CORM-3 Attenuates Oxidative Stress-Induced Bone Loss via the Nrf2/HO-1 Pathway

Bone metabolism occurs in the entire life of an individual and is required for maintaining skeletal homeostasis. The imbalance between osteogenesis and osteoclastogenesis eventually leads to osteoporosis. Oxidative stress is considered a major cause of bone homeostasis disorder, and relieving excessive oxidative stress in bone mesenchymal stem cells (BMSCs) is a potential treatment strategy for osteoporosis. Carbon monoxide releasing molecule-3 (CORM-3), the classical donor of carbon monoxide (CO), possesses antioxidation, antiapoptosis, and anti-inflammatory properties. In our study, we found that CORM-3 could reduce reactive oxygen species (ROS) accumulation and prevent mitochondrial dysfunction thereby restoring the osteogenic potential of the BMSCs disrupted by hydrogen peroxide (H2O2) exposure. The action of CORM-3 was preliminarily considered the consequence of Nrf2/HO-1 axis activation. In addition, CORM-3 inhibited osteoclast formation in mouse primary bone marrow monocytes (BMMs) by inhibiting H2O2-induced polarization of M1 macrophages and endowing macrophages with M2 polarizating ability. Rat models further demonstrated that CORM-3 treatment could restore bone mass and enhance the expression of Nrf2 and osteogenic markers in the distal femurs. In summary, CORM-3 is a potential therapeutic agent for the treatment of osteoporosis.


Introduction
Osteoporosis is one of the most common metabolic bone diseases worldwide, characterized by bone mass reduction and bone microarchitecture deterioration. This weakens the bones, making them fragile and prone to fracture. The related treatment costs and weak human resources present substantial clinical and economic impacts [1][2][3]. The pathogenesis of osteoporosis is caused by numerous factors, with oxidative stress being central [4,5]. Hydrogen peroxide and superoxide ions are thought to interfere with the bone environment in two main ways: suppressing osteoblastic functions and increasing osteoclastic activity [6,7]. Hence, modulating ROS production is a potential strategy for osteoporosis therapies [8].
CORM-3, a safe and readily available alternative to CO, has been widely used to investigate the working mechanisms of living systems [16,17]. Also, it can be used in treating several diseases. Recently, CORM-3 has been reported to ameliorate spinal cord-blood barrier disruption following injury to the spinal cord and alleviate neuron death after spinal cord injury via inflammasome regulation by Zheng et al. [18]. Also, it has been proposed that CORM-3 promotes the osteogenic differentiation of BMSCs and enhances the osteogenic differentiation of human periodontal ligament stem cells [19,20]. However, how CORM-3 ameliorates oxidative stress-induced damage is unclear. In the present study, oxidative stress was induced in BMSCs using H 2 O 2 as previously reported [21].
Therefore, we first measured the changes in CO concentration and the expression of HO-1 in normal and OVX rats. We found that in the OVX rats, the changes in CO contents correlated with HO-1 expression after OVX. Furthermore, we found that exogenously increasing CO by CORM-3 treatment rescued their osteogenic differentiation capacity and reduced apoptosis in the ROS environment. In addition, the injection of CORM-3 restored the bone microstructure and protected ovariectomized rats against bone loss in vivo. This study showed the mechanism and potential application of CO for osteoporosis therapy.

Isolation and BMSC Culture.
BMSCs extracted from the femur of Sprague-Dawley rats were cultured as previously described [22]. Briefly, BMSCs were isolated from the femur bone of euthanized Sprague-Dawley rats, sterilized for 15 minutes in 70% ethanol, and rinsed three times with phosphate-buffered saline (PBS). The BMSCs were then suspended in α-MEM supplemented with 10% (v/v) FBS and 104 IU/mL penicillin/streptomycin. Nonadherent BMSCs after 12 hours of incubation were removed, whereas the adherent cells were cultured in fresh media for follow-up experiments, which were changed every two to three days.

Cell Proliferation and Apoptosis
Assay. The effects of CORM-3 and H 2 O 2 on the viability of BMSCs were assessed using the Cell Counting Kit 8 (CCK-8) kit (Beyotime Institute of Biotechnology). The cells were inoculated in 96-well plates at a density of 5 × 10 3 and treated for 2 hours with varying CORM-3/iCORM concentrations (0, 12.5, 25, 50, 100, 200, and 400 μm) with or without H 2 O 2 treatment. After 24 hours, 10 μL CCK-8 reagent was added to each well before further culturing for another 1 hour. The optical density in each was measured at 450 nm using a spectrophotometer (Thermo Fisher Science, Waltham, MA, USA).
The proliferation of cells was performed using EDU Cell Proliferation Kit (C0071L, Beyotime, China) according to manufacturer's instructions [23]. Briefly, BMSCs were incubated for 4 hours with EDU and fixed for 30 minutes using 4% polyformaldehyde. Nuclear staining was performed using Hoechst. The stained cells were observed and photographed using a microscope.
The effect of CORM-3 and H 2 O 2 on the apoptosis of BMSCs was assessed using the FITC Annexin V/PI Apoptosis Detection Kit (Beyotime Institute of Biotechnology). Briefly, 1 × 10 5 cells were washed twice using cold PBS and resuspended in 1× binding buffer. The cells were then stained using propidium iodide and Annexin V FITC. Flow cytometry was used to detect early and late cellular apoptosis. The data were analyzed using the FlowJo software.
2.4. Osteogenic Differentiation of BMSCs. BMSCs were first seeded into 24-well plates for osteogenic differentiation at an initial density of 5 × 10 4 cells/well. After treatment, the culture medium was replaced with an osteogenic medium. The medium was changed every day. Alkaline phosphatase (ALP) and Alizarin red S (ARS) staining were performed as previously described [24]. Staining was performed on osteogenic days 7 and 21 using the ALP Staining Kit (Beyotime Institute of Biotechnology; Jiangsu, China) and the ARS solution (Solarbio Science & Technology). The absorbance of ALP and ARS was measured using a microplate reader at 520 nm and 570 nm, respectively.

Carbon Monoxide Content Detection in Femur.
The CO content in different groups of rats was detected using the Endogenous Carbon Monoxide Assay Kit (Jiancheng Biotechnology, Nanjing, China) as previously described [25]. Briefly, the distal femur bones for rats in both groups (sham group and OVX group) were cut into small pieces and washed in PBS. After homogenizing in PBS, 0.5 mL of the tissues were added into the Hb solution (1 mL). Following vortexing and quiescence, mixtures were measured at 541 nm (as absorbance) and 555 nm (as a reference). The CO content in samples was based on the ratio of x and y measured at 541 nm and 555 nm.
2.6. Western Blot Analysis. Western blot analysis was performed as previously described [26]. BMSCs were washed for 30 minutes at 4°C with ice-cold PBS and lysed with 2 Oxidative Medicine and Cellular Longevity radioimmunoprecipitation (RIPA) lysis buffer (Beyotime Institute of Biotechnology) supplemented with phosphatase and protease inhibitors (1 mM) (Sigma-Aldrich). The cell lysate was centrifuged at 12000 g to separate the proteins. The concentration of the proteins was evaluated using the BCA Protein Assay Kit (Beyotime Institute of Biotechnology). SDS-polyacrylamide gel electrophoresis (SDS-PAGE) was performed using 20 μg of the extracted proteins. Briefly, the proteins were first transferred to a polyvinylidene difluoride membrane through overnight incubation at 4°C. After three washes with 0.1% Tween 20 (TBST), the membranes were incubated for 4 hours at room temperature with corresponding HRP-conjugated secondary antibodies. The proteins were visualized using a ChemiDoc XRS+ system equipped with an augmented chemiluminescence reagent (Bio-Rad; Hercules, CA, USA). The various protein bands were quantitatively analyzed using the Image Lab software V3.0 (Bio-Rad).

Quantification of the Activities of CAT and GSH.
After treatment with H 2 O 2 , the cells were rinsed three times with PBS and lysed for 30 minutes on ice-cold lysis buffer. The activities of CAT and GSH were assessed using commercial assay kits (Jiancheng Biotechnology, Nanjing, China).

Lentivirus
Transfection. Nrf2 knockdown in BMSCs was performed through transfection with lentivirus (GeneChem, Shanghai, China) at a density of 30-50% density. The culture media of the cells was replaced after twelve hours after transfection and replaced every three days replaced until the growth of cells reached 80-90% confluency. The efficacy of transfection was analyzed using western blot.
2.9. Reactive Oxygen Species Assay. The levels of intracellular ROS were quantified using the dihydroethidium (DHE) test (Yeasen Biotech Co., Ltd., Shanghai, China) based on a fluorescent probe. Treated BMSCs were washed three times using PBS buffer and incubated at 37°C for 30 minutes in darkness with DHE. The cells were observed immediately under a microscope (Olympus Life Science; Tokyo, Japan). Three random photos were then captured and analyzed using the Image-Pro Plus software version 6.0.

Mitochondrial Function
Assays. The levels of superoxide ions and the mitochondrial membrane potential (MMP) in treated BMSCs were determined by MitoSox and JC-1 probes, respectively, according to the manufacturer's instructions (Beyotime, Shanghai, China). The cells were observed using a fluorescent microscope (Olympus Life Science; Tokyo, Japan).
2.14. Immunofluorescence. BMSCs were cultured on glass slides at 37°C for 20 minutes and permeabilized with 0.1% (v/v) Triton X-100 in PBS for 15 minutes. Then, the cells were incubated for 1 hour at room temperature with 1% goat serum to block nonspecific antibody binding sites. The cells were then incubated at 4°C overnight with primary antibodies rinsed three times with PBS and further incubated at room temperature for 1 hour in the dark with fluorescence-conjugated secondary antibodies. After staining the cell nuclei with DAPI for 5 minutes at room temperature, the cells were visualized using an Olympus fluorescent microscope (Olympus Life Science; Tokyo, Japan). The IF results were analyzed quantitatively using the Image-Pro Plus 2D Software (Rockville, MD, USA).

Treatment and Animal
Model. The protocol for animal experiments was approved by the Ethics Committee of the Second Affiliated Hospital and Yuying Children's Hospital 3 Oxidative Medicine and Cellular Longevity of Wenzhou Medical University. A total of 40 female SD rats (three months of age and weighed 250 ± 20 g) used in this study were purchased from Shanghai Laboratory Animal Center (SLACCAS; Shanghai, China) and were maintained under specific pathogen-free (SPF) conditions at 22-25°C under a 12-hour light/dark cycle. The rats were provided with enough food and water. Inactive CORM-3 (iCORM-3) was generated by overnight incubation of CORM-3 in PBS at room temperature to assess the effect of CO. The rats were divided into four groups. The sham group consisted of ten randomly picked rats. The remaining 30 rats underwent bilateral OVX using the double dorso-lateral method described by Park et al. OVX rats were randomly divided into OVX+CORM-3, OVX+iCORM-3, and OVX+saline groups (OVX group). The OVX+CORM-3 group received 15 mg/kg of CORM-3 diluted with normal saline for 8 weeks. At the same time, the OVX group and OVX +iCORM-3 group received an intraperitoneal injection of an equivalent volume of saline and 15 mg/kg of iCORM-3, respectively, for 8 weeks. The femur bones of the rats were collected and preserved in 4% paraformaldehyde after sacrificing the rats.
2.16. The Microstructure of the Distal Femur. The microstructure of the distal femur was analyzed using a micro-CT imaging system (μCT 100, Scanco Medical; Bruttisellen, Switzerland). The microarchitectural parameters of the volume of interest (VOI) were obtained using software in the micro-CT workstation, based on three-dimensional reconstructed images. Finally, further bone quality assessment was performed by quantitatively measuring region of interest-(ROI-) related parameters, including the bone volume to tissue volume (BV/TV), trabecular thickness (Tb.Th, mm), trabecular number (Tb.N, 1/mm), and trabecular separation (Tb.Sp, mm).
2.17. Calcein Labeling Assay. Rats were injected intraperitoneally with 20 mg/kg calcein (Sigma-Aldrich, USA) on days 14 and 2 before being sacrificed. After sacrifice, the left femur bone was removed and fixed in 4% formaldehyde and embedded in methyl methacrylate. After hardening, the bone tissue was dissected into 30 μm thin sections using a microtome (SP1600; Leica, Germany). The cortical endosteum surfaces were evaluated immediately using a fluorescence microscope (STP6000; Leica, Germany) with an excitation wavelength of 488 nm.

Histomorphological and Immunohistochemical (IHC)
Analyses. After CT scanning, the femur bones were decalcified for 3 weeks in 10% EDTA solution, embedded in paraffin wax, fixed in ethanol, cleared in xylene, and embedded in paraffin wax. The bones were then cut longitudinally into 4 μm thick sections and mounted on polylysine-coated glass slides before hematoxylin and eosin (H&E) and Masson's trichrome staining (Solarbio Science & Technology). For IHC analysis, 4 μm thick sections were incubated with primary antibodies against Nrf2, COL1A1, HO-1, and RUNX2, rinsed three times using PBS, and incubated with horseradish peroxidase-conjugated secondary antibodies. The level of Nrf2, COL1A1, HO-1, and RUNX2 expression were analyzed using the Image-Pro Plus software.
2.19. The Level of ROS In Vivo. The level of ROS in femurs was detected using fluorescent dye DHE as previously described [27]. Briefly, femurs bones preserved in 4% paraformaldehyde (PFA) for four weeks were delicately decalcified in 10% EDTA solutions. The fixed bone tissues were cryoprotected through overnight incubation at 4°C with 20% sucrose and 2% polyvinylpyrrolidone solution. The bones were then immersed in optimum cutting temperature solution (O.C.T.) and stored at -80°C till further analyses. The bone tissues were cut into thin 5 m sections and incubated with DAPI solution and with an antifluorescence quencher (Yeasen Biotech Co., Ltd., Shanghai, China). The DHE images of the newly frozen segments were observed using a fluorescent microscope (Olympus Life Science; Tokyo, Japan).
2.20. TUNEL Assay. The TUNEL staining assay was performed using Cell Death Detection Kit (Yeasen Biotech Co., Ltd., Shanghai, China), following the manufacturer's instructions. Briefly, the tissues were incubated at 37°C in a humid environment with a mixture of TUNEL solution hydrolyzed with 15 g/mL proteinase K, and thereafter treated with 3% hydrogen peroxide. The percentage of apoptosis was then determined.
2.21. Data Analysis. Continuous normally distributed data were expressed as a mean ± SEM. All experiments were performed at least 3 times. Data were analyzed using the SPSS software (IBM, Armonk, NY, USA). Differences between groups were analyzed using the Tukey test and one-way ANOVA. Statistical significance was set at p value of < 0.05 (p < 0:05).

HO-1 Expression and CO Content in the Femur.
We detected the expression of HO-1 in the distal femur, the most pivotal enzyme of autogenous CO generation. Western blotting results revealed that HO-1 expression level decreased after OVX (Figures 1(a) and 1(b)). Further, IHC and IF staining revealed that HO-1 expression decreased in the distal femur of OVX rats compared to the Sham group (Figures 1(c) and 1(d)). To ascertain the change in CO level after OVX, we quantified the CO content in different groups. We found that the CO concentration was reduced in the distal femur of the OVX rat (Figure 1(e)). These results indicated that HO-1 expression and endogenous CO were decreased after OVX. Similarly, Cytochrome C (Cyto-c), as previously reported, is upregulated early during apoptosis while antiapoptotic Bcl-2 was downregulated [28]. Notably, the levels of these proteins returned to normal levels upon CORM-3 treatment (Figures 3(a)-3(e)). The percentage of early apoptotic plus late apoptotic cells was calculated following flow cytometric analysis (Figures 3(f) and 3(g)).
Considering these results, we demonstrated that CORM-3 ameliorated H 2 O 2 -induced damage on BMSCs but iCORM-3 had no effect on the H 2 O 2 -induced damage on BMSCs.

CORM-3 Modulates H 2 O 2 -Induced Oxidative Stress in
BMSCs. Oxidative stress disrupts the normal functioning of cells. Therefore, this study assessed the impact of CORM-3 on ROS production, antioxidant enzyme activity, and mitochondrial dysfunction in H 2 O 2 -treated BMSCs. Further, DHE staining was performed to determine ROS levels in cells, whereas MitoSox and JC-1 probes were utilized to assess mitochondrial ROS and mitochondrial membrane potential (MMP), respectively. H 2 O 2 exposure enhanced ROS production relative to untreated cells, which was neutralized by pretreatment with CORM-3 (Figure 4(a)). Compared to untreated cells, H 2 O 2 -treated cells had significantly reduced MMP and higher superoxide anion content and both recovered to near-physiological levels by CORM-3 treatment (Figure 4(b)). TEM analyses supported this finding, and we observed mitochondrial morphology deterioration following H 2 O 2 exposure, including shrunken mitochondria, ruptured membranes, and mitochondrial recovery after CORM-3 treatment (Figure 4   Oxidative Medicine and Cellular Longevity 7 Oxidative Medicine and Cellular Longevity model of myocardium cell injury, suggesting that the protective effects of CORM-3 are mediated through this pathway. NAD(P)H quinone oxidoreductase 1 (NQO1) is an essential homodimer flavin protease that is against endogenous oxidative stress by conserving the reduced form of ubiquinone and tocopherol. CORM-3 treatment reversed the H 2 O 2 inhibitory effects on Nrf2 nuclear translocation and expression of HO-1 and NQO1 (Figures 6(a)-6(g)). Interestingly, suppressing Nrf2 expression inhibited the protective effect of CORM-3 against H 2 O 2 . Western Blot shows that Nrf2 was knocked out successfully (Figures 7(a) and 7(b)). Seven days after osteogenic induction, the level of expression of COL1A1, RUNX2, and OCN in BMSCs treated with CORM-3 was partially reduced by Nrf2 knockout (Figures 7(g)-7(j)) and reversed CORM-3's beneficial impacts on BMSCs' mineralization and differentiation (Figures 7(c)-7(f)). These findings demonstrated that CORM-3 exerts its effects by inducing nuclear translocation of Nrf2 and activating its downstream pathways.      9 Oxidative Medicine and Cellular Longevity M1/M2 polarization of macrophages may be a new approach to treating osteoporosis. Therefore, we speculated that CORM-3 reduces inflammation and inhibits osteoclastosis by decreasing the M1/M2 ratio. Immunofluorescence analysis revealed that H 2 O 2 treatment significantly increased the proportion of iNOS-positive cells, but CORM-3 treatment reversed this phenomenon (Figure 8(c)). Also, CORM-3 treatment significantly increased the proportion of the CD206positive cells (Figure 8(d)). In addition, the CORM-3 treatment modulated the mRNA expression for iNOS, TNF-α, IL-1β, Arg-1, Ym1, and CD206 (Figure 8(g)). The expression of iNOS and CD206 proteins exhibited a similar trend    8(b)). These findings suggested that CORM-3 promotes the polarization of macrophages from M1 to M2 phenotypes and inhibits osteoclastosis.

CORM-3 Treatment Restored Bone Microstructure and
Promoted Bone Synthesis In Vivo. The OVX rat model was used to evaluate the effect of CORM-3 on osteoporosis in vivo. Analysis of 3-D micro-CT scan images revealed that 8 weeks of CORM-3 treatment restored the architecture of trabecular bone in distal femurs (Figure 9(a)). Additionally, OVX rats were administered with CORM-3 for 8 weeks also showed significant improvements in BV/TV, Tb.N, and Tb.Sp (Figures 9(b)-9(d)). However, no significant change in the Tb.Th was observed (Figure 9(e)). Consistently, H&E and Masson trichrome staining of the distal femur sections indicated more numerous trabeculae and distinct lower trabecular separation in the OVX+CORM-3 group compared to OVX and OVX+iCORM-3 group (Figure 9(f)). The effect of CORM-3 on bone formation in OVX rats was analyzed using calcein double labels. calcein was injected twice (on days 2 and 14) before sample extraction and analysis. We found that compared to control, the interlabel dis-tance was significantly increased in the CORM-3 group (Figure 9(g)).

CORM-3 Injection Restored Nrf2 Expression and
Osteogenic Indices of the Distal Femur and Reduced the Production of ROS and Tissue Apoptosis In Vivo. To test whether Nrf2 has the same trend in rats, we determined the level of Nrf2 expression in rats' distal femurs using IHC and IF. Nrf2 levels were significantly elevated in CORM-3-treated distal femurs compared to OVX-treated distal femurs (Figures 10(a), 10(b), 10(f), and 10(g)). Next, IHC staining was performed on the distal femur to detect osteogenesis in vivo. Similar to in vitro results, RUNX2 and COL1A1 levels were significantly higher in the CORM-3 group than in the OVX group (Figures 10(c), 10(h), and 10(i)). Furthermore, TUNEL assays were utilized to determine the level of apoptosis in the distal femurs of the various treatment groups. The results reveal that the OVX group had a significantly greater index of apoptotic cells in the distal femur than the sham group. The TUNELpositive cell percentage was also considerably lowered in the CORM-3 treatment group (Figures 10(d) and 10(j)). Finally, the femur ROS level was evaluated by DHE staining. DHE staining revealed that COMR-3 modulated the

Discussion
The correlations between osteoporosis and ROS and underlying molecular mechanism have generated tremendous interest over the years. Skulachev et al. found high ROS levels in almost all osteoporosis forms, including senile osteoporosis, menopausal osteoporosis and diabetic osteoporosis, suggesting that ROS independently promotes osteoporosis [29,30]. As reported, ROS can be endogenously eliminated by the HO-1/CO pathway [31]. Up to date, little is known regarding the effect of CO in OVX rats. To the best of our knowledge, we are the first to report the relationship between HO-1 and its product CO in OVX. We found that HO-1 and its product CO significantly decreased in the OVX rat model compared to the controls. In addition, our experiments showed that supplementing exogenous CO reversed H 2 O 2 -induced mitochondrial dysfunction by activating the Nrf2/HO-1 signaling pathway and inhibited osteoclast differentiation by regulating the macrophage polarization and, thus, preventing osteoporosis. Given that iCORM-3 does not exert the same effects as CORM-3, directly implicating CO as the active mediator in these responses.
As one of the degradation products of heme, CO possesses diverse biological roles [14]. However, because of its high affinity to Hb, CO is poisonous to living animals [31]. CORM-3 is a novel compound that carriers of CO and reproduce its biological actions [32]. Numerous studies have demonstrated that CORM-3 is non-toxic in low doses. For instance, Liu et al. found that CORM-3 ameliorates acute pancreatitis, and Portal et al. reported that CORM-3 protects adult cardiomyocytes against hypoxia-reoxygenation by modulating pH restoration [33,34]. Accordingly, we speculated that CORM-3 plays a vital role in the metabolism of bone. CORM-3 inhibits osteoclastogenic differentiation by the HO-1 pathway [35]. Similarly, Li et al. found that CORM-3 promotes the osteogenic differentiation of BMSCs [19]. Therefore, we performed animal experiments to verify the therapeutic effect of CORM-3 in vivo. We found that CORM-3 ameliorated bone loss in OVX rats in vivo.
The mitochondrion is a major site for ROS production, and excessive ROS accumulation disrupts the mitochondrial membrane potential and impairs the mitochondrial function, triggering apoptosis [24]. Growing evidence shows that maintaining mitochondrial homeostasis is critical for osteogenic differentiation of BMSCs [36]. According to Chen et al.'s discovery, BMSCs with compromised mitochondria displayed lower osteogenic protein expression levels, accompanied by decreased ALP activities and reduced mineralized nodule formation [37]. Pal et al. reported that restoring the mitochondrial membrane potential promoted osteogenic differentiation [38]. In the present study, we found that CORM-3 treatment attenuated the deleterious impact of H 2 O 2 and downregulated the expression of apoptosisrelated proteins, including Cyto-C in mitochondria, Bax, and cleaved caspase-3, but upregulated that of Bcl2, an antiapoptotic protein.  Significant differences between groups are indicated as * * * p < 0:001, * * p < 0:01, and * p < 0:05; ns: no significance. 16 Oxidative Medicine and Cellular Longevity

17
Oxidative Medicine and Cellular Longevity COL1A1, RUNX2, and OCN, accompanied by concomitant loss of ALP and ARS osteogenic phenotypes. However, CORM-3 treatment upregulated the expression of osteogenic phenotypes of these proteins and restored mitochondrial function and structure, consistent with previous findings [39]. ROS has previously been reported to induce macrophage polarization to the M1 phenotype, promoting osteoclastogenesis [40,41]. Herein, qPCR, immunofluorescence, and TRAP staining confirmed that CORM-3 regulates H 2 O 2 -induced macrophage polarization and inhibits osteoclastogenesis. Taken together, we suppose that CORM-3 prevents BMSCs apoptosis due to its inhibition of mitochondrial dysfunction and inhibits osteoclastogenesis by regulating macrophage polarization.