Circular RNA HIPK3 downregulation mediates hydrogen peroxide-induced cytotoxicity in human osteoblasts

Hydrogen peroxide (H2O2) induces oxidative injury to human osteoblasts. The expression and potential function of circular RNA HIPK3 (circHIPK3) in H2O2-treated human osteoblasts were tested. We show that H2O2 significantly downregulated circHIPK3 in OB-6 cells and primary human osteoblasts. Furthermore, circHIPK3 levels were decreased in the necrotic femoral head tissues of dexamethasone-treated patients. In OB-6 osteoblastic cells and primary human osteoblasts, forced overexpression of circHIPK3 by a lentiviral construct alleviated H2O2-induced viability reduction, cell death and apoptosis. Contrarily, circHIPK3 silencing by targeted shRNA potentiated H2O2-induced cytotoxicity in OB-6 cells and primary human osteoblasts. Moreover, circHIPK3 downregulation by H2O2 induced miR-124 accumulation in OB-6 cells and primary human osteoblasts. On the contrary, miR-124 inhibition by transfection of the miR-124 inhibitor protected human osteoblasts from H2O2. Importantly, forced overexpression of miR-124 by transfection of the miR-124 mimic induced significant cytotoxicity in OB-6 cells and primary human osteoblasts. H2O2 downregulated miR-124’s targets, cyclin dependent kinase 6 and Rho-Associated Protein Kinase 1, in human osteoblasts. In conclusion circHIPK3 downregulation mediates H2O2-induced cytotoxicity in human osteoblasts.


INTRODUCTION
In the pathogenesis of osteoporosis and osteonecrosis, increased reactive oxygen species (ROS) production and oxidative injury will lead to severe damage to human osteoblasts and bone cells [1][2][3][4]. To the cultured human osteoblasts or osteoblastic cells hydrogen peroxide (H2O2) was added, as an in vitro cellular model of osteoporosis/osteonecrosis [5][6][7][8]. H2O2 induces profound oxidative stress, protein damage, lipid peroxidation and DNA breaks in human osteoblasts, leading to cell death and apoptosis. Further understanding the pathological mechanisms of H2O2-induced osteoblast injury is important for the development of possible intervention strategies [5][6][7][8].
Circular RNAs (circRNAs) are a large family of conserved and stable non-coding RNAs (ncRNAs) exclusively in the cytoplasm of eukaryotic cells [9,10]. Compared with linear RNAs, circRNAs have covalentlyclosed loop structures, but without a free 3′ or 5′ end nor poly-adenylated tails [9,10]. circRNAs function as microRNA (miRNA) sponges to sequester and competitively inhibit miRNA expression and activity [9,10]. The potential functions of circRNAs in the pathogenesis of osteoporosis and osteonecrosis have not been extensively studied.
Derived from homeodomain-interacting protein kinase 3 (HIPK3) gene Exon2, the circular RNA HIPK3 (circHIPK3) has the sequence length of 1099 base-pair [11]. circHIPK3 could possibly exert pro-survival functions in a number of cancer cells, partially mediated through as sponges of cancer-suppressive miRNAs [11][12][13]. A very recent study has shown that circHIPK3 levels are downregulated in high glucose (HG)-treated human umbilical vein endothelial cells (HUVECs) and in primary human aortic endothelial cells (HAECs) from the diabetic patients [14]. More importantly, circHIPK3 downregulation mediated in vitro endothelial cell injury by HG [14]. The results of the current study will show that H2O2 downregulates circHIPK3 to promote human osteoblast cell death and apoptosis.

Forced overexpression of circHIPK3 alleviates H2O2-induced death and apoptosis in human osteoblasts
The results in Figure 1 indicate a potential activity of circHIPK3 in H2O2-induced cytotoxicity. To test this hypothesis, circHIPK3-expressing lentivirus ("LV-circHIPK3", from Dr. Lu at Nanjing University of Traditional Chinese Medicine [14]) was transduced to OB-6 osteoblastic cells. Following selection by puromycin two stable cell lines with LV-circHIPK3 were established: "OE-circHIPK3-L1 and OE-circHIPK3-L2". Analyzing circHIPK3 expression, by qPCR, confirmed that circHIPK3 levels increased over ten folds in the LV-circHIPK3-expressing OB-6 cells (Figure 2A), even with H2O2 treatment (Figure 2A).
It has been previously shown that H2O2 could induce both programmed necrosis and apoptosis in human osteoblasts and osteoblastic cells [18,19]. Significantly, H2O2induced cell viability (MTT OD) reduction ( Figure 2B) and death (increased medium LDH release, Figure 2C) were significantly inhibited in circHIPK3-overexpressed stable OB-6 cells. Furthermore, H2O2-induced apoptosis activation in OB-6 cells was attenuated by circHIPK3 overexpression as well ( Figure 2D and 2E). Apoptosis activation in H2O2-treated OB-6 cells was evidenced by cleavages of caspase-3, caspase-9 and ploy ADP ribose treated with hydrogen peroxide (H2O2, at applied concentrations) and cultured for indicated time periods, relative circHIPK3 expression was tested by qPCR (A-C) qPCR analysis of the relative circHIPK3 expression in the surgery-isolated femoral head tissues (both normal and necrotic) from ten (10) different dexamethasone-treated patients (D) "Veh" stands for vehicle control (PBS, same for all Figures). Quantified values were mean ± standard deviation (SD). * P < 0.05 vs. "Veh" treatment (A-C) * P < 0.05 vs. "S" tissues (surrounding normal femoral head tissues) (D; n=10). Experiments were repeated three times, with similar results obtained. polymerase (PARP) ( Figure 2D) as well as the increased nuclear TUNEL staining ratio ( Figure 2E). Furthermore, ectopic overexpression of circHIPK3 largely inhibited H2O2-induced increase in Annexin V staining ( Figure 2F), further supporting the anti-apoptosis activity by circHIPK3. Additionally, H2O2 treatment in vector control OB-6 cells induced mitochondrial depolarization, tested by JC-1 green intensity increase ( Figure 2G). The actions by H2O2 were again inhibited in circHIPK3overexpressed OB-6 cells ( Figure 2G).
In the primary human osteoblasts, LV-circHIPK3 similarly resulted in an increase of circHIPK3 expression, regardless of H2O2 stimulation ( Figure 2H). H2O2-induced cell death ( Figure 2I, tested by LDH medium release) was significantly alleviated by LV-circHIPK3 in the primary osteoblasts. Furthermore, circHIPK3 overexpression potently inhibited H2O2-induced apoptosis activation, decreasing cell numbers with positive TUNEL ( Figure 2J) and Annexin V ( Figure 2K) staining. Together, these results showed that forced overexpression of circHIPK3 alleviated H2O2-induced death and apoptosis in human osteoblasts.
Based on the results we proposed that H2O2-induced downregulation of circHIPK3 caused miR-124 accumulation, mediating osteoblast cell death and apoptosis. Thus, forced expression of miR-124 should induce the similar action of H2O2. To test this hypothesis, the miR-124 mimic was transfected to OB-6 cells and primary human osteoblasts, resulting in significant increase in miR-124 expression ( Figure 4J). Significantly, the miR-124 mimic induced viability reduction ( Figure 4K), cell death ( Figure 4L) and apoptosis ( Figure  4M) in OB-6 cells and primary human osteoblasts.

DISCUSSION
CircRNAs are formed from exon transcripts through nonlinear reverse splicing or gene re-arrangements [9,10,24]. Dysregulation of circRNAs could be important for oxidative stress-induced osteoblast injury and pathogenesis of osteoporosis/osteonecrosis. The results of the present study show that circHIPK3 is downregulated in the necrotic femoral head tissues of dexamethasonetreated human patients, indicating a possible association between circHIPK3 reduction and pathophysiology of femoral head necrosis.
In vitro results of this study show that H2O2 downregulated circHIPK3 in OB-6 cells and primary human osteoblasts. Importantly, forced overexpression of circHIPK3, by a lentiviral construct, alleviated H2O2-induced viability reduction, cell death and apoptosis. Contrarily, circHIPK3 silencing by targeted shRNAs potentiated H2O2induced cytotoxicity in OB-6 cells and primary human osteoblasts. These results imply that circHIPK3 downregulation mediates H2O2-induced cytotoxicity in human osteoblasts.
In the current study, we show that circHIPK3 possibly acts as the sponge of miR-124 in human osteoblasts. miR-124 levels were significantly increased in circHIPK3-silenced OB-6 cells, but downregulated with circHIPK3 overexpression. Moreover, circHIPK3 downregulation by H2O2 induced miR-124 accumulation in OB-6 cells and primary human osteoblasts. On the contrary, miR-124 inhibition by a miR-124 inhibitor protected osteoblasts from H2O2. Forced expression of miR-124, by the miR-124 mimic, induced significant cytotoxicity in human osteoblasts. Importantly, H2O2 downregulated verified miR-124's targets, including CDK6 and ROCK1, in human osteoblasts. These results imply that miR-124 accumulation by circHIPK3 downregulation possibly mediated H2O2-induced cytotoxicity in human osteoblasts.
Together, we show that circHIPK3 downregulation mediates H2O2-induced cytotoxicity in human osteoblasts. Targeting circHIPK3-miR-124 cascade could be a novel Table 1. Primers of the qPCR assay in this study.
strategy to protect human osteoblasts from oxidative injury.

Cell culture
Established OB-6 human osteoblastic cells were provide Dr. Cui [29,30], cells were cultured as described previously [29,30]. The primary human osteoblasts were provided by Dr. Ji [17], cultured under a previouslydescribed condition [17,31]. Primary human osteoblasts at passage 3-10 were utilized for in vitro biomedical studies. All protocols were approved by Ethics Committee of authors institutions, and according to the Declaration of Helsinki.

Human tissues
The lysate samples of necrotic femoral head tissues and the surrounding normal femoral head tissues from ten (10) dexamethasone-taking patients with femoral head resection surgery were provided by Dr. Cui [32]. All clinical investigations were conducted according to the criteria set by the Declaration of Helsinki.

Quantitative real-time polymerase chain reaction assay (qPCR)
OB-6 cells or the primary human osteoblasts were seeded into six-well plates at 1.5×10 5 cells per well. Following the treatments, TRIzol reagent was added to extract total cellular RNA. qPCR was performed by a SYBR Green PCR kit (Applied Biosystems, Shanghai, China) under a 7500H FAST Real-Time PCR System (Takara, Osaka, Japan) [33]. Melting curve analysis was always performed to calculate product melting temperature. Using a ΔΔ Ct method, target gene expression was quantified. U6 RNA was tested to normalize expression levels of listed genes. All the primers for qPCR assay were purchased from Origene (Beijing, China) Table 1. qPCR primers of CDK6 and ROCK1 were provided by Dr. Wu from Medical School of Nanjing University [20].

Western blotting
At a density of 1.5 × 10 5 cells per well OB-6 cells or primary human osteoblasts were seeded into six-well plates. Following the treatments, the cell lysis buffer (Biyuntian, Wuxi, China) was added. The lysates (30-40 μg per lane) were separated by 10-12% SDS-PAGE gels, and transferred to polyvinylidene difluoride (PVDF) blots (Millipore, Bedford, MA). After blocking in PBST with 10% non-fat milk, the blots were probed with the designated primary and secondary antibodies. The enhanced chemiluminescence (ECL) reagents (Amersham Bioscience, Piscataway, NJ) were added to visualize the targeted protein signals. Image J software (National Institutes of Health) was utilized for the data quantification.

Ectopic circHIPK3 overexpression
The lentivirus with pGLV3-U6-GFP-Puro vector encoding circHIPK3 ("LV-circHIPK3") was provided by Dr. Lu [14], and added to OB-6 cells and primary human osteoblasts. Afterwards, cells were cultured in the fresh complete medium for another 48h. When necessary, puromycin (5 μg/mL) was added to select stable cells for another 10 days. CircHIPK3 overexpression was verified by qPCR.

circHIPK3 shRNA
Two circHIPK3 shRNAs, with non-overlapping and unique sequence ("S1/S2"), were designed by Shanghai Genechem Co. The shRNA was sub-cloned into a GV248 lentiviral construct to general lentivirus. OB-6 cells and primary human osteoblasts were seeded into the six-well plates at a 50% confluence, and the shRNA lentivirus was added. Afterwards, cells were cultured in the fresh complete medium for another 48h. When necessary puromycin (5 μg/mL) was added to the medium to select stable cells. CircHIPK3 knockdown was confirmed by qPCR.

Cell death assay
At a density of 1.5 × 10 5 cells per well OB-6 cells or primary human osteoblasts were seeded into six-well plates. Following the treatments, cell death was tested by examining lactate dehydrogenase (LDH) release in the conditional medium, by a simple two-step LDH kit (Takara, Tokyo, Japan). Medium LDH contents were always normalized to the total LDH contents.

JC-1 assaying of mitochondrial depolarization
In stressed cells mitochondrial depolarization will cause JC-1 aggregating in mitochondria, thereby forming green monomers [34]. OB-6 cells or primary human osteoblasts were seeded into 12-well tissue-culture plates (5 × 10 4 cells in each well). Following the applied treatments cells were stained with JC-1 (5 μg/mL) and tested immediately by a fluorescence spectrofluorometer at 550 nm. The representative JC-1 images, merging both the green fluorescence image (at 550 nm) and the red fluorescence image (at 650 nm), were presented.
Annexin V assay OB-6 cells or primary osteoblasts were seeded into sixwell plates (3 × 10 5 cells per well). Following the applied treatment cells were incubated with Annexin V (10 μg/mL) and PI (10 μg/mL), and analyzed by a fluorescent-activated cell sorting (FACS) machine. The Annexin V ratio was recorded.

Transfection of miR mimic and miR inhibitors
OB-6 cells and primary human osteoblasts were seeded into the six-well plates at a 40-50% confluence. Cells were transfected with 500 nM of the applied miR inhibitor, control miR inhibitor or miR-124 mimic by Lipofectamine 2000 (Thermo-Fisher) for 24h. The siRNA/mimic transfection was repeated another round (total 48h). Afterwards, miRNA expression was tested by qPCR.

Statistical analysis
Data were presented as the mean ± standard deviation (SD). ANOVA with multiple comparisons through Bonferroni post-hoc test, analyzed by SPSS version 18.0 (SPSS Co., Chicago, IL), was utilized to test statistical differences. Values of P < 0.05 were considered statistically significant.