microRNA-216b enhances cisplatin-induced apoptosis in osteosarcoma MG63 and SaOS-2 cells by binding to JMJD2C and regulating the HIF1α/HES1 signaling axis

Cancer tissues and paracancerous tissues were collected from a total of 60 OS patients (aged 10–58 years, with an average age of 21.63 ± 9.70 years old) at the Shanghai Tenth People’s Hospital from April 2010 to April 2013. All included patients had not undergone chemotherapy or radiotherapy prior to specimen collection. Patient follow-up was conducted over a period of 60 months. The time from specimen collection to tumor recurrence or death (overall survival) was recorded. The last follow-up date was recorded if no tumor recurrence or death occurred. All included patients received the same treatment regimen at the Shanghai Tenth People’s Hospital, with no metastasis occurring during the follow-up period.

Human osteosarcoma cell lines U2OS, HOS, SaOS-2, MG-63 and human embryonic immortalized osteoblasts hFOB1.19 cells (American Type Culture Collection, Manassas, VA, USA) were cultured in Dulbecco’s modified Eagle’s medium (DMEM; HyClone, Logan, UT USA) supplemented with 10% fetal bovine serum, 100 μg/mL streptomycin, and 100 U/mL penicillin at 37 °C with 5% CO₂ [28].

The MG63 and SaOS-2 cells were subsequently transfected with agomir negative control (NC), miR-216b agomir, agomir NC + over-expression (oe)-NC, miR-216b agomir + oe-NC, miR-216b agomir + oe-JMJD2C, miR-216b agomir + oe-HES1, oe-NC + lentivirus (LV)-short hairpin RNA (sh)NC, oe-JMJD2C + LV-shNC, oe-JMJD2C + LV-shHIF1α, LV-shJMJD2C + oe-NC, and LV-shJMJD2C + oe-HIF1α. The cDNA constructs (pcDNA3.1-NC, pcDNA3.1-JMJD2C, pcDNA3.1-HIF1α and pcDNA3.1-HES1) were all constructed by HanBio Biotechnology Co., Ltd. (Shanghai, China). Meanwhile, agomir NC (5′-UUCUCCGAACGUGUCACGUTT-3′), miR-216b agomir (5′-AAAUCUCUGCAGGCAAAUGUGA-3′), LV-shNC, LV-shHIF1α, LV-shJMJD2C were all designed and constructed by GenePharma (Shanghai, China). Transfection was performed using Lipofectamine 2000 reagents (Invitrogen, Carlsbad, CA, USA) in accordance with the manufactures’ instructions. JMJD2C (gene ID: 23081), HIF1 (gene ID: 3091) and HES1 (gene ID: 3280) were identified by the National Center for Biotechnology Information (NCBI) database.

Total RNA content was extracted from the fresh tissues using RNeasy Mini kits (Qiagen, Valencia, CA, USA). For mRNA detection, the extracted total RNA was reverse transcribed into complementary DNA (cDNA) with the help of reverse transcription kits (RR047A, Takara, Japan). For miRNA detection, cDNA was obtained using miRNA First Strand cDNA Synthesis (Tailing Reaction) kits (B532451–0020, Sangon Biotech Co., Ltd., Shanghai, China). The samples were then loaded using a SYBR® Premix Ex TaqTM II (Perfect Real Time) kit (DRR081, Takara, Japan). RT-qPCR was performed on an ABI 7500 instrument (Applied Biosystems, Foster City, CA, USA), and each sample was evaluated in triplicate. The universal negative primers for miRNA, while the internal reference U6 were provided by miRNA First Strand cDNA Synthesis (Tailing Reaction) kit. The remaining primers were provided by Sangon Biotech Co., Ltd. (Shanghai, China). The primer sequences are depicted in Table 1. The Ct value of each target gene was recorded and normalized to an internal reference, namely, β-actin or U6. Relative expression of all target genes was calculated by means of relative quantification (2^-ΔΔCt method) [29].

Total protein content was extracted from the tissues or cells using radioimmunoprecipitation assay (RIPA) lysis buffer containing phenylmethylsulphonyl fluoride (PMSF) on ice for 30 min, followed by centrifugation at 8000 g and 4 °C for 10 min. The supernatant was then collected for protein concentration determination using a bicinchoninic acid (BCA) protein assay kit (Thermo Fisher Scientific, Rockford, IL, USA, #23250). Protein (50 μg) was subsequently dissolved in 2 × sodium dodecyl sulfate (SDS) loading buffer, boiled at 100 °C for 5 min, and then separated using SDS-polyacrylamide gel electrophoresis (PAGE). The protein was subsequently transferred onto polyvinylidene fluoride (PVDF) membranes, which were blocked by 5% skim milk at room temperature for 1 h and incubated with primary rabbit antibodies against JMJD2C (ab85454, dilution ratio of 1:1000), HIF1α (ab51608, dilution ratio of 1:500), HES1 (ab71559, dilution ratio of 1:500), Ki67 (ab92742, dilution ratio of 1:5000), B-cell lymphoma 2 (Bcl-2) (ab32124, dilution ratio of 1:1000), and β-actin (ab227387, dilution ratio of 1:5000) overnight at 4 °C. After three Tris-buffered saline Tween-20 (TBST) washes (10 min per wash), the membranes were then incubated with horseradish peroxidase (HRP)-labeled secondary goat anti-rabbit immunoglobulin G (IgG) H&L (ab97051, dilution ratio of 1:2000) for 1 h. All the aforementioned antibodies were purchased from Abcam Inc. (Cambridge, UK). The immunocomplexes on the membrane were visualized using enhanced chemiluminescence (ECL) reagents (BB-3501, Amersham, UK), with the band intensities quantified with the Bio-Rad Image Analysis System (Bio-Rad, Inc., Hercules, CA, USA), and Quantity One v4.6.2 software. The ratio of the gray value of the target band to β-actin was regarded as a reflection of the relative protein expression.

Wild type (wt) (pGL3-wt-JMJD2C-3’untranslated region [UTR]) or mutant (mut) (pGL3-mut-JMJD2C-3’UTR) reporter plasmids (GenePharma, Shanghai, China) were co-transfected with agomir NC or miR-216b agomir into MG63 and SaOS-2 cells. Following a 48 h period of transfection, the cells were collected and lysed. The subsequent procedures were performed with using a luciferase assay kit (K801–200, Biovision, Milpitas, CA, USA) following the manufacturer’s instructions on the luciferase reporter assay system (Promega, Madison, WI, USA). The relative luciferase activity was calculated based on the ratio of the luciferase activity of firefly luciferase to that of renilla luciferase. The sequences of wt-JMJD2C-3’UTR and mut-JMJD2C-3’UTR were 5′-GCAUGUAUGCUAAUGAGAUUU-3′ and 5′-GCUGUAAACGACGUCUCUAAA-3′, respectively [30].

The cells were lysed for 30 min at 4 °C in RIPA buffer (Thermo Scientific, Waltham, MA, USA), and then centrifuged at 13,000 g for 30 min at 4 °C. The supernatant was subsequently collected and incubated with specific antibody at 4 °C overnight, followed by the addition of pierce protein A/G Magnetic Beads (#88803, Thermo Scientific) and 4 h of incubation by shaking at 4 °C. The beads were then centrifuged, washed three times, and mixed with the loading buffer, followed by protein evaluation using SDS-PAGE and immunoblotting [31].

ChIP was performed using EZ-Magna ChIP A/G Chromatin Immunoprecipitation kits (17–371, Millipore, Billerica, MA, USA). Briefly, the cells were sonicated and centrifuged at 12,000 g for 10 min at 4 °C to remove insoluble precipitates. The cells were subsequently incubated with protein G agarose beads at 4 °C for 1 h, and centrifuged at 5000 g for a total of 1 min. Next, 10 μL (1%) of the supernatant was collected as the ‘Input’, while the remaining supernatant was divided into 2 portions and incubated with antibodies against H3K9me3 (ab8898, dilution ratio of 1:20, Abcam Inc., Cambridge, UK), H3K9me2 (ab1220, dilution ratio of 1:10, Abcam Inc., Cambridge, UK), H3 (ab213257, dilution ratio of 1:5, Abcam Inc., Cambridge, UK), and NC rabbit anti-human IgG (ab2410, dilution ratio of 1:25, Abcam Inc., Cambridge, UK) at 4 °C overnight. The precipitated protein-DNA complex was then incubated with protein G agarose beads for 1 h at 4 °C. After centrifugation at 5000 g for 1 min, the supernatant was discarded. The protein-DNA complex was fragmented overnight at 65 °C. The DNA fragment was recovered and used as an amplification template for RT-qPCR. The primers employed to detect the quantity of HES1 promoter were agomir NC (5′-UUCUCCGAACGUGUCACGUTT-3′) and miR-216b agomir (5′-AAAUCUCUGCAGGCAAAUGUGA-3′).

Cell proliferation was measured using a cell counting kit-8 (CCK-8) assay (Kumamoto, Japan). MG63 and SaOS-2 cells were firstly seeded in 96-well plates. After 24 h of transfection, various concentrations of cisplatin (0, 0.2, 0.4, 0.8, 1.6 μM) were added to the cells and incubated for 72 h. The CCK-8 reagent (10 μL) was then added to each well and incubated for 1 h. Absorbance was finally measured at a wavelength of 450 nm using a microplate reader.

After 24 h of transfection, the cells were treated with 1 μM cisplatin [32] for 72 h. An apoptosis assay was subsequently performed using Annexin V-fluorescein isothiocyanate/propidium iodide (FITC/PI) kits (KeyGEN Biotech. Co., Ltd., Nanjing, China). The cells were subsequently observed and analyzed using a flow cytometer (FACSCalibur, BD Biosciences, Franklin Lakes, NY, USA). The experiment was performed in triplicate, with the respective average values obtained and recorded.

A total of 30 BALB/c nude mice (aged 5–6 weeks old, weighing 15–18 g, acquired from the Shanghai Experimental Animal Center of Chinese Academy of Sciences) were randomly divided into the 5 following groups (6 mice per group): control group, Lv-oe-NC group, Lv-oe-NC + Cis group, Lv-oe-miR-216b + Lv-oe-NC + Cis group, and Lv-oe-miR-216b + Lv-oe-HES1 + Cis group. The lentiviruses carrying oe-NC, oe-miR-216b and oe-HES1 were purchased from Shanghai Gene Pharma Co., Ltd. (Shanghai, China). Following lentivirus infection, the stably-transfected MG63 cells were subcutaneously injected into the axillary region of nude mice at a density of 1 × 10⁶ cells. From the 7th day of inoculation, cisplatin (8 mg/kg, [33]) was intraperitoneally administered to the mice in the Lv-oe-NC + Cis, Lv-oe-miR-216b + Lv-oe-NC + Cis, and Lv-oe-miR-216b + Lv-oe-HES1 + Cis groups twice a week. On the 14th day, the mice were euthanized using carbon dioxide. The tumors were subsequently removed, after which the tumor volume was measured using the following formula: (a × b²)/2 [34], where a represented the length of the tumor and b was the width of the tumor. The volume of the tumor was calculated and a tumor growth curve was then plotted [35].

Data analyses were processed using the SPSS 21.0 statistical software (IBM Corp. Armonk, NY, USA). Measurement data were expressed as mean ± standard deviation. Data with normal distribution and homogeneity of variance between two groups were compared using paired t-test (paired data) or unpaired t-test (unpaired data). Data comparisons between multiple groups were performed using one-way analysis of variance (ANOVA), followed by a Tukey’s test. Data comparison between groups at different time points was performed by repeated measures ANOVA with Bonferroni post-hoc test. Correlation of two variants was analyzed by Pearson correlation coefficient. Patient survival rate was analyzed using the Kaplan-Meier method followed by log-rank test. A p < 0.05 value was considered to be indicative of statistical significance.

Firstly, the expression of miR-216b was found to be markedly lower in OS tissues compared to that in paracancerous tissues (fold = 0.315) (Fig. 1a). The median miR-216b relative expression pattern in cancer tissues from 60 patients was subsequently determined, after which the patients with values above the median were placed into the high miR-216b expression group, while those below the median were regarded as the low miR-216b expression group. In addition, Kaplan-Meier survival analysis revealed that elevated expression of miR-216b was correlated with higher overall survival in OS patients (Fig. 1b). However, no significant correlation was identified between the miR-216b expression and factors such as age, sex, location of the tumor, and metastasis (Table 2). Additionally, results of multivariate analysis demonstrated that miR-216b served as an independent prognostic factor for OS (Table 3). Lastly, the expression of miR-216b was determined to be significantly lower in the MG63 and SaOS-2 cell lines relative to other OS cell lines and control (hFOB1.19 cell line) (U2OS: fold = 0.436, HOS: fold = 0.614, SaOS-2: fold = 0.248, MG63: fold = 0.161) (Fig. 1c). Hence, MG63 and SaOS-2 cells were selected for subsequent experimentation.

We subsequently set out to evaluate the sensitivity of MG63 and SaOS-2 cells to cisplatin as OS resistance against chemotherapy remains a significant obstacle during the treatment of OS. Transfection of miR-216b agomir, as expected, elevated the expression of miR-216b in MG63 and SaOS-2 cells, indicating successful transfection (MG63: fold = 2.218, SaOS-2: fold = 3.094) (Supplementary Fig. 1). Meanwhile, it was found that cisplatin decreased the cell viability in MG-63 and SaOS-2 cells but much more so in cells transfected with miR-216b agomir (MG63: 0.25 fold = 0.73, 0.5 fold = 0.64, 1 fold = 0.39; SaOS-2: 0.25 fold = 0.77, 0.5 fold = 0.72, 1 fold = 0.46) (Fig. 2a). Cisplatin was found to also augment cell apoptosis, which was further elevated following miR-216b agomir treatment in cells (MG63: miR-216b agomir fold = 1.39; SaOS-2: miR-216b agomir fold = 1.26) (Fig. 2b, c). These results suggested that miR-216b enhanced the promotive effects of cisplatin on OS cell apoptosis, while inhibiting cell viability in vitro.

The online bioinformatic website Starbase [36] predicted the presence of binding sites of miR-216b to the 3’UTR of the histone demethylase JMJD2C mRNA (KDM4C). Subsequently, a dual luciferase reporter assay was performed, which revealed that co-transfection of miR-216b agomir with wt-JMJD2C-3’UTR triggered a reduction in the luminescence intensity (MG63 fold = 0.407; SaOS-2 fold = 0.442), whereas no alterations were detected upon co-transfection with mut-JMJD2C-3’UTR in cells (MG63 fold = 0.974; SaOS-2 fold = 0.981) (Fig. 3a). Furthermore, Pearson correlation analysis highlighted a negative correlation between the expression of miR-216b and JMJD2C in OS tissues (Fig. 3b). In addition, miR-216b agomir was identified to increase the expression of miR-216b, while reducing the mRNA and protein expression of JMJD2C in MG-63 and SaOS-2 cells (MG63-RT-qPCR miR-216b agomir + oe-NC fold = 2.263; miR-216b agomir + oe-JMJD2C fold = 2.155) (SaOS-2-RT-qPCR miR-216b agomir + oe-NC fold = 3.043, miR-216b agomir + oe-JMJD2C fold = 2.955) (MG63-western blot analysis miR-216b agomir + oe-NC fold = 0.359; miR-216b agomir + oe-JMJD2C fold = 3.316) (SaOS-2-western blot analysis miR-216b agomir + oe-NC fold = 0.426; miR-216b agomir + oe-JMJD2C fold = 2.862) (Fig. 3c). It was further found that miR-216b diminished the cisplatin-induced decrease in cell viability (MG63 0.25 fold = 0.47, 0.5 fold = 0.45, 1 fold = 0.41) (SaOS-2 0.25 fold = 0.45, 0.5 fold = 0.41, 1 fold = 0.37), the effect of which was weakened following JMJD2C over-expression (MG63 0.25 fold = 2.55, 0.5 fold = 2.80, 1 fold = 3.68) (SaOS-2 0.25 fold = 2.71, 0.5 fold = 3.24, 1 fold = 3.81) (Fig. 3d). Similarly, miR-216b was observed to enhance cisplatin-induced apoptosis (MG63 fold = 2.32; SaOS-2 fold = 2.73), the effect of which was weakened by JMJD2C over-expression (MG63 fold = 0.26; SaOS-2 fold = 0.24) (Fig. 3e). Based on these results, it could be inferred that miR-216b could bind to JMJD2C and inhibit its expressions, ultimately enhancing cisplatin-induced apoptosis in OS cells.

In order to characterize the effect of JMJD2C/HIF1α on HES1, we silenced HIF1α in cells, and subsequently identified that si-HIF1α-3 exhibited the best HIF1α silencing efficiency (MG63 si-HIF1α-1 fold = 0.504, si-HIF1α-2 fold = 0.401, si-HIF1α-3 fold = 0.328) (SaOS-2 si-HIF1α-1 fold = 0.512, si-HIF1α-2 fold = 0.392, si-HIF1α-3 fold = 0.250), and was thus chosen for the following experiments (Supplementary Fig. 2). HIF1α silencing was found to down-regulate the expression of HES1 in MG-63 and SaOS-2 cells (MG63 fold = 0.403, SaOS-2 fold = 0.445) (Fig. 4a). HIF1α silencing also reversed the elevation in HES1 expression caused by JMJD2C over-expression, demonstrating that HIF1α functioned as a crucial factor in the up-regulation of HES1 by JMJD2C and HIF1α (MG63 fold = 2.287, SaOS-2 fold = 2.260) (Fig. 4b). ChIP-qPCR assay was then employed to detect the quantity of the HES1 promoter in the immune complexes in MG-63 and SaOS-2 cells, the results of which revealed reduced amounts of HES1 promoter in the immune complexes upon JMJD2C over-expression (MG63 fold = 0.13, SaOS-2 fold = 0.15), while no significant differences were found regarding the quantity of the HES1 promoter (MG63 fold = 1.08, SaOS-2 fold = 0.93) following combined treatment with oe-JMJD2C and shHIF1α(Fig. 4c). Meanwhile, HIF1α over-expression failed to reverse the down-regulation of HES1 brought about by JMJD2C silencing (MG63-LV-shJMJD2C + oe-NC JMJD2C fold = 0.416, HIF1α fold = 1.144, HES1 fold = 0.325) (MG63-LV-shJMJD2C + oe-HIF1α JMJD2C fold = 0.742, HIF1α fold = 1.908, HES1 fold = 0.820) (SaOS-2-LV-shJMJD2C + oe-NC JMJD2C fold = 0.496, HIF1α fold = 1.135, HES1 fold = 0.396) (SaOS-2-LV-shJMJD2C + oe-HIF1α JMJD2C fold = 0.833, HIF1α fold = 1.832, HES1 fold = 0.831) (Fig. 4d). These results demonstrated that JMJD2C and HIF1α were both essential in the up-regulation of HES1. ChIP-qPCR assay was further applied to detect the quantity of the HES1 promoter enriched by IgG and H3K9me3 antibodies in MG-63 and SaOS-2 cells, and the results demonstrated that HIF1α over-expression (MG63 fold = 1.01, SaOS-2 fold = 0.94) exhibited no effect on the elevated H3K9me3 levels caused by JMJD2C silencing (MG63 fold = 2.61, SaOS-2 fold = 2.37) (Fig. 4e). In addition, the quantity of HES1 promoter in the immune complexes with anti-H3K9me2 and anti-H3 antibodies was determined in MG-63 and SaOS-2 cells following different treatments (Fig. 4f-g). The quantity of HES1 promoter in the immune complexes with anti-H3K9me2 was found to be diminished (MG63 fold = 0.52, SaOS-2 fold = 0.45) in response to JMJD2C silencing, whereas combined treatment with JMJD2C silencing and HIF1α over-expression did not negate the effect of JMJD2C silencing (MG63 fold = 0.92, SaOS-2 fold = 0.89) (Fig. 4f). Besides, JMJD2C silencing and HIF1α over-expression alone or in combination did not affect the quantity of HES1 promoter in the immune complexes with anti-H3 (MG63 fold = 1.02, SaOS-2 fold = 1.04) (Fig. 4g). These findings highlighted the ability of JMJD2C to up-regulate HES1 expression via HIF1α interaction to diminish H3K9 methylation modification in vitro.

The transfection of miR-216b agomir and oe-HES1 in MG-63 and SaOS-2 cells was efficient for further experimentation (MG63-RT-qPCR miR-216b agomir + oe-NC fold = 2.315, miR-216b agomir + oe-HES1 fold = 2.721) (SaOS-2-RT-qPCR miR-216b agomir + oe-NC fold = 2.845, miR-216b agomir + oe-HES1 fold = 3.126) (MG63-western blot analysis miR-216b agomir + oe-NC fold = 0.222, miR-216b agomir + oe-HES1 fold = 5.532) (SaOS-2-western blot analysis miR-216b agomir + oe-NC fold = 0.259, miR-216b agomir + oe-HES1 fold = 4.310) (Fig. 5a). Over-expression of miR-216b was found to diminish the cisplatin-induced reduction in the viability in MG-63 and SaOS-2 cells (MG63 0.25 fold = 0.38, 0.5 fold = 0.34, 1 fold = 0.24), (SaOS-2 0.25 fold = 0.51, 0.5 fold = 0.42, 1 fold = 0.33), the effect of which was abrogated by HES1 over-expression (MG63 0.25 fold = 2.36, 0.5 fold = 3.04, 1 fold = 4.02) (SaOS-2 0.25 fold = 3.32, 0.5 fold = 3.71, 1 fold = 5.20) (Fig. 5b). In addition, miR-216b over-expression further enhanced the cisplatin-induced apoptosis in MG-63 and SaOS-2 cells (MG63 fold = 2.76, SaOS-2 fold = 3.12), whereas HES1 over-expression brought about the opposite effects (MG63 fold = 0.19, SaOS-2 fold = 0.25) (Fig. 5c). These results in concert with those discussed above, demonstrated the effect of miR-216b/JMJD2C/HIF1α/HES1 signaling axis on enhancing the effects of chemotherapy in OS cells.

Representative tumor xenografts isolated from nude mice are depicted in Fig. 6a. Tumor growth (day 7 fold = 0.53, day 14 fold = 0.71) (Fig. 6b) and weight (fold = 0.46) (Fig. 6c) were found to be reduced by cisplatin, while this effect was further increased upon miR-216b over-expression (tumor growth: day 7 fold = 0.69, day 14 fold = 0.70) (tumor weight: fold = 0.51). Meanwhile, the addition of a HES1 over-expression plasmid weakened the effect of miR-216 over-expression on tumor growth (day 7 fold = 3.30, day 14 fold = 2.24) and weight (fold = 6.70). Moreover, in tumor tissues, miR-216b expression was observed to be increased, and further increased following miR-216b over-expression. However, no changes were observed in miR-216b expression following the simultaneous over-expression of miR-216b and HES1 in tumor tissues (Fig. 6d). Furthermore, cisplatin was found to diminish the protein expressions of Ki67, Bcl-2, JMJD2C, HIF1α and HES1 (Ki67 fold = 0.399, Bcl-2 fold = 0.593, JMJD2C fold = 0.649, HIF1α fold = 0.529, HES1 fold = 0.547), all of which were further down-regulated by miR-216b over-expression (Ki67 fold = 0.581, Bcl-2 fold = 0.437, JMJD2C fold = 0.535, HIF1α fold = 0.640, HES1 fold = 0.548) (Fig. 6e). HES1 over-expression was also identified to weaken the effects associated with miR-216b over-expression (Ki67 fold = 3.172, Bcl-2 fold = 3.051, JMJD2C fold = 2.274, HIF1α fold = 2.373, HES1 fold = 2.662). These in vivo results provided evidence confirming the promotive role of miR-216b on the effects of cisplatin via the JMJD2C/HIF1α/HES1 signaling axis.