microRNA-455 targets cullin 3 to activate Nrf2 signaling and protect human osteoblasts from hydrogen peroxide

Over-production of hydrogen peroxide (H2O2) will lead to human osteoblast dysfunction and apoptosis, causing progression of osteoporosis and osteonecrosis. NF-E2-related factor 2 (Nrf2) is a well-characterized anti-oxidant signaling. Cullin 3 (Cul3) ubiquitin E3 ligase dictates Nrf2 degradation. We demonstrate that microRNA-455 (“miR-455”) is a putative Cul3-targeting microRNA. Forced-expression of miR-455 in both hFOB1. 19 osteoblast cell line and primary human osteoblasts induced Cul3 degradation and Nrf2 protein stabilization, which led to subsequent transcription of ARE (anti-oxidant response element)-dependent genes (NQO1, HO1 and GCLC). Cul3 silencing, by expressing miR-455 or targeted-shRNA, protected human osteoblasts from H2O2. Reversely, miR-455 anti-sense caused Cul3 accumulation and Nrf2 degradation, which exacerbated H2O2 damages in hFOB1. 19 cells. Moreover, forced over-expression of Cul3 in hFOB1. 19 cells silenced Nrf2 and sensitized H2O2. Together, we propose that miR-455 activated Nrf2 signaling and protected human osteoblasts from oxidative stress possibly via targeting Cul3.


Exogenous over-expression of Cul3 causes Nrf2 degradation
Based on the above results, we speculate that Cul3 over-expression should cause Nrf2 protein degradation. To test this hypothesis, the Cul3 expression vector was constructed, and was introduced to hFOB1. 19 cells. 19 cells (puromycin-selected), expressing miRNA-455 Vector [two lines, "Vec (1)/(2)"], microRNA-control ("miRC") or the Cul3-shRNA ("shCul3"), as well as the parental control hFOB1. 19 cells ("PAR") were treated with/out H 2 O 2 (250 μM, for applied time), relative NQO1, HO1 and GCLC mRNA expressions were tested by qRT-PCR assay (A). Cellular ROS intensity was tested by DCFH-DA fluorescent dye assay (B). Cell viability, cell death and apoptosis were tested by MTT assay (C), Trypan blue staining assay (D) and histone DNA ELISA assay (E), respectively. Data were shown as mean (n=5) ± standard deviation (SD). *p<0.05 vs. "PAR" cells. Experiments in this figure were repeated three times, and similar results were obtained. www.impactjournals.com/oncotarget Via puromycin selection, two hFOB1. 19 cell lines with the Cul3 construct were established, namely "Cul3 Vec(1)/(2)". As shown in Figure 4A, Cul3 mRNA level was significantly increased in the two stable lines. Western blotting assay results in Figure 4B confirmed the expression of exogenous Cul3 (Flag-tagged) and  endogenous Cul3 in the two lines. Notably, exogenous over-expression of Cul3 indeed led to Nrf2 protein degradation in hFOB1. 19 cells. Nrf2 mRNA ( Figure  4C) and Keap1 expression ( Figure 4B and 4D) were unchanged with Cul3 over-expression.
Cul3 is a member of the cullin-based ubiquitin ligase family, which is required for Nrf2 degradation [43]. Cul3 forms a complex with Hrt1 and BTB-domain containing proteins, which functions as an E3 ligase to bring Keap1 to ubiquitination and degradation [43]. On the other hand, Cul3 inhibition, silence or mutation will cause inhibition of Nrf2 degradation, and Nrf2 protein stabilization [44]. Recent studies have proposed that miRNA could be a novel and promising strategy to provoke Nrf2 signaling activation (mostly by targeting Keap1) [28,29]. Very few have focused on miRNA-mediated targeting of Cul3.
Our results here demonstrated that miR-455 is a Cul3-targeting miR in human osteoblasts. Forcedexpression of miR-455 in human osteoblasts led to Cul3 degradation, Nrf2 protein stabilization and subsequent transcription of ARE-dependent genes (NQO1, HO1 and GCLC). Remarkably, Cul3 silencing by miR-455 expression or targeted-shRNA protected human osteoblasts from H 2 O 2 . On the other hand, miR-455 depletion by miR-455 anti-sense led to Cul3 upregulation and Nrf2 protein degradation, which then exacerbated H 2 O 2 damages in human osteoblasts. These results together indicate that miR-455 expression could be a novel strategy to provoke Nrf2-ARE signaling activation in human osteoblasts. It will also be interesting to test the in vivo function of miR-455 against oxidative-damaged human osteoblasts. Expressions of miR-455 and Cul3 in human osteoporosis and osteonecrosis tissues should also be tested in future studies.

CONCLUSIONS
Together, our results suggest that miR-455 activates Nrf2 signaling via silencing Cul3, and protects human osteoblasts from oxidative stress.

Reagents
Puromycin was purchased from Sigma Aldrich (St. Louis, MO). All the antibodies were purchased from Cell Signaling Tech (Beverly, MA). Cell culture reagents were obtained from Gibco (Nantong, China).

Culture of osteoblastic cell line
The hFOB1.19 human osteoblastic cell line [45,46] was obtained from the Cell Bank of Shanghai Institute of Biological Science (Shanghai, China). Cells were maintained in α-modified essential medium (α-MEM) supplemented with 10% FBS, under 37°C in the presence of 5% CO 2 . Cells were fully differentiated as described [47].

Primary culture of human osteoblasts
The trabecular bone fragments from healthy donors were minced into small pieces, which were digested by incubation with 5 mg/mL collagenase D (Sigma) for 90 min at 37 °C with agitation. The resulting trabecular bone fragments were further digested with 0.5 mg/mL collagenase D overnight at 37 °C. Cells were then filtered through a 70-μm nylon mesh, and were placed onto the culture flasks with the described medium [48]. Medium was changed three times a week until reaching confluence, and were fully differentiated as described [47]. Primary human osteoblasts were used for further experiments stating at passage 3. The protocols of using human tissues and cells were approved by Ethics Board of Nantong University. Written-informed consent was obtained from each donor.

Cell viability assay
Human osteoblasts (5000 cells per each well) were initially seeded onto 96-well plates. Following the applied treatment, the MTT dye (20 μL/per well, 5 mg/mL, Sigma) was added to the supernatant for two hours. Afterwards, the optic density (OD) absorbance of MTT at 450 nm was measured by a microplate reader to reflect cell viability.

Cell death assay
Trypan blue staining assay was performed to test cell death after applied treatment. Cells excluding the dye were considered alive. Trypan blue positive cells were considered dead, and the ratio was recorded using an automatic cell counter.

Apoptosis quantification by ELISA assay
Thenucleosomal histone-bound DNA fragmentation is the characteristic marker of cell apoptosis, which was examined by the commercial available ELISA kit (Roche, Shanghai, China), using the anti-histone antibody and a secondary anti-DNA antibody. The ELISA OD at 450 nm was tested as the quantitative measurement of cell apoptosis. www.impactjournals.com/oncotarget

miR-455 anti-sense expression
The hFOB1. 19 osteoblasticcells were transfected with 20 nM of miR-455 anti-sense ("Anti-miR-455", Ambion, Shanghai, China) by Lipofectamine 2000 (Invitrogen). After two days, cells were split and were transfected with Anti-miR-455 again. This process was repeated for five rounds for a total of 10 days. Expression of miR-455 in the stable cells was examined by qRT-PCR assay. The Ambion miRNA anti-sense negative control ("Anti-miRC") was transfected to hFOB1. 19 cells as the control cells.

Western blotting assay
Equivalent amount of total cellular proteins (30 μg per lane) were extracted by a RIPA buffer (Biyuntian, Wuxi, China), and were separated by the 10% SDS gel, prior to transfer onto polyvinylidene difluoride (PVDF) membranes (Millipore, Shanghai, China). The blots were then blocked in 5 % (m/v) milk dissolved in Trisbuffered saline with 0.05 % (w/v) Tween-20 (TBS-T), and were probed with the designated primary and secondary antibodies. The protein signals were visualized under an enhanced chemiluminescence (ECL) system (Amersham Bioscience, Shanghai, China). β-Tubulin ("Tubulin") was always tested as the loading control. The images were analyzed with Image J software.

Cul3-shRNA
The lentiviral Cul3-shRNA particles (with GFPtag) were purchased from Santa Cruz Biotech (sc-35130-V, Nanjing, China). The lentiviral particles (20 μL/mL, per each well) were added to cultured human osteoblasts for 48 hours. Stable cells were again selected by puromycin (2.5 μg/mL, Sigma) for 96 hours. Over 98% of stable cells were GFP positive. Cul3 knockdown in the stable cells was verified by Western blotting assay and qRT-PCR assay. For the control cells, the lentiviral scramble control shRNA particles (Santa Cruz Biotech) were added.

Exogenous Cul3 over-expression
The full-length human Cul3 cDNA was synthesized by Genepharm (Shanghai, China), which was inserted to the pSuper-puro-GFP-Flag vector (Addgene, Shanghai, China). Lipofectamine 2000 was applied to transfect the Cul3 construct to human osteoblasts. Puromycin (2.5 μg/mL, Sigma) was added to select stable cells for 96 hours. Over 98% of cells were GFP positive. Expression of endogenous and exogenous (Flag-tagged) Cul3 in the stable cells was verified by Western blotting assay and qRT-PCR assay. The empty pSuper-puro-GFP-Flag vector was transfected to the control cells.

Quantitative RT-PCR
Trizol reagents (Invitrogen) were utilized to extract total cellular RNA, and the High Capacity cDNA Reverse Transcription Kit was applied to synthesize cDNA from 0.5 μg mRNA per treatment. Quantitative real-time PCR ("qRT-PCR") assay was performed by the Power SYBR Green RT-PCR Reagents Kit using the ABI-7500 system [49]. We utilized 2 ΔΔCt method to yield relative mRNA fold expression (as compare to GAPDH mRNA). mRNA primers for HO-1 and GCLC were described previously [50]. mRNA primers for Nrf2, Keap1, Cul3, NQO1 and GAPDH were described early [51]. miR-455 (-3p) expression was tested via the TaqMan microRNA assay [52] (Applied Biosystems, Shanghai, China), from 5 ng of total RNA [53].

Statistics
Data were expressed as the mean ± SD [45,57]. Comparisons between groups were performed via oneway ANOVA and then Student-Newman-Keuls test (SPSS 18.0). The p values < 0.05 were considered statistically significant.

Author contributions
All authors carried out the experiments, participated in the design of the study and performed the statistical analysis, participated in its design and coordination and helped to draft the manuscript.