Pyrroloquinoline quinone alleviates natural aging‐related osteoporosis via a novel MCM3‐Keap1‐Nrf2 axis‐mediated stress response and Fbn1 upregulation

Abstract Age‐related osteoporosis is associated with increased oxidative stress and cellular senescence. Pyrroloquinoline quinone (PQQ) is a water‐soluble vitamin‐like compound that has strong antioxidant capacity; however, the effect and underlying mechanism of PQQ on aging‐related osteoporosis remain unclear. The purpose of this study was to investigate whether dietary PQQ supplementation can prevent osteoporosis caused by natural aging, and the potential mechanism underlying PQQ antioxidant activity. Here, we found that when 6‐month‐old or 12‐month‐old wild‐type mice were supplemented with PQQ for 12 months or 6 months, respectively, PQQ could prevent age‐related osteoporosis in mice by inhibiting osteoclastic bone resorption and stimulating osteoblastic bone formation. Mechanistically, pharmmapper screening and molecular docking studies revealed that PQQ appears to bind to MCM3 and reduces its ubiquitination‐mediated degradation; stabilized MCM3 then competes with Nrf2 for binding to Keap1, thus activating Nrf2‐antioxidant response element (ARE) signaling. PQQ‐induced Nrf2 activation inhibited bone resorption through increasing stress response capacity and transcriptionally upregulating fibrillin‐1 (Fbn1), thus reducing Rankl production in osteoblast‐lineage cells and decreasing osteoclast activation; as well, bone formation was stimulated by inhibiting osteoblastic DNA damage and osteocyte senescence. Furthermore, Nrf2 knockout significantly blunted the inhibitory effects of PQQ on oxidative stress, on increased osteoclast activity and on the development of aging‐related osteoporosis. This study reveals the underlying mechanism of PQQ's strong antioxidant capacity and provides evidence for PQQ as a potential agent for clinical prevention and treatment of natural aging‐induced osteoporosis.


| INTRODUC TI ON
Aging has been recognized as an important risk factor for many chronic diseases, including osteoporosis. Cellular senescence, a fundamental aging mechanism implicated in multiple age-related disorders, also drives bone loss with aging (Farr et al., 2017). Indeed,

Farr and colleagues have reported that senescent cells accumulate
in the bone microenvironment with aging (Farr et al., 2016), and selectively targeting senescent cells and their senescence-associated secretory phenotype (SASP) could prevent age-related bone loss (Farr et al., 2017). The increase in the production of ROS (Reactive Oxygen Species) with aging has long been considered a driver of cellular senescence and aging (López-Otín et al., 2013), and maintaining redox balance has been reported to alleviate aging-related osteoporosis .
Pyrroloquinoline-quinine (PQQ) was first shown to function as a cofactor for oxidoreductases in bacteria (Killgore et al., 1989). PQQdeficient food can induce different systemic alterations in mice, including growth disorders, immune dysfunction, and abnormal reproductive capacity (Killgore et al., 1989;Steinberg et al., 1994). In contrast, PQQ supplementation has been reported to alleviate the progression of multiple diseases in mice (Dai et al., 2022;Jonscher et al., 2017). The effects of PQQ largely rely on its antioxidant function, as a previous study demonstrated that its radical-scavenging activity is 7.4-fold higher than that of vitamin C (Akagawa et al., 2016); however, the specific mechanism underlying its antioxidant function is still unclear. Although PQQ cannot be synthesized in mammals, it is available in foods such as milk, vegetables, and meat (Mitchell et al., 1999;Noji et al., 2007), making it an ideal regent for alleviating oxidative stress-induced diseases, including osteoporosis. Indeed, we have previously reported that PQQ supplementation can alleviate estrogen deficiency-induced osteoporosis (Geng et al., 2019). However, given the transition from estrogen deficiency to aging and oxidative stress (Manolagas, 2010), and a recent study indicating that estrogen and aging-inducing cellular senescence might play independent roles in the pathogenesis of osteoporosis (Farr et al., 2019), we sought in the current study to investigate the role and underlying mechanism of PQQ in the prevention of agerelated osteoporosis in mice.
Nuclear factor erythroid 2-related factor 2 (Nrf2) is a broadly expressed transcription factor and a key regulator of antioxidantresponsive genes that maintain cellular redox homeostasis, thus protecting against oxidative stress (Niture et al., 2014). Under basal conditions, Nrf2 is bound to its inhibitor, Kelch-like ECH-associated protein 1 (Keap1), which acts as an adapter for E3-ubiquitin ligase and promotes the degradation of Nrf2 by the 26S-proteasome (Cullinan et al., 2004); in contrast, under electrophilic/oxidative stress, the activity of Keap1 is inhibited, causing the release of Nrf2, which is translocated from the cytoplasm to the nucleus and binds to antioxidant response elements (AREs) in genes encoding antioxidant enzymes that protect cells against ROS (Suzuki & Yamamoto, 2015;Tonelli et al., 2018). Mounting evidence indicates that there is crosstalk between Keap1-Nrf2 and other proteins, such as p62, iASPP2, and MCM3, which can disrupt the normal Keap1-Nrf2 signaling and facilitate Nrf2-mediated cellular redox homeostasis (Ge et al., 2017;Komatsu et al., 2010;Mulvaney et al., 2016). The Nrf2 signaling pathway has also been shown to be protective in several diseases, including osteoporosis Park et al., 2014;Sánchezde-Diego et al., 2021;Sun et al., 2015;, although the effect of Nrf2 activation on natural aging-induced osteoporosis is unclear.
The purpose of this study was to investigate the role and mechanism of dietary PQQ supplementation in preventing natural age-related osteoporosis. This study will provide insight into the mechanism whereby PQQ inhibits oxidative stress and prevents age-related osteoporosis by indirectly inhibiting osteoclastic bone resorption via MCM3-Keap1-Nrf2 signaling activation in osteoblastlineage cells and provides evidence for PQQ as an attractive agent for the prevention and treatment of natural aging-induced osteoporosis.

| Animal experiments
A total of 24 six-month-old C57BL/6 male wild-type mice were randomly divided into three groups. Young (6-month-old) and aging (18-month-old) wild-type mice on a normal diet were used as the control groups, and the other two groups of mice were given a PQQcontaining diet (4 mg/kg diet) from 12 months of age and 6 months of age, respectively. Nrf2 +/− mice were purchased from Cyagen Biosciences Inc. To determine whether Nrf2 is a key effector of PQQ, 6-month-old wild-type and Nrf2 −/− mice were given the normal diet or the PQQ-containing diet for 6 months. All animal experiments were performed in compliance with the guidelines approved by the Institutional Animal Care and Use Committee of Nanjing Medical University.
Briefly, CT scanning was performed using the SkyScan1176 Micro-CT system (Bruker) at a resolution of 9 μm for quantitative analysis. A region 1.8 mm wide and 2.0 mm long in trabecular bone K E Y W O R D S aging, Fbn1, Keap1-Nrf2 signaling, osteoporosis, pyrroloquinoline quinone of lumbar vertebrae (50 slices) was quantitatively analyzed for bone mineral density (BMD), trabecular bone volume per tissue volume (BV/TV), trabecular number (Tb.N), thickness (Tb.Th), and separation (Tb.Sp) using CTAn software. The region from 8.25 mm (500 slices) to 8.43 mm (520 slices) below the growth plate of the femurs was used for cortical thickness (Ct.Th) analysis using CTAn.
The region from 10.05 mm (690 slices) to 10.5 mm (740 slices) below the growth plate of the femurs was analyzed for BMD, BV/ TV, Tb.N, Tb.Th and Tb.Sp.

| Histology and bone histomorphometry
At the time of euthanasia, lumbar vertebrae were fixed in PLP fixative buffer for 48 h. For routine histology analysis, decalcified bone using EDTA was dehydrated and embedded in paraffin, after which 5μm sections were prepared and bone sections were stained with the osteoclast marker tartrate-resistant acid phosphatase (TRAP) and total collagen (T-Col) as previously described (Yang et al., 2020).
For assay of dynamic bone formation, mice were intraperitoneally injected with calcein (10 mg/kg, Sigma), 10 days after which they were given an injection of xylenol orange (XO) (90 mg/kg, Sigma).
Dynamic bone formation was analyzed as previously described (Yang et al., 2020).

| Cell cultures and treatment
Human and mouse BM-MSCs were isolated and cultured as described previously . Primary osteoblasts from calvaria were isolated and cultured as described previously (Nollet et al., 2014). Briefly, calvaria were wished with PBS and cut into small fragments, which were then incubated in collagenase II and trypsin solution at 37°C in a shaking water bath. Bone pieces were then transferred into dishes in DMEM containing 10% FBS and 1% penicillin/streptomycin. Third-passaged mouse osteoblasts were treated with PQQ for the indicated times.

| Osteoclast formation assay
For osteoclast differentiation, mouse bone marrow monocytes (BMMs) isolated from femurs and tibias were cultured in DMEM supplemented with 10% FBS, 1% Penicillin/Streptomycin, and 50 ng/ mL M-CSF (PeproTech) for 3 days. Subsequently, adherent cells were further cultured in the presence of 50 ng/mL M-CSF and 50 ng/mL RANKL with indicated treatment. For osteoblast and osteoclast coculture, osteoblasts isolated as described above were seeded into a 12-well plate. BMMs were collected and cultured with osteoblasts in the presence of 1,25-dihydroxyvitamin D3 (10 nM; Sigma) and PGE2 (1 μM; Sigma). The culture medium was changed every other day. Cells were fixed and stained for TRAP as previously described .

| Prediction of potential targets of PQQ and molecular docking
The SDF (Structure Data File) file of PQQ was obtained from PubChem (https://pubch em.ncbi.nlm.nih.gov/compo undco mpoun d/1024), and the effective targets were predicted using the PharmMapper Server (http://www.lilab -ecust.cn/pharm mappe r/) by employing the "All Targets" model. Autodock Vina 1.1.2 was employed to investigate the binding affinity and binding sites between PQQ and MCM. Briefly, the SDF file of PQQ was converted into PDBQT format by using AutoDock tools. The 3D crystal structure of the MCM3 was obtained from the Uniprot (http://www.unipr ot.org, PDB ID: AF-P25206-F1). The docking protocol was generated as described previously , and the partial diagram of molecular docking was generated using the PyMol software.
For ARE-driven luciferase reporter assay, a total of 8 ARE copies was cloned to a pGL3-promoter as we previously described . All constructs were verified by DNA sequencing (Tsingke).

| Lentivirus production
Lentivirus particles were generated from HEK293T cells as previously described . Briefly, HEK293T cells were co-transfected with transfer plasmid and packaging plasmids by Lipofectamine 2000 transfection reagent (Invitrogen). Viral particle-containing supernatants were harvested 24 and 48 h later and were concentrated by ultra-centrifugation (4000 × g for 3 h).

| Detection of ROS
ROS evaluation of living cells in vitro using Dihydroethidium (DHE, ApexBio) was performed as previously described . For the detection of ROS in vivo, bone was sectioned without fixing using a transparent film, and sections were incubated in DHE at 37°C for 30 mins. The fluorescence was captured under a microscope, and the intracellular intensity of DHE was quantified using Image J.

| ChIP-qPCR
ChIP-qPCR was performed using the ChIP kit (Millipore) as previously described (Yang et al., 2020). The antibodies used in ChIP-qPCR were as follows: rabbit anti-IgG (Abcam, ab172730) and rabbit anti-Nrf2 (Proteintech). The enriched DNA was used for qPCR to detect the putative ARE of Fbn1. The sequences of primers were listed in Table S1.

| Dual luciferase assay
The indicated length of the Fbn1 promoter and its mutant were directly synthesized and cloned to pGL3-basic (Promega). For AREdriven luciferase reporter assay, BM-MSCs were co-transfected with ARE-driven luciferase, pcDNA3.1-Nrf2, and 0.05 μg of renilla followed without or with PQQ treatment. Forty-eight hours later, relative luciferase activity was measured using a kit (Promega) according to the manufacturer's instructions. The promoter sequences of Fbn1 and its mutant sequence cloned to pGL4.17 are listed in Table S3.

| RNA isolation and real-time RT-PCR
Total RNA extraction, cDNA synthesis, and real-time RT-PCR were performed as previously described .
Gapdh was amplified at the same time to normalize gene expression.
The PCR primer sequences used in this study are shown in Table S2.

| Statistical analysis
Measured data were described as mean ± SD. The statistical analyses were performed using GraphPad Prism (Version 8.0). Two-tailed Student's t-test was used to compare differences between groups.

| PQQ supplementation prevents natural aging-related bone loss and skeletal aging in mice
To assess whether PQQ supplementation can prevent natural aginginduced osteoporosis, wild-type male mice were given a PQQcontaining diet (4 mg PQQ/kg diet) from the age of 6 and 12 months, respectively, and the other two groups (6-and 18-month-old male mice) were given a normal diet as the control groups. All mice were sacrificed, and skeletal phenotypes were analyzed using micro-CT and total collagen (T-Col) staining (Figure 1a-c). Compared with age-matched littermates given the normal diet, natural aging-related bone loss of vertebrae and femurs were significantly alleviated upon dietary supplementation of PQQ for 6 or 12 months, as determined F I G U R E 1 PQQ supplementation alleviates the bone aging phenotype in aging wild-type mice. (a) Experimental design for investigating the effects of PQQ on natural aging-induced bone loss; wild-type mice at the age of 6 and 12 months were given the PQQ diet (4 mg/kg standard feed) for 12 and 6 months, respectively. Control mice were given the normal diet. Mice were sacrificed and skeletal phenotypes analyzed at indicated ages. (b) Representative μCT scans and (c) Total collagen staining of lumbar vertebrae from indicated groups of mice. Microtomography indices were measured as (d) BMD, (e) BV/TV, (f) Tb.N, (g) Tb.Th, and (h) Tb.Sp. (i) Representative μCT images of bone microarchitecture at the femur diaphysis of indicated groups of mice. (j) Quantification of μCT-derived cortical thickness (Ct.Th). One-way ANOVA. *p < 0.05, **p < 0.01, ***p < 0.001. ns, not significant. Figure S1a-g). Femur cortical thickness was significantly increased in aged mice following PQQ treatment for 12 months as compared to controls (Figure 1i,j). To further confirm our hypothesis, we also administered PQQ to 12-month-old wild-type mice for 12 months, analyzed bone phenotypes at the age of 24 months and results showed that aging-related bone loss of vertebrae was also significantly alleviated upon PQQ treatment as determined by BMD, BV/TV, Tb.N, Tb.Th, and Tb.Sp ( Figure S2a-f). Interestingly, natural aginginduced bone loss phenotypes were much more substantially prevented when PQQ was supplemented at a younger age i.e., starting from 6 months of age than if PQQ was initiated at 12 months of age

| PQQ rescues aging-induced reduction of osteoblastic bone formation and augmentation of osteoclastic bone resorption by inhibiting oxidative stress
We next sought to determine the relative contributions of osteoblast and osteoclast activity to the alleviation of bone loss in PQQsupplemented aged mice. We firstly performed histomorphometric analysis to evaluate static and dynamic parameters of bone forma-

| PQQ increases stress response capacity and reduces Rankl secretion of osteoblasts from aged mice by activating Nrf2-ARE signaling
Next, we sought to investigate in greater detail the mechanism whereby PQQ inhibits osteoclast activity. BMMs isolated from 16-month-old wild-type mice were induced to differentiate into osteoclasts by adding Rankl and M-CSF with or without  PQQ. Additionally, we induced osteoclasts by OB-OC co-culture.
Osteoblasts isolated from long bones of 16-month-old wild-type mice were treated with PQQ for 3 days, following which they were co-cultured with BMMs in the presence of 1,25(OH) 2 D 3 and Prostaglandin E2 (PGE2). Interestingly, we found that direct formation of osteoclasts from BMMs was not significantly changed in the presence of PQQ ( Figure S5a,b), whereas osteoclast formation of BMMs in the co-culture system was significantly decreased by treating osteoblasts with PQQ (Figure 3a,b). Furthermore, the mRNA and protein expression levels of Rankl were significantly reduced in PQQ-treated osteoblasts compared to vehicle-treated control Rankl production was reported to be associated with increased oxidative stress in osteoblasts (Baek et al., 2010;Choi, 2012;Nollet et al., 2014;Rana et al., 2012). Senescent cells appear to have decreased capacity to initiate antioxidant signaling in the presence of stress stimulation, in association with decreased expression of Nrf2, the master regulator of oxidative stress (Meng et al., 2017).
Consistent with this observation, we found here that senescent osteoblasts exhibited decreased Nrf2 expression and stress response capacity following treatment with paraquat (an oxidative stress inducer), as well as increased oxidative stress and Rankl production ( Figure S6a-i). Furthermore, Nrf2-deficient osteoblasts resulted in higher Rankl production ( Figure S7a,b) and increased osteoclast formation ( Figure S7c,d) in the co-culture system. Moreover, DHE staining showed that ROS levels in osteoblasts were significantly reduced in PQQ-treated osteoblasts relative to vehicle-treated controls in the absence or presence of paraquat (Figure 3g,h). Furthermore, the protein, but not mRNA expression levels of Nrf2 were significantly increased in PQQ-treated osteoblasts relative to vehicle-treated control osteoblasts (Figure 3i,j). In addition, in the presence of paraquat, the mRNA levels of antioxidant-related genes downstream of Nrf2 were not significantly changed in vehicle-treated senescent osteoblasts; however, these genes were significantly up-regulated in PQQ-treated senescent BM-MSCs (Figure 3i,j). These results suggested that PQQ increases stress response capacity and inhibits Rankl production of osteoblasts, thus directly reducing osteoclastogenesis via activating Nrf2-ARE signaling.

| PQQ binds to and promotes protein stability of MCM3, competing with Nrf2 for binding to Keap1
To investigate the mechanism underlying how PQQ activates Nrf2-ARE signaling, human BM-MSCs isolated from aged osteoporosis patients were treated with indicated concentrations of PQQ for 24 h, and the expression levels of Nrf2 and Keap1 (the main regulator of Nrf2) were detected using Western blot and real-time PCR. Results showed that the protein expression level of Nrf2, but not its mRNA expression level was significantly increased in PQQ-treated hBM-MSCs relative to vehicle-treated control (Figure 4a; Figure S8a), indicating that PQQ might regulate Nrf2 at a post-translational level.
Indeed, further results showed that Nrf2 protein degradation was delayed (Figure 4b), and the expression level of ubiquitin-Nrf2 was significantly decreased in PQQ-pretreated hBM-MSCs as compared with vehicle-pretreated control (Figure 4c). In addition, ARE-driven luciferase activity was significantly increased in hBM-MSCs treated with PQQ for indicated times (Figure 4d), indicating that the transcriptional activity of Nrf2 was increased upon PQQ treatment.
Keap1 is the master regulator of Nrf2 at the post-translational level; however, neither the protein nor the mRNA expression of Keap1 was increased in PQQ-treated hBM-MSCs relative to vehicletreated control (Figure 4a; Figure S8a). To determine the direct target of PQQ, the structural data file of PQQ was submitted to the Pharmmapper, selecting the "All Targets" model, and the top 5 potential targets ranked by the normalized fit score in descending order were listed (job ID: 220330013617) (Figure 4e). The detailed results were uploaded as a separate file (Table S4) Figure S9a,b). Keap1 is considered as the main regulatory effector of Nrf2, and it has been reported that MCM3 can compete with Nrf2 for binding to Keap1, thus inhibiting Nrf2 degradation (Mulvaney et al., 2016;Tamberg et al., 2018). To test the effect of PQQ-MCM3 binding on MCM3 expression and PQQ-induced Nrf2 signaling activation, we examined the expression of MCM3 and the Keap1-Nrf2 interaction using F I G U R E 3 PQQ increases stress response and decreases Rankl production in osteoblasts from aged mice by activating Nrf2-ARE signaling. (a, b) Osteoclastogenesis by OB-OC co-culture in vitro using osteoblasts from 18-month-old wild-type mice treated with or without PQQ for 3 days: (a) TRAP staining in the co-culture system and (b) the quantification of the area of TRAP + multinucleated cells (MNCs, nuclei ≥3). (c-e) Rankl expression in vitro using osteoblasts from 18-month-old wild-type mice treated with or without PQQ for 48 h: (c) Real-time PCR detection of Rankl; (d) Western blot detection and (e) quantification of Rankl protein levels. (f) The effect of PQQ on aged osteoblasts viability was determined using CCK8 assay. Two-tailed Student's t-test. *p < 0.05, **p < 0.01, ***p < 0.001. (g) ROS levels in paraquat-treated aged osteoblasts in the presence or absence of PQQ were detected using DHE staining. (h) Quantitative analysis of (g). Two-way ANOVA. *p < 0.05, **p < 0.01. (i) Western blot detection of Nrf2 and HO1 protein levels in paraquat-treated aged osteoblasts in the presence or absence of PQQ at indicated times. (j) Real-time PCR detection of Nrf2, HO1, Nqo1, CAT, and GPX7 mRNA levels in the presence or absence of PQQ in paraquat-treated aged osteoblasts at indicated times. Two-way ANOVA. *p < 0.05, **p < 0.01. Western blots and co-immunoprecipitation. We found that the protein levels of MCM3 were significantly decreased in aging bones (3 vs. 18 M) and in passage-induced replicative senescent (P4 vs. P14) hBM-MSCs ( Figure S9c-f). In addition, the protein, but not mRNA expression level of MCM3 was significantly higher in PQQ-treated human BM-MSCs than in vehicle-treated control (Figure 4g). Protein degradation experiments showed that the ubiquitination and protein degradation rate of MCM3 were significantly inhibited in PQQtreated hBM-MSCs relative to vehicle-treated controls (Figure 4h; Figure S8b). Furthermore, co-immunoprecipitation results showed that PQQ significantly increased the interaction between Keap1 and MCM3 and decreased the interaction between Keap1 and Nrf2. (Figure 4i). Moreover, the reduced interaction between Nrf2 and Keap1 and the transcriptional activity of Nrf2 induced by PQQ were significantly blocked upon MCM3 knockdown in hBM-MSCs (Figure 4j-l). These results indicated that PQQ could maintain the expression levels of Nrf2 in repeatedly passaged BM-MSCs by directly binding to MCM3 to augment its protein stability; MCM3 thus competes with Nrf2 for binding to Keap1 and reduces the degradation of Nrf2 by the ubiquitin system.

| Nrf2 activation partly inhibits osteoclast differentiation by upregulating Fbn1 in BM-MSCs
To examine whether PQQ-mediated Nrf2 activation can directly inhibit osteoclast differentiation in the OB-OC co-culture system independent of stress response regulation, RNA-seq was performed on primary BM-MSCs isolated from WT and Nrf2 −/− mice. The results showed that aside from antioxidant-related genes, fibrillin-1 (Fbn1) was significantly decreased in Nrf2 −/− BM-MSCs as compared to wild-type controls (Figure 5a). Fbn1 deficiency has been reported to result in bone loss with constitutively enhanced bone resorption in mice (Smaldone et al., 2016), and Fbn1-deficient osteoblasts have been shown to stimulate pre-osteoclast differentiation more than wild-type osteoblasts (Tiedemann et al., 2013). Real-time qPCR results showed that PQQ significantly increased Fbn1 expression in aged wild-type BM-MSCs, but not in age-matched Nrf2 −/− BM-MSCs, indicating that the increase of Fbn1 induced by PQQ depends on Nrf2 (Figure 5b,c). Based on these results, we speculated that PQQ might regulate the transcription of Fbn1 through Nrf2. A predicted ARE was detected in the promoter of Fbn1 (Figure 5d), and the chromatin immunoprecipitation (ChIP) assay was used to confirm its involvement. Thus, ChIP results showed that Nrf2 bound to the promoter region of Fbn1 under physiological conditions; this binding was significantly increased by treatment with tertbutylhydroquinone (tBHQ, a classic Nrf2 activator) (Figure 5e). Dual luciferase assays showed that luciferase activity, driven by a Fbn1 promoter containing the predicted ARE, was significantly increased following Nrf2 overexpression, while this effect was abolished when the predicted ARE was mutated (Figure 5f). Furthermore, the antioxidant reagent NAC did not rescue the decreased Fbn1 mRNA level in Nrf2 −/− BM-MSCs (Figure 5g), indicating that the increased Fbn1 induced by Nrf2 activation was independent of decreased oxidative stress. Meanwhile, we detected Fbn1 expression in mouse bone tissues and found that the mRNA expression level of Fbn1 was also significantly down-regulated in aging bone tissues relative to young controls ( Figure 5h). Moreover, co-culture experiments and TRAP staining results showed that Fbn1 knockdown in mouse calvarial osteoblasts significantly increased Rankl production in osteoblasts, and enhanced osteoclast differentiation of BMMs ( Figure S10a-d).
Notably, Fbn1 knockdown in osteoblasts significantly blocked PQQinduced reduction of Rankl production in osteoblasts and osteoclast differentiation of BMMs (Figure 5i-k). These results demonstrated that Nrf2 activation can inhibit osteoclast differentiation by transcriptionally upregulating Fbn1 in osteoblast-lineage cells.

| Nrf2 depletion largely blocks the inhibitory effects of PQQ on oxidative stress, osteoclast activity, and aging-related osteoporosis
In order to determine whether Nrf2 is the key mediator of the inhibitory effects of PQQ on oxidative stress, osteoclast activity, and aging-related osteoporosis, 6-month-old wild-type and Nrf2 −/− male mice were fed with a PQQ-or vehicle-containing diet for 6 months, and skeletal phenotypes of these mice were analyzed using μCT and histomorphometry at the age of 12 months. We firstly found that   Figure S11g-p). Furthermore, results showed that BMD, trabecular number, volume, and thickness were significantly elevated in PQQ-supplemented wild-type mice, but not in PQQ-supplemented Nrf2 −/− mice (Figure 6a,e-h), whereas trabecular separation, oxidative stress in osteoblasts, osteoclast activity in vivo, and osteoclastogenesis in co-culture system were significantly reduced in PQQ-supplemented wild-type mice, but not in PQQ-supplemented Nrf2 −/− mice (Figure 6b-d,i-l). These results indicated that Nrf2 is essential to mediate the inhibitory effects of PQQ on oxidative stress, osteoclast activity, and aging-related osteoporosis.

| DISCUSS ION
Estrogen and aging-inducing cellular senescence have been reported to play independent roles in the pathogenesis of osteoporosis (Farr et al., 2019). We have previously reported that PQQ can prevent estrogen deficiency-induced osteoporosis (Geng et al., 2019); however, whether PQQ can alleviate aging-related osteoporosis and the specific mechanism remains unclear. Here, we showed that dietary PQQ supplementation could not only promote bone formation of osteoblasts, but also inhibit bone resorption through the activation of stress response and the upregulation of Fbn1 mediated by MCM3-Keap1-Nrf2, thus preventing natural aging-related osteoporosis.
PQQ was identified as a redox cofactor with strong radicalscavenging activity, and it has been reported to alleviate several disorders (Jonscher et al., 2017). Our previous studies have demonstrated that PQQ can prevent oophorectomy-induced osteoporosis (Geng et al., 2019). Aging has been recognized as the most important inducer of osteoporosis in both males and females. Here, we administered PQQ to wild-type mice from 6 or 12 months of age for 12 or 6 months, respectively, and found that PQQ alleviated natural aging-induced osteoporosis. Of note, PQQ has served as a safe food supplement in some countries, making it an ideal option for the prevention of aging-related diseases, including osteoporosis. Indeed, we found here that PQQ treatment did not alter liver and kidney damage-related markers. Overall, our findings further confirmed that PQQ could be a potential drug beneficial for the prevention and treatment of aging-related osteoporosis. Interestingly, this skeletal protective effect of PQQ appears to be more pronounced when PQQ is provided at a relatively younger age. This suggests that PQQ needs to be assessed in humans in future studies and might be useful in delaying the onset of osteoporosis by increasing peak bone mass in youth.
In our studies, PQQ alleviated natural aging-related osteoporosis not only by stimulating osteoblastic bone formation, but also by inhibiting osteoclastic bone resorption. Previous studies have demonstrated that upregulation of Nrf2 in osteoblasts can indirectly reduce osteoclast formation by inhibiting the secretion of Rankl in osteoblasts (Narimiya et al., 2019;Rana et al., 2012).
Increased Rankl production has been reported to be associated with increased oxidative stress in osteoblasts (Baek et al., 2010;Choi, 2012;Nollet et al., 2014;Rana et al., 2012). PQQ also significantly increased Nrf2 expression and decreased osteoclastic bone resorption in aging mice. It has been reported that the stress response capacity declined in senescent cells (Meng et al., 2017), and we found that PQQ could partially restore the oxidative stress response capacity of repeatedly passaged BM-MSCs by upregulating Nrf2 expression, while inhibiting osteoclast formation by decreasing Rankl secretion. Overall, our results suggest that PQQ can increase the protein level of Nrf2 to enhance the oxidative stress response and decrease Rankl production in osteoblasts, thus inhibiting osteoclast activity. Interestingly, here we did not observe a direct inhibitory effect of PQQ on osteoclast differentiation of BMMs, which is inconsistent with a previous study (Kong et al., 2013). This may be caused by the different sources of BMMs used, with the previous study using BMMs from young rats and this study using BMMs from aging mice.
In our study, PQQ treatment not only decreased oxidative stress and inhibited osteoclast activity in vivo in aging mice, but also increased osteoblastic bone formation. This appears to be consistent with the crucial role for Nrf2-antioxidant signaling for redox homeostasis and osteoblast survival. Indeed, a recent study reported that Nrf2 protects osteoblasts from cell death caused by glucocorticoid-induced oxidative stress and promotes osteogenesis (Rai et al., 2022). In addition, we here found that PQQ-mediated  Mut: 5' * * * * * * * * * * * * * * * 3' Mut: 5' * * * * * * * * * * * * * * * 3' WT: 5' TGTCTAGTCATGGTC 3' WT 1 # 2 # 3 # 1 # 2 # 3 #  inhibiting oxidative stress-induced osteoclast activity (59), but also maintains osteoblast cell survival to increase osteoblast activity in response to oxidative stress with aging. Interestingly, the effects of Nrf2 on bone seem to be different between young and aging mice in bone-related cells. Based on previous studies (Sánchez- de-Diego et al., 2021; and the present study, Nrf2 exerts more protective effects in response to aging-induced stress in osteoblasts, and targeted activation of Nrf2 in osteoblasts in aged mice is needed to test this hypothesis. Here, we also detail a novel mechanism whereby PQQ stabi- has been reported to reduce osteoclast activity in vitro and in vivo by antagonizing Rankl and inhibiting NFATc1 signaling (Smaldone et al., 2016;Tiedemann et al., 2013). Consistently, we found that for Nrf2 activation in decreasing osteoclast differentiation.
To further determine whether Nrf2 could serve as a key target of PQQ in correcting natural aging-induced osteoporosis, 6-month-old WT and Nrf2 −/− mice were supplemented with PQQ in vivo; osteoblasts isolated from WT and Nrf2 −/− mice were treated with PQ Q , and their phenotypes were analyzed in vivo and in vitro. Here, we report that aged Nrf2-deficient mice displayed severe aging-related bone loss and a skeletal aging phenotype, although the effect of Nrf2 deficiency on bone in young mice was inconclusive in previous studies Park et al., 2014;Yin et al., 2020). Moreover, we found that PQQ supplementation corrected the increased bone resorption and bone loss in aging WT mice; however, these responses were largely blocked in Nrf2 −/− mice. In addition, we found that PQQ inhibits oxidative stress in osteoblasts and decreased osteoclastogenesis in OB-OC co-culture systems; however, these responses were significantly blocked by Nrf2 deficiency. Overall, these results suggest that Nrf2 is essential for PQQ to alleviate aging-induced osteoporosis by increasing an oxidative stress response and reducing Rankl production in osteoblasts.
In summary, the results of this study reveal that PQQ can

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors declare no competing interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data and materials used in the study are available to any researcher for purposes of reproducing or extending the analysis.