Effect of miR-34a on the expression of clock and clock-controlled genes in DLD1 and Lovo human cancer cells with different backgrounds with respect to p53 functionality and 17β-estradiol-mediated regulation.

Colorectal cancer (CRC) is one of the leading causes of cancer death in Europe and the U.S.A. [ 1 , 2 ]. It is the third most frequently diagnosed oncological disease and has the second highest mortality rate [ 3 ]. Moreover, an increase in CRC incidence in patients below 50 years of age recently emerged [ 4 , 5 ]. Therefore, novel strategies for CRC treatment are being studied extensively. Among them, gene silencing and personalisation of therapy are promising approaches to replace or modify classical disease management. Interestingly, a higher incidence of colorectal cancer has been demonstrated in male than in female patients [ 6 , 7 ]. Sex-dependent features of CRC progression are attributed, in addition to factors such as inflammatory diseases occurrence, genetic predisposition and/or social habits, as well as differences in levels of sex hormones between males and females [ 8 – 10 ]. In particular, 17β-estradiol (E2) has been shown to exert beneficial effects with respect to CRC [ 11 – 13 ]. E2 signalling is mediated via two nuclear receptors, ESR1 (ERα) and ESR2 (ERβ), and one membrane-bound G protein-coupled receptor-1 (GPER1). The most of beneficial E2 effects in the gastrointestinal tract of CRC patients are executed by ESR2 receptors [ 14 ]. The first implications of E2’s capacity to influence CRC management were based on epidemiological studies including patients subjected to hormone replacement therapy where E2 administration showed a negative association with CRC incidence [ 15 , 16 ]. A protective effect of E2 was demonstrated in ovariectomised (OVX) mice bearing a mutation in the Apc gene that produced many more intestinal tumours compared to those that were not subjected to OVX. In this model, E2 administration reversed the effect of OVX and induced ESR2 expression [ 17 ]. Similarly, APC-deficient mice with deletion of ESR2 receptors produced larger adenomas compared to those with wild-type ESR2 [ 18 ]. The protective role of E2 mediated by ESR2 receptors was also demonstrated in azoxymethane-treated OVX mice [ 19 ]. Implantation of SW480 cells overexpressing ESR2 to immunodeficient mice issued in lower tumour growth compared to mice implanted with control SW480 cells [ 20 ]. In male and female mice, E2 significantly reduced the number of azoxymethane/dextran sulphate sodium induced tumours and attenuated inflammation [ 21 ]. The oncostatic potential of E2 has been shown in a wide range of cellular models of colorectal cancer (for a review see [ 22 ]). Protective E2 effects mediated by ESR2 receptors are usually executed by induction of DNA repair, initiation of cell cycle arrest, apoptosis and/or epigenetic control mediated via miRNA and DNA methylation [ 13 ]. As cell cycle regulation is a complex process, E2 interferes with other control systems in this respect. In particular, the circadian system seems to influence cancer cell progression and patient survival in a sex-dependent manner [ 12 , 23 , 24 ]. The generation of circadian oscillators is based on a feedback loop of mRNA transcription and translation of protein products of three homologous per ( per1 , per2 and per3 ) and two homologous cry ( cry1 and cry2 ) clock genes. Clock gene expression is induced by a heterodimer composed of the transcription factors BMAL1 and CLOCK via E-box. After accumulation of PER and CRY proteins in the cytoplasm, they create an inhibitory complex that, after translocation to the nucleus, interferes with the stimulatory action of the heterodimer CLOCK/BMAL1 and suppresses their corresponding mRNA transcription [ 25 ]. Clock genes can influence the transcriptome directly via E-box [ 26 ] or indirectly via clock-controlled genes, including rev-erbα [ 27 ], sirt1 [ 28 ] and cyclin D1 ( ccnd1 ) [ 29 – 31 ]. Numerous studies have reported changes in clock gene expression in CRC patients [ 32 – 35 ]. Moreover, it was shown that CRC-related dysregulation of the circadian molecular feedback loop exerts sex-dependent effects. Expression of the clock genes cry2 and per2 showed a more pronounced decrease in tumours compared to adjacent healthy tissue in males than in females [ 12 ]. Expression of cry1 differed between left- and right-sided CRC tumours, and these differences exerted sex-dependent patterns [ 36 ]. Higher cry1 expression was observed in colorectal liver metastases [ 37 ] and in CRC tumour tissue of males compared to that of females [ 33 , 34 ]. In female CRC patients, better five-year survival was associated with low cry1 and cry2 expression, and this association was not observed in males [ 12 , 36 ]. The tumour oncogenic role is usually attributed to the cry1 gene [ 34 , 38 , 39 ], whereas per2 is recognized as a tumour suppressor [ 40 – 43 ]. Regulatory interference between the circadian system and E2 signalling has been described previously. E2 effects on the circadian oscillator are mediated via the ERE region that was identified in per1 [ 44 ], per2 [ 44 – 46 ] and clock [ 47 ], and an ESR1 binding site was revealed near the bmal1 gene promoter [ 48 ]. E2 signalling also influences the 24-h pattern of clock gene expression; E2 administration caused a phase advance in per2 expression in the rat colon. In the same tissue, E2 induced an increase in bmal1 expression in the middle of the D phase of the LD cycle. This effect was confirmed under in vitro conditions, where E2 administration induced bmal1 mRNA expression in DLD1 cells [ 49 ]. In addition to E-box and E2-mediated control, clock gene expression is under the regulatory influence of microRNAs (miRNAs). miRNAs are small non-coding RNAs that usually inhibit gene expression after targeting mRNA [ 50 – 52 ]. miR-34a is an extensively studied miRNA that was identified as a tumour suppressor [ 53 , 54 ] and whose expression is induced by p53 [ 55 ]. The expression of per2 , a key component of the molecular feedback loop [ 25 ], is also modulated by p53 [ 56 ]. As the majority of CRC tumours express mutated forms of p53 [ 57 ], the aim of the present study was to reveal whether p53 status influences the effects of miR-34a on clock and clock-controlled gene expression involved in cell cycle control. To achieve the above-mentioned aim, we analysed for the first time the effects of miR-34a on the circadian oscillator using the cell lines DLD1 and LoVo CRC, which differ with respect to p53 functionality [ 58 ]. Our experimental approach was based on manipulation of intracellular miR-34a levels and consequent changes in clock and clock-controlled gene expression, cell proliferation and the rate of wound healing. The effect of E2 on DLD1 cells was investigated to reveal possible interference between miR-34a and E2-mediated regulation. miR-34a caused p53-independent inhibition of per2 expression and changes in the expression of other clock genes ( bmal1 , clock ). Ectopic miR-34a also inhibited the expression of the clock-controlled genes rev-erbα , sirt1 and cyclin D1 . Our results do not support a synergic effect of miR-34a and E2 in regulation of DLD1 cell progression.

To determine the success of miRNA mimic transfection, the expression of miR-34a in control and miR-34a-transfected cells was measured. The expression of miR-34a was significantly higher in samples transfected with the miR-34a mimic ( p < 0.001) compared to the control ( S1 Fig ). After validation of transfection, the expression of the clock genes per2 , clock and bmal1 were analysed. In response to miR-34a administration, the expression of the per2 gene was significantly decreased in DLD1 ( p < 0.05) ( Fig 1A ) and LoVo cell lines ( p < 0.001) ( Fig 1B ). Fluorescent staining of per2 protein in DLD1 cells ( S2A and S2B Fig ) supports these findings. The intensity of fluorescence was significantly decreased in pre-miR-34a-transfected cells ( S2C Fig ). Expression of clock genes in DLD1 (A, C, E) and LoVo (B, D, F) cell lines. mRNA levels of per2 (A, B), clock (C, D) and bmal1 (E, F) were analysed 48 h after miR-34a mimic transfection. Results are displayed as mean ± SEM (n = 6). Statistical differences were determined by unpaired Student‘s t-test (* p < 0.05, *** p < 0.001). NC–negative control. Unlike per2 , we detected a significant increase in clock gene expression after miR-34a transfection in DLD1 and LoVo cells ( p < 0.05) ( Fig 1C and 1D , respectively). The expression of bmal1 was not significantly altered in response to miR-34a treatment in DLD1 cells ( Fig 1E ); however, a significant decrease in bmal1 expression was revealed in LoVo cells after miR-34a administration compared to the control ( p < 0.05) ( Fig 1F ). The expression of sirt1 significantly decreased after miR-34a mimic transfection in DLD1 cells ( Fig 2A ), ( p < 0.01) and LoVo cells ( Fig 2B ) ( p < 0.05). Expression of the clock-controlled genes sirt1 (A, B) and rev-erbα (C, D) 48 h after transfection of DLD1 (A, C) and LoVo (B, D) cells with the miR-34a mimic or the corresponding control. Results are displayed as mean ± SEM (n = 6). Statistical differences were determined by unpaired Student‘s t-test (* p < 0.05, ** p < 0.01). NC–negative control. Expression of the clock-controlled gene rev-erbα was significantly decreased after miR-34a mimic transfection in DLD1 cells ( Fig 2C ) ( p < 0.05). In the LoVo cell line, we observed only a slight decreasing trend after miR-34a administration ( Fig 2D ). Transfection of the miR-34a mimic-induced decreasing trend in cyclin D1 expression in DLD1 ( p = 0.1559) ( Fig 3A ). A significant decrease in cyclin D1 expression was observed in the LoVo cell line in response to miR-34a treatment (*** p < 0.001) ( Fig 3B ). Expression of cyclin D1 48 h after transfection of DLD1 (A) and LoVo (B) cell lines with the miR-34a mimic or the corresponding control. Results are displayed as mean ± SEM (n = 6). Statistical differences were determined by unpaired Student‘s t-test (* p < 0.05). NC–negative control. Two-day incubation of DLD1 cells with E2 administered in concentrations from 0.1 nM to 100 nM caused a significant increase in cyclin D1 mRNA expression in DLD1 cells exposed to 100 nM E2 compared to control cells ( Fig 4 ). E2 was administered in concentrations from 0.1 nM to 100 nM (grey columns). Results are displayed as mean ± SEM (n = 3–4). Statistical differences were determined by one-way ANOVA followed by Tukey‘s post hoc test. Different letters above columns indicate significant differences between groups at the level of p < 0.05. NC–negative control (white column). The effect of ectopic miR-34a on migration of DLD1 cells was analysed by a wound closure assay. The migration of transfected cells was recorded 24 h after transfection. We observed a trend ( p = 0.115) of decreasing migration after miR-34a administration in DLD1 cells ( Fig 5 ). Results are displayed as mean ± SEM (n = 4). Statistical differences were determined by an unpaired Student‘s t-test. NC–negative control. To determine possible reciprocal regulatory interference between miR-34a and E2, we evaluated the expression of esr2 mRNA in control and miR-34a-treated DLD1 cells. According to our results, miR-34a mimic transfection did not affect esr2 expression ( Fig 6A ). Expression of esr2 mRNA 48 h after transfection of DLD1 cells with miR-34a mimic or the corresponding control (A). Effect of 17β-estradiol administration on miR-34a expression in the DLD1 cell line (B). Results are displayed as mean ± SEM (n = 3–6). NC–negative control, esr2 –estrogen receptor 2. We also tested the possible effect of E2 on the levels of miR-34a in DLD1 cells. E2 administered in four concentrations up to 100 nM did not significantly influence the levels of mature miR-34a after 48 h of incubation with E2 ( Fig 6B ). Subsequently, we determined the effect of E2 on DLD1 cell proliferation and migration. We found that E2 administration in concentrations of 60, 80 and 100 nM significantly decreased the proliferation activity of DLD1 cells compared to control cells ( Fig 7 ). E2 was administered in concentrations of 10, 20, 40, 60, 80 or 100 nM (grey columns). White column indicates control group without E2. Results are displayed as mean ± SEM (n = 6–12). Statistical differences were determined by one-way ANOVA followed by Tukey‘s post hoc test. Different letters above columns indicate significant differences between groups at the level of p < 0.05. To further elucidate possible regulatory interference between miR-34a and E2, the anti-proliferative effect of E2 was tested with or without the presence of ectopic miR-34a. E2 at a concentration of 100 nM was added to the cell culture 48 h after miR-34a transfection. The absorbance at 490 nm was measured when the cells were exposed to E2 for 24 h. Transfection of the mir-34a mimic significantly decreased the proliferation of DLD1 cells ( Fig 8 ). However, the simultaneous administration of miR-34a and E2 did not change the proliferation rate compared to cells treated with only miR-34a ( Fig 8 ). Results are displayed as mean ± SEM (n = 8). Statistical differences were determined by one-way ANOVA followed by Dunnett’s multiple comparison test (* p < 0.05). NC–negative control, E2–17β estradiol. The effect of E2 on DLD1 migration was measured after 24 h and 48 h by a wound closure test in control cells and cells cultured in the presence of 1 nM, 10 nM or 100 nM concentrations of E2. Administration of 100 nM E2 significantly decreased migration of DLD1 cells 48 h after wound scratch compared to cells treated with 10 nM E2 ( Fig 9 ). There was a pronounced trend ( p = 0.058) between control cells and cells treated with 100 nM E2 ( Fig 9C ). Upper panel shows wound closure in control (A) and 100 nM E2-treated DLD1 cells (B, magnification 40x). Lower panel shows the effect of E2 on migration of DLD1 cells after 24 h and 48 h of incubation. Results are displayed as mean ± SEM (n = 4). Statistical differences were determined by one-way ANOVA followed by Tukey‘s post hoc multiple comparison test (* p < 0.05). C–control without E2. Next, relative expression of nuclear E2 receptors in DLD1 cells was measured. As is shown in Fig 10 , expression of esr2 is more than 60 times higher than that of esr1 . A cycle threshold (Ct) of 35 was considered the limit of PCR sensitivity in esr1 measurement. When Ct < 35 or the esr1 concentration was undetectable, Ct 35 was used for calculation of esr1 expression. Results are displayed as mean ± SEM (n = 6). Statistical differences were determined by an unpaired Student‘s t-test (*** p < 0.001). esr1 –estrogen receptor type 1, esr2 –estrogen receptor type 2.

To conclude, miR-34a strongly influences the expression of clock ( per2 , bmal1 and clock ) and clock controlled ( rev-erbα , sirt1 and cyclin D1 ) genes. Ectopic mir-34a inhibits the expression of per2 without respect to p53 status, but its oncostatic effects are more likely caused by inhibition of sirt1 and cyclin D1 expression. E2 administration inhibits the growth of DLD1 cells; however, this effect seems to be independent of miR-34a-mediated action. With respect to the possible use of miR-34a in cancer treatment, per2 can be considered as an off-target gene. Its possibly ambiguous oncogenic character should be taken into consideration in future clinical studies focused on miR-34a.

Human colorectal carcinoma cell lines DLD1 and LoVo were obtained from the ATCC (USA). DLD1 cells were cultured in RPMI 1640 GlutaMax medium (Gibco, USA) supplemented with 10% FBS, 100 U/mL penicillin and 100 μg/mL streptomycin (Biosera, France). LoVo cells were cultured in F-12K medium (Bioconcept, Switzerland) supplemented with 10% FBS, 100 U/mL penicillin and 100 μg/mL streptomycin (Biosera, France). Cells were cultivated at 37°C in a HF90 humidified incubator (Heal Force, China) containing 5% CO 2 . Experiments were performed three times. Transfection of miRNA was performed using mirVana™ miRNA mimic hsa-miR-34a (Invitrogen, ThermoFisher Scientific, USA) or mirVana™ miRNA mimic Negative control (Invitrogen, ThermoFisher Scientific, USA). For gene expression analysis, cells were seeded on 12-well plates (DLD1 10 6 cells /well; LoVo 5 x 10 5 cells/well), and immediately after seeding, the transfection mixture consisting of Lipofectamine RNAiMAX transfection reagent (8 μl/ml; Invitrogen, ThermoFisher Scientific, USA), Opti-MEM TM (Invitrogen, ThermoFisher Scientific, USA) and micro-RNA mimic or their negative control at a concentration of 50 nM were added to the cells and incubated for 48 h. Total RNA isolation from DLD1 or LoVo cells was performed using RNAzol RT (Molecular Research Center, USA). In particular, we used the ‘Protocol for isolation of large RNA and small RNA fraction’ that allows the separation of long and short RNA molecules. The RNA concentration and purity were determined using a Simply Nano spectrometer (Ge Healthcare, USA). To synthesise cDNA from mRNA, we used 0.3 μg of the fraction containing long molecules and a mixture consisting of ImProm-II reverse transcriptase, ImProm reaction buffer, 10 mM dNTP, RNasin ribonuclease inhibitor and random hexamer primers (Promega, USA). Quantitative PCR was performed with a QuantiTec SYBR Green kit (Qiagen, Germany) and the CFX Connect Real-Time Detection System (Bio-Rad Laboratories, USA). Data were analysed using CFX Manager Software (Bio-Rad Laboratories, USA) and normalized to u6 , rplp13a and/or β-actin . The list of primers is provided in S1 Table . miRNAs from the DLD1 cell line were isolated using RNAzol RT (Molecular Research Center). The miRNA concentration and purity were determined using a Simply Nano spectrometer (Ge Healthcare, USA). Synthesis of cDNA was performed after polyadenylation (Poly(A) Tailing Kit, ThermoFisher Scientific, USA) of isolated miRNAs using a reaction mixture consisting of ImProm-II reverse transcriptase, ImProm reaction buffer, 10 mM dNTP, RNasin Ribonuclease Inhibitor (Promega, USA) and a primer with a universal tag [ 59 ]. Quantitative PCR was performed using QuantiTec SYBR Green (Qiagen, Germany) in the CFX Connect Real-Time Detection System (Bio-Rad Laboratories, USA). Data were analysed using CFX Manager Software (Bio-Rad Laboratories, USA) [ 60 ]. The sequences of primers for the measurement of miR-34a-5p are provided in S1 Table . DLD1 cells were transfected using mirVana™ hsa pre-miR-34a-5p (Invitrogen; ThermoFisher Scientific, USA) or mirVana™ pre-miRNA mimic Negative control (Invitrogen; ThermoFisher Scientific, USA). The transfection mixture consisted of Lipofectamine RNAiMAX Reagent (8 μl/ml; Invitrogen; ThermoFisher Scientific, USA), Opti-MEMTM (Invitrogen; ThermoFisher Scientific, USA) and pre-micro-RNA or pre-negative control (Invitrogen; ThermoFisher Scientific, USA, 100 nM). Cells were seeded into an eight-well chamber slide that was covered with 0.1% gelatine. After 24 h of incubation, fixation was performed with 4% paraformaldehyde (PFA). After 2 min of incubation, the 4% PFA was discarded and replaced with 300 μl 2% PFA and incubated for 20 min at laboratory temperature. After fixation, cell monolayers were washed with Dulbecco′s phosphate buffered saline (DPBS). Permeabilization of DLD1 cells was performed by adding 0.1% Triton, and the chamber slide was gently mixed for 2 min. This step was repeated four times. After permeabilization, the slide was washed three times with DPBS. Background staining was blocked by incubation of cells with 10% bovine serum albumin for 30 min at laboratory temperature. Next, monolayers were washed with DPBS three times for 5 min. The PER2 primary antibody (sc-377290, Santa Cruz Biotechnology, US) was added to each well in a final concentration of 100 nM, and the cells were incubated overnight at 4°C. Then, DLD1 cells were washed with DPBS for 5 min. The m-IgGk secondary antibody (sc-516176, Santa Cruz Biotechnology, US) was added to each well at a concentration of 100 nM, and the cells were incubated for 60 min at laboratory temperature in the dark. After incubation, the cells were washed with DPBS three times. Finally, a drop of mounting medium with DAPI (ibidi, Germany) was added to each well, and the cells were visualised using an inverse fluorescent microscope NIB-100F (NOVEL, China) and programme BEL Capture 3.2 (NOVEL, China). The intensity of fluorescence was measured using ImageJ 1.53a (NIH, USA). DLD1 cells were seeded in 96-well plates at a density of 2 x 10 4 cells/well. The next day, the cells were treated with E2 at concentrations of 10, 20, 40, 60, 80 or 100 nM for 24 h. After treatment, the cells were incubated with 10 μM of MTS reagent according to the manufacturer’s instructions (CellTiter 96 ® Aqueous MTS Reagent, Promega, USA). The absorbance was measured at a wavelength of 490 nm (Epoch2, BioTek instruments, USA). DLD1 cells were seeded in a 96-well plate at a density of 2 x 10 4 cells/well. Immediately after seeding, the cells were transfected with mirVana™ miRNA mimic has-miR-34a-5p (Invitrogen; ThermoFisher Scientific, USA) or mirVana™ miRNA mimic Negative control (Invitrogen; ThermoFisher Scientific, USA), according to the manufacturer’s instructions. The transfection mixture consisted of Lipofectamine RNAiMAX transfection reagent (8 μl/ml; Invitrogen; ThermoFisher Scientific, USA), Opti-MEM TM (Invitrogen; ThermoFisher Scientific, USA), and micro-RNA or negative control at a concentration of 50 nM was added to the cells and incubated for 48 h. After the transfection, E2 at a concentration of 100 nM was added to the medium for 24 h. Next, the cells were incubated with 10 μM of MTS reagent according to the manufacturer’s instructions (CellTiter 96 ® Aqueous MTS Reagent, Promega, USA). The absorbance was measured at a wavelength of 490 nm (Epoch2, BioTek instruments, USA). A wound healing assay was used to determine migration of DLD1 cells after miR-34a transfection or after E2 treatment. The cells were cultured in 24-well plates. After 24 h of incubation with miR-34a or after incubation with E2 (1 nM, 10 nM, 100 nM) for 24 h or 48 h, the monolayer of each well was wounded with a 200 μl pipette tip. The recovered area was observed with a NIB-100F inverted microscope (NOVEL, China) and recorded at time zero and after 24 h and 48 h of incubation. Photographs were taken from the same areas as those recorded at time zero and evaluated by ImageJ 1.53a (NIH, USA). All results are presented as the mean ± SEM in arbitrary units (a.u.). Statistical differences among groups of independent samples were determined by Student‘s unpaired t-test or by one-way ANOVA followed by Tukey‘s post hoc test or Dunnett’s multiple comparison test. A p -value < 0.05 was considered statistically significant. ANOVA and t-tests were calculated using Graph Pad Prism 6 (GraphPad Software, Inc., San Diego, California, USA). Image analysis was carried out using ImageJ software [ 61 ].

Results are displayed as mean ± SEM (n = 5–6). Statistical differences were determined by a Student’s unpaired t-test (*** p < 0.001). NC–negative control. (TIF) Click here for additional data file. Fluorescent staining of PER2 protein in control (A) and pre-miR-34a-transfected DLD1 cells (B). Fluorescence intensity (C) is expressed as mean ± SEM (n = 12). Images were taken with an inverted fluorescence microscope with a magnification of 100x. Statistical differences were determined by an unpaired Student‘s t-test (* p < 0.05). preNC–negative control. (TIF) Click here for additional data file. (DOC) Click here for additional data file.