1.Introduction

In many organisms, the skeletal system supports and protects the internalsofttissue.Damagetothebonetissuemayleadtoseveralissues. There are two mechanisms of bone development: endochondral and intramembrane osteogenesis [1,2]. Osteoclasts serve a crucial function in forming bone and maintaining bone homeostasis throughout devel-opment [3]. During osteogenesis, various cells, hormones, and factors collaborate closely to coordinate osteoblast and osteoclast activity to maintain bone homeostasis [4]. In the elderly, dysfunction between osteoblast and osteoclast activity is the major cause of osteoporosis [5]. There are several causes of osteoporosis include endocrine disorders or chronic disease [6]. In our recent study, we showed that iron overload raises reactive oxygen species (ROS) levels, consequently inducing osteoporosis [7].

ROSareanelectronreductionproductofoxygenspeciesinvivo.They are generated by electrons leaking out of the respiratory chain. ROS consume approximately 2% of cellular oxygen before delivery to the terminal oxidase [8,9]. Importantly, ROS are involved in various com-plex cellular processes, including infammation, proliferation, and differentiation [10]. However, under adverse conditions in vivo, a sud-den increase in ROS levels may cause irreversible damage to cellular structure and function. Bone tissue is a target organ for excessive ROS production. Recent studies indicated that excessive ROS inhibit osteo-blast differentiation and enhance osteoclast differentiation and prolif-eration [11,12].

Inthe body, ROS originatesfrom Nicotinamide adenine dinucleotide phosphate (NADPH) oxidase (Nox). The Nox family has seven members, including Nox1-5 and double oxidase 1–2(Doux1-2). Noxs expression is tissue- and cell-type specifc [13]. Further, Nox proteins participate in several biochemical reactions. Nox4 is one of the main source of ROS production and increased Nox4 expression is a main inducer of ROS production. Protein structure analysis of Nox4 revealed three key do-mains near the C-terminus: ferric reductase-like transmembrane com-ponents, Flavin adenine dinucleotide(FAD)-binding domains, and Nicotinamide adenine dinucleotide(NAD)-binding domains. These do-mains are involved in electron transfer in the respiratory chain and su-peroxide production. Previous studies indicated that Nox4 plays an important role in different cells and is closely related to angiogenesis, chronic kidney disease, and bone homeostasis [11,14–19]. Recent studies showed that deleting Nox4 also induces solid tumor formation [20].

The involvement of Nox4 in bone metabolism is well known. In 2000, Su et al. reported that Nox4 is expressed in osteoclasts [21]. Subsequent studies reported that Nox4 is also present in osteoblasts. In 2013, Claudia et al. found that Nox4 reduces bone mass by promoting osteoclast activity. Nox4-defcient mice show signifcantly higher bone mineral density and reduced osteoclast numbers [19]. In humans, Nox4 expression increases in osteoporosis patients, which leads to increased boneturnover.However,Claudiaetal.foundthatNox4doesnotseemto be involved in bone formation [11]. Another study found that Nox4 plays an indispensable role in osteoblast proliferation and differentia-tion. Further, the production of ROS originating from Nox4 indirectly mediates the Bone morphogenetic protein 2(Bmp2) transcription and osteoblast differentiation [19]. The Transforming growth factor-β(TGF-β) pathway is the main signaling pathway that promotes osteoblast differentiation [22]. TGF-β regulates Smad2/3 phosphoryla-tion, promotes Runx2 expression, and participates in osteoblast differ-entiation [23–25]. In addition, many studies showed that BMP mediates the phosphorylation of receptor-specifc Smad1, 5, and 8, which form complexes with Smad4, translocates to the nucleus [26]; stimulates the transcription of osteoblast differentiation genes runx2, osx, and ocn; and induces osteoblast differentiation [27–29]. Although numerous studies reported that Nox4 is involved in osteoblast and osteoclast formation, how Nox4 is involved in bone metabolism remains unclear,especially in the early stages of bone development.

Zebrafsh are increasingly used in studies of genetics, development, disease models, and drug screening. Zebrafsh are a good model for studying bone-related diseases because bone development and regulatory mechanisms in these organisms are similar to those in mammals. In addition,the development of the craniofacial bone in zebrafsh is similar to that in mammals, including endochondral and intramembrane osteogenesis. Because zebrafsh develop in vitro, which is different from mouse development in vivo,they are a good model organism for studying early bone development [30,31]. Zebrafsh has only one nox4 gene, which facilitates studies of Nox4-related functions. Previous studies found that Nox4 is limits bone mass by promoting osteoclastogenesis. However, whether Nox4 deletion leads to early bone metabolic damage has not been reported, and it’s hard to solve these problems with mice. In this study, we constructed nox4−/−zebrafsh using CRISPR/Cas9 technology to study the role of Nox4 in bone development and to investigate the role of Nox4 in early bone metabolism.

2.Materials and methods  

2.1.Zebrafsh

WT, Tg(sox9-EGFP), and Tg(sp7-EGFP) zebrafsh were provided by Suzhou Murui Biotechnology Company. The expression of these genes was observed under a fuorescence microscope. ImageJ software was used to quantify the fuorescence and to analyze bone development.

Adultzebrafshwereraisedincirculatingsystemwaterat28◦Canda 14 h/10 h light/dark cycle. The fsh were fed three times per day with fresh harvested shrimp. In zebrafsh mating experiments, male and fe-male zebrafsh were placed into partitions overnight. At 9:00 a.m. the next day, the partitions were removed. After 30 min, the embryos were collected, placed in a Petri dish containing methylene blue (0.3 μg/ml), andculturedinalight-controlledincubator(14h/10hlight/darkcycle) at 28 ◦C. Larvae were fed with Paramecium every day starting at 5 dpf, and the state of the larvae was observed. At 7 dpf, Paramecium was gradually replaced with shrimp. At 20 dpf, the zebrafsh were placed in circulating system water.

2.2.Construction of nox4−/−zebrafsh

nox4−/−zebrafsh were constructed using CRISPR/Cas9 technology. Thetargetsitewasdesignedonexon4ofnox4.gRNA(100pg)(Thenox4 target sequence: 5′-CCTGTGTGTGGCCGGAGGGATC-3′)was synthesized in vitro and microinjected into zebrafsh embryos together with Cas9- capped mRNA at the single-cell stage. The injected embryos were placed in E3 medium (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl2, 0.33 mM MgSO4) in a 28 ◦C constant temperature incubator. At 3 dpf, DNA from10 larvae was extracted by Animal tissue Genomic DNA Kit(ZP307,  CN),theregioncontainingthetargetsitewasamplifedviaPCR(forward primer: 5′- TGCACGTGGATGGTCCGTT-3’; reverse primer: 5′- GTGCTTGAAGGACACAAGCG-3′), and the nox4 deletion was confrmed  by DNA sequencing. Primary embryos(F0) were raised to adulthood and mated with WT zebrafsh to obtain F1 generation embryos, which were raised to adulthood. The tails of F1 adult zebrafsh were collected, and the DNA was extracted, amplifed via PCR, and sequenced to determine whether the F1 generation was mutated. F2 embryos were obtained by mating F1 embryos with the fsh with the same mutation. The F2 embryos were raised to adulthood. DNA was extracted from the zebrafsh tail, amplifed via y PCR, and compared with DNA from the control group to confrm the mutation. The mutated PCR products sequenced to verify the mutation (Fig. 2).

2.3.Zebrafsh drug treatment

GLX351322 (MCE, USA) is a Nox4 inhibitor that inhibits tetracycline-induced Nox4 overexpressing cells to produce hydrogen  peroxide. The working concentration used in the animal experiments  was 20 μM. SB431542 (Selleck, USA) is a TGF-β inhibitor and was used  at 10 μM in animal experiments. All reagents were prepared in dimethyl  sulfoxide (DMSO) and diluted in E3 medium to the working concentration. In the drug treatment experiment, the control group was treated  with the same amount of DMSO (0.01%).

2.4. ROS probe detection  

A ROS detection kit (Solarbio, China) was used to detect and quantify  ROS in cells. At 4 dpf, the zebrafish larvae were treated with Tricaine  (50 mg/L) and washed with PBS. The larvae were incubated with DCFHDA (10 μM) fluorescent probe for 30 min at 37 ◦C, washed with E3  medium, and imaged using a fluorescence microscope within 30 min.  Fluorescence intensity was quantified using ImageJ software.  

2.5. Neutral red staining  

Sudan black (Solarbio, China) (0.6 g) was dissolved in 200 mL pure  alcohol, filtered, and stored at 4 ◦C. Then, 30 mL Sudan black solution  was diluted with 20 mL buffer solution (16 g phenol, 30 mL 100%  ethanol, and 0.3 g disodium hydrogen phosphate in 100 mL distilled  water), filtered, and stored until use. The larvae or tails were fixed with  4% paraformaldehyde (PFA) overnight, washed with PBS, and placed in  Sudan black staining solution in the dark for 30–60 min. The tissue was  rinsed with 70% ethanol 3 times for 15 min each. After staining,  Neutral red N8160 (Beyotime, China) (0.125 g) was dissolved in 50  mL of E3 medium and stored in the dark. Before the experiment, E3  medium was diluted 1000 × . At 3 dpf, larvae were cultured in a 12 well  plate with 2.5 mg/L neutral red dye at 28.5 ◦C in the dark for 8 h. Then,  the larvae were washed with PBS, treated with anesthetics (50 mg/L),  and imaged with a microscope. Neutral red staining was analyzed using  ImageJ software.  

2.6. Sudan black staining  

Sudan black (Solarbio, China) (0.6 g) was dissolved in 200 mL pure  alcohol, filtered, and stored at 4 ◦C. Then, 30 mL Sudan black solution  was diluted with 20 mL buffer solution (16 g phenol, 30 mL 100%  ethanol, and 0.3 g disodium hydrogen phosphate in 100 mL distilled  water), filtered, and stored until use. The larvae or tails were fixed with  4% paraformaldehyde (PFA) overnight, washed with PBS, and placed in  Sudan black staining solution in the dark for 30–60 min. The tissue was  rinsed with 70% ethanol 3 times for 15 min each. After staining, the  tissue was rinsed with 3% H2O2/0.5% KOH for 10 min. Then, 25%  glycerol/0.1% KOH was added to remove bubbles and the tissue was  placed in 50% glycerol. The tissue was imaged with a microscope and  the dyed area was analyzed using ImageJ software.

2.7. Alcian blue staining  

Alcian blue staining is a common staining method that detects  cartilage in bone tissue and cultured bone cells. Zebrafish larvae were  collected, placed in 4% PFA, and incubated overnight at 4 ◦C. The larvae  were washed with 100 mM Tris-HCl pH 7.5/10 mM MgCl2 for 10 min.  Then, the larvae were incubated with 1 mL 0.04% alcian/10 mM MgCl2  in 100 mM Tris-HCl pH 7.5/10 mM MgCl2 overnight at room temperature while shaking. After staining, the tissue was sequentially rinsed  with 80%, 50%, and 25% ethanol/100 mM Tris pH7.5/10 mM MgCl2 for  5 min each. After rinsing, the pigment was bleached with an equal  amount of 3% H2O2 and 0.5% KOH, until the pigment was bleached  clean. The tissue was rinsed with 25% glycerol/0.01% KOH for 10 min  until there were no bubbles. Then, the samples were incubated in 50%  glycerol/0.1% KOH and imaged with a microscope. ImageJ software  was used to analyze the stained area.

2.8. Alizarin red staining  

Zebrafish larvae or tails were collected and placed in 4% PFA overnight at 4 ◦C. The tissue was washed with 100 mM Tris pH 7.5/10 mM  MgCl2 for 10 min. After rinsing, the body pigment was bleached with an  equal amount of 3% H2O2 and 0.5% KOH, until the body pigment was  bleached clean. The solution was rinsed with 25% glycerol/0.1% KOH  for 10 min until there were no bubbles. 0.01% alizarin red was added for  30 min at room temperature, washed with 50% glycerin/0.1% KOH for  10 min, and incubated in fresh 50% glycerin/0.1% KOH solution. The  tissue was imaged and analyzed using ImageJ software.

2.9. TRAP staining  

TRAP staining was used to detect the bone tissue and osteoclasts.  Zebrafish larvae or tails were collected and placed in 4% PFA overnight  at 4 ◦C. Then, the tissue was washed with 1 × PBST for 30 min. Next, the  tissue was placed in incubation buffer (80 μL acetic acid and 40 μL  tartaric acid in 2 mL deionized water) (sigma,USA)at room temperature  for 40 min. Then, 10 μL naphthol AS-BI phosphoric acid solution(sigma,  USA) was added to 1 mL incubation buffer. The samples were incubated  in this solution in the dark for 10 min. Then, the tissue was incubated in  10 μL nitrite solution and 10 μL rapid garnet type alkali solution for  30–60 min. The tissue was sequentially washed with 80%, 50%, and  25% ethanol/100 mM Tris-HCl pH 7.5/10 mM MgCl2 for 5 min. After  rinsing, the pigment was bleached with an equal amount of 3% H2O2  and 0.5% KOH. The samples were rinsed twice with 25% glycerol/0.1%  KOH for 10 min each, rinsed with 50% glycerol/0.1% KOH for 10 min,  and stored in 50% glycerol. The stained area was observed using a microscope and analyzed using ImageJ software.

2.10. μCT analysis  

μCT (Aolong, China) is a nondestructive 3D imaging technology. WT  and nox4− /− zebrafish were placed in a μCT instrument, and the layers of  various zebrafish tissues were scanned. Bone development was analyzed  using the following bone tissue parameters: BV/TV, BS/BV, TB Th, and  TB SP.

2.11. Transcriptome analysis  

At 9 dpf, WT and nox4− /− 30 larvae per group were collected. Total  RNA was extracted using an RNA extraction kit (Invitrogen, Carlsbad,  CA, USA), stored in liquid nitrogen, and sent to BEJ (China) for transcriptome analysis. The data were analyzed for differential expression in  the WT and nox4− /− zebrafish. The differentially expressed genes were  analyzed using GO and KEGG to elucidate downstream molecular  mechanisms.

2.12. Cell culture and mineralization  

Saos-2 human osteosarcoma cells were purchased from Sun Cell,  China. The cells were cultured in McCoy’s 5a medium supplemented  with 10% fetal bovine serum and 1% penicillin-streptomycin at 37 ◦C  and 5% CO2. Then, 100 μg/mL ascorbic acid and 5 mM β-glycerophosphate were added to the culture for 2 weeks to induce cell  mineralization. Then, 5 μM GLX351322 (MCE, USA) and/or 10 μM  SB431542 (Selleck, USA) were added to the cell culture medium. The  medium was changed every two days. After drug treatment, the cells  were washed three times with PBS, fixed in 70% ethanol at 4 ◦C for 1 h,  washed with PBS, and stained with 40 mM alizarin red (pH 4.0). After  10 min, the cells were washed with PBS, dried, and imaged with a  microscope.

2.13. RT-qPCR  

Total RNA from 30 larvae or 1 × 106 cells was extracted using an  RNA extraction kit (Invitrogen, Carlsbad, USA. Then, 1 μg RNA was  reverse transcribed into cDNA with oligo(dT)18 primer and M-MLV  reverse transcriptase. The cDNA was diluted 10 times and used as an  amplification template with β-actin as an internal reference gene. SYBR  Premix ExTaq (Takara, China) and a Step1 real-time PCR detection  system was used for RT-qPCR. Relative gene expression was calculated  using the 2− ΔΔCT method. Each group included at least three independent samples. See Supplementary Table 1 for primer sequences.

2.14. CCK-8 assays  

Saos-2 cells were divided into control, induction, GLX351322, and  inducer co-treatment groups and cultured for four days as described  above. Then, 100 μL of cell suspension was added to each well of a 96-  well plate. Three replicates were used per condition. The culture plate  was placed in an incubator for 4 h (37 ◦C, 5% CO2) to allow the cells to  adhere to the well. Then, 10 μL CCK-8 solution was mixed into each well  and the culture plate was incubated for 2 h(37 ◦C, 5% CO2). Absorbance  at 450 nm was measured using a microplate reader.

2.15. Western blotting  

At 9 dpf, WT and nox4− /− larvae were collected. Thirty larvae were  mixed with 1 mL lysis buffer (800 μL RIPA, 100 μL protease inhibitor,  100 μL phosphatase inhibitor, and 10 μL PMSF) (Service,CN)and incubated at 4 ◦C for 30 min. Then, the supernatant was removed by  centrifugation at 15000 rpm for 15 min at 4 ◦C. The larval proteins were  fractionated via sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE).Equal amount of protein samples was resolved on  a 12% SDS-PAGE gel and transferred onto a PVDF membrane (Millipore). Proteins were detected using mouse monoclonal anti-Smad2  (1:1000,Cell Signaling Technology,D43B4) and rabbit polyclonal antiphosphorylated Smad2(1:1000,Cell Signaling Technology,138D4).  Western blot analysis was performed for three times. Quantitative  analysis for western blotting was conducted using ImageJ software.  

2.16. Data analysis  

All data are expressed as the mean ± standard deviation. Student’s ttests and Student–Newman–Keuls tests were used to analyze the differences between groups. Each experiment was repeated at least three  times. P < 0.05 indicated significant differences.

3. Results  

3.1. Zebrafish Nox4 is highly conserved  

To verify that Nox4 is evolutionarily conserved, we compared the Nox4 amino acid sequence homology in humans, mice, and zebrafish.  Mouse and human nox4 were 91% homologous, while zebrafish and  human nox4 were 64% homologous. Human, mouse, and zebrafish Nox4  contains functional domains such as ferric reductase-like transmembrane components, FAD-binding domains, and NAD-binding domains. The amino acids in these functional domains are completely  conserved in humans, mice, and zebrafish. These three functional domains allow electrons to pass through the plasma membrane and  participate in various biochemical reactions in vivo (Fig. 1A–C). Using  conservative synteny analysis, we found that nox4 was located at 89.59  Mbp on chromosome 11 in humans, 86.89 Mbp on chromosome 5 in  mice, and 43.90 Mbp on chromosome 15 in zebrafish. Compared with  humans, the region upstream of nox4 in zebrafish contains the genes  ctsc, grm5b, and tyr. In mice, Ctsc, Grm5b, and Tyr are arranged downstream of nox4. In addition, the gene arrangement at the other end  of nox4 is not conserved (Fig. 1D). It showed that the surrounding genes  (CTSC,GRM5,TYR) of NOX4 in human and zebrafish have the similar  positions. These data suggest that zebrafish nox4 gene is syntenic to  human NOX4 gene.

图1  斑马鱼 nox4 高度保守

3.2. Construction of nox4− /− zebrafish  

In zebrafish, nox4 contains seven exons. The transcription start site is  located in exon 2 and the stop codon is located in exon 7. The gene  encodes 564 amino acids. The gRNA target sites in exon 4 were predicted and selected using online software (https://www.benchling.  com/crispr). Zebrafish mutants were constructed using CRISPR/Cas9  technology (Fig. 2A). We obtained two homozygous nox4 knockout mutants (Fig. 2B): the first mutation was a 4bp deletion and the other  one was a 142 bp increased (Fig. 2C and D). nox4 encodes the main  enzyme that produces ROS. To verify Nox4 depletion, 2′ ,7′ -Dichlorofluorescin diacetate(DCFH-DA) probe was used to detect the ROS content in wild-type (WT) and nox4− /− zebrafish. The fluorescence intensity  in nox4− /− zebrafish was significantly reduced compared with that of  WT zebrafish (Supplementary Figs. 1A and B). Previous studies showed  that Nox4 is involved in inflammatory processes. Using neutral red and  Sudan black staining of macrophages and neutrophils, we observed that  the number of macrophages and neutrophils in nox4− /− zebrafish was  significantly lower than that in WT zebrafish (Supplementary Figs. 1E  and F). These results show that nox4− /− zebrafish were successfully  constructed.

图2 nox4^−/−斑马鱼的构建

3.3. Bone mineralization in nox4− /− zebrafish is significantly reduced in  early development  

To study how Nox4 deletion affects early bone development, alcian  blue staining was performed 5 days postfertilization (dpf). All craniofacial cartilage and bone (Meckel’s cartilage, the ceratobranchial bone,  the ceratohyal bone, the ethmoid bone, and the palato-quadrate bone)  were clearly visible. Compared with WT zebrafish, the developmental of  the craniofacial cartilage in nox4− /− zebrafish significantly delay.  Compared with WT zebrafish, nox4− /− zebrafish exhibited significantly  lower jaw length (LJL), ceratohyal cartilage length (CCL), and intercranial distance (ICD) (Supplementary Figs. 2A and B).  Zebrafish were stained with alizarin red on 9, 15, and 30 dpf. On day  9 and 15, bone mineralization in nox4− /− zebrafish decreased significantly, especially in the parasphenoid bone, operculum, notochord,  otolith, occipital bone, and vertebrae. The formation of the parasphenoid bone and operculum occur via intramembrane osteogenesis,  while the occipital bone, notochord, and otolith develop by endochondral osteogenesis, suggesting that nox4 knockout significantly affects  osteogenesis (Fig. 3A–D). At 30 dpf, bone mineralization in nox4− /− zebrafish remained significantly lower than in control zebrafish (Fig. 3Eand F). At 6 months postfertilization, microcomputed tomography (μCT)  analysis of bone development in adult fish showed that there was no  significant difference in the bone tissue of nox4− /− zebrafish compared  with the control group. Further analysis of the bone volume fraction  (BV/TV), bone surface area to bone volume ratio (BS/BV), trabecular  thickness (TB Th), and trabecular separation (TB SP) showed no significant differences (Fig. 3G and H). These data suggest that the loss of  Nox4 dysregulates early bone development in zebrafish.

图3 nox4−/−斑马鱼的骨矿化在早期发育时受到干扰

3.4. Treatment with a Nox4 inhibitor results in early skeletal  abnormalities similar to those observed in nox4− /− zebrafish  

GlX351322, a nitrogen oxide, is a specific Nox4 inhibitor. We  investigated whether GlX351322 reduces early bone development in  zebrafish. WT zebrafish were treated with GLX351322 at 8 h postfertilization (hpf) and stained with alizarin red at 9 dpf. Compared with  the control group, mineralization in the parasphenoid bone, operculum,  notochord, otolith, and occipital bone was significantly reduced in the  GLX351322 treatment group. Whole-body mineralization in the  GLX351322 treatment group was similarly reduced (Fig. 4A and B).  Additionally, zebrafish cartilage was stained with alcian blue at 5 dpf. Compared to the control group, the staining in the GLX351322 treatment group was weaker. However, there was no significant difference in  LJL, CCL, and ICD compared to the control group (Supplementary  Figs. 3A and B). These data demonstrate that treatment with the Nox4  inhibitor GLX351322 results in a nox4− /− -like phenotype.  

sox9 and sp7 are chondrocyte- and osteoblast-specific marker genes,  respectively. sox9 plays an important role in chondrocyte differentiation  and participates in the regulation of chondrocyte differentiation. sp7 is  specifically expressed in osteoblasts.Tg(sox9-EGFP) and Tg(sp7-EGFP)  zebrafish were treated with GLX351322 at 8 hpf. At 5 dpf, the fluorescence intensity of Tg(sox9-EGFP) zebrafish treated with GLX351322 was  reduced in Meckel’s cartilage, the basihyoid bone, the ceratobranchial  bones, and the ceratohyal bone, compared to controls (Fig. 4C). The  fluorescence intensities in Tg(sox9-EGFP) zebrafish in the GLX351322  treatment group were significantly lower than those in the control group  (Fig. 4D).The number of chondrocytes was also significantly decreased,  especially in the cervical vertebrae (Supplementary Figs. 3C and D). At 7  dpf, the fluorescence intensity in Tg(sp7-EGFP) zebrafish treated with  GLX351322 also significantly decreased in the ceratobranchial bone,  branchial ray, and operculum (Fig. 4E). The fluorescence intensities in  Tg(sp7-EGFP) zebrafish in the GLX351322 treatment group were  significantly lower than those in the control group (Fig. 4F).  

Previous studies showed that Nox4 promotes osteoclast differentiation. To examine whether Nox4 deletion affects osteoclast growth and  differentiation, Tartrate resistant acid phosphatase(TRAP) staining was  performed to determine osteoclast content. Compared with the control  group, the number of osteoclasts in nox4-/- zebrafish was significantly  decreased at 9 dpf. A similar result was found in zebrafish treated with  GLX351322 (Fig. 4G and H). These data suggest that deleting Nox4 leads  to chondrocyte, osteoblast, and osteoclast dysfunction.

图4 用Nox4抑制剂GLX351322治疗可减少早期发育的骨矿化

3.5. Nox4 deletion disrupted the regeneration ability of zebrafish tail  

The zebrafish tail can regenerate through a process similar to that of  mammalian fracture repair. To study whether nox4 deletion affects tail  regeneration, the tails of adult WT and nox4-/- zebrafish were removed,  and alizarin red staining and TRAP staining were performed at 3, 6, 10,  and 13 days postamputation (dpa). Compared with the control group,  mineralization and TRAP staining of nox4− /− zebrafish tail regeneration  were significantly reduced (Fig. 5A and B). The length of zebrafish tail  regeneration at different time periods was quantified. There was no  significant difference in the regeneration length of the nox4-/- zebrafish  tail at 3 dpa, but the regenerated length in nox4-/- zebrafish was  significantly shorter at 6, 10, and 13 dpa (Fig. 5C). TRAP staining in the  zebrafish tail was analyzed at 3, 6, 10, and 13 dpa. We observed that the  osteoclast content in nox4− /− zebrafish was significantly reduced  (Fig. 5D). To investigate whether reduced regeneration in the nox4-/-  zebrafish was due to reduced osteoblasts, Tg(sp7-EGFP) zebrafish were  treated with GLX351322 after tail cutting. At 3 dpa, GLX351322 treatment significantly decreased the Tg(sp7-EGFP) fluorescence intensity  (Fig. 5E and F) and reduced the alizarin red staining area (Supplementary Fig. 4A). These results suggest that nox4 deletion disrupts regeneration in the zebrafish tail.

图5 Nox4缺失减少了尾部再生

3.6. nox4 knockout impairs TGF-β signaling  

Background ROS play an important role as signaling molecules in  many physiological processes. To verify whether the damage to early  bone development in nox4-/- zebrafish was caused by reduced ROS  levels, nox4− /− mice were treated with the ROS inducer AAPH (2,2′ -  azobis-2-methyl-propanimidamide, dihydrochloride) and alizarin red  staining was used to examine bone development. AAPH treatment  increased ROS levels but did not reduce bone mineralization caused by  nox4 deletion (Supplementary Figs. 5A and B). These data suggest that in  early development, nox4 deletion reduces bone mineralization in  zebrafish, but not by reducing ROS levels.

To further study the internal molecular mechanism of the reduction  of bone mineralization in zebrafish caused by nox4 deletion, WT and  nox4-/- zebrafish larvae were collected 9 dpf and used for transcriptome  sequencing. We detected 226 upregulated and 668 downregulated genes  in nox4-/- zebrafish compared to WT (Fig. 6A). Gene ontology (GO)  analysis of the differentially expressed genes showed that they were  mainly involved in skeletal system development, immune system regulation, signal receptor activity regulation, ossification, oxidative stress  response, and biomineral tissue development (Fig. 6B). Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of differentially expressed genes showed that the differentially expressed genes  were mainly involved in ferroptosis, apoptosis, the Wnt signaling  pathway, the TGF-β signaling pathway, the MAPK signaling pathway,  the mTOR signaling pathway, and cytokine-cytokine receptor interactions (Fig. 6C).  

Previous studies have shown the TGF-β signaling pathway to play an  important role in bone development and osteoblast differentiation. Our  transcriptome data revealed that the TGF-β signaling pathway was  disordered in nox4-/- zebrafish. Therefore, we hypothesized that the  TGF-β signaling pathway is involved in bone injury caused by nox4  deletion. We evaluated TGF-β signaling pathway-related genes in WT  and nox4− /− zebrafish using RT-qPCR at 9 dpf. There were no significant  differences in the expression of BMP-related genes in nox4− /− zebrafish  compared to the control group (Fig. 6D), but the expression of tgfb1b,  tgfbr1b, tgfbr2b, tgfbr3, and tgfbr2a was significantly downregulated. We  observed no significant differences in tgfb1a, tgfbr1a, or tgfbrap1  (Fig. 6E). These data suggest that the TGF-β signaling pathway in nox4− /  − zebrafish is dysregulated.  

Activating TGF-β signaling leads to Smad2 phosphorylation, which  causes Smad2 translocation to the nucleus to regulate gene expression.  We founded that Smad2 and p-Smad2 protein expression in nox4− /− zebrafish to be significantly higher than that in the WT (Fig. 6F and G).  However, there was no significant change in Smad1/5/9 and p-Smad1/  5/9 expression (Supplementary Figs. 6A and B). These data support the  notion that TGF-β signaling is disrupted in nox4− /− zebrafish.

图6 nox4^−/−斑马鱼中TGF-β信号传导异常

3.7. TGF-β pathway inhibitors can rescue nox4 deletion-mediated bone  dysplasia in early development  

We used SB431542, an effective and selective TGF-βRI(ALK5) inhibitor, to verify whether the bone injury caused by nox4 deletion occurs  through the TGF-β signaling pathway. nox4− /− zebrafish were treated  with SB431542 during early development. At 5 dpf, nox4− /− zebrafish  were treated with SB431542 and then stained with alizarin red at 9 dpf.  Compared with WT zebrafish, bone mineralization in the parasphenoid  bone, operculum, notochord, otolith, and occipital bone in nox4− /− zebrafish was rescued after SB431542 treatment (Fig. 7A). Alizarin red  staining revealed that bone mineralization in nox4− /− zebrafish significantly increased after SB431542 treatment and exceeded the normal  level (Fig. 7B). Then, cartilage in nox4− /− zebrafish treated with  SB431542 was stained with alcian blue at 5 dpf. Compared with WT  zebrafish, the LJL, CCL, and ICD of nox4− /− zebrafish treated with  SB431542 restored to normal (Supplementary Figs. 7A and B).

At 5 dpf, Tg(sp7-EGFP) zebrafish were treated with GLX351322 and  SB431542. The intensity of Tg(sp7-EGFP) fluorescence in the  GLX351322 and SB431542 co-treatment groups significantly recovered  at 8 dpf. Quantitative analysis of the fluorescence intensity showed that  the osteoblast content significantly increased and exceeded normal  levels (Fig. 7C and D). Additionally, nox4− /− zebrafish were treated with  SB431542 at 5 dpf. TRAP staining analysis showed that at 9 dpf, the  stained area in nox4− /− zebrafish in SB431542 treatment group was not  rescued and the osteoclast content significantly decreased (Fig. 7E and  F). These results show that blocking TGF-β signaling reduces the damage  to chondrocytes and osteoblasts caused by nox4 deletion in early  development, but it cannot reduce the damage to osteoclasts.

图7 阻断TGF-β信号通路可挽救早期发育中Nox4丢失引起的骨发育不良

3.8. NOX4 inhibitors damage Saos-2 cell proliferation and differentiation  

To determine whether nox4 plays a similar role in zebrafish and  human osteoblasts, we examined mineralization in Saos-2 cells. Ascorbic acid and β-sodium glycerophosphate are required for complete Saos-  2 cell mineralization. After culturing Saos-2 cells for 12 days, alizarin  red staining showed that mineralization of Saos-2 cells treated with  ascorbic acid and β-sodium glycerophosphate increased significantly.  However, GLX351322 treatment significantly reduced Saos-2 cell  mineralization, even when treated with ascorbic acid and β-sodium  glycerophosphate (Fig. 8A). Based on our previous results, we speculated that the proliferation of Saos-2 cells treated with GLX351322  would be impaired. CCK-8 assays were used to detect the proliferation of  Saos-2 cells treated with GLX351322. Saos-2 cell proliferation was  significantly inhibited by GLX351322 (Supplementary Fig. 8A). Next,  we examined the expression of genes related to early, middle, and late osteoblast differentiation to determine which stage of osteoblast differentiation is affected by nox4 deletion. The cells treated with ascorbic  acid and β-sodium glycerophosphate were treated with GLX351322 on  days 1–4 (early differentiation), 5–8 (middle stage of differentiation),  and 9–12 (late differentiation). Alizarin red staining was performed 12 h  after treatment. Osteogenic mineralization was the smallest when  GLX351322 was added during early mineralization, though osteogenic  mineralization was also significantly reduced in middle and late osteoblast differentiation (Fig. 8B). RT-qPCR was used to analyze gene  expression at each osteoblast differentiation stage. Compared with the  control group, Osterix(OSX) expression in the inducer treatment group  increased significantly during early differentiation; OCN expression  increased significantly during the middle stage of differentiation; and  RUNX2, OSX, and OCN expression increased significantly during late  differentiation. However, RUNX2, OSX, and OCN expression was  significantly downregulated during all differentiation stages after treated by GLX351322 (Fig. 8C). When the TGF-β signaling pathway was  blocked with SB431542, mineralization in Saos-2 cells treated with  inducer, GLX351322, and SB431542 increased significantly and  returned to the normal level compared with cells treated with inducer  and GLX351322 (Fig. 8D). These results suggest that the loss of nox4  impairs early osteoblast differentiation. Further, nox4 may play a role in  every stage of osteoblast differentiation. Finally, nox4 affects osteoblast  differentiation through the TGF-β signaling pathway.

图8 GLX351322处理会损害Saos-2细胞的增殖和分化

4. Discussion  

In this study, we used zebrafish as a model to study the effect of nox4  deletion on early bone development. nox4− /− zebrafish were constructed using CRISPR/Cas9. Mineralization in nox4− /− zebrafish  decreased and tail regeneration was significantly disrupted in the early  developmental stage. Transcriptome sequencing revealed that the TGF-β  pathway in nox4− /− zebrafish was significantly disrupted, Smad2  phosphorylation increased, and the expression of the osteoblast  differentiation-related genes RUNX2, OSX, and OCN was reduced.  Additionally, TGF-β signaling blockade ameliorated the abnormal bone development caused by nox4 loss and increased the number of osteoblasts (Fig. 9). However, this had no effect on the number of osteoclasts.  In Saos-2 cells, we elucidated that the loss of nox4 during early development lead to impaired osteogenic differentiation and abnormal bone  development. These findings provide a theoretical basis underlying the  bone metabolic diseases caused by nox4 deficiency in early bone  development.

There is a close relationship between ROS and bone homeostasis.  Redox imbalances and oxidative stress are important causes of tissue  damage. However, ROS also act as cell messengers and play an important role in cell signal transduction [32]. The expression of nox4 in  undifferentiated Vascular Smooth Muscle Cells(VSMCs) is significantly  higher than that in differentiated cells [33]. Further, nox4 and its  superoxidation products(hydrogen peroxides,H2O2) plays a role in long-term adaptive signaling and cell differentiation [34]. Many factors  can indirectly increase cellular ROS content by increasing nox4  expression. This is also a reasons for osteoblast differentiation injury. In  our earlier studies, we found that iron overload can lead to osteoporosis  by inducing osteoblast death related to ROS, which also indicates that  excessive ROS is an adverse factor for osteogenesis [7]. ROS at baseline  levels act as intracellular signaling molecules for cell proliferation and  differentiation. In this study, we found that bone mineralization in  nox4− /− zebrafish was decreased in the early developmental stage.  Treatment with the ROS inducer AAPH increased ROS levels but did not  reduce the decreased bone mineralization observed in nox4− /− zebrafish  in the early developmental stage, indicating that the bone damage  caused by the loss of nox4 may not be caused by low ROS levels, but  rather by the compartmentalized superoxide product of nox4.

Osteoclasts play a role in shaping the bones during development. The  main reason for osteoporosis is that increased osteoclast activity in bone  tissue causes decreased bone mass. ROS are one factor that induce  osteoclast differentiation. Other factors such as heavy metals and  dexamethasone increase the ROS content. Increasing ROS levels induces  expression of Rankl, which is a key factor in osteoblast differentiation  [35–37]. Peroxisome proliferator-activated receptor gamma coactivator  1-beta(PGC-1β) promotes osteoclast differentiation by increasing the  number of mitochondria and ROS [38]. NOX4 is the main enzyme that  catalyzes ROS production. Nox4 promotes osteoclast formation and affects bone mass in mice(11). Our results are similar to those of previous  studies showing that the number of osteoclasts in nox4− /− zebrafish was  significantly reduced. Similarly, treatment with the Nox4 inhibitor  GLX351322 led to a significant reduction in the number of osteoclasts  during early development. Su et al. first clarified that nox4 expression is  upregulated in mouse osteoclasts and induces osteoclast formation [21].  Goettsch et al. also found that the number of osteoclasts in nox4− /− mice  decreases significantly. In ovariectomized mice, Nox4 expression was  significantly upregulated, resulting in bone loss [11]. Although upregulated nox4 expression leads to increased osteoclast differentiation, it is  not known whether Nox4 promotes osteoclast differentiation directly by  increasing ROS levels or through other mechanisms.

Zebrafish are vertebrates that develop in vitro. Thus, they allow a  convenient way to study the role of genes in early development.  Zebrafish have recently been used to study early bone development. Sun  et al. found that Rmrp(RNA component of mitochondrial RNA processing  endoribonuclease) mutation disrupts chondrogenesis and bone ossification in a zebrafish model of cartilage-hair hypoplasia via enhanced  Wnt/β-catenin signaling [39]. Fgfr3(Fibroblast growth factor receptor 3)  mutation disrupts chondrogenesis and bone ossification in a zebrafish  model that mimics Camptodactyly-tall stature-scoliosis-hearing loss  syndrome(CATSHL) partially via enhanced Wnt/β-catenin signaling  [40]. In this study, we found that loss of Nox4 decreases osteoblast  differentiation and bone mineralization, indicating that Nox4 also plays  an important role in early bone development. The loss of Nox4 results in  severely reduced bone mineralization in the parasphenoid bone, operculum, notochord, occipital bone, and vertebrae. The parasphenoid and  operculum are formed via intramembranous osteogenesis, while the  notochord, occipital bone, and vertebrae develop by endochondral  osteogenesis. Thus, our findings suggest that reduced Nox4 activity is  involved in reduced osteoblast differentiation and bone mineralization.  Additionally, our results indicate that Nox4 is present in bones during  the early stages of development and plays an important role in development. Similarly, Nox4 expression promotes osteoblast differentiation  in mice. During the development of newborn mice, the femur undergoes  endochondral osteogenesis. Nox4 is closely related to the formation of  the bone matrix, which is mainly synthesized by osteoblasts. However,  in adult mice, Nox4 loss has no effect on osteoblast activity [11]. These  results show that Nox4 is an essential factor for normal bone  development.

Unlike in humans, almost all tissues and organs of the zebrafish can  regenerate. Tail regeneration has been widely studied in zebrafish, as  tail regeneration is similar to mammalian fracture repair [41]. nox4  deletion led to regeneration injury in the zebrafish tail, with a corresponding decrease in the number of osteoclasts and osteoblasts. The role  of Nox4 in fracture repair has not been reported, but a study in rats  showed that Nox and ROS are produced in large quantities at the healing  site, which hinders fracture healing in a rat forelimb fracture model  [42]. We inhibited Nox4 during tail regeneration, resulting in reduced  osteoblasts, which may explain the slow regeneration we observed.

Knockout of nox4 reduced basal ROS levels, but exogenous induction  of ROS cannot ameliorate the bone phenotype caused by Nox4 deletion,  indicating that the bone injury caused by Nox4 deletion is not caused by  dysregulated oxidative stress-related signaling. To explore the underlying mechanism, we analyzed the transcriptome data of WT and nox4− /  − zebrafish in early development and found that genes associated with  TGF-β signaling pathway. The TGF-β signaling pathway is a classical  signaling pathway that promotes osteoblast proliferation and differentiation. The classic TGF-β signaling pathway includes the BMP signaling  pathway, but we observed no significant difference in the expression of  genes related to BMP signaling in nox4− /− zebrafish. However, the  expression of genes encoding TGF-β-mediated cytokines and receptor  changed significantly, indicating that the TGF-β signaling pathway was  disrupted. Activation of TGF-β receptors regulates Smad2/3 phosphorylation, which increased significantly in nox4− /− zebrafish. Nox4  regulation of the TGF-β signaling pathway activity has been reported in  previous studies. For example, TGF-β promotes epithelial-mesenchymal  transitions via NF-κB/NOX4/ROS signaling in lung cancer cells [45]. In  another study, TGF-β-induced tissue fibrosis was abrogated by a global  genetic deletion of Nox4 [46]. These results indicate that the TGF-β  signaling pathway is downstream of Nox4. In recently, two report shows  that NOX4 regulated the osteoblats though the MAPK signaling pathway  [43,44]. The MAPK signaling pathway was enriched in our RNA-seq by  DEGs, Those results shows that NOX4 regulated osteoblast maybe via  differently pathway to produce “cross-talk”.

Additionally, TGF-β signaling pathway inhibits RUNX2 expression  though recruitment of class II histone deacetylases by enhancing  Smad2/3 activity [47]. RUNX2 is a transcription factor that promotes osteogenic differentiation. However, recent studies showed that TGF-β  signaling also regulates the expression of many bone metabolism-related  genes, including many osteoclast genes [22]. SB431542 is an effective  and selective TGF-βRI(ALK5) inhibitor, that inhibits ALK5 as well as the  activin type I receptor ALK4 and the nodal type I receptor ALK7, which  share kinase domains with ALK5, also it efficiently inhibits Smad  phosphorylation induced by TGF-β but not BMP4 [48]. We observed that  TGF-β signaling pathway inhibitors reduce damage to osteoblasts, but  not osteoclasts, indicating that Nox4 may regulate osteoclasts through  another mechanism.

Loss of Nox4 by the TGF-β signaling pathway affects bone dysplasia.  Saos-2 cells are osteosarcoma cell lines that are often used to study  osteoblast differentiation. Nox4 inhibitors inhibit the differentiation of  Saos-2 cells into osteoblasts, significantly reducing the expression of  osteoblast differentiation genes. However, when TGF-β signaling is  blocked, the differentiation of Saos-2 cells into osteoblasts rescues. In a  study by Bai et al., Nox4 expression promoted apoptosis by activating  the MAPK pathway, while MAPK-derived ROS mediated osteoblast  apoptosis by increasing phosphorylated ASK1 and p38 levels [49].  TGF-β signaling plays an important role in cell proliferation. Nox4 inhibitor reduced Saos-2 cell proliferation. However, it is unknown  whether the reduced osteoblast proliferation leads to impaired osteogenic mineralization during early zebrafish development. In this study,  Saos-2 cells were treated with the inhibitor of TGF-β at three developmental stages, which inhibited osteogenic differentiation. However, the  phenomenon was more pronounced in early differentiation, which also  indicates that Nox4 may play a larger role in earlier developmental  stages.One of Chen’s latest studies confirms our view, the early osteoprogenitors and skeletal development are dependent on Nox4’s activity  [50].  

5. Conclusion  

In this study, we found that Nox4 is highly conserved in zebrafish.  Using the CRISPR/Cas9 system, nox4− /− zebrafish were successfully  generated. We found that in the early development stage in nox4− /− zebrafish, bone mineralization decreased, the number of osteoblasts and  osteoclasts decreased, and tail regeneration was damaged. Treatment  with the Nox4 inhibitor GLX351322 resulted in bone dysplasia similar to  that observed in nox4− /− zebrafish. Transcriptome sequencing revealed  that the TGF-β signaling pathway was dysregulated in nox4− /− zebrafish. Blocking TGF-β signaling reduced bone dysplasia caused by nox4  deletion in early development but had no effect on the number of osteoclasts. Saos-2 cells were used to assess the role of Nox4 in human  osteoblasts. The findings of the current study show that nox4 is involved  in osteoblast growth and differentiation during early development.  Finally, Nox4 may be a target to treat early stages of bone metabolic  diseases.  

Funding  

This work is supported by: National Key R&D Program of China  (2021YFC2501702); Natural Science Foundation of China (82072474);  Gusu Health Talent Program (GSWS2020024); Geriatrics Clinical  Technology Application Research Project (LR2021022); Jiangsu Province Applied Engineering Research Center of Physical And Medical  Fusion Promoting Bone Health In The Elderly Natural Science Foundation of China (SYS2018050).

Author contributions  

Authors’roles: Study design: ZC, GL,MW and YX. Technical and  material support:HZ. Data analysis and interpretation: ZC,GL and MW.  Drafting and revising manuscript:ZC,GL,MW and YX. ZC,GL and MW  take responsibility for the integrity of the data analysis.All authors  approved the final version of the manuscript.

Data and materials availability  

Additional supporting information may be found online in the Supporting Information section.

Declaration of competing interest  

The authors have declared no conflicts of interest.

Acknowledgments  

We thank Zhang Wen from Institute of Orthopaedics at Soochow  University for his help of μCT, Zhang Ruizhi,Zhu Keyu,Jin Xingyu,Li  Guangfei,Wei Qi and Zheng Miao from the Department of Orthopaedics,  the Second Affiliated Hospital of Soochow University,for helpful  discussion.

Appendix A. Supplementary data  

Supplementary data to this article can be found online at https://doi.org/10.1016/j.freeradbiomed.2022.11.016.

References  

[1] F. Long, D.M. Ornitz, Development of the endochondral skeleton, Cold Spring  Harbor Perspect. Biol. 5 (2013) a008334.  [2] W.R. Atchley, B.K. Hall, A model for development and evolution of complex  morphological structures, Biol. Rev. 66 (1991) 101–157.  [3] Y. Yahara, T. Barrientos, Y.J. Tang, V. Puviindran, P. Nadesan, H. Zhang, J. Gibson,  S.G. Gregory, Y. Diao, Y. Xiang, Y.J. Qadri, T. Souma, M.L. Shinohara, B.A. Alman,  Erythromyeloid progenitors give rise to a population of osteoclasts contributing to  bone homeostasis and repair, Nat. Cell Biol. 22 (2020) 49–59.  [4] J.-M. Kim, C. Lin, Z. Stavre, M.B. Greenblatt, J.-H. Shim, Osteoblast-osteoclast  communication and bone homeostasis, Cells 9 (2020) 2073.  [5] X. Chen, Z. Wang, N. Duan, G. Zhu, E.M. Schwarz, C. Xie, Osteoblast-osteoclast  interactions, Connect. Tissue Res. 59 (2018) 99–107.  [6] L.A. Fitzpatrick, Secondary causes of osteoporosis, Mayo Clin. Proc. 77 (2002)  453–468.  [7] P. Jia, Y.J. Xu, Z.L. Zhang, K. Li, B. Li, W. Zhang, H. Yang, Ferric ion could facilitate  osteoclast differentiation and bone resorption through the production of reactive  oxygen species, J. Orthop. Res. 30 (2012) 1843–1852.  [8] B. Chance, H. Sies, A. Boveris, Hydroperoxide metabolism in mammalian organs,  Physiol. Rev. 59 (1979) 527–605.  [9] D.B. Zorov, M. Juhaszova, S.J. Sollott, Mitochondrial reactive oxygen species  (ROS) and ROS-induced ROS release, Physiol. Rev. 94 (2014) 909–950.  [10] K. Schr¨oder, NADPH oxidases in bone homeostasis and osteoporosis, Cell. Mol. Life  Sci. 72 (2015) 25–38.  [11] C. Goettsch, A. Babelova, O. Trummer, R.G. Erben, M. Rauner, S. Rammelt,  N. Weissmann, V. Weinberger, S. Benkhoff, M. Kampschulte, B. ObermayerPietsch, L.C. Hofbauer, R.P. Brandes, K. Schr¨ oder, NADPH oxidase 4 limits bone  mass by promoting osteoclastogenesis, J. Clin. Invest. 123 (2013) 4731–4738.  [12] J. Tan, X. Xu, Z. Tong, J. lin, Q. Yu, Y. Lin, W. Kuang, Decreased osteogenesis of  adult mesenchymal stem cells by reactive oxygen species under cyclic stretch: a  possible mechanism of age related osteoporosis, Bone Res. 3 (2015), 15003.  [13] K. Schr¨oder, NADPH oxidases in bone homeostasis and osteoporosis, Free Radic.  Biol. Med. 132 (2019) 67–72.  [14] X. Tang, J.J. Wang, J. Wang, H.E. Abboud, Y. Chen, S.X. Zhang, Endotheliumspecific deletion of Nox4 delays retinal vascular development and mitigates  pathological angiogenesis, Angiogenesis 24 (2021) 363–377.  [15] K. M. O’Neill, D. C. Campbell, K. S. Edgar, E. K. Gill, A. Moez, K. J. McLoughlin, C.  L. O’Neill, M. Dellett, C. J. Hargey, R. A. Abudalo, M. O’Hare, P. Doyle, T. Toh, J.  Khoo, J. Wong, C. M. McCrudden, M. Meloni, C. Brunssen, H. Morawietz, M. C.  Yoder, C. J. Watson, A. W. Stitt, A. Margariti, R. J. Medina, D. J. Grieve, NOX4 IS A  MAJOR REGULATOR OF CORD BLOOD-DERIVED ENDOTHELIAL COLONYFORMING CELLS WHICH PROMOTES POSTISCHAEMIC REVASCULARISATION,  vol. 28.  [16] M. Mu˜noz, M.E. L´opez-Oliva, C. Rodríguez, M.P. Martínez, J. S´ aenz-Medina,  A. S´ anchez, B. Climent, S. Benedito, A. García-Sacrist´ an, L. Rivera, M. Hern´ andez,  D. Prieto, Differential contribution of Nox1, Nox2 and Nox4 to kidney vascular  oxidative stress and endothelial dysfunction in obesity, Redox Biol. 28 (2019),  101330.  [17] R.D. Rajaram, R. Dissard, A. Faivre, F. Ino, V. Delitsikou, V. Jaquet, T. Cagarelli,  M. Lindenmeyer, P. Jansen-Duerr, C. Cohen, S. Moll, S. de Seigneux, Tubular NOX4  expression decreases in chronic kidney disease but does not modify fibrosis  evolution, Redox Biol. 26 (2019), 101234.  [18] J. Sun, W. Chen, S. Li, S. Yang, Y. Zhang, X. Hu, H. Qiu, J. Wu, S. Xu, T. Chu, Nox4  promotes RANKL-induced autophagy and osteoclastogenesis via activating ROS/  PERK/eIF-2α/ATF4 pathway, Front. Pharmacol. 12 (2021), 751845.[19] C.C. Mandal, S. Ganapathy, Y. Gorin, K. Mahadev, K. Block, H.E. Abboud, S.  E. Harris, G. Ghosh-Choudhury, N. Ghosh-Choudhury, Reactive oxygen species  derived from Nox4 mediate BMP2 gene transcription and osteoblast  differentiation, Biochem. J. 433 (2011) 393–402.  [20] V. Helfinger, F. Freiherr von Gall, N. Henke, M.M. Kunze, T. Schmid, F. Rezende,  J. Heidler, I. Wittig, H.H. Radeke, V. Marschall, K. Anderson, A.M. Shah, S. Fulda,  B. Brüne, R.P. Brandes, K. Schr¨ oder, Genetic deletion of Nox4 enhances  cancerogen-induced formation of solid tumors, Proc. Natl. Acad. Sci. U. S. A. 118  (2021), e2020152118.  [21] S. Yang, P. Madyastha, S. Bingel, W. Ries, L. Key, A new superoxide-generating  oxidase in murine osteoclasts, J. Biol. Chem. 276 (2001) 5452–5458.  [22] I. Grafe, S. Alexander, J.R. Peterson, T.N. Snider, B. Levi, B. Lee, Y. Mishina, TGF-В  family signaling in mesenchymal differentiation, Cold Spring Harbor Perspect.  Biol. 10 (2018) a022202.  [23] A. Chaikuad, A.N. Bullock, Structural basis of intracellular TGF-β signaling:  receptors and smads, Cold Spring Harbor Perspect. Biol. 8 (2016) a022111.  [24] A. Hata, Y.-G. Chen, TGF-В signaling from receptors to smads, Cold Spring Harbor  Perspect. Biol. 8 (2016) a022061.  [25] P. Xu, X. Lin, X.-H. Feng, Posttranslational regulation of smads, Cold Spring Harbor  Perspect. Biol. 8 (2016) a022087.  [26] C.S. Hill, Transcriptional control by the SMADs, Cold Spring Harbor Perspect. Biol.  8 (2016) a022079.  [27] T. Komori, Functions of the osteocyte network in the regulation of bone mass, Cell  Tissue Res. 352 (2013) 191–198.  [28] N.A. Sims, C. Vrahnas, Regulation of cortical and trabecular bone mass by  communication between osteoblasts, osteocytes and osteoclasts, Arch. Biochem.  Biophys. 561 (2014) 22–28.  [29] K. Nakashima, X. Zhou, G. Kunkel, Z. Zhang, J.M. Deng, R.R. Behringer, B. de  Crombrugghe, The novel zinc finger-containing transcription factor osterix is  required for osteoblast differentiation and bone formation, Cell 108 (2002) 17–29.  [30] I.J. Marques, E. Lupi, N. Mercader, Model systems for regeneration: zebrafish,  Development 146 (2019) dev167692.  [31] K. Dietrich, I.A. Fiedler, A. Kurzyukova, A.C. L´opez-Delgado, L.M. McGowan,  K. Geurtzen, C.L. Hammond, B. Busse, F. Knopf, Skeletal biology and disease  modeling in zebrafish, J. Bone Miner. Res. 36 (2021) 436–458.  [32] G. Leonarduzzi, B. Sottero, G. Poli, Targeting tissue oxidative damage by means of  cell signaling modulators: the antioxidant concept revisited, Pharmacol. Ther. 128  (2010) 336–374.  [33] R.E. Clempus, D. Sorescu, A.E. Dikalova, L. Pounkova, P. Jo, G.P. Sorescu,  B. Lass`egue, K.K. Griendling, Nox4 is required for maintenance of the differentiated  vascular Smooth Muscle cell phenotype, Arterioscler. Thromb. Vasc. Biol. 27  (2007) 42–48.  [34] S. Guo, X. Chen, The human Nox4: gene, structure, physiological function and  pathological significance, J. Drug Target. 23 (2015) 888–896.  [35] M.M. Brz´oska, J. Rogalska, Protective effect of zinc supplementation against  cadmium-induced oxidative stress and the RANK/RANKL/OPG system imbalance  in the bone tissue of rats, Toxicol. Appl. Pharmacol. 272 (2013) 208–220.  [36] Y. Zhang, X. Liu, Y. Li, M. Song, Y. Li, A. Yang, Y. Zhang, D. Wang, M. Hu, AUcubin  slows the development of osteoporosis by inhibiting osteoclast differentiation via  the nuclear factor erythroid 2-related factor 2-mediated antioxidation pathway.  Pharm. Biol.. 59, 1556–1565. [19] C.C. Mandal, S. Ganapathy, Y. Gorin, K. Mahadev, K. Block, H.E. Abboud, S.  E. Harris, G. Ghosh-Choudhury, N. Ghosh-Choudhury, Reactive oxygen species  derived from Nox4 mediate BMP2 gene transcription and osteoblast  differentiation, Biochem. J. 433 (2011) 393–402.  [20] V. Helfinger, F. Freiherr von Gall, N. Henke, M.M. Kunze, T. Schmid, F. Rezende,  J. Heidler, I. Wittig, H.H. Radeke, V. Marschall, K. Anderson, A.M. Shah, S. Fulda,  B. Brüne, R.P. Brandes, K. Schr¨ oder, Genetic deletion of Nox4 enhances  cancerogen-induced formation of solid tumors, Proc. Natl. Acad. Sci. U. S. A. 118  (2021), e2020152118.  [21] S. Yang, P. Madyastha, S. Bingel, W. Ries, L. Key, A new superoxide-generating  oxidase in murine osteoclasts, J. Biol. Chem. 276 (2001) 5452–5458.  [22] I. Grafe, S. Alexander, J.R. Peterson, T.N. Snider, B. Levi, B. Lee, Y. Mishina, TGF-В  family signaling in mesenchymal differentiation, Cold Spring Harbor Perspect.  Biol. 10 (2018) a022202.  [23] A. Chaikuad, A.N. Bullock, Structural basis of intracellular TGF-β signaling:  receptors and smads, Cold Spring Harbor Perspect. Biol. 8 (2016) a022111.  [24] A. Hata, Y.-G. Chen, TGF-В signaling from receptors to smads, Cold Spring Harbor  Perspect. Biol. 8 (2016) a022061.  [25] P. Xu, X. Lin, X.-H. Feng, Posttranslational regulation of smads, Cold Spring Harbor  Perspect. Biol. 8 (2016) a022087.  [26] C.S. Hill, Transcriptional control by the SMADs, Cold Spring Harbor Perspect. Biol.  8 (2016) a022079.  [27] T. Komori, Functions of the osteocyte network in the regulation of bone mass, Cell  Tissue Res. 352 (2013) 191–198.  [28] N.A. Sims, C. Vrahnas, Regulation of cortical and trabecular bone mass by  communication between osteoblasts, osteocytes and osteoclasts, Arch. Biochem.  Biophys. 561 (2014) 22–28.  [29] K. Nakashima, X. Zhou, G. Kunkel, Z. Zhang, J.M. Deng, R.R. Behringer, B. de  Crombrugghe, The novel zinc finger-containing transcription factor osterix is  required for osteoblast differentiation and bone formation, Cell 108 (2002) 17–29.  [30] I.J. Marques, E. Lupi, N. Mercader, Model systems for regeneration: zebrafish,  Development 146 (2019) dev167692.  [31] K. Dietrich, I.A. Fiedler, A. Kurzyukova, A.C. L´opez-Delgado, L.M. McGowan,  K. Geurtzen, C.L. Hammond, B. Busse, F. Knopf, Skeletal biology and disease  modeling in zebrafish, J. Bone Miner. Res. 36 (2021) 436–458.  [32] G. Leonarduzzi, B. Sottero, G. Poli, Targeting tissue oxidative damage by means of  cell signaling modulators: the antioxidant concept revisited, Pharmacol. Ther. 128  (2010) 336–374.  [33] R.E. Clempus, D. Sorescu, A.E. Dikalova, L. Pounkova, P. Jo, G.P. Sorescu,  B. Lass`egue, K.K. Griendling, Nox4 is required for maintenance of the differentiated  vascular Smooth Muscle cell phenotype, Arterioscler. Thromb. Vasc. Biol. 27  (2007) 42–48.  [34] S. Guo, X. Chen, The human Nox4: gene, structure, physiological function and  pathological significance, J. Drug Target. 23 (2015) 888–896.  [35] M.M. Brz´oska, J. Rogalska, Protective effect of zinc supplementation against  cadmium-induced oxidative stress and the RANK/RANKL/OPG system imbalance  in the bone tissue of rats, Toxicol. Appl. Pharmacol. 272 (2013) 208–220.  [36] Y. Zhang, X. Liu, Y. Li, M. Song, Y. Li, A. Yang, Y. Zhang, D. Wang, M. Hu, AUcubin  slows the development of osteoporosis by inhibiting osteoclast differentiation via  the nuclear factor erythroid 2-related factor 2-mediated antioxidation pathway.  Pharm. Biol.. 59, 1556–1565. [37] L.-Z. Ding, X. Teng, Z.-B. Zhang, C.-J. Zheng, S.-H. Chen, Mangiferin inhibits  apoptosis and oxidative stress via BMP2/Smad-1 signaling in dexamethasoneinduced MC3T3-E1 cells, Int. J. Mol. Med. 41 (2018) 2517–2526.  [38] H. Zhang, A. Wang, G. Shen, X. Wang, G. Liu, F. Yang, B. Chen, M. Wang, Y. Xu,  Hepcidin-induced reduction in iron content and PGC-1β expression negatively  regulates osteoclast differentiation to play a protective role in postmenopausal  osteoporosis, Aging 13 (2021) 11296–11314.  [39] X. Sun, R. Zhang, M. Liu, H. Chen, L. Chen, F. Luo, D. Zhang, J. Huang, F. Li, Z. Ni,  H. Qi, N. Su, M. Jin, J. Yang, Q. Tan, X. Du, B. Chen, H. Huang, S. Chen, L. Yin,  X. Xu, C. Deng, L. Luo, Y. Xie, L. Chen, Rmrp mutation disrupts chondrogenesis and  bone ossification in zebrafish model of cartilage-hair hypoplasia via enhanced  Wnt/β-catenin signaling, J. Bone Miner. Res. 34 (2019) 2101–2116.  [40] X. Sun, R. Zhang, H. Chen, X. Du, S. Chen, J. Huang, M. Liu, M. Xu, F. Luo, M. Jin,  N. Su, H. Qi, J. Yang, Q. Tan, D. Zhang, Z. Ni, S. Liang, B. Zhang, D. Chen, X. Zhang,  L. Luo, L. Chen, Y. Xie, Fgfr3 mutation disrupts chondrogenesis and bone  ossification in zebrafish model mimicking CATSHL syndrome partially via  enhanced Wnt/β-catenin signaling, Theranostics 10 (2020) 7111–7130.  [41] S.P. Singh, J.E. Holdway, K.D. Poss, Regeneration of amputated zebrafish fin rays  from de novo osteoblasts, Dev. Cell 22 (2012) 879–886.  [42] H. Yeler, F. Tahtabas, F. Candan, Investigation of oxidative stress during fracture  healing in the rats, Cell Biochem. Funct. 23 (2005) 137–139.  [43] J.-R. Chen, K. Shankar, S. Nagarajan, T.M. Badger, M.J.J. Ronis, Protective effects  of estradiol on ethanol-induced bone loss involve inhibition of reactive oxygen  species generation in osteoblasts and downstream activation of the extracellular  signal-regulated kinase/signal transducer and activator of transcription 3/receptor  activator of nuclear factor-kappaB ligand signaling cascade, J. Pharmacol. Exp.  Therapeut. 324 (2008) 50–59.  [44] J. Watt, A.W. Alund, C.F. Pulliam, K.E. Mercer, L.J. Suva, J.-R. Chen, M.J.J. Ronis,  NOX4 deletion in male mice exacerbates the effect of ethanol on trabecular bone  and osteoblastogenesis, J. Pharmacol. Exp. Therapeut. 366 (2018) 46–57.  [45] M. Ma, F. Shi, R. Zhai, H. Wang, K. Li, C. Xu, W. Yao, F. Zhou, TGF-β promote  epithelial-mesenchymal transition via NF-κB/NOX4/ROS signal pathway in lung  cancer cells, Mol. Biol. Rep. 48 (2021) 2365–2375.  [46] P.J. Wermuth, F.A. Mendoza, S.A. Jimenez, Abrogation of transforming growth  factor-β-induced tissue fibrosis in mice with a global genetic deletion of Nox4, Lab.  Invest. 99 (2019) 470–482.  [47] J.S. Kang, T. Alliston, R. Delston, R. Derynck, Repression of Runx2 function by  TGF-β through recruitment of class II histone deacetylases by Smad3, EMBO J. 24  (2005) 2543–2555.  [48] G.J. Inman, F.J. Nicol´ as, J.F. Callahan, J.D. Harling, L.M. Gaster, A.D. Reith, N.  J. Laping, C.S. Hill, SB-431542 is a potent and specific inhibitor of transforming  growth factor-beta superfamily type I activin receptor-like kinase (ALK) receptors  ALK4, ALK5, and ALK7, Mol. Pharmacol. 62 (2002) 65–74.  [49] S.-C. Bai, Q. Xu, H. Li, Y.-F. Qin, L.-C. Song, C.-G. Wang, W.-H. Cui, Z. Zheng, D.-  W. Yan, Z.-J. Li, D. Li, X. Wan, H.-F. Zhang, NADPH oxidase isoforms are involved  in glucocorticoid-induced preosteoblast apoptosis, Oxid. Med. Cell. Longev. 2019  (2019), 9192413.  [50] J.-R. Chen, O.P. Lazarenko, M.L. Blackburn, J.F. Chen, C.E. Randolph, J. Zabaleta,  K. Schroder, K.B. Pedersen, M.J.J. Ronis, Nox4 expression in osteo-progenitors  controls bone development in mice during early life, Commun. Biol. 5 (2022) 583.