Graphical abstract graphic file with name fx1.jpg [49]Open in a new tab Highlights * • PINK1/Drp1-mediated mitophagy contributes to osteoblastogenesis * • Hyperglycemia suppresses PINK1/Drp1-mediated mitophagy, causing impaired osteoblastogenesis * • BMP9 reverses hyperglycemia-induced impaired osteoblastogenesis by activating mitophagy * • BMP9 treats diabetic osteoporosis in mice by activating PINK1/DRP1-mediated mitophagy __________________________________________________________________ Physiology; Molecular biology; Cell biology Introduction As with many other diabetic complications, the poor bone quality and increased fracture risk in diabetes has already aroused great concerns⁠.[50]^1^,[51]^2 It was well recognized that hyperglycemia exposure inhibits proliferation and differentiation of osteoblasts,[52]^3 but its specific mechanism is still not clear, and no definitive pharmaceutical remedies for diabetic osteoporosis are available except for those traditional medications to treat primary osteoporosis. Energy metabolism is an important physiological basis for maintaining cell function. The function of mitochondria, as the hub of energy metabolism, is the key to maintaining normal life activities. Mitochondrial dysfunction-related diseases have expanded from rare monogenic disorders to common polygenic diseases, including metabolic, cardiovascular, neurodegenerative, and neuromuscular diseases in nowadays.[53]^4^,[54]^5 Lots of effort was made to explore signaling pathways to reverse mitochondrial dysfunction.[55]^6 Recent studies have suggested a mutual cross-talk between bone remodeling and glucose homeostasis.[56]^7^,[57]^8 So far, the best characterized metabolic pattern for osteoblast maturation is the glucose metabolism reprogramming from oxidative phosphorylation (OXPHOS) toward aerobic glycolysis, coupling with diminished mitochondrial respiration.[58]^9^,[59]^10 Suppression of glucose metabolism restricts osteoblast differentiation.[60]^11 Thus, it is interesting to investigate whether and how the disturbed metabolic reprogramming causes impaired osteoblastogenic differentiation and poor bone quality in diabetes, and how to reverse it. Mitophagy is a mechanism that eliminates damaged mitochondria and controls mitochondrial quality and quantity. Simultaneously, mitophagy mediates metabolic reprogramming, either into an OXPHOS phenotype or into a glycolytic phenotype, to satisfy the metabolic requirements of cells.[61]^12^,[62]^13 Mitophagy dictates stem cell fate by facilitating glycolytic phenotype swift.[63]^14 The deficiency of mitophagy induces reactive oxygen species (ROS) stress, and leads to the development of neurodegenerative disorder,[64]^15 cardiomyopathy,[65]^16 and acute kidney injury.[66]^17 In type 2 diabetes patients, mitophagy pathway gene expression was downregulated.[67]^18 In β-cells, decreased glucose-responsive insulin secretion was associated with impaired mitophagy.[68]^19 Mitophagy is mediated by PTEN-induced kinase 1 (PINK1)/Parkin pathway. PINK1 is a mitochondrial serine/threonine-protein kinase. It can sense mitochondrial stress, accumulate on defective mitochondria, recruit Parkin, and finally trigger mitophagy.[69]^20 Loss of PINK1/Parkin may contribute to several diseases, such as Parkinson’s disease.[70]^21 As to bone metabolism, PINK1-mediated mitophagy is involved in cartilage degeneration seen in osteoarthritis.[71]^22^,[72]^23^,[73]^24 Therefore, our research interest is what role mitophagy plays in hyperglycemia-induced osteoblastogenic defects and whether the intervention of related mechanisms can restore diabetic osteoporosis. Results High glucose level inhibited osteogenic differentiation and respiration rate of MC3T3-E1 cells To investigate the effect of high glucose level on the osteoblastogenesis of MC3T3-E1 cells in vitro, alkaline phosphatase (ALP) and Alizarin red staining were performed. Alizarin red staining in H-Glu group showed significantly fewer mineralization nodules compared with L-Glu group in the late stage (14 days, 21 days) of osteogenic differentiation. However, there was no significant difference in ALP staining between the two groups in the early stage (1 day, 7 days) of osteogenic differentiation ([74]Figure 1A). Furthermore, the mRNA expression levels of Runx2, Osx, Col1a1, and ALP were reduced in H-Glu group at day 14, while there was no significant difference between H-Glu and L-Glu group at day 1, and only Osx expression decreased significantly in the H-Glu group at day 7 ([75]Figures 1B and 1D). Consistently, the western blot (WB) results showed that the protein levels of Runx2 and Osx were decreased in H-Glu group at day 14 ([76]Figure 1E). These results suggest that high glucose inhibits the late stage of osteogenic differentiation in MC3T3-E1 cells. Figure 1. [77]Figure 1 [78]Open in a new tab High glucose inhibited osteoblastogenic differentiation and respiration rate of MC3T3-E1 cells The concentration of glucose in the medium of the L-Glu group was 5.5 mmol/L, and that in the medium of the H-Glu group was 25 mmol/L. (A) ALP staining was performed in L-Glu and H-Glu cultured MC3T3-E1 cells after 1 and 7 days of osteogenic induction. Alizarin red S staining was performed in L-Glu and H-Glu cultured MC3T3-E1 cells after 14 and 21 days of osteogenic induction. Scale bars, 500 μm. (B–D) MC3T3-E1 cells were cultured in an osteogenic medium for 1, 7, and 14 days. Expressions of osteogenic markers at different time points were determined by RT-qPCR. (E) MC3T3-E1 cells were cultured in osteogenic medium for 14 days. Expressions of Runx2 and Osx were determined by WB. (F–Q) MC3T3-E1 cells were cultured in osteogenic medium for 1, 7, and 14 days. Mitochondrial oxidative capacity and glycolytic capacity were measured in real time. Basal respiration, ATP-linked respiration, maximal respiration, spare respiratory capacity, basal, proton leak, glycolysis, glycolytic capacity, and glycolytic reverse were calculated by WAVE software. Data presented as mean ± SD. n = 3 biological replicates. One-way ANOVA was used for comparison among multiple groups. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. Ns, no significance. Since the disruption of energy metabolism in osteoblasts may lead to impaired osteoblast function,[79]^25^,[80]^26 we then detected two key energy metabolic pathways, aerobic respiration and glycolysis in MC3T3-E1 cells using the Seahorse extracellular flux analyzer. On day 1 of osteogenic differentiation, no noteworthy distinctions of basal respiration, maximal respiration, spare respiration capacity, and ATP-linked respiration, which are pivotal parameters that underscore mitochondrial functionality, were observed between H-Glu group and L-Glu group ([81]Figures 1F and 1G). As to cellular glycolytic function, H-Glu group exhibited elevated glycolysis and glycolytic capacity compared to the L-Glu group, with no statistically significant variance in glycolytic reversal between the two groups ([82]Figures 1H and 1I). Subsequent to 7 days of osteogenic induction, exclusive reductions in maximal respiration and glycolysis were discerned in H-Glu group relative to L-Glu group ([83]Figures 1J–1M). Following a 14-day osteogenic differentiation period, the mitochondrial stress test revealed a substantial attenuation in basal, maximal, and ATP-linked respiration in H-Glu cells as contrasted with those in L-Glu group ([84]Figures 1N and 1O). Simultaneously, outcomes from the glycolytic rate examination exhibited notable diminutions across glycolysis, glycolytic capacity, and glycolytic reversal in H-Glu cells in comparison to L-Glu counterparts ([85]Figures 1P and 1Q). Collectively, these findings provide evidence that the late stage of osteogenic differentiation is notably affected by a discernible inhibition of osteoblastic mitochondrial function induced by the high glucose level. High glucose level did not affect mitochondrial biogenesis in MC3T3-E1 cells Mitochondrial proliferation may indicate the engagement of feedback loops to compensate for compromised mitochondrial function. Alterations in gene expression profile involved in mitochondrial biogenesis, and changes of concentrations and activities of candidate respiratory complexes in cells may lead to energetic dysfunction.[86]^27 Hence, the peroxisome proliferator-activated receptor gamma coactivator 1a (PGC-1α), nuclear respiratory factors (NRF-1 and NRF-2), and mitochondrial transcription factor A (TFAM) that are the key regulators of mitochondrial biogenesis were determined in osteoblasts. The results showed that there was no difference in the expression of the aforementioned genes between H-Glu and L-Glu groups ([87]Figure 2A). Figure 2. [88]Figure 2 [89]Open in a new tab High glucose did not affect mitochondrial biogenesis but elevated ROS production and decreased mitochondrial membrane potential (MMP) in MC3T3-E1 cells (A) MC3T3-E1 cells were cultured in osteogenic medium for 14 days. The total DNA of the cells was extracted. Using HK2 as reference gene, expression ND1 and 16s were determined by RT-qPCR. (B) MC3T3-E1 cells were cultured in osteogenic medium for 14 days. Expressions of mitochondrial biogenesis markers were determined by RT-qPCR. (C) MC3T3-E1 cells were cultured in osteogenic medium for 14 days. Expressions of respiratory chain complexes were determined by WB. (D–F) MC3T3-E1 cells were cultured in osteogenic medium for 14 days. ROS generation was observed by using flow cytometry following staining with DCFH-DA. (G and H) MC3T3-E1 cells were cultured in osteogenic medium for 14 days. MMP of living cells was assessed by the Image-iT TMRM Reagent. Scale bars, 50 μm. Data presented as mean ± SD. n = 3 biological replicates. A t test was used for comparison between two groups. One-way ANOVA was used for comparison among multiple groups. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. Ns, no significance. Mitochondrial biogenesis involves the integration of multiple transcriptional pathways controlling both nuclear and mitochondrial gene expression. For example, mitochondrial DNA (mtDNA) transcription is activated by the family of PGC-1 proteins (PGC-1α and PGC-1β), from which PGC-1α is considered the master regulator of mitochondrial biogenesis. The pathway is initiated by PGC-1α activation, followed by stimulation of a series of nuclear transcription factors, that is the NRF-1 and NRF-2 and by the increase in expression of TFAM, the final effector of mtDNA transcription and replication.[90]^28 To further measure mitochondrial biogenesis, we quantitated relative mtDNA copy number by qPCR in MC3T3-E1 cells. As shown in [91]Figure 2B, mtDNA copy number in osteoblasts of H-Glu group remained similar to L-Glu group. Additionally, we examined the relative levels of the 5 OXPHOS complexes in mitochondria by WB. The result was in accordance with the mitochondrial biogenesis gene expression. Mitochondrial respiratory chain complex protein levels were not significantly different between the two groups ([92]Figure 2C). These results demonstrate that high glucose level does not affect the mitochondrial biogenesis of MC3T3-E1 cells and suggest that high glucose level may affect mitochondrial function through other mechanisms. High glucose level elevated ROS production and decreased MMP in MC3T3-E1 cells Mitochondria is one of the main sources of ROS, as it utilizes oxygen for energy production. Mitochondrial dysfunction may lead to increased radical production and consequent “oxidative stress.” Therefore, cell ROS level is one of the recognized indicators to evaluate mitochondrial function.[93]^29 The flow cytometry results in this study showed that the number of positively stained cells in H-Glu group was significantly higher than that in L-Glu group ([94]Figures 2D and 2E). Fluorescence intensity quantitative analysis revealed that the ROS level in H-Glu group was significantly higher than that in L-Glu group ([95]Figure 2F). Mitochondrial function can also be assessed through monitoring changes in MMP. A decrease in the MMP indicates mitochondrial dysfunction.[96]^30 We examined the MMP using a cell-permeant dye tetramethylrhodamine (TMRM), which can accumulate in active mitochondria with intact membrane potentials. The TMRM signals in living cells were observed by confocal laser microscopy, quantitative analysis of TMRM fluorescence intensity in images showed that the average fluorescence intensity of cells in H-Glu group was 36% lower than that in L-Glu group. The MMP of H-Glu group cells were reduced ([97]Figures 2G and 2H). These results suggest that high glucose level impairs mitochondrial function in MC3T3-E1 cells. RNA-seq data analysis of differentially expressed genes in MC3T3-E1 cells cultured in L-Glu and H-Glu The aforementioned results demonstrated that high glucose level inhibited the late stage of osteogenic differentiation and impaired mitochondrial function of these cells. Then, what is its molecular mechanism? To address this question, MC3T3-E1 cells cultured in low glucose or high glucose osteoblastogenic induction media for 14 days were examined by RNA sequencing (RNA-seq) analysis ([98]Figure 3A). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis showed that the major pathways involved in the differential genes between the two groups were mitophagy, autophagy, and cell cycle ([99]Figures 3B and 3D). Gene ontology (GO) functional enrichment showed that the major pathways involved in the differentially expressed genes were ubiquitin proteasome system and mitophagy ([100]Figure 3C). As shown in [101]Figure 3E, further analysis of mitophagy genes revealed that transcription levels of PINK1 and dynamic-associated protein 1 (Drp1) were significantly reduced in H-Glu group compared to L-Glu group. We further verified the expression of Drp1 and PINK1 by qPCR ([102]Figure 3F). It is worth mentioning that we also detected mitophagy marker genes of non-ubiquitination-dependent pathways by qPCR and there were no significantly difference between the two groups ([103]Figures 3G–3I). These findings suggest that marker gene PINK1 of mitophagy signaling pathway and Drp1 may be important factors involved in the mechanism by which high glucose inhibits osteogenic differentiation. Figure 3. [104]Figure 3 [105]Open in a new tab RNA-seq data analysis of differentially expressed genes in MC3T3-E1 cells cultured in L-Glu and H-Glu MC3T3-E1 cells were cultured in osteogenic medium for 14 days, then eukaryotic transcriptome sequencing was performed. (A) The volcano map showed significant differences in gene expression between the two groups. (B) The enrichment analysis of KEGG pathway based on the significantly different genes in the volcano map showed that the main pathways involved in the damage of osteoblasts by high sugar were mitochondrial autophagy, cell autophagy and cell cycle. (C) GO ontology functional enrichment showed that the main pathways involved in the damage of osteoblasts by high glucose were ubiquitin-proteasome system, ubiquitin-proteasome system, and mitochondrial autophagy. (D) The KEGG pathway loop map showed the main mechanisms of hyperglycemic damage: mitochondrial autophagy, autophagy, and cell cycle. (E) Genes of interest with significant differences. n = 4 biological replicates. (F) MC3T3-E1 cells were cultured in osteogenic medium for 14 days. Expressions of PINK1 and Drp1 were determined by RT-qPCR. (G–I) MC3T3-E1 cells were cultured in osteogenic medium for 1, 3, 7, and 14 days. Expressions of marker genes involved in ubiquitin-dependent mitophagy pathway at different time points were determined by RT-qPCR. Data presented as mean ± SD. n = 3 biological replicates. One-way ANOVA was used for comparison among multiple groups. A t test was used for comparison between two groups. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. Ns, no significance. High glucose inhibited PINK1/Drp1 signaling, leading to impaired osteogenic differentiation To further validate the functionality of PINK1 and Drp1 in osteogenic differentiation, we inhibited the expression of PINK1 and Drp1 in MC3T3-E1 cells by siRNA. [106]Figures 4A and 4B demonstrated that the expression of PINK1 and Drp1 genes was significantly knocked down. We substantiated the involvement of PINK1/Drp1 signaling in the process of osteogenic differentiation through gene knockdown under normal glucose culture. Evidently, the targeted suppression of PINK1 or Drp1 significantly inhibited the expression levels of Runx2 and Osx ([107]Figure 4C). We also observed that inhibiting the expression of PINK1 resulted in the downregulation of Drp1 protein. Conversely, inhibiting Drp1 did not affect the level of PINK1 protein, implying that PINK1 acts as the upstream signaling molecule of Drp1 ([108]Figure 4C). Consistently, the quantification of mineralized nodules within osteoblasts, following the downregulation of PINK1 and Drp1, exhibited a substantial reduction compared to the control group ([109]Figure 4D). Figure 4. [110]Figure 4 [111]Open in a new tab High glucose inhibited PINK1/Drp1 signaling, leading to impaired osteoblastogenic differentiation (A and B) Expressions of PINK1 and Drp1 genes were determined by RT-qPCR after siRNA transfection. (C) MC3T3-E1 cells were cultured in osteogenic medium (L-Glu) for 14 days after siRNA transfection. Expressions of PINK1, Drp1, Runx2, and Osx were determined by WB. (D) MC3T3-E1 cells were cultured in osteogenic medium (L-Glu) for 21 days after siRNA transfection. Alizarin red S staining was performed. Scale bars, 500 μm. (E) Expression of PINK1 gene was determined by RT-qPCR after PINK1 plasmid transfection. (F) MC3T3-E1 cells were cultured in osteogenic medium (H-Glu) for 14 days after plasmid transfection. Expressions of PINK1, Drp1, Runx2, and Osx were determined by WB. (G) MC3T3-E1 cells were cultured in osteogenic medium (H-Glu) for 21 days after plasmid transfection. Alizarin red S staining was performed. Scale bars, 500 μm. (H) MC3T3-E1 cells were cultured in osteogenic medium (L-Glu) for 14 days after siRNA transfection. MMP of living cells was assessed by the Image-iT TMRM Reagent. Scale bars, 20 μm. (I) MC3T3-E1 cells were cultured in osteogenic medium (L-Glu) for 14 days after siRNA transfection. Mitochondrial reactive oxygen species of living cells were assessed by the MitoSox Reagent. Scale bars, 50 μm. (J) MC3T3E1 cells were transfected with the mito-Keima plasmid and cultured in osteogenic medium (L-Glu) for 14 days. The fluorescent dots of mito-Keima were observed by confocal microscopy. Scale bars, 20 μm. (K) MC3T3-E1 cells were cultured in osteogenic medium (H-Glu) for 14 days after plasmid transfection and siRNA transfection. Expression of PINK1, Drp1, Runx2, and Osx were determined by WB. (L) MC3T3-E1 cells were cultured in osteogenic medium (H-Glu) for 14 days after plasmid transfection and siRNA transfection. Alizarin red S staining was performed. Scale bars, 500 μm. (M) MC3T3-E1 cells were cultured in osteogenic medium (H-Glu) for 14 days after plasmid transfection and siRNA transfection. MMP of living cells was assessed by the Image-iT TMRM Reagent. Scale bars, 50 μm. (N) MC3T3-E1 cells were cultured in osteogenic medium (H-Glu) for 14 days after plasmid transfection and siRNA transfection. Mitochondrial reactive oxygen species of living cells were assessed by the MitoSox Reagent. Scale bars, 50 μm. (O) MC3T3-E1 cells were cultured in osteogenic medium (H-Glu) for 14 days after plasmid transfection and siRNA transfection. The fluorescent dots of mito-Keima were observed by confocal microscopy. Scale bars, 50 μm. Data presented as mean ± SD. A t test was used for comparison between two groups. One-way ANOVA was used for comparison among multiple groups. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. Ns, no significance. We further constructed an overexpressed PINK1 plasmid and transfected osteoblasts. Then, we induced osteogenic differentiation of MC3T3-E1 cells in high glucose medium. [112]Figures 4E and 4F showed that the expression of PINK1 gene and protein were significantly increased after transfection. Notably, this augmentation in PINK1 expression coincided with noteworthy enhancements in the expressions of Runx2 and Osx ([113]Figure 4F). Alizarin red staining showed that the number of mineralized nodules of osteoblasts overexpressing PINK1 increased significantly under high glucose medium culture ([114]Figure 4G). We also examined the mitochondrial function of osteoblasts that had knocked down PINK1 and Drp1, respectively. In [115]Figure 4H, TMRM staining was performed on living cells in each group to detect MMP. The results showed that inhibiting the expression of either Drp1 or PINK1 could decrease the MMP of cells. In [116]Figure 4I, MitoSox was used to detect mitochondrial reactive oxygen species, and the results showed that knockdown of Drp1 or PINK1 could increase the superoxide level of cellular mitochondria. In [117]Figure 4J, MCT3E1 cells were transfected with pMito-Keima to detect mitophagy, and the results showed that mitophagy was weakened in cells with the Drp1 or PINK1 knocked down. To verify Drp1 is a downstream target of PINK1, we further knocked down Drp1 in the presence of overexpression of PINK1 in MCT3E1 cells and examined the osteogenic phenotype and mitochondrial function of osteoblasts in high glucose medium. [118]Figures 4K and 4L showed that the protein levels of Runx2 and Osterix of cells in high glucose medium were significantly increased when PINK1 overexpressed, and the formation of mineralized nodules was increased; while Drp1 depletion significantly weakened this phenotype. Correspondingly, the MMP decreased and the level of mitochondrial superoxide of PINK1 overexpressed cells in high glucose medium increased significantly after Drp1 knockdown ([119]Figures 4M and 4N). Furthermore, the level of mitophagy of PINK1-overexpressed cells in high glucose medium was also significantly decreased after Drp1 was knocked down ([120]Figure 4O). Collectively, these findings indicate that the PINK1/Drp1 signaling exerts positive influence on mitochondrial function of osteoblasts and osteogenic differentiation. Overexpression of PINK1 can rescue the impaired osteogenic differentiation under high glucose medium. BMP9 reversed both osteogenic differentiation impairment and mitochondrial dysfunction induced by high glucose in MC3T3-E1 cells The aforementioned results show that high glucose level inhibits osteogenic differentiation and impairs mitochondrial function of MC3T3-E1 cells, so we are interested in whether there are molecules that can protect the mitochondrial function and promote osteogenic differentiation of MC3T3-E1 cells exposed to high glucose level. Our previous mice studies have demonstrated that bone morphogenetic protein 9 (BMP9) is effective in improving both ovariectomized (OVX) and senile osteoporosis.[121]^31^,[122]^32 In this study, an obvious less mineralized nodule formation was observed in H-Glu group compared to L-Glu group. However, the impaired osteogenic differentiation induced by high glucose treatment was partially reversed by BMP9 treatment: mineralized nodule formation was increased after BMP9 treatment relative to L-Glu group ([123]Figure 5A). Moreover, expression levels of ALP, Runx2, Osx, Ocn, and Col1a1 were significantly increased after BMP9 treatment ([124]Figure 5B). Consistently, the WB results showed that the protein levels of Osx and Runx2 were increased after BMP9 treatment ([125]Figure 5C). Figure 5. [126]Figure 5 [127]Open in a new tab BMP9 reversed both the osteoblastogenic differentiation impairment and mitochondrial dysfunction induced by high glucose in MC3T3-E1 cells The concentration of glucose in the medium of the L-Glu group was 5.5 mmol/L, and that in the medium of the H-Glu and H-Glu+BMP9 group was 25 mmol/L. The final concentration of BMP9 was 100 ng/ml. (A) Alizarin red S staining was performed in MC3T3-E1 cells after 21 days of osteogenic induction. Scale bars, 500 μm. (B) MC3T3-E1 cells were cultured in osteogenic medium for 14 days. Expressions of osteogenic markers were determined by RT-qPCR. (C) MC3T3-E1 cells were cultured in osteogenic medium for 14 days. Expressions of Runx2 and Osx were determined by WB. (D and E) MC3T3-E1 cells were cultured in osteogenic medium for 14 days. Mitochondrial oxidative capacity was measured in real time. Basal respiration, ATP-linked respiration, maximal respiration, and spare respiratory capacity were calculated by WAVE software. (F and G) MC3T3-E1 cells were cultured in osteogenic medium for 14 days. MMP of living cells was assessed by the Image-iT TMRM Reagent. Scale bars, 50 μm. (H and I) MC3T3-E1 cells were cultured in osteogenic medium for 14 days. Mitochondrial reactive oxygen species of living cells were assessed by the MitoSox Reagent. Scale bars, 50 μm. Data presented as mean ± SD. n = 3 biological replicates. One-way ANOVA was used for comparison among multiple groups. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. Ns, no significance. We used extracellular flux analysis to real-time monitor mitochondrial respiratory capacity in vitro. BMP9 treatment dramatically increased basal, ATP production-coupled respiration, maximal respiration, and the spare respiratory capacity ([128]Figures 5D and 5E). Furthermore, the result of TMRM signal showed that the MMP was significantly increased after BMP9 treatment ([129]Figures 5F and 5G). We also examined superoxide in the mitochondria of live MC3T3-E1 cells using mitochondrial superoxide indicator MitoSOX Red. In the MC3T3-E1 cells cultured in high glucose, superoxide production was higher than that in control cells, and BMP9 significantly reduced superoxide production ([130]Figures 5H and 5I). Collectively, these data indicate that BMP9 could improve mitochondrial function of MC3T3-E1 cells under high glucose medium culture. BMP9 activated PINK1/Drp1-mediated mitophagy, leading to the reversal of high glucose-induced osteogenic differentiation impairment in MC3T3-E1 cells To explore the regulatory mechanism of BMP9 on mitophagy, we inhibited the expression of PINK1 and Drp1 in cells by siRNA. After Drp1 or PINK1 depletion, the protein levels of Runx2 and Osterix of cells in H-Glu+BMP9 group were significantly decreased ([131]Figure 6A), and the formation of mineralized nodules was inhibited, as compared with only treated with BMP9 ([132]Figure 6B). Correspondingly, after PINK1 or Drp1 knockdown, the MMP decreased, and the level of mitochondrial superoxide of cells in H-Glu+BMP9 group increased significantly ([133]Figures 6C and 6D). Furthermore, the level of mitochondrial autophagy of cells in H-Glu+BMP9 group was also significantly decreased after Drp1 or PINK1 was knocked down ([134]Figure 6E). Figure 6. [135]Figure 6 [136]Open in a new tab BMP9 activated PINK1/Drp1-mediated mitophagy, leading to the reversal of high glucose-induced osteoblast differentiation impairment in MC3T3-E1 cells The concentration of glucose in the medium of the L-Glu group was 5.5 mmol/L, and that in the medium of the H-Glu, H-Glu+BMP9, Drp1 siRNA and PINK1 siRNA groups was 25 mmol/L. The final concentration of BMP9 in H-Glu+BMP9, Drp1 siRNA and PINK1 siRNA groups was 100 ng/ml. (A) MC3T3-E1 cells were cultured in osteogenic medium for 14 days after siRNA transfection. Expressions of PINK1, Drp1, Runx2, and Osx were determined by WB. (B) MC3T3-E1 cells were cultured in osteogenic medium for 21 days after siRNA transfection. Alizarin red S staining was performed. Scale bars, 500 μm. (C) MC3T3-E1 cells were cultured in osteogenic medium for 14 days after siRNA transfection. MMP of living cells was assessed by the Image-iT TMRM Reagent. Scale bars, 20 μm. (D) MC3T3-E1 cells were cultured in osteogenic medium for 14 days after siRNA transfection. Mitochondrial reactive oxygen species of living cells were assessed by the MitoSox Reagent. Scale bars, 20 μm. (E) MC3T3E1 cells were transfected with mito-Keima plasmid and siRNA and cultured in osteogenic medium for 14 days. The fluorescent dots of mito-Keima were observed by confocal microscopy. Scale bars, 20 μm. (F) Representative electron micrographs show mitochondria and mitophagosomes (arrowhead) in MC3T3-E1 cells of different groups. Scale bars, 500 nm. Data presented as mean ± SD. A t test was used for comparison between two groups. One-way ANOVA was used for comparison among multiple groups. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. Ns, no significance. We further observe the ultrastructure of mitochondria of MC3T3-E1 cells by transmission electron microscope. [137]Figure 6F showed mitochondria of L-Glu cells showed a bilayer-membrane rod-like structure with clear mitochondrial ridge. Mitochondria of cells in H-Glu group were swollen, and mitochondrial ridge structure was destroyed and disappeared. The ridge structure of some mitochondria in H-Glu+BMP9 group was blurred, and the damaged mitochondria could be seen to phagocytosed and degraded by double-layer membrane structure ([138]Figure 6F). While mitochondrial membrane structure was damaged but mitochondrial autophagy was not observed in PINK1 or Drp1 knockdown groups. Our results indicate that BMP9 effectively ameliorates osteogenic differentiation of cells under high glucose condition by activating PINK1/Drp1 signaling axis. BMP9 activated PINK1/Drp1-mediated mitophagy, leading to the reversal of high glucose-induced osteogenic differentiation impairment in BMSCs We conducted experiments to explore whether high glucose inhibits osteogenic differentiation of bone marrow mesenchymal stem cells (BMSCs) through the mitophagy-associated PINK1/Drp1 signaling pathway and whether BMP9 has a reversal effect. After Drp1 or PINK1 depletion, the protein levels of Runx2 and Osterix of cells in H-Glu+BMP9 group were significantly decreased ([139]Figure 7A), and the formation of mineralized nodules was inhibited, as compared with only treated with BMP9 ([140]Figure 7B). Correspondingly, the MMP of cells in H-Glu+BMP9 group significantly decreased after PINK1 or Drp1 knockdown ([141]Figure 7C). Furthermore, the level of mitophagy of cells in H-Glu+BMP9 group was also significantly decreased after Drp1 or PINK1 was knocked down ([142]Figure 7D). The results indicate that BMP9 can also effectively ameliorate osteogenic differentiation of BMSCs under high glucose by augmenting the activity of the PINK1/Drp1 signaling axis. Figure 7. [143]Figure 7 [144]Open in a new tab BMP9 activated PINK1/Drp1-mediated mitophagy, leading to the reversal of high glucose-induced osteoblast differentiation impairment in BMSCs The concentration of glucose in the medium of the L-Glu group was 5.5 mmol/L, and that in the medium of the H-Glu, H-Glu+BMP9, Drp1 siRNA and PINK1 siRNA groups was 25 mmol/L. The final concentration of BMP9 in H-Glu+BMP9, Drp1 siRNA and PINK1 siRNA groups was 100 ng/ml. (A) BMSCs cells were cultured in osteogenic medium for 14 days after siRNA transfection. Expressions of PINK1, Drp1, Runx2, and Osx were determined by WB. (B) BMSCs cells were cultured in osteogenic medium for 21 days after siRNA transfection. Alizarin red S staining was performed. Scale bars, 1 mm. (C) BMSCs cells were cultured in osteogenic medium for 14 days after siRNA transfection. MMP of living cells was assessed by the Image-iT TMRM Reagent. Scale bars, 20 μm. (D) BMSCs cells were transfected with mito-Keima plasmid and siRNA and cultured in osteogenic medium for 14 days. The fluorescent dots of mito-Keima were observed by confocal microscopy. Scale bars, 20 μm. BMP9 treatment in diabetic mice enhances bone mineral density and improves bone microarchitecture To further verify the effects of BMP9 on bone mineral density (BMD) and bone quality in vivo, we constructed an streptozotocin (STZ) induced diabetes mice model. The design of the animal experiment is shown in [145]Figure 8A. All diabetic mice remained hyperglycemic until the end of the study. Although serum BMP9 concentration was higher in STZ+BMP9 group, blood glucose was similar between STZ and STZ+BMP9 group. ([146]Figures 8B–8D). Figure 8. [147]Figure 8 [148]Open in a new tab Diabetic mice models induced by STZ were treated with BMP9 (A) Flow charts for the establishment of STZ mice model. (B) Fasting blood glucose concentrations were measured and compared every two weeks from week 2 to week 18. (C) Fasting blood glucose concentrations at week 18 were compared between groups. (D) Serum BMP9 levels were detected at week 18. Data presented as mean ± SD. n = 8 biological replicates. One-way ANOVA was used for comparison among multiple groups. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. Ns, no significance. As shown in [149]Figure 9A, a 3D reconstruction of the cortical bone and trabecular bone at the distal femurs clearly showed thinner cortical bone and the breakage of cancellous bone in STZ diabetic mice, and the administration of BMP9 improved the cortical bone thickness and microarchitecture of the distal femur trabecular bone. Moreover, micro-CT analysis of the distal femur showed decreased cortical BMD in the STZ mice and improved BMD in the STZ+BMP9 group (p < 0.05) ([150]Figure 9B). Other cortical bone parameters, including BV/TV and Ct.Th, were significantly decreased in STZ mice as compared with controls (p < 0.05). The Ct.Th value of STZ mice treated with BMP9 increased compared to that in the STZ group (p < 0.05) ([151]Figure 9D), while the difference in BV/TV value between the two groups was not statistically significant ([152]Figure 9C). Micro-CT analysis also showed that compared with mice in the control group, STZ mice exhibited obvious decrease in BV/TV, Tb.N, and BMD and increased Tb.Sp of the distal femur trabecular bone, while the BV/TV, Tb.N, and BMD of mice treated with BMP9 increased by 43% (p < 0.05), 30% (p < 0.05), and 29% (p < 0.01), respectively ([153]Figures 9E–9I). Consistent with the micro-CT analysis results, H&E staining of the distal femur sections confirmed a reduction in trabecular number and trabecular thickness in STZ mice compared with that of the control, and these reductions were attenuated by BMP9 treatment ([154]Figure 9J). Figure 9. [155]Figure 9 [156]Open in a new tab BMP9 promoted mitophagy markers to the control levels in diabetic mice in vivo (A) Representative images derived from micro-CT analysis, including 2D image construction of distal femur, 3D images reconstruction of trabecular bone of distal femur, and 3D image reconstruction of the femoral midshaft corticoid bone. (B–D) Quantitative analysis of the vBMD, BV/TV and Ct.Th of corticoid bone by micro-CT. (E–I) Quantitative analysis of the BV/TV, Tb.N, Tb.Sp, Tb.Th, and vBMD of trabecular by micro-CT. (J) Representative images of HE-stained decalcified femur sections. Scale bars, 200 μm. (K–M) The right femur was isolated and subjected to biomechanical properties analysis. The maximum load, elastic modulus and SMI were evaluated for each group. (N) Representative images derived from micro-CT analysis, including 2D image construction and 3D image reconstruction of L3 lumbar vertebra. (O–S) Quantitative analysis of the vBMD, BV/TV, Tb.N, Tb.Th, and Tb.Sp of L3 by micro-CT. (T and U) The levels of serum bone turnover parameters PINP and CTX-I were detected by ELISA. (V) WB analysis of protein levels of mitophagy marker in skull. (W) Expressions of ALP, Osx, Drp1, and PINK1 in skull were determined by RT-qPCR. (X) Immunofluorescence analysis of Drp1 and PINK1 expression in femur sections. Scale bars, 10 μm. (Y) Immunofluorescence analysis of PINK1 expression in femur sections. Scale bars, 20 μm. (Z) Immunofluorescence results of double labeling of osteoblasts with PINK1 and Drp1 in femur sections. Scale bars, 20 μm. Data presented as mean ± SD. n = 8 biological replicates. One-way ANOVA was used for comparison among multiple groups. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. Ns, no significance. Furthermore, compared with the control group, the maximum load and Young’s modulus of femur in STZ group were significantly decreased (p < 0.05), the maximum load and elastic modulus of femur in STZ+BMP9 group were increased by 27.1% (p < 0.01), and the value of elastic modulus was increased by 21.6% compared with STZ mice, but there was no statistical difference. The structural simulation index was not significantly different from that of STZ group ([157]Figures 9K and 9L). Additionally, we examined the parameters of lumbar vertebrae in mice. The bone density and 3D reconstruction images of the third lumbar vertebra in mice were shown in [158]Figure 9M. Compared with the normal control group, the lumbar-spine BMD in the STZ group decreased by 16.9% (p < 0.01) and increased by 14.1% (p < 0.01) in STZ+BMP9 group. The BV/TV, Tb.N, and Tb.Th values of lumbar cancellous bone in the STZ group were significantly decreased compared with the control group, and the Tb.Sp value was significantly increased in the STZ mice. Compared with the STZ group, the BV/TV and Tb.N values of lumbar cancellous bone in the STZ+BMP9 group were increased by 10.7% and 7.5%, respectively, and there was no significant difference compared with the control group. The Tb.Sp and Tb.Th values of lumbar vertebrae in STZ+BMP9 group were not significantly different from those in STZ group ([159]Figures 9N–9S). We then examined bone turnover markers in mice. Compared with the normal control group, the serum procollagen I N-terminal propeptide (PINP) level of mice in STZ group was significantly decreased (p < 0.01), but was significantly increased in STZ+BMP9 group compared with the STZ mice group (p < 0.01) ([160]Figure 9T). There was no significant difference in serum C-terminal telopeptide of type 1 collagen (CTX-I) concentration between three groups ([161]Figure 9U). These results suggest that BMP9 has the ability to ameliorate diabetic osteoporosis. BMP9 promoted mitophagy markers to the control levels in diabetic mice in vivo The favorable bone effect of BMP9 in diabetic STZ mice was investigated mechanistically. The protein from parietal bone in mice was quantified by WB. The results demonstrated that the expressions of mitophagy related proteins PINK1 and Drp1 were significantly increased in mice treated with BMP9, and the ratio of autophagy marker proteins LC3A/BⅡ and LC3A/BⅠ were significantly increased in STZ+BMP9 group ([162]Figure 9V). Moreover, qPCR results showed expression levels of ALP, Osx, Drp1, and PINK1 were significantly increased in BMP9 treatment group ([163]Figure 9W). We further stained the bone sections of the femur in mice by immunofluorescence. As shown in [164]Figure 9X, the expression of Drp1 protein in osteoblasts located on the bone trabecular surface of mice in STZ group was significantly decreased compared with that of control mice, and was significantly increased after BMP9 administration, which was consistent with the variation trend of PINK1 in [165]Figure 9Y. In order to further confirm the aforementioned results, double-label immunofluorescence staining was performed on the femur section. Osx was used to label osteoblasts and the expression of PINK1 in osteoblasts was detected. Compared with the control group, the expression of PINK1 in osteoblasts in the STZ group was significantly decreased. The expression of PINK1 in osteoblasts in STZ+BMP9 group was significantly higher than that in STZ group ([166]Figure 9Z). These results suggested that BMP9 can promote the expression of mitophagy protein in bone tissue of diabetic mice, promoting mitophagy may be an effective means to improve the bone mass and quality in diabetic mice. Discussion In this study, we revealed that the suppressed PINK1/Drp1-mediated mitophagy is a molecular pathway responsible for impaired osteogenic differentiation in diabetes. Activation of mitophagy by BMP9 can mitigate osteoporosis induced by hyperglycemia. Osteoblast-mediated bone formation requires enough ATP, which is generated through OXPHOS and glycolysis.[167]^33 Our research found that high glucose condition lasting to the late stage of differentiation could induce the decrease in basal and maximal respiration, ATP synthesis, and glycolytic rate of osteoblast. We also found mitochondrial dysfunctions, including the increased ROS, and decreased MMP in osteoblast cultured under high glucose level. Our RNA sequencing analysis suggested that impaired mitophagy might be the molecular mechanism for high glucose-induced inhibition of osteogenic differentiation. Upon mitochondrial stress, mitophagy is activated and can selectively eliminate the damaged mitochondria, and thus stabilizing cellular homeostasis.[168]^34 Autophagy and mitophagy play a vital role in regulating osteoblasts’ proliferation and differentiation. Thus, we further explored the role of mitophagy in high glucose-induced inhibition of osteogenic differentiation and its related signaling pathways. PINK1 is an important promotor of mitochondrial autophagy. When MMP declines, the Ser228 and Ser402 groups of PINK1 will be phosphorylated,[169]^35 upon activation of PINK1 phosphorylation, ubiquitin on the mitochondrial outer membrane can be phosphorylated, then initialing mitophagy. We found that siRNA knockdown of PINK1 inhibited the ameliorative effect of BMP9 to impaired mitophagy and osteogenic differentiation in high glucose conditions, while overexpression of PINK1 can reverse the inhibitory effect of high glucose level on osteogenic differentiation. Mitophagy and mitochondrial dynamics are close related. Drp1, a GTPase of the dynein superfamily, is the most important molecule in promoting mitochondrial fission. After Drp1 is recruited into the outer membrane of mitochondria, it will form helical oligomers, thereby inducing contraction and severing of the mitochondrial membrane.[170]^36 Our RNA sequencing analysis showed the decreased expression level of Drp1 in osteoblasts cultured in high glucose medium, suggesting the potential interplay between Drp1 and PINK1-induced mitophagy in impaired osteogenic differentiation caused by hyperglycemia. Then, we conducted a series of experiments to address the regulatory effect between Drp1 and PINK1. We found that the expression of Drp1 was downregulated after the knockdown of PINK1. In contrast, the expression of PINK1 did not change following the inhibition of Drp1, suggesting Drp1 as a downstream signaling molecule of PINK1. Our results also demonstrated that Drp1 knockdown can inhibit osteogenic differentiation and mitophagy and impair mitochondrial function. PINK1 was shown to directly phosphorylate S616 of Drp1 to regulate mitochondrial fission,[171]^37 which was consistent with our results. Taken together, we hypothesized that PINK1/Drp1-mediated mitophagy plays a critical role in osteogenic differentiation, and its suppression contributes to hyperglycemia-induced impaired osteoblastogenesis. To further verify this hypothesis, we then tried to find a molecule to activate this signaling pathway to maintain mitophagy of osteoblast in hyperglycemia and treat diabetic osteoporosis in vivo. BMP9, a member of the transforming growth factor beta (TGF-β) superfamily, regulates various physiological processes such as angiogenesis, glucose metabolism, neurogenesis, and tumorigenesis. Our previous research found that BMP9 promotes osteoblastogenesis by upregulating LGR6 and activating the Wnt/β-catenin pathway and suppresses osteoclast differentiation by inhibiting the nuclear factor-κB (NF-κB) pathway.[172]^31 Our team’s other research also found that BMP9 reduces expression of senescent genes and alleviates senescence-associated secretory phenotype in bone microenvironment, simultaneously increasing bone mass and bone biomechanical properties in aged mice.[173]^32 As a continuation of work, we investigated the regulatory role of BMP9 on mitophagy in osteoblast and its protective effect on diabetic osteoporosis in this study. In vitro, we observed that BMP9 improved the osteogenic differentiation of MC3T3-E1 cells under high glucose conditions. This protective effect of BMP9 on osteoblasts could be mediated by the regulation of mitophagy through its capacity to upregulate PINK1/Drp1 expression. Whereas knockdowns of PINK1 and Drp1 can abolish the bone protective function of BMP9 under high glucose conditions. Mitophagy dysfunction participates in pathogenesis diabetes and diabetes-related complications that include β-cell damage,[174]^38 glucose intolerance,[175]^39 cardiomyopathy,[176]^40 insulin resistance,[177]^41 and diabetes-associated cognitive dysfunction.[178]^42 Previous study has found that non-imprinted protein 2 in the Prader-Willi/Angelman syndrome region (NIPA2), a mitophagy-related molecule, was decreased in the bone tissue of diabetic mice.[179]^43 Our research found that PINK1/Drp1-mediated mitophagy was inhibited in the bone of diabetic mice and its activation by BMP9 ameliorated diabetic osteoporosis, indicating that PINK1/Drp1-mediated mitophagy may be a therapeutic target for diabetic osteoporosis. BMP9 is known to be a powerful promoter of osteogenic differentiation. Our research found that BMP9 rescued diabetic osteoporosis by addressing various suboptimal bone phenotypes, including promoting bone formation, improving bone density, enhancing the microstructure of cancellous and cortical bones, and ameliorating femur biomechanics. Notably, all these favorable bone effects of BMP9 in diabetes are not related to changes in blood glucose. These results suggest that BMP9 may act directly and specifically on bone tissue in diabetic mice, and this advantage makes BMP9 a potential therapeutic drug for diabetic osteoporosis. In conclusion, our findings suggest the suppressed PINK1/Drp1-mediated mitophagy is responsible for the impairment of osteogenic differentiation in diabetes. BMP9 facilitates osteogenic differentiation and maturation under diabetic conditions through activating PINK1/Drp1-mediated mitophagy, ultimately exerting a bone-protective effect in diabetic mice. Further research on PINK1/Drp1-mitophagy pathway and its agonists, such as BMP9, is a promising way to explore the potential pathogenesis and therapeutic strategy for osteoporosis in diabetes. Limitations of the study This study had several limitations. Firstly, the precise mechanism through which BMP9 activates PINK1/Drp1 signaling pathway requires further elucidation. Secondly, the late onset of skeletal phenotype in diabetic mice led to a limited number of extracted skull primary cells or mesenchymal stem cells with diminished activity. Thirdly, we did not use osteoblast conditional gene knockout mouse models to verify the impaired PINK1/Drp1-mediated mitophagy as a pathogenesis of diabetic osteoporosis. Resource availability Lead contact Further information and requests for the resources and reagents should be directed to and will be fulfilled by the lead contact, Hong-yan Zhao (hyanzhao@163.com). Materials availability This study did not generate any new unique reagents and components. The siRNA oligonucleotide sequences were provided in [180]Table S1, and primer sequences of target genes were provided in [181]Table S2. Data and code availability * • RNA sequencing data have been deposited at Zenodo and are publicly available as of the date of publication. Accession numbers are listed in the [182]key resources table. * • This paper does not report the original code. * • Any additional information required to reanalyze the data reported in this work paper is available from the [183]lead contact upon request. Acknowledgments