human Panel sequence‑specific primers. The 20 μl PCR included 10 μl of 2x Universal PCR Master Mix (No AmpErase, UNG), 1 μl of each 20x TaqMan MicroRNA Assay Mix and 1.5 μl of reverse transcription product. The reaction mixtures were incubated in a 96‑well plates at 95°C for 10 min, followed by 40 cycles of 15 sec at 95°C, and 1 min at 60°C. The expression level of each miRNA was measured using the threshold cycle method using the expression of RNU44 (endogenous control) as reference and the expression of miRNAs under normoxia as calibrator. All reagents were supplied by Applied Biosystems.

cDNA was synthesized with 2 μg of total RNA using SuperScript II Reverse Transcriptase (Roche, Basel, Switzerland) and random hexamers primers for IGF2BP1 gene or oligodT for Myc gene according to the manufacturer’s instructions. Amplification reaction assays contained 1x SYBR Green PCR Mastermix (Applied Biosystem) and primers (IDT, Coralville, IA, USA) at the optimal concentrations. Primers for detection of mRNA were as follows IGF2BP1: 5′-CAGGAGATGGTGCAGGTGTTTATCC-3′ and 5′-GTTTGCCATAGATTCTTCCCTGAGC-3′, c-myc: 5′-CAGCTGCTTAGACGCTGGATT-3′ and 5′-GTAGAAATACGGCTGCACCGA-3′ [26]. The RNA expression level was measured using the threshold cycle method [45] using 23 kDa as the normalization reference.

Chemically synthesized miRNA duplexes, called pre-miR-210 and pre-miR-196b and negative control miRNA (pre-miR CTL-), were purchased from Ambion (Life Technologies, Carlsbad, CA, USA). HepG2 transfection was performed in 6-well plates for caspase 3/7 activity assay (50 000 cells per well) or in 25-cm² polystyrene flasks for protein extraction (750 000 cells per flask) with Dharmafect 1 transfection reagent used at a 1:500 dilution (Dharmacon, Lafayette, CO, USA) and with pre-miR at a final concentration of 50 nM. Cells were incubated for 24 h with transfection medium. The medium was then replaced by fresh medium with 10% of FCS at least 6 h before being incubated under hypoxia. Hypoxia and/or etoposide exposure during 16 h was performed 24 or 48 h after transfection.

IGF2BP1 silencing was achieved using the following siRNA sequence: 5′-CCGGGAGCAGACCAGGCAA(dTdT)-3′. RISC-free (RF) siRNA purchased from Dharmacon was used as negative control. For siRNA experiments, 750 000 cells were seeded in 25 cm² polystyrene flasks (Costar). The next day, cells were transfected with 50 nM siRNA using Dharmafect 1 (Dharmacon) transfection reagent according to manufacturer’s instructions. The transfection medium was removed 24 h after transfection and replaced by fresh culture medium. Cells were then incubated for 24, 48, 72 and 96 h and then incubated under normoxia or hypoxia with or without etoposide for 16 h.

HepG2 cells, seeded in 25 cm² flasks, were scrapped in 200 μl of lysis buffer (40 mM Tris pH 7.5, 150 mM KCl, 1 mM EDTA, 1% triton X-100) containing a protease inhibitor mixture (« Complete » from Roche Molecular Biochemicals, 1 tablet in 2 ml H₂O, added at a 1: 10 dilution) and phosphatase inhibitors (25 mM NaVO₃, 250 mM PNPP, 250 mM α-glycerophosphate and 125 mM NaF, at a 1: 10 dilution). Proteins were separated by SDS-PAGE on 4–12% NuPAGE Bis-Tris gels with MES or MOPS buffer (Invitrogen) and then transferred onto a low fluorescence background polyvinylidene fluoride (PVDF) membrane (Immobilon-FL, Millipore, Billerica, MA, USA). For fluorescence detection, membranes were blocked 1 h at room temperature in Odyssey blocking buffer (BD Biosciences, Franklin Lakes, MJ, USA) diluted 2X in PBS and then overnight at 4°C in Odyssey blocking buffer with 0.1% Tween containing a specific antibody: mouse anti-PARP-1 monoclonal antibody (#556493 from BD Pharmingen, Franklin Lakes, MJ, USA), rabbit anti-caspase3 antibody (#9662 from Cell Signaling, Dancer, MA, USA), rabbit anti-IGF2BP1 antibody (#2852 from Cell Signaling), mouse anti-Myc (#sc-42 from SantaCruz Technologies, Dallas, TX, USA) and mouse anti-α-tubulin monoclonal antibody (#T5168 from Sigma-Aldrich). After 3 washes of 10 min in PBS-0.1% Tween, the incubation with the secondary antibody (goat anti-rabbit or anti-mouse IR Dye labelled antibody 1/7 500 dilution) (Licor, Lincoln, NE, USA,) was performed for 1 h in Odyssey blocking buffer with 0.1% Tween followed by 3 washes of 10 min in PBS-Tween and 2 washes in PBS. The membrane was then dried 30 min at 37°C and scanned using Odyssey infrared imaging system (Licor). Immunodetection of α-tubulin was used as the loading control.

The fluorogenic substrate Ac-DEVD-AFC (BD Pharmingen) was used to measure caspase-3/7 activity according to Lozano et al. [46]. Protein extraction and the assay for caspase-3/7 activity were performed according to Notte et al. [47] using 20 μg of protein extracts.

HepG2 cells were seeded at a density of 5 500 000 cells in 75-cm² flasks. The next day, cells were then transfected with pre-miR negative control or pre-miR-196b (50 nM) for 24 h. Cells were harvested in 0.25 M saccharose containing a protease inhibitor mixture (Roche Molecular Biochemicals) and phosphatase inhibitors (25 mM NaVO₃, 250 mM PNPP, 250 mM α-glycerophosphate and 125 mM NaF, at a 1: 10 dilution) and homogenized by 5 strokes of a B plunger in a Dounce. The homogenate was centrifuged for 10 min at 1 000 rpm (Labofuge 400R Heraeus Instruments). The pellet was resuspended in 0.25 M saccharose and named the N (Nuclear) fraction. The supernatant was recovered and centrifuged at 8 000 rpm (Beckman L7-35, rotor 50 Ti) for a duration dependent on the volume. The pellet was resuspended in 0.25 M saccharose and named the MLP fraction. The supernatant was centrifuged at 35 000 rpm (Beckman L7-35, rotor 50 Ti) for a duration dependent on the volume. The supernatant was recovered and named the S fraction, which contains the cytosolic proteins.

Labeling with the cyanine dyes (GE Healthcare, Uppsala, Sweden) was performed as described in details in the supplementary methods (see Additional file 2). An amount corresponding of 25 μg of each individual labeled samples (pre-miR CTL-, pre-miR-196b and pooled standard samples from four replicate experiments) were mixed and subjected to isoelectric focusing on a IPGphor isoelectric focusing unit (GE Healthcare) along a continuous 3–11 pH gradient using IPG strips (18 cm, nonlinear 3 – 11 pH, GE Healthcare) (details in supplementary methods, see Additional file 2). Gels were scanned with a Typhoon 9400 imager (GE Healthcare) to detect cyanin labeled proteins. Determination of protein spot abundance and statistical analyses were carried out automatically using the DeCyder 2D Differential Analysis software 6.0 (GE Healthcare). Protein spots with more than 1.2-fold change in volume after normalization (between pre-miR CTL- and pre-miR-196b in the 4 replicates) and statistically significant Student’s t-test (p < 0.05) were considered as differentially abundant.

Each spot of interest was excised automatically using the Ettan spot picking system (GE Healthcare). Detailed trypsin digestion and mass spectrometry procedures by using nanoLC-MS-MS maXis UHR-TOF (ultra-high resolution time-of-flight) coupled with a 2D-LC Dionex UltiMate 3 000 (Bruker, Bremen, Germany) are given in the section of the supplementary methods (see Additional file 2).

A biocomputational approach was used based on a stepwise guide to identify miR‑196b targets. We focused on mRNA predicted by either TargetScan 6.2 [48], Microcosm [49] or miRDB [50] softwares.

The pmiRGLO plasmid vector contains cDNA sequences encoding the Firefly luciferase (F-luc) reporter gene and the Renilla luciferase (R-luc) gene, which acts as an internal control reporter. The pmiRGLO IGF2BP1 3′UTR wild-type and mutated were kindly obtained from Yang lab [51]. The wild-type 2 701-bp fragment contains two of three putative miR-196b seed match sites (4 268 – 4 274 nt and 4 922–4 928 nt). The pmiRGLO IGF2BP1 3′UTR mutated contains mutations at 2 sites cloned into the pmiRGLO vector. One day before the transfection, HepG2 cells were seeded at 30 000 cells per well into 24-well plates. The next day, HepG2 cells were co-transfected with 0.1 μg of pmiRGLO IGF2BP1 3′UTR wild-type or mutated and either pre-miR negative control or pre-miR-196b (50 nM final concentration). Transfection was performed with Lipofectamine 2000 reagent (Invitrogen) according to the co-transfection manufacturer’s protocol. At 72 h post-transfection, firefly and renilla luciferase activities were measured using Dual-luciferase Reporter Assay System (Promega, Madison, WI, USA) according the manufacturer’s protocol. The F-luc activity was normalized to the R-luc activity.

The methylthiazol tetrazolium (MTT, Sigma-Aldrich) spectrophotometric dye assay was used to detect the number of viable cells. HepG2 cells were seeded at 20 000 cells per well into 24-well plates 24 h before transfection. Cells were transfected for 24 h with either pre-miR CTL- or pre-miR-196b (50 nM) either RISC Free negative control (RF) or IGF2BP1 siRNA (50 nM) with Dharmafect. 24, 48, 72 and 96 h post-transfection, cells were incubated at 37°C for 2 h with 500 μL MTT solution (2.5 mg/mL; Sigma-Aldrich). Culture medium was then removed and formazan crystals in the cells were solubilized using lysis buffer (30% SDS, N,N-dimethyl-formamide; pH 4.7) with plate shaking for 2 h. Absorbance was measured at 570 nm.

The BrdU Cell Proliferation ELISA kit (Roche) has been used to measure cell proliferation according to manufacturer’s protocol. HepG2 cells were seeded at 2 500 cells per well into 96-well plates. 48 h after transfection, the BrdU solution (10 μM) was added during 24 h at 37°C. Absorbance was measured at 370 nm (reference wavelength 492 nm).

SigmaStat software (Jandle Scientific, Germany) was used for the statistical analysis. Data are presented as means ± standard deviation (SD) and were evaluated by two-way ANOVA, using the Holm-Sidak method. For some analyses values did not follow a Gaussian distribution. In order to deal with this absence of normality, statistical analyses were performed on log-transformed data. In order to facilitate interpretation, untransformed data are shown.