Rheb1 deletion impairs osteoclast function, resulting in growth restriction and delayed bone fracture healing

Given the critical role of Rheb1 in nutrient sensing, we generated osteoclast Rheb1 KO mice (Rheb1OC) (Fig. 1a) to investigate its role in osteoclast function. Analysis of skeletal preparations showed that Rheb1 deletion led to suppressed growth during the postnatal periods prior to weaning (Fig. 1b), with significant reductions in both body weight and length at 2 and 6 weeks of age (Fig. 1c, d). Toluidine blue staining of the growth plate revealed a narrowed height in Rheb1OC mice (Fig. S1a, c), indicating that osteoclast Rheb1 is crucial for bone development before weaning.

Fig. 1

Rheb1 deletion impairs osteoclast function, resulting in growth restriction and delayed bone fracture healing. a Immunoblots of the Rheb1 level and mTORC1 activity (S6 and phosphorylated S6) in osteoclasts induced from Rheb1OC or control mice BMDMs, with ɑ-tubulin as an internal control (n = 3 technical replicates from six biological replicates for each strain). Osteoclast formation was induced for 4 days. b Skeletal preparation of Rheb1OC or control mice at newborn (P0), aged 1 week (P7) and aged 2 weeks (P14). c, d Body weight and length of Rheb1OC mice aged 2 weeks (body weight n = 12 for control, n = 11 for Rheb1OC, body length n = 9 for control, n = 8 for Rheb1OC) and 6 weeks (n=6) compared to littermate controls. Unpaired t test, *P < 0.05, ****P < 0.000 1. e Immunofluorescence staining images of type I or type II collagen deposition in 2-week-old Rheb1OC mice compared to littermate controls (scale bar, 100 µm and 50 µm). f Von Kossa staining images of calcified bone matrix in 2-week-old Rheb1OC mice compared to littermate controls (scale bar, 100 µm and 50 µm). g, h Serum CTX-I levels (n = 5 for control, n = 6 for Rheb1OC) and P1NP (n = 12 for control, n = 11 for Rheb1OC) in 2-week-old Rheb1OC mice compared to littermate controls. Unpaired t test, *P < 0.05; n.s., no significance difference. i Toluidine blue staining of bone sections from 7 or 14 days after the on-set in drill-hole modeled mice. (n = 3 technical replicates from five biological replicates for each strain. Scale bar, 50 µm. Callus: black dashed line). j–l ELISA detection analysis of CTX-I (n = 9 for control, n = 10 for Rheb1OC), TRAP5b (n = 9 for control, n = 9 for Rheb1OC) or MMP-9 (n = 15 for control, n = 12 for Rheb1OC) levels in the supernatants from Rheb1-deficient or control osteoclasts cultured on bone slices. Multinucleated osteoclasts were maintained in fresh medium for 2 days before sample collection. Unpaired t test, *P < 0.05, **P < 0.01; n.s., no significance difference. All data are presented as mean ± SEM

To understand the mechanism behind the growth restriction observed in Rheb1OC mice, we first examined the in vivo number of osteoclasts. TRAP staining of the metaphyseal spongy bone in 2-week-old mice revealed no significant difference in mature osteoclast numbers between Rheb1OC and control animals (Fig. S1b, d). However, osteoclast function is required for degrading the cartilaginous matrix, leading to the formation of cartilaginous trabeculae and the subsequent deposition of calcified bone matrix. This process is essential for longitudinal bone growth.2 We compared the bony and cartilage trabecular networks between Rheb1OC and control mice. Type I and type II collagens, major components of the bone extracellular matrix,37 showed increased deposition in 2-week-old Rheb1OC mice (Fig. 1e and Fig. S1e, f). Von Kossa staining also indicated increased bone mineralization in Rheb1OC mice (Fig. 1f and Fig. S1g). Notably, CTX-I levels, a marker of osteoclast activity in degrading type I collagen, were significantly reduced in Rheb1OC mice (Fig. 1g), while the serum marker of bone formation, procollagen type 1 N-terminal pro-peptide (P1NP), remained unaffected (Fig. 1h). These results suggest dysregulated extracellular matrix decomposition rather than synthesis, leading to altered bone matrix composition in Rheb1OC mice. Consistently, bone callus removal was impaired, and bone defect healing was delayed in Rheb1OC mice (Fig. 1i).

To further elucidate the impact of Rheb1 on osteoclast function, we measured the levels of various resorption degradation products released by osteoclasts. A significant decrease in CTX-I, a fragment released by CTSK cleavage, was detected in the culture supernatants of Rheb1OC osteoclasts cultured on bone slices (Fig. 1j). In contrast, the level of TRAP5b,38 a product indicative of tartrate-resistant acid phosphatase (TRAP) activity, remained comparable to that of control osteoclasts (Fig. 1k). Additionally, the level and activity of matrix metalloproteinase-9 (MMP-9), a type IV collagenase highly expressed in osteoclasts and crucial for cartilage degradation,39 were unchanged in Rheb1-deleted osteoclasts (Fig. 1l and Fig. S1h). These results highlight the specific role of Rheb1 in regulating osteoclast function and its importance in the degradation of bone extracellular matrix components.

Rheb1 deletion impairs osteoclast function via suppressing CTSK production in an mTORC1-independent manner

The decreased CTX-I level indicates a potential defect in osteoclast resorption function. To elucidate Rheb1’s role in osteoclast resorption, we performed RNA sequencing on mature osteoclasts generated from both Rheb1OC and control mice bone marrow-derived monocytes/macrophages (BMDMs). Our analysis revealed that CTSK, a key enzyme for collagen degradation in osteoclasts,19 was significantly downregulated in Rheb1-deficient multinucleated osteoclasts (Fig. 2a, b). Further examination showed that CTSK protein levels, including both the proenzyme and mature forms,40 were markedly reduced in Rheb1-deficient osteoclasts cultured ex vivo (Fig. 2c–e). In contrast, the levels of other resorption enzymes, such as Tartrate-resistant acid phosphatase type 5 (ACP5), MMP-9 and Cathepsin B (CTSB), remained unaffected. This finding aligns with the observed significant reduction of CTSK in osteoclasts on cancellous bone surfaces in Rheb1OC mice (Fig. S2a–c). Consistently, increased CTSK mRNA levels were detected in Rheb1-overexpressing osteoclasts (Fig. 2f). To confirm that reduced CTSK production underlies the defective osteoclast function, we incubated cellular contents from control and Rheb1-deficient osteoclasts with type I collagen ex vivo. As anticipated, when compared with the control group, Rheb1-deficient osteoclasts exhibited an impaired ability to degrade type I collagen under equivalent collagen concentrations. Supplementation with active murine CTSK enzyme partially rescued the degradation defect observed with Rheb1 deletion (Fig. 2g). Given that procathepsin K processing depends on lysosome activity41,42 and mTORC1 is known to regulate lysosome function,43 we found a substantial reduction in procathepsin K in Rheb1-deficient osteoclasts with inhibited lysosome biogenesis (Fig. 2h), suggesting that reduced procathepsin K level is not due to impaired CTSK processing. Additionally, CTSK has a short half-life (~60 min) in both Rheb1-deficient and control osteoclasts (Fig. S2d), indicating that Rheb1 loss does not affect CTSK protein degradation. These findings collectively suggest that Rheb1 specifically regulates procathepsin K production in osteoclasts.

Fig. 2

Rheb1 deletion impairs osteoclast function via suppressing CTSK production in an mTORC1-independent manner. a Number of differentially expressed genes (DEGs) in the indicated osteoclasts in RNA-seq data. b Volcano plot of RNA-seq data to show CTSK is a significantly down-regulated gene in Rheb1-deficient osteoclasts compared to controls (n = 3). c Immunoblot analysis of CTSK, MMP-9, ACP5 and CTSB protein levels in Rheb1-deficient or control osteoclasts. Approximately 40 kD indicates the pre-proenzyme form, while approximately 35 kDa indicates the active CTSK form. Cells were collected at times when a large majority of multinuclear cells had formed (n = 3 technical replicates from more than three biological replicates for each strain). d, e Immunofluorescence staining and analysis of the CTSK level in multinucleated Rheb1-deficient or control osteoclasts (scale bar, 20 µm, n = 5 for control, n = 10 for Rheb1OC). Unpaired t test. **P < 0.01. f Q-PCR analysis of CTSK and Rheb1 relative mRNA expression in Rheb1-deficient or Rheb1-overexpressing (Rheb1-OE) osteoclasts, compared to the negative controls, respectively (n = 4). g Immunoblotting analysis to detect type I collagen (0.4 mg/mL) ex vivo degradation in cellular contents from Rheb1-deficient or control osteoclasts. Active CTSK (4E-04 mg/mL) or type I collagenase (1.2 mg/mL) administration as a positive control (n = 3 for control, n = 4 for Rheb1OC, n = 3 for Rheb1OC + active CTSK). One-way ANOVA. *P < 0.05. h Immunoblots of the levels of the pre-proenzyme and active CTSK forms in Rheb1-deficient or control osteoclasts, with or without lysosome inhibitor bafilomycin A1 (10 ng/mL) treatment (n = 3 technical replicates from more than three biological replicates). i–j Staining for TRAP and analysis of ex vivo osteoclast differentiation in Rheb1OC or littermate control mice (scale bar, 250 µm); osteoclast differentiation was induced by RANKL and MCSF for 4 days (n = 8). k Q-PCR analysis to determine relative CTSK, ACP5, MMP-9 and Nfatc1 mRNA expression levels in Rheb1-deficient or Raptor-deficient osteoclasts compared to their controls (n = 4). Unpaired t test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.000 1; n.s., no significant difference. l Immunoblotting for the protein levels of the CTSK pre-proenzyme and active forms in osteoclasts which were treated with 10 nmol/L rapamycin for 0, 2, 6 or 10 h. A decreased phosphorylated-S6 level indicates mTORC1 inhibition. All data are presented as mean ± SEM

Interestingly, ex vivo assays revealed that osteoclast formation was enhanced in the absence of Rheb1 (Fig. 2i, j), similar to observation in mTORC1-inhibited osteoclasts,44,45 suggesting that Rheb1 may have a role in mTORC1-dependent regulation of osteoclast differentiation. However, upon analyzing the expression levels of several osteoclast-specific genes, we found that CTSK mRNA levels were significantly decreased in Rheb1-deficient osteoclasts, while the expression levels of ACP5 and MMP-9 remained unaffected (Fig. 2k). Furthermore, the increased nuclear factor of activated T-cells, cytoplasmic 1 (Nfatc1) level was consistent with the enhanced osteoclast formation observed in Rheb1-deficient osteoclasts. In contrast, the expression levels of these genes were increased in Raptor-deficient osteoclasts (Fig. 2k). These findings suggest that insufficient CTSK might be a marker of osteoclast function defects rather than a determinant of differentiation suppression. Consistent with this, pharmacological inhibition of mTORC1 using the allosteric inhibitor rapamycin significantly reduced the phosphorylation of ribosomal protein S6, a downstream substrate and major effector of mTORC1,46 but did not alter CTSK proenzyme and mature form levels in osteoclasts (Fig. 2l). Taken together, these data demonstrate a crucial mTORC1-independent role of Rheb1 in regulating CTSK.

Rheb1 deletion mainly impairs osteoclast mitochondrial respiration

Notably, our previous findings demonstrated that Raptor-deficient osteoclasts caused osteoporosis in adult mice.47 In contrast, micro-computed tomography (micro-CT) analysis of adult Rheb1OC mice showed a significant reduction in bone mineral content (BMC) without affecting bone microstructure parameters compared to littermate controls (Fig. S3a–c). This discrepancy suggests a specific role for Rheb1 in osteoclast function. Recent studies indicate that Rheb1 localizes in the mitochondrial matrix or outer membrane, regulating mitochondrial respiration.48,49,50 We propose that the metabolic state of osteoclasts likely determine their distinct resorption behaviors. Gene set enrichment analysis (GSEA) revealed that OxPhos is one of the most significantly down-regulated pathways in Rheb1-deficient osteoclasts (Fig. S4a, b). In contrast, glycolysis is inhibited in Raptor-deficient osteoclasts, while OxPhos remains largely unaffected (Fig. S4c, d). These findings imply a mTORC1-independent regulatory role for Rheb1 in mitochondrial function.

To confirm Rheb1’s metabolic role in osteoclast function, we measured energy metabolic pathways in Rheb1-deficient osteoclasts using the Seahorse XF24 Extracellular Flux analyzer. Rheb1-deficient osteoclasts showed a significantly reduced oxygen consumption rate (OCR) (Fig. 3a, c), indicating suppressed mitochondrial respiration. However, the basal extracellular acidification rate (ECAR), an indicator of aerobic glycolysis, was unaffected (Fig. 3b, d). The reduced maximal ECAR level suggests potential involvement of mTORC1.51,52 Consistently, mitochondrial ATP production rate was significantly decreased, while glycolytic ATP production rate remained similar to controls, resulting in a reduced total ATP production rate (Fig. 3e). Despite the inhibition of mitochondrial respiration, Rheb1 deletion did not induce oxidative stress (Fig. S5a) or significantly elevate mitophagy (Fig. S5b, c) in osteoclasts.

Fig. 3

Rheb1 deletion mainly impairs osteoclast mitochondrial respiration. a Mitochondrial OCR in the routine, noncoupled (oligomycin), and maximal (FCCP) steady state in Rheb1-deficient osteoclasts or control cells were determined by the real-time Seahorse assay. b Extracellular acidification rate (ECAR) in Rheb1-deficient osteoclasts or their respective control cells was determined by the real-time Seahorse assay. c The basal respiration and FCCP-induced maximal respiration in the OCR shown in a were compared between Rheb1-deficient osteoclasts and their controls (n = 26). Unpaired t test. ****P < 0.000 1. d The basal and maximal glycolysis levels in the ECAR shown in b were compared between Rheb1-deficient osteoclasts and their controls (n = 26). Unpaired t test. *P < 0.05; n.s., no significant difference. e The mitochondrial and glycolytic ATP production rates were compared between Rheb1-deficient osteoclasts and their respective control cells (n = 26). Unpaired t test. ****P < 0.000 1; n.s., no significant difference. f Schematic of the type I collagen-based ex vivo model. g OCR in Rheb1-deficient osteoclasts adherent on type I collagen-coated substrates was determined by the Seahorse assay in real time, which was compared to control cells. h The basal and maximal respiration in the OCR shown in g (n = 4). i Schematic of the type II collagen-based ex vivo model. Unpaired t test. **P < 0.01, ***P< 0.001. j OCR in Rheb1-deficient osteoclast adherent on type II collagen-coated substrates was determined by the Seahorse assay in real time, which was compared to control cells. k The basal and maximal respiration in the OCR shown in j (n = 5 for control, n = 8 for Rheb1OC). Unpaired t test. **P < 0.01, ****P < 0.000 1. All data are presented as mean ± SEM

During endochondral ossification, type I collagen is associated with bone matrix production, while type II collagen is associated with cartilage formation.53 Since approximately 90% of the organic matrix of bone is composed of collagens,19 we investigated whether Rheb1 modulates the metabolic status of osteoclasts adherent to type I (Fig. 3f) or type II collagen (Fig. 3i). We found that both OCR (Fig. 3g, h) and ECAR (Fig. S5d, e) were suppressed in Rheb1-deficient osteoclasts adherent to type I collagen compared to controls. In contrast, while OCR was also suppressed in Rheb1-deficient osteoclasts on type II collagen (Fig. 3j, k), ECAR remained unaffected (Fig. S5f, g). The varying ECAR levels in Rheb1-deficient osteoclasts adherent to different collagen types suggest distinct underlying mechanisms. However, these findings indicate a specific regulatory role for Rheb1 in osteoclast mitochondrial respiration and underscore the importance of mitochondrial function in osteoclast behavior.

Mitochondrial respiration is critical to osteoclast procathepsin K production

Given that Rheb1 deletion leads to metabolic changes and reduced CTSK in osteoclasts, and that CTSK degrades both type I and type II collagen,54 we aimed to determine whether the osteoclast metabolic state depends on the matrix composition. To evaluate whether the osteoclast metabolic state is influenced by the adherence substrate, we established ex vivo models with osteoclasts grown on collagen-coating substrates or bone slices. Fluorescence imaging with Mito-Tracker showed a significant increase in mitochondrial content in mature osteoclasts grown on type I or type II collagen matrices, compared to those grown on bone slices or plastic plates (Fig. 4a, b), suggesting enhanced mitochondrial respiration to support osteoclast activity on collagen-rich matrices. Next, we investigated the metabolic changes in osteoclasts adhering to different substrates. Both basal and maximal OCRs were significantly enhanced in osteoclasts grown on collagen I- or II-coated substrates compared to plastic plates (Fig. 4c, d). Basal and maximal ECARs were unchanged in osteoclasts on collagen I but increased in those on collagen II (Fig. S6a, b), suggesting collagen augments both OCR and ECAR. Notably, when osteoclasts adhered to type I or type II collagen, CTSK protein levels in the perinuclear area increased in response to fresh culture medium (Fig. 4e), indicating a heightened requirement for newly synthesized CTSK proenzyme in osteoclasts on collagen-containing substrates.

Fig. 4

Mitochondrial respiration is critical to osteoclast procathepsin K production. a Labeling of mitochondria with Mito Tracker Dyes in osteoclasts adherent on type I collagen-coated or type II collagen-coated interfaces, bone slices or plastic interfaces without collagen-coating, respectively (scale bar, 10 µm). b Statistical analysis of immunofluorescence signal density in a (n = 5 for type I collagen-coating, n = 4 for type II collagen-coating, n = 5 for bone slice, n = 4 for plastic). One-way ANOVA. *P < 0.05, **P < 0.01; n.s., no significant difference. c OCR in osteoclasts adherent on type I or type II collagen-coated substrates compared to cells culture on plastic substrates were determined by the Seahorse assay in real-time. d The basal respiration and FCCP-induced maximal respiration in the OCR shown in c were compared between osteoclasts adherent on collagen-coated substrates and plastic substrates (n = 3 for collagen-coating, n = 6 for plastic). One-way ANOVA. *P < 0.05, **P < 0.01, ****P < 0.000 1. e Immunofluorescence staining analysis to show CTSK protein abundance in osteoclasts adherent on different substrates (scale bar, 100 Pixel, n = 3 for plastic, n = 4 for type I collagen-coating, n = 3 for type II collagen-coating). One-way ANOVA. *P < 0.05. f Immunofluorescence staining to show the CTSK expression level in osteoclasts that were treated with oligomycin A (10 nmol/L), rotenone (100 nmol/L), FCCP (1 μmol/L), 2DG (1 μmol/L) or 3BrPA (1 μmol/L) for 6 h (scale bar, 100 µm). g Immunoblots of the pre-proenzyme and active CTSK levels in osteoclasts treated with 2DG (1 μmol/L), FCCP (1 μmol/L), rotenone (100 nmol/L) or 3-BrPA (1 μmol/L) for 10 h. The MMP-9 and CTSB expression levels were also detected. DMSO-treated osteoclasts were used as controls. h, i Immunoblots of the levels of the CTSK pre-proenzyme and active forms in osteoclasts treated with 10 nmol/L oligomycin A or 100 nmol/L rotenone, with or without lysosome inhibitor bafilomycin A1 (10 ng/mL) treatment for 4 h. Bafilomycin A1 administration was performed followed 2 h later by treatment with OxPhos inhibitors. All data are presented as mean ± SEM

Glycolysis is generally considered the preferred energetic pathway for bone resorption in osteoclasts,11 while mitochondrial OxPhos is the main energy source for osteoclast differentiation.55 To determine the metabolic pathway supporting CTSK production, we measured CTSK levels in osteoclasts treated with glycolysis or mitochondrial stress components (2-deoxy-D-glucose (2DG), 3-bromopyruvate (3BrPA), carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP), rotenone and oligomycin A). Surprisingly, CTSK expression levels were not significantly affected by 3BrPA (Fig. S6c) or 2DG treatment (Fig. S6d) but were profoundly reduced by rotenone or oligomycin A (Fig. 4f and Fig. S6e, f), suggesting that glycolysis is not the preferred bioenergetics pathway for CTSK production. Consistently, mitochondrial inhibitor reduced CTSK levels in osteoclasts attached to collagen I or II substrates (Fig. S6g, h). Notably, rotenone treatment specifically reduced CTSK levels without affecting other resorption enzymes like MMP-9 and CTSB (Fig. 4g), indicating specific regulation by mitochondrial respiration. Oligomycin A (Fig. S6i) or rotenone (Fig. S6j) treatment significantly reduced both procathepsin K and active CTSK levels in a time-dependent manner, whereas FCCP likely suppressed CTSK processing without similar reductions (Fig. 4g). The different effects of FCCP and oligomycin/rotenone suggest that mitochondrial uncoupling is not crucial for procathepsin K production. Consistent with findings in Rheb1-deficient osteoclasts, oligomycin or rotenone treatment suppressed CTSK proenzyme production (Fig. 4h, i). These ex vivo data indicate that osteoclasts have a particular preference for the mitochondrial metabolic pathway when producing CTSK.

Acetyl-CoA availability is essential in supporting osteoclastic CTSK production

Rotenone and oligomycin both reduce the OCR and metabolite availability,56,57 while FCCP acts as a potent uncoupler of mitochondrial OxPhos, thereby elevating the OCR.58 The inhibition of CTSK production under mitochondrial stress implies that metabolites generated from mitochondrial respiration are crucial for osteoclast CTSK production. Examination of central metabolites revealed a decreased pyruvate content in Rheb1-deficient osteoclasts (Fig. 5a), while lactate levels remained similar to controls (Fig. 5b). These results suggest that the glycolysis-tricarboxylic acid cycle (TCA) cycle flux is attenuated to a certain extent in Rheb1-deficient osteoclasts. A significantly lower level of acetyl-CoA was detected in Rheb1-deficient osteoclasts compared to controls (Fig. 5c), indicating down-regulation of acetyl-CoA generation due to Rheb1 deficiency. Given that pyruvate and glutamine are two major carbon sources entering the TCA cycle, we supplemented Rheb1-deficient osteoclasts with pyruvic acid or glutamine to identify what stimulates CTSK production. To exclude nutrient effects on osteoclast differentiation, pyruvate or glutamine was supplied after multinuclear osteoclasts had formed (Fig. S7a). Both pyruvate (Fig. S7b) and glutamine (Fig. S7c) significantly elevated the expression of proenzyme and active CTSK in multinucleated osteoclasts. However, these treatments did not restore CTSK level in Rheb1-deficient osteoclasts when compared to controls (Fig. 5d). Similarly, the addition of dimethyl-α-ketoglutarate (DMKG), a key intermediate in glutamine metabolism, failed in restoring CTSK levels in Rheb1-deficient osteoclasts (Fig. 5d and Fig. S7d). Interestingly, the addition of acetate (Fig. S7e, f) or citrate (Fig. S7g, h) increased CTSK protein levels, but not MMP-9 or ACP5, in mature osteoclasts. Administering acetate or citrate to Rheb1-deficient multinucleated osteoclasts increased both procathepisn K and mature CTSK protein levels (Fig. 5e). Given that acetate and citrate can be converted to acetyl-CoA, incubation with citrate restored the spare respiration capacity (Fig. 5f), while acetate restored cellular acetyl-CoA levels in Rheb1-deficient osteoclast (Fig. 5g). The transcriptional regulation of enzymes involved in acetyl-CoA conversion remained unaffected by Rheb1 deletion (Fig. S7i). These findings indicate that mitochondrial malfunction resulting from Rheb1 deletion primarily impairs acetyl-CoA availability (Fig. 5h). To validate the centrality of acetyl-CoA availability for CTSK production, we examined CTSK levels in osteoclasts incubated with acetyl-CoA synthase 2 (ACCS2) or ATP citrate lyase (ACLY) inhibitors. Notably, administration of ACCS2 inhibitor reduced both procathepsin K and mature CTSK production (Fig. 5i). Similarly, ACLY inhibitor SB 204990 also decreased CTSK levels (Fig. 5j). These findings indicate that mitochondrial respiration is essential for CTSK expression in osteoclasts, potentially through providing acetyl-CoA via the TCA cycle.

Fig. 5

Acetyl-CoA availability is essential in supporting osteoclastic CTSK production. a Intracellular pyruvate level was determined and analyzed in Rheb1-deficient osteoclasts compared to controls (n = 16 for control, n = 15 for Rheb1OC). Unpaired t test. *P < 0.05. b Intracellular lactate level was determined and analyzed in Rheb1-deficient osteoclasts compared to controls (n = 8 for control, n = 10 for Rheb1OC). Unpaired t test, n.s., no significant difference. c Intracellular acetyl-CoA level was measured and analyzed in Rheb1-deficient osteoclasts compared to controls (n = 10). Unpaired t test. *P < 0.05. d Immunoblots of the pre-proenzyme and active CTSK protein levels in Rheb1-deficient or respective control osteoclasts grown in media containing additional 10 mmol/L pyruvate, 10 mmol/L glutamine or 100 μmol/L DMKG for 6 h, respectively. e Immunoblots of pre-proenzyme and active CTSK protein levels in Rheb1-deficient or respective control osteoclasts incubated with 1 mmol/L acetate or 100 μmol/L citrate for 6 h, respectively. f The spare respiration capacity was compared between Rheb1-deficient osteoclasts and their controls, with or without citrate incubation for 6 h (n = 3 or 4). One-way ANOVA. *P < 0.05. g Intracellular acetyl-CoA levels were measured and analyzed in Rheb1-deficient osteoclasts compared to controls, with or without 1 mmol/L acetate incubation for 6 h (n = 11). One-way ANOVA. *P < 0.05, n.s., no significant difference. h Schematic of acetyl-CoA availability defect in Rheb1-deficient osteoclasts. i Immunoblots of the pre-proenzyme and active CTSK protein levels in osteoclasts incubated with or without different doses of ACCS2 inhibitor for 6 h. j Immunoblots of the pre-proenzyme and active CTSK protein levels in osteoclasts incubated with or without different doses of ACLY inhibitor SB 204990 for 6 h. All data are presented as mean ± SEM

Acetyl-CoA availability influences histone acetylation, thereby regulating transcription.59 However, high concentrations of acetate significantly reduced both CTSK mRNA and protein levels in osteoclasts (Fig. S8a, b), while increasing global histone H3 acetylation (Fig. S8c). These data align previous studies showing that global acetylation suppresses osteoclast formation.60 Given the overall reduction in acetyl-CoA levels in Rheb1-deficient osteoclasts and the restoration of CTSK protein levels with specific acetate concentrations, it is likely that the insufficient CTSK levels in Rheb1-deficient osteoclasts are due to a decreased nucleocytosolic pool of acetyl-CoA.

Mitochondrial respiration suppression in Rheb1-deficient osteoclasts might also be due to decreased levels of NAD-dependent protein deacetylase sirtuin-3 (SirT3), which enhances mitochondrial enzyme activity.61,62,63 Consistent with our hypothesis, mitochondrial SirT3 levels decreased in both Rheb1-deficient osteoclasts (Fig. S9a) and tumor necrosis factor ligand superfamily member 11 (RANKL)-incubated Rheb1 KO Raw264.7 cells (Fig. S9b, c), with additional acetate partially reversing this reduction. However, SirT3 levels were not reduced in Rheb1 KO Raw264.7 cells without RANKL stimulation (Fig. S9d), indicating a specific regulation by Rheb1 on SirT3 in osteoclasts, distinct from their precursors. These results suggest that acetate helps maintain osteoclast mitochondrial activity when Rheb1 is absent.

In conclusion, our data suggest that acetyl-CoA availability is crucial for procathepisn K production in osteoclasts and is regulated by Rheb1. The ability of osteoclasts to generate acetyl-CoA from internal or external sources is vital for maintaining CTSK levels.

Ethanol intake causes a poor fracture healing due to increased CTSK production

Acetate, a byproduct of hepatic alcohol metabolism, elevates circulating acetate levels following alcohol consumption.64,65 To investigate whether acetate affects CTSK during endochondral ossification in the developing skeleton, we modeled alcohol consumption with equivalent caloric intake in postnatal day 7 (P7) mice (Fig. 6a). Our observations indicated that intake of 5% and 1% ethanol did not significantly impact body weight (Fig. S10a, b). However, both procathepsin K and its active form were abundantly expressed in the epiphyseal regions of all alcohol-fed mice, while other osteoclast matrix-degrading enzymes were not (Fig. 6b). This occurred despite no changes in serum CTX-I and P1NP levels (Fig. S10c). In adult mice (Fig. 6c), CTSK levels were also specifically upregulated (Fig. 6d). To further confirm the sensitivity of skeletal CTSK production to acetyl-CoA availability in adults, we modeled alcohol consumption in mice with drill-hole bone defects (Fig. 6e). We observed non-healing fractures in mice offered 1% ethanol, with severity increasing at 3% ethanol intake (Fig. S10d), without affecting body weight (Fig. S