REGγ is essential to maintain bone homeostasis by degrading TRAF6, preventing osteoporosis

Significance

Primary osteoporosis torments numerous older adults, which is a common systemic bone disease. Based on the potential pathogenesis of osteoporosis by proteomics analysis for osteoporosis samples and experiments in biochemistry and molecular biology, we demonstrated that TRAF6 could be degraded by REGγ via ubiquitin/ATP (Adenosine Triphosphate)-independent degradation manner directly and the dephosphorylation of NIP30 modulated by the inhibitor of CKII TTP22 promoted the degradation of TRAF6 by REGγ-20S proteasome to postpone osteoporosis. Based on the pivotal mechanism of the NIP30/REGγ/TRAF6 axis, we defined that the CKII inhibitor TTP22 could alleviate osteoporosis by promoting the ubiquitin-independent proteasomal degradation of TRAF6 and provided a strategy for the treatment of osteoporosis.

Abstract

Primary osteoporosis, manifesting as decreased bone mass and increased bone fragility, is a “silent disease” that is often ignored until a bone breaks. Accordingly, it is urgent to develop reliable biomarkers and novel therapeutic strategies for osteoporosis treatment. Here, we identified REGγ as a potential biomarker of osteoporotic populations through proteomics analysis. Next, we demonstrated that REGγ deficiency increased osteoclast activity and triggered bone mass loss in REGγ knockout (KO) and bone marrow-derive macrophage (BMM)-conditional REGγ KO mice. However, the osteoclast activity decreased in BMM-conditional REGγ overexpression mice. Mechanistically, we defined that REGγ-20S proteasome directly degraded TRAF6 to inhibit bone absorption in a ubiquitin-independent pathway. More importantly, BMM-conditional Traf6 KO with REGγ KO mice could “rescue” the osteoporosis phenotypes. Based on NIP30 (a REGγ “inhibitor”) dephosphorylation by CKII inhibition activated the ubiquitin-independent degradation of TRAF6, we selected TTP22, an inhibitor of CKII, and defined that TTP22 could alleviate osteoporosis in vitro and in vivo. Overall, our study reveals a unique function of NIP30/REGγ/TRAF6 axis in osteoporosis and provides a potential therapeutic drug TTP22 for osteoporosis.

Primary osteoporosis, which affects numerous older adults, especially postmenopausal women, is a common systemic bone disease that mainly manifests in decreased bone mass and increased bone fragility, which could lead to bone microstructure degradation, such as low back pain and spinal deformation, and fragile fractures (1–6). Furthermore, fragile fractures secondary to osteoporosis are a major concern because of their frequency and health hazards (7–9). The silent progression of osteoporosis disrupts early diagnosis and effective treatment. Bone mineral density (BMD), as the diagnostic gold standard, gives partial information on bone strength (10). Hence, complementary methods, such as biomarkers, are essential for early diagnosis. However, the existing biomarkers are not accurate for early diagnosis (11, 12). Current pharmacologic therapies, such as RANKL (Receptor Activator of Nuclear Factor Kappa-B Ligand) antibodies and parathyroid hormone-related peptide analogs, have only reduced fragile fractures to some extent (7, 13–16). At present, bisphosphonate and hormone therapy are the commonly used, but bisphosphonate is not easily absorbed by the human body and causes adverse reaction to the digestive tract. Hormone therapy tends to increase the incidence rate of other diseases, such as cancer and cardiovascular disease. Therefore, doubts about the long-term efficacy and adverse events of these drugs have prevented individuals from adhering to antiosteoporosis therapy (17, 18). Accordingly, it is still urgent to explore the underlying mechanisms of osteoporosis to develop reliable biomarkers and novel agents.

Enhanced osteoclastic bone resorption is one of the principal causes and therapeutic targets of osteoporosis (19). RANKL plays a key regulatory role in osteoclast maturation and activation (20–22). The binding of RANKL to RANK recruits the crucial adaptor molecule tumor necrosis factor receptor (TNFR)-associated factor 6 (TRAF6) to orchestrate subsequent signaling, including the NF-κβ, AP-1, and MAPK pathways, activating NFATC1 and resulting in excessive activation of osteoclasts (23–26). Traf6 knockout (KO) in mice resulted in osteoclast deficiency and osteopetrosis (27), which indicates that the regulation of TRAF6 could be the direction for osteoporosis therapy. Recent studies found that the proteasome system could regulate the function of TRAF6 via degradation (28, 29), indicating that understanding the regulation of TRAF6 degradation could be a new therapeutic strategy for osteoporosis. However, to our knowledge, whether the ubiquitin-independent proteasome participates in TRAF6 turnover and osteoclastogenesis remains unknown.

REGγ, also named Ki antigen, PSME3, PA28γ, or 11Sγ, was first found in the serum of patients with systemic lupus erythematosus and could regulate proteolysis in a ubiquitin-independent pathway by activating the 20S proteasome (30, 31). REGγ KO cells display retarded growth, decreased proliferation, and increased apoptosis (32, 33). Our previous research also found that REGγ-deficient mice exhibited an aging phenotype and osteoporosis symptoms (34). In addition, NIP30, also known as PIP30, attenuates REGγ-20S proteasome activation through its direct binding to REGγ. The evolutionarily conserved serine-rich domain in the C-terminus of NIP30 is a posttranslational modification site that can be phosphorylated by CKII to enhance its inhibitory binding to REGγ (35, 36). Hence, we propose that CKII inhibition could be the strategy for augmenting the function of the REGγ-20S proteasome.

In this study, we identified REGγ as a candidate biomarker in osteoporosis and demonstrated that REGγ regulated bone metabolism by degrading TRAF6 in a ubiquitin-independent manner. Furthermore, we found that TTP22 could postpone osteoporosis by regulating theCKII/NIP30/REGγ/TRAF6 axis and could thereby provide a therapeutic approach for osteoporosis treatment. Collectively, our study reveals that the NIP30/REGγ/TRAF6 axis is critical in osteoporosis and TTP22 is a potential unique drug for osteoporosis treatment.

Results

REGγ Is Identified as a Potential Biomarker in Osteoporosis.

To investigate candidate biomarkers of osteoporosis, we collected osteoporosis and control samples from patients and measured the BMD by dual-energy X-ray absorptiometry (DXA) (Fig. 1 A–C and SI Appendix, Table S1). Through proteomic analysis of trabecular bone specimens from the above groups, we found 4,015 proteins, including 536 downregulated proteins and 69 upregulated proteins (SI Appendix, Fig. S1A), in the first proteomic analysis of osteoporosis samples from patients. Screening the differentially expressed proteins (OP/Ctrl ratio ≥ 1.2 or ≤ 0.83, P value ≤ 0.001) (Fig. 1D) demonstrated that REGγ was markedly downregulated in osteoporosis patients. Interestingly, our previous study reported that the deficiency of REGγ could promote premature aging (34). However, it was noteworthy that the downregulation of REGγ in osteoporosis was independent of age, and its expression levels positively correlated with bone mineral density (BMD) (Fig. 1 E and F and SI Appendix, Fig. S1B). Moreover, we observed the same phenomenon in serum samples from both healthy individuals and patients (SI Appendix, Fig. S1 C and D). This suggests that REGγ might be a potential biomarker for osteoporosis. Next, we generated an ovariectomy (OVX)-induced osteoporosis mouse model and found lower trabecular BMD and bone volume (BV/TV) in the OVX group than in the sham control group (Fig. 1 G and H). Similarly, the expression of REGγ in the OVX group was lower than that in the sham control group (Fig. 1 I and J and SI Appendix, Fig. S1E). These observations also indicated that REGγ is a potential biomarker in osteoporosis.

Fig. 1.

REGγ is identified as a potential biomarker in osteoporosis. (A) Workflow for the discovery of differentially expressed protein proteins by MS using control and osteoporosis (OP) lumbar cancellous bone tissues. (B) Representative spine CT images of Ctrl and OP patients. (C) Quantitative analysis of the BMD and T values of Ctrl and OP patients (n = 4). BMD, bone mineral density. (D) Heatmap of the differentially expressed proteins in Ctrl and OP samples. (E and F) Western blotting analysis of REGγ protein levels in Ctrl and OP samples (E), with quantification using ImageJ (F) (Ctrl n = 17, OP n = 11). (G) Schematic diagram of the OVX-induced osteoporosis mouse model (Left), with representative micro-CT images displaying the 3D bone structures of femurs from 4-mo-old sham and OVX mice (similar results were obtained in all mice, n = 5). (Scale bar: 1 mm.) (H) Micro-CT measurements of BMD and BV/TV in femurs from 4-mo-old sham and OVX mice. BV/TV, bone volume as a fraction of total bone volume (n = 5). (I and J) Western blotting analysis of REGγ protein levels in sham and OVX mice (I), with quantification using ImageJ (J) (n = 8). Markers of significance are as follows: N.S, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001. Control: Ctrl; osteoporosis: OP. Sham: Sham operation; OVX: ovariectomy.

Deficiency of REGγ Triggers the Loss of Bone Mass by Enhancing the Activity of Osteoclasts.

Bone homeostasis is maintained by coordinated cycles of osteoclast bone resorption and osteoblast bone formation (18, 37–39). To further explore the function of REGγ in bone homeostasis, we analyzed the bone mass of wild-type (WT) and REGγ KO mice. The results showed that REGγ KO mice displayed bone loss compared to WT mice, as evidenced by the trabecular BMD and BV/TV (Fig. 2 A and B). However, the cortical BMD and BV/TV were not significantly different between WT and REGγ KO mice (SI Appendix, Fig. S2 A and B). Moreover, the Hematoxylin and Eosin (H&E) staining assay revealed the same results in the trabeculae of REGγ KO mice (Fig. 2C and SI Appendix, Fig. S2C). Taken together, these data demonstrated that REGγ deficiency leads to low bone mass.

Fig. 2.

Deficiency of REGγ triggers the loss of bone mass by enhancing the activity of osteoclasts. (A) Representative micro-CT images showing the 3D bone structures of femurs from 6-mo-old WT littermates and REGγ KO mice (n = 6). (Scale bar: 1 mm.) (B) Micro-CT measurements of BMD and BV/TV in femurs from 6-mo-old WT littermates and REGγ KO mice (n = 6). (C) Representative images of H&E staining of femurs of 6-mo-old REGγ KO mice and their WT littermates (n = 6). (Scale bar: 100 μm.) (D) Representative immunofluorescence images of TRAP expression in 6-mo-old WT littermates and REGγ KO mice (n = 6). (Scale bar: 100 μm.) (E) Histomorphometrical analysis of TRAP staining in (D) (n = 6). (F) Quantification of Trap, Nfatc1, and Ctsk expression in femurs of 6-mo-old REGγ KO mice and WT littermates (n = 6). Markers of significance are as follows: N.S, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001. WT: wildtype; KO: knockout.

To investigate whether the loss of bone mass in REGγ KO mice is caused by increased bone resorption or decreased bone formation, we detected the osteoclast marker tartrate-resistant acid phosphatase (TRAP) and the osteoblast marker osteocalcin (OCN). We found the presence of an increasing number of osteoclasts in the trabecular bone of REGγ KO mice (Fig. 2 D and E), whereas there was no difference in OCN levels between WT and REGγ KO mice (SI Appendix, Fig. S2 D and E). Subsequently, osteoclast markers, including Trap, Nfatc1, and Ctsk, were significantly increased in femoral samples from REGγ-deficient mice (Fig. 2F). In contrast, osteogenic markers showed no significant differences (SI Appendix, Fig. S2F), and REGγ deficiency did not affect osteoclast-to-osteoblast coupling factors (SI Appendix, Fig. S2G).

To further assess the role of REGγ in bone resorption and bone formation, we isolated bone marrow–derived macrophages (BMMs) and mesenchymal stem cells (MSCs) from both WT and REGγ KO mice. BMMs and MSCs were utilized for osteoclast differentiation (SI Appendix, Fig. S2H) and osteoblast differentiation (SI Appendix, Fig. S2K), respectively. Ex vivo osteoclast differentiation assay revealed an increase osteoclast formation of REGγ KO mice compared with that of WT mice (SI Appendix, Fig. S2 I and J). However, ALP staining, Von Kossa staining, and Alizarin Red S staining indicated no significant difference in osteoblast differentiation between WT and REGγ KO MSCs at various stages (SI Appendix, Fig. S2L). Additionally, qRT-PCR results demonstrated no significant changes in osteoblast markers, including Alp, Osteocalcin (Ocn), and Osterix (SI Appendix, Fig. S2M). Interestingly, we found that the expression of REGγ gradually decreases during osteoclast differentiation, and this reduction largely depends on Nrf2. (SI Appendix, Fig. S2 N–Q). Collectively, these results suggest that REGγ deficiency triggers the loss of bone mass by enhancing the activity of osteoclasts.

REGγ Specifically Suppresses Osteoclast Activity.

To verify whether REGγ deficiency triggers the loss of bone mass in an osteoclast-specific manner, we generated BMM-conditional REGγ-KO (REGγ cKO) mice and REGγ-overexpressing (REGγ cOE) mice by crossing LysM-Cre mice with floxed alleles of REGγ^fl/fl and LSL-REGγ mice, respectively (40, 41). We successfully obtained REGγ cKO and cOE mice. Both REGγ cKO and REGγ cOE mice showed no changes in size and weight compared with their controls at 2 mo (SI Appendix, Fig. S3 A, B, F, and G). However, the trabecular bone mass of the REGγ cKO mice decreased compared to that of the controls, while the trabecular bone mass of REGγ cOE mice showed the opposite phenomenon, as evidenced by an obvious change in BMD and BV/TV (Fig. 3 A and B). However, there was no difference in cortical bone (SI Appendix, Fig. S3 C–E and H–J). The H&E staining assay further revealed decreased trabecular bone mass in REGγ cKO mice and increased trabecular bone mass in REGγ cOE mice compared to their respective control groups (Fig. 3C and SI Appendix, Fig. S3K). The number of TRAP⁺ osteoclasts was significantly increased in bone sections of the REGγ cKO mice compared to their control littermates and were decreased in REGγ cOE mice (Fig. 3 D and E).

Fig. 3.

REGγ specifically suppressed osteoclast activity. (A) Representative micro-CT images showing the 3D bone structures of femurs from 2-mo-old REGγ^fl/fl (control), REGγ cKO, LSL-REGγ (control), and REGγ cOE mice (n = 6). (Scale bar: 1 mm.) (B) Micro-CT measurements of BMD and BV/TV in femurs from 2-mo-old cKO Ctrl, REGγ cKO, cOE Ctrl, and REGγ cOE mice (n = 6). (C) Representative images of H&E staining of femurs of 2-mo-old cKO Ctrl, REGγ cKO, cOE Ctrl, and REGγ cOE mice (n = 6). (Scale bar: 100 μm.) (D) Representative TRAP staining of femurs from 2-mo-old cKO Ctrl, REGγ cKO, cOE Ctrl, and REGγ cOE mice. (Scale bar: 1 mm.) (E) Histomorphometrical analysis of TRAP staining in (E) (n = 6). (F) Quantification of Trap, Nfatc1, and Ctsk expression in femurs of 2-mo-old cKO Ctrl, REGγ cKO, cOE Ctrl, and REGγ cOE mice (n = 6). Markers of significance are as follows: N.S, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001. Control: Ctrl; REGγ cKO: bone marrow-derive macrophage (BMM)-conditional REGγ KO; REGγ cOE: bone marrow-derive macrophage (BMM)-conditional REGγ OE.

Molecular analysis of bone samples showed increased osteolytic markers, including Trap, Nfatc1, and Ctsk, in REGγ cKO mice compared with the controls, while these markers were markedly decreased in REGγ cOE mice (Fig. 3F). Ex vivo osteoclast differentiation assay revealed an increase osteoclast formation in REGγ cKO mice compared to controls, while a decrease formation in REGγ cOE mice (SI Appendix, Fig. S3 L and M). These results demonstrate that REGγ regulates osteoclast activity.

REGγ Suppresses RANKL-Induced Osteoclastogenesis.

RANKL plays a key regulatory role in osteoclast maturation and activation (20–22). To further assess the role of REGγ in osteoclast differentiation, we set up a RANKL concentration gradient for osteoclast induction of BMMs in vitro (Fig. 4A). Our results showed that REGγ deficiency promoted the RANKL sensitivity of BMMs (Fig. 4 B–D). At the same time, we also set a time gradient in vitro for osteoclast induction of BMMs (Fig. 4E). REGγ deficiency promoted earlier differentiation of BMMs into osteoclasts (Fig. 4 F–H). To identify whether REGγ can affect bone resorption, we removed mature osteoclasts to visualize bone resorption pits using Versene (Fig. 4I). After 24 h of culture, REGγ-null osteoclasts could resorb larger cavities than WT osteoclasts (Fig. 4 J and K).

Fig. 4.

REGγ suppressed RANKL-induced osteoclastogenesis. (A) Schematic illustration depicting osteoclast differentiation of BMMs under varying concentrations of RANKL: BMMs were exposed to 10 ng/mL M-CSF along with different concentrations of RANKL to induce osteoclast differentiation over a period of 6 d. The concentrations of RANKL used were 0, 6.25, 12.5, 25, and 50 ng/mL. (B) Representative TRAP staining of osteoclasts from WT and REGγ KO BMMs treated with RANKL-gradient stimulation. (Scale bar: 100 μm.) (C and D) Quantification of osteoclast number and TRAP-positive area in (B) (n = 5). (E) Schematic diagrams illustrating osteoclast differentiation of BMMs at different treatment durations: BMMs were treated with 10 ng/mL M-CSF and 50 ng/mL RANKL to induce osteoclast differentiation over varying time intervals. The time gradients were set to 0, 2, 4, 6, and 8 d. (F) Representative TRAP staining of osteoclasts from WT and REGγ KO BMMs at various treatment durations. (Scale bar: 100 μm.) (G and H) Quantification of osteoclast number and TRAP-positive area in (F) (n = 5). (I) Schematic diagram of pit formation assay: BMMs were treated with 10 ng/mL M-CSF and 50 ng/mL RANKL to induce differentiation for 6 d. Subsequently, mature osteoclasts were harvested and reseeded onto bone slices for a 48-h culture period. (J) Representative images of bone erosion pits from WT and REGγ KO osteoclasts. (Scale bar: 75 μm.) (K) Quantification of the osteoclast resorption area in (J) (n = 5). (L) Quantification of Trap, Nfatc1, and Ctsk gene expression levels in WT and REGγ KO osteoclasts (n = 3). (M) Schematic illustration depicting BMM stimulation with RANKL at different times: BMMs were treated with 10 ng/mL M-CSF and 100 ng/mL RANKL to promote activation of the MAPK and NF-κB signaling pathways. The time points were set at 0, 15, 30, and 60 min. (N) Western blotting images showing the changes in the MAPK and NF-κB signaling pathways in WT and REGγ KO BMMs during RANKL stimulation. This experiment was repeated independently three times. Markers of significance are as follows: N.S, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001.

Molecular analysis of osteoclasts showed significantly increased osteolytic markers, including Trap, Nfatc1, and Ctsk, in REGγ KO mice (Fig. 4L). It was reported that RANKL-induced osteoclast differentiation involved activation of the NF-κβ and MAPK pathways downstream of TRAF6 (42–47). We found that the deletion of REGγ significantly increased phosphorylated p65, phosphorylated p38, phosphorylated ERK, and phosphorylated JNK after RANKL stimulation in BMMs (Fig. 4 M and N).

Therefore, we further analyzed the expression of the aforementioned proteins in the tibia tissues of REGγ KO, REGγ cKO, and REGγ cOE mice (SI Appendix, Fig. S4 A, E, and I). We observed that the levels of phosphorylated p65, phosphorylated p38, phosphorylated ERK, and phosphorylated JNK were elevated in REGγ-deficient mice (SI Appendix, Fig. S4 B–D and F–H), whereas the expression of these proteins was reversed in REGγ-overexpressing mice (SI Appendix, Fig. S4 J–L). Taken together, these findings demonstrated that REGγ can suppress RANKL-induced osteoclastogenesis both in vitro and vivo (SI Appendix, Fig. S4M).

REGγ Increases Bone Mass through the Ubiquitin-Independent Degradation of TRAF6.

During RANKL-induced osteoclast differentiation, RANKL binds to its receptor RANK, which recruits the adaptor molecule TRAF6, an E3 ubiquitin ligase that regulates downstream NF-kB and MAPK/AP-1 (46). According to the correlation between TRAF6 involved in the ubiquitin-dependent proteasome pathway and REGγ’s function in ubiquitin-independent protein degradation (48–50), we speculate that the positive regulation of REGγ in bone mass is likely through the nonubiquitinated degradation of TRAF6. Therefore, we examined the expression of TRAF6 in the tibia and fibula tissues of REGγ KO mice, REGγ cKO mice, REGγ cOE mice, and their littermates. We observed an upregulation of TRAF6 expression in the tissues of REGγ KO mice and REGγ cKO mice, while a downregulation was observed in REGγ cOE mice (Fig. 5 A–C and SI Appendix, Fig. S5D). Similar alterations were detected in the BMMs of mice with above genotypes (SI Appendix, Fig. S5 A and B). In addition, the mRNA levels of Traf6 were unchanged (SI Appendix, Fig. S5C). Interestingly, we found that the plasma concentration of TRAF6 was dramatically upregulated in OP patients compared with that in control patients (SI Appendix, Fig. S5E and Table S2), indicating that the upregulation of TRAF6 could cause osteoporosis. Furthermore, we utilized confocal microscopy to reveal the colocalization of REGγ and TRAF6 in both the nucleus and cytoplasm of osteoclasts (Fig. 5D). The nucleoplasmic separation experiments also confirmed this phenomenon (SI Appendix, Fig. S5F). Then, we generated a series of truncations of TRAF6 to analyze the REGγ-interactive domain of TRAF6 (Fig. 5E). We found that the fourth zinc finger of TRAF6 plays an essential role in the interaction of REGγ and TRAF6 (Fig. 5F).

Fig. 5.

REGγ increased bone mass through ubiquitin-independent degradation of TRAF6. (A–C) Western blotting analysis of TRAF6 protein levels in WT, REGγ KO, cKO Ctrl, REGγ cKO, cOE Ctrl, and REGγ cOE hindlimb bones. (D) Representative images of IF analysis. IF showed overlapping REGγ (red) and TRAF6 (green) in osteoclasts. (Scale bar: 50 μm.) (E) Schematic diagram of TRAF6 mutation construction: Domain organization of TRAF6. Z1–Z4: zinc fingers 1 to 4. CC: coiled. FL stands for full-length TRAF6 (residues 1 to 552), RZ₁ (residues 1 to 159), RZ₁₂ (residues 1 to 187), RZ₁₂₃ (residues 1 to 211), N-terminal region (residues 1 to 332), and C-terminal region (residues 333 to 522). (F) The interactions between REGγ and TRAF6 were determined by reciprocal coimmunoprecipitation and Western blotting analysis. (G and H) Western blotting images showing that REGγ can regulate TRAF6 protein stability in BMMs. WT and REGγ KO BMMs were treated with cycloheximide (CHX, 100 μg/mL) for the indicated times (G), with quantification using ImageJ (H) (n = 3). (I) Schematic illustration depicting the process of TRAF6 protein degradation in vitro: In vitro proteolytic analyses were performed using purified REGγ, 20S proteasome, and Traf6 protein at 37 °C for 3 h (Up). Western blotting images showing that REGγ directly promoted the degradation of TRAF6 in vitro. (J) Representative micro-CT images showing the 3D bone structures of femurs from 3-mo-old REGγ WT, REGγ KO, BMM-conditional Traf6 KO (Traf6 cKO), and Traf6 cKO with REGγ KO (DKO) mice. (K) Quantification of BMD and BV/TV in femurs by Micro-CT measurement software (n = 6). (Scale bar: 1 mm.) (L) Western blotting images depict alterations in TRAF6, RANK, as well as the MAPK and NF-κB signaling pathways in WT, REGγ KO, Traf6 cKO, and DKO BMMs during RANKL stimulation. (M) Representative micro-CT images showing the 3D bone structures of femurs from 12-mo-old WT, REGγ KO, C25-140-treated WT, and C25-140-treated REGγ KO mice. (Scale bar: 1 mm.) (N) Micro-CT measurements of BMD and BV/TV in femurs from 12-mo-old WT, REGγ KO, C25-140-treated WT, and C25-140-treated REGγ KO mice (n = 6). The above experiments were repeated independently three times. Markers of significance are as follows: N.S, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001. Bone marrow macrophage (BMM)-conditional Traf6 KO with REGγ KO: DKO.

To determine whether REGγ mediates the degradation of TRAF6, we performed a cycloheximide (CHX) chase assay and found that TRAF6 was significantly stabilized in REGγ KO BMMs in the presence of CHX (Fig. 5 G and H). These results further indicated that TRAF6 might be a substrate of REGγ. To verify that REGγ directly degrades TRAF6, we examined the ability of REGγ to direct cell-free proteolysis. Incubation of purified TRAF6 with either the latent 20S proteasome or REGγ alone resulted in no significant degradation of TRAF6 except for nonspecific decay. However, the combination of REGγ and the 20S proteasome promoted significant degradation of TRAF6 (Fig. 5I). Taken together, these data revealed that REGγ could regulate the degradation of TRAF6 in a ubiquitin-independent manner.

To ascertain the TRAF6 dependency of the REGγ-osteoclast phenomenon, we generated double KO (DKO) mice by crossing bone marrow-derive macrophage (BMM)-conditional Traf6 KO mice with REGγ KO mice (SI Appendix, Fig. S5G). We found that DKO mice exhibited higher bone mass compared to REGγ KO mice (Fig. 5J). This observation was further confirmed by H&E staining. In addition, the TRAP staining results indicated a decreased TRAP-positive rate in the femur sections of DKO mice (SI Appendix, Fig. S5 H and I). Ex vivo osteoclast differentiation assay revealed a decrease osteoclast formation in DKO mice compared to REGγ KO mice (SI Appendix, Fig. S5 J and K).

Furthermore, we also found that the double deletion of REGγ and Traf6 significantly decreased phosphorylated p65, phosphorylated p38, phosphorylated ERK, and phosphorylated JNK after RANKL stimulation in BMMs compared with deletion of REGγ (Fig. 5L and SI Appendix, Fig. S5L). The same results were obtained in bone tissue (SI Appendix, Fig. S6 A and B). Collectively, these data indicated that TRAF6 deletion rescued the enhanced osteoclast activity and reduced bone mass caused by REGγ deficiency, suggesting that the REGγ-osteoclast phenomenon was TRAF6-dependent. However, another REGγ substrate, p21, had minimal impact on this phenomenon (SI Appendix, Fig. S6 C–F).

C25-140 is a TRAF6 inhibitor that disrupts the interaction between TRAF6 and Ubc13, directly affecting the activity of TRAF6 (51). To investigate its potential in treating osteoporosis, we administered C25-140 treatment and observed its inhibitory effect on osteoclast differentiation in both WT and REGγ KO BMMs (SI Appendix, Fig. S6 G–J). Considering that in vivo experiments are more illustrative, we administered C25-140 to REGγ KO mice via intraperitoneal injection for 1 mo (SI Appendix, Fig. S6K). We found that bone mass was higher in the treated mice than in the saline-injected group with no weight change (Fig. 5 M and N and SI Appendix, Fig. S6L), and the number of osteoclasts in the treated mice was lower (SI Appendix, Fig. S6 M and N).

In summary, REGγ degrades TRAF6 to maintain physiological homeostasis. In the absence of REGγ, the expression level of TRAF6 increases, triggering the activation of osteoclast differentiation pathways and subsequent bone resorption. The deficiency of TRAF6 rescues this phenomenon. Similarly, the TRAF6 inhibitor C25-140 exhibits analogous effects, suggesting its potential as a unique therapeutic strategy for treating osteoporosis (SI Appendix, Fig. S6O).

NIP30 Dephosphorylation Activates the Ubiquitin-Independent Degradation of TRAF6 to Alleviate Osteoporosis.

Considering that REGγ is a positive regulator of bone mass, we tried to find molecules that could augment the function of the REGγ-20S proteasome. Our previous work found that NIP30 cannot bind to REGγ to promote the function of REGγ-20S proteasome after mutating four serine lines at positions 226-230 to alanine, suggesting that inhibiting the phosphorylation of NIP30 is an "activator" of the REGγ-20S proteasome (35, 36). It may be possible to upregulate bone mass by inhibiting NIP30 phosphorylation (Fig. 6A). To investigate whether the NIP30/REGγ pathway regulates osteoporosis, we first constructed Nip30 4A (mimic loss function of NIP30) transgenic mice (Fig. 6B and SI Appendix, Fig. S7A). We found that the BMD and bone volume of Nip30^4A/4A mice showed an increase in cancellous bone compared to that of WT mice (Fig. 6 C and D), but there was little change in cortical bone (SI Appendix, Fig. S7 B and C). The same phenomenon was demonstrated by H&E staining (Fig. 6E and SI Appendix, Fig. S7D). Moreover, the number of osteoclasts and osteoclastic markers in the bones of Nip30^4A/4A mice appeared to be significantly lower compared to those in WT mice (Fig. 6 F and G) and revealed a decrease osteoclast differentiation of Nip30^4A/4A BMMs compared to that of WT mice (Fig. 6 H–J and SI Appendix, Fig. S7 E and F). Meanwhile, there is no difference of bone formation in vitro and in vivo (SI Appendix, Fig. S7 G–I).

Fig. 6.

NIP30 dephosphorylation activates the ubiquitin-independent degradation of TRAF6 to alleviate osteoporosis. (A) Hypothetical schematic diagram illustrating the involvement of NIP30 in the regulation of osteoporosis: NIP30 acts as an upstream regulator of REGγ. When serine residues at positions 226 to 230 of NIP30 are dephosphorylation, it fails to bind to REGγ and promote the function of the REGγ-20S proteasome. This alteration may impact the ubiquitin-independent degradation of TRAF6, thereby regulating the development of osteoporosis. (B) Strategy diagram illustrating the generational design of Nip30^4A/4A mice. (C) Representative micro-CT images showing the 3D bone structures of femurs from 2-mo-old WT and Nip30^A/4A mice (n = 6). (Scale bar: 1 mm.) (D) Micro-CT measurements of BMD and BV/TV in femurs from 2-mo-old WT and Nip30^4A/4A mice (n = 6). (E) Representative images of H&E staining of femurs of 2-mo-old WT and Nip30^4A/4A mice (n = 6). (Scale bar: 100 μm.) (F) Representative immunofluorescence images of TRAP expression in 2-mo-old WT and Nip30^4A/4A mice (n = 6). (Scale bar: 100 μm.) (G) Histomorphometrical analysis of TRAP staining in (F) (n = 6). (H) Representative TRAP staining of osteoclasts from WT and Nip30^4A/4A BMMs treated with RANKL stimulation. (Scale bar: 100 μm.) (I and J) Quantification of osteoclast number and TRAP-positive area in (H) (n = 5). (K–M) Western blotting images depict alterations in REGγ, TRAF6, NIP30, as well as the MAPK and NF-κB signaling pathways in the hindlimb bones of WT and Nip30^4A/4A mice (K). The above results were analyzed using ImageJ (L and M) (n = 5). (N) Schematic diagram illustrating the mechanism of bone resorption in Nip30^4A/4A mice: The NIP30 4A mutation inhibits the function of the REGγ-20S proteasome, thereby promoting ubiquitin-independent degradation of TRAF6, leading to inhibition of the MAPK and NF-κB signaling pathways, and consequently suppressing the occurrence of osteoporosis. Markers of significance are as follows: N.S, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001.

Next, we examined TRAF6 expression in bone tissues and BMMs of both Nip30^4A/4A and WT mice. We observed a downregulation of TRAF6 expression in Nip30^4A/4A mice (SI Appendix, Fig. S7 J and K), suggesting that NIP30/REGγ/TRAF6 pathway regulated osteoporosis. Furthermore, we also found that phosphorylation of p65, p38, ERK, and JNK was reduced in the bone tissue of Nip30^4A/4A mice compared to WT mice (Fig. 6 K–M).

To ascertain the REGγ dependency of the Nip30 4A-osteoclast phenomenon, we generated REGγ KO Nip30^4A/4A mice by crossing Nip30^4A/4A mice with REGγ KO mice. Ex vivo osteoclast differentiation assay revealed an increase osteoclast formation of REGγ KO Nip30^4A/4A mice compared to that of Nip30^4A/4A mice (SI Appendix, Fig. S7 L–N). Additionally, we found that compared to Nip30^4A/4A mice, REGγ KO Nip30^4A/4A mice exhibited increased expression of TRAF6 in bone tissue, accompanied by elevated phosphorylation levels of p65, p38, ERK, and JNK (SI Appendix, Fig. S7 O–Q). Taken together, these results suggested that NIP30 dephosphorylation inhibits the function of the REGγ-20S proteasome, thereby promoting ubiquitin-independent degradation of TRAF6 to inhibit the development of osteoporosis (Fig. 6N).

TTP22 Alleviates Osteoporosis by Regulating the CKII/NIP30/REGγ/TRAF6 Axis.

Several studies have shown that CKII is an upstream kinase of NIP30 and can promote NIP30 phosphorylation (35, 36). Considering that the NIP30/REGγ/TRAF6 axis is critical in osteoporosis, we used TTP22 to determine whether it could inhibit the phosphorylation of NIP30 to affect the function of REGγ (SI Appendix, Fig. S7R). Interestingly, we found that TTP22 inhibited the phosphorylation of NIP30 and attenuated the expression of TRAF6 in BMMs of WT mice (Fig. 7A). In addition, we found that another inhibitor of CKII, CX4945, could also inhibit the phosphorylation of NIP30 and attenuate the expression of TRAF6 in BMMs of WT and REGγ KO mice (SI Appendix, Fig. S8A). However, the function of CX4945 was not REGγ dependent. Subsequently, TTP22 intervention was performed during osteoclast differentiation, and TTP22 was found to inhibit the ability of BMMs to differentiate toward osteoclasts (Fig. 7 B and C and SI Appendix, Fig. S8H). These results may imply that TTP22 is able to regulate bone mass by affecting osteoclast differentiation.

Fig. 7.

TTP22 alleviates osteoporosis by regulating the CKII/NIP30/REGγ/TRAF6 axis. (A) Western blotting analysis of the protein levels of P-NIP30 and TRAF6 with or without TTP22 treatment (0, 30, 60, 120 min). (B) Representative TRAP staining of BMMs from WT mice treated with or without TTP22. (Scale bar: 100 μm.) (C) Quantification of osteoclast number and TRAP-positive area in (B). (D) Schematic diagram illustrating the experimental design of TTP22 treatment in WT mice: One week after OVX surgery, mice were subjected to intraperitoneal injections of TTP22 every other day for a total of 8 wk. (E) Representative micro-CT images showing the 3D bone structures of femurs from 5-mo-old Sham, OVX, and TTP22-treated WT mice (n = 6). (Scale bar: 1 mm.) (F) Micro-CT measurements of BMD and BV/TV in femurs from 5-mo-old Sham, OVX, and TTP22-treated WT mice (n = 6). (G) Representative images of H&E staining of femurs of 5-mo-old Sham, OVX, and TTP22-treated WT mice (n = 6). (Scale bar: 100 μm.) (H) Representative immunofluorescence images of TRAP expression in 5-mo-old Sham, OVX, and TTP22-treated WT mice (n = 6). (Scale bar: 100 μm.) (I and J) Western blotting images depict alterations in TRAF6, NIP30, and P-NIP30 as well as the MAPK and NF-κB signaling pathways in the hindlimb bones of OVX and TTP22-treated WT mice (I). The above results were analyzed using ImageJ (J) (n = 5). Markers of significance are as follows: N.S, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001.

To determine the in vivo therapeutic potential of TTP22 in bone resorption-related diseases, TTP22 was intraperitoneally injected every other day in OVX mice (Fig. 7D). Subsequently, we analyzed the drug metabolism by pharmacokinetics and found that the drug concentration in blood was still at an effective concentration 24 h after injection, indicating that TTP22 has good potential for application (SI Appendix, Fig. S8B). Two months after the first injection, the body weights of the mice were barely affected (SI Appendix, Fig. S8C). H&E staining of internal tissues also showed no significant changes in tissue morphology (SI Appendix, Fig. S8D), indicating that TTP22 had no significant side effects. Interestingly, micro-CT analysis showed that the bone mass of TTP22-treated OVX mice significantly increased compared to that of untreated OVX mice (Fig. 7E). The quantitative values of the BV/TV and BMD significantly increased in the femurs of TTP22-treated mice (Fig. 7F). However, there was no difference in cortical bone (SI Appendix, Fig. S8 E and F). Bone histomorphometric analysis with TRAP staining showed that the number of osteoclasts in OVX mice decreased after TTP22 treatment (Fig. 7 G and H). Additionally, compared to OVX mice, OVX mice treated with TTP22 exhibited a noticeable reduction in osteoclastic markers in their bones (SI Appendix, Fig. S8G). Meanwhile, there were no differences in bone formation observed both in vitro and in vivo (SI Appendix, Fig. S8 I–K).

Furthermore, Western blot results also demonstrated a decrease in TRAF6 and P-NIP30 expression after TTP22 treatment, accompanied by a reduction in the phosphorylation levels of proteins associated with the MAPK and NF-κB signaling pathways (Fig. 7 I and J). Taken together, our data demonstrate that TTP22 can alleviate osteoporosis by inhibiting NIP30/REGγ/TRAF6.

Discussion

The proteasome system plays essential roles in numerous cellular processes, and its dysfunction may result in many diseases, such as multiple myeloma and autoimmune diseases, making it a considerable therapeutic target (52–54). The proteasome system can be divided into two groups: the ubiquitin-dependent 19S proteasome activators and the ubiquitin-independent 11S proteasome family (PSME1, PSME2, and PSME3, also named REGγ), the latter of which is a set of conserved proteins that promote ATP-independent protein degradation (55, 56). Previous studies have given much attention to the ubiquitin–proteasome system, but the ubiquitin-independent proteasome system has received increasing attention in recent years. In this study, we investigated the role of REGγ in osteoporosis in detail, demonstrating that REGγ could be a potential predictive biomarker for the early diagnosis and modulation of REGγ activity by interfering with NIP30-REGγ interactions, which may provide unique venues for the development of pharmacological agents.

To the best of our knowledge, in addition to our previous finding that REGγ deficiency manifests as premature aging, including osteoporosis (34), few studies have explored the regulatory mechanism between the ubiquitin-independent proteasome system and osteoporosis. In this study, we revealed that the expression of REGγ is negatively related to osteoporosis and that REGγ deficiency aggravates osteoporosis by promoting the activation of osteoclasts. Interestingly, we found the expression of REGγ gradually decreases in osteoclast differentiation process, thereby exacerbating osteoporosis. Based on our results, we believe that REGγ could be a biomarker predicting individual susceptibility to osteoporosis and provide a different venue for treating osteoporosis by interfering with the REGγ-20S proteasome.

Our team provided the evidence that the REGγ-20S proteasome can degrade intact cellular proteins (57). To date, multiple important regulatory proteins have been found to degrade in this ubiquitin-independent manner, such as p21, p16, p19, and p53 (31, 56–58). However, the downstream targets of the REGγ-20S proteasome involved in the progression of osteoporosis and osteoclastogenesis remain unclear. One of the most striking features of the present study is that we identified TRAF6 as a pivotal downstream target of the REGγ-20S proteasome involved in regulating osteoclast activation in the context of osteoporosis. TRAF6 is essential for many biological processes, including the activation of osteoclasts (59). As a central upstream adaptor, TRAF6 regulates osteoclastogenesis by orchestrating RANKL/RANK and downstream signaling, such as the NF-κβ, MAPKs, and PI3K axes (23, 25, 27, 42, 60–62). Although there is evidence that the degradation of TRAF6 requires the ubiquitin-dependent proteasome system (63), another study suggested that TRAF6 could be selectively degraded via autophagy (64). Our results further illustrated that the REGγ-20S proteasome directly degrades the E3 ligase TRAF6, indicating that the alternative ubiquitin-independent proteasome pathway is an extremely attractive protein degradation mechanism and that the regulation of the REGγ-20S proteasome/TRAF6 pathway could be a potential strategy for osteoporosis treatment.

NIP30, recently identified as a REGγ “inhibitor,” functions to modulate REGγ in cells (35, 36). The highly specific binding of NIP30 to REGγ inhibits ubiquitin-independent proteasome-degrading proteins and depends on the phosphorylation of NIP30. There are no effective molecular agents regulating the function of NIP30 or REGγ. Our previous study showed that CKII could phosphorylate the C-terminus of NIP30 (36), indicating that fine-tuning the NIP30/REGγ/TRAF6 axis is a potential strategy to treat osteoporosis. In our study, we utilized TTP22, a selective inhibitor of CKII, to test the hypothesis that it could suppress osteoclastogenesis and treat osteoporosis. The results showed that TTP22 effectively restrained the progression of osteoporosis in the OVX mouse model with favorable biological safety, indicating potential for its clinical application in the treatment of osteoporosis.

In conclusion, our study reveals REGγ as a therapeutic biomarker of osteoporosis. Based on the pivotal mechanism of the NIP30/REGγ/TRAF6 axis, we determined that TTP22 can alleviate osteoporosis by promoting the ubiquitin-independent proteasomal degradation of TRAF6 and provide a different strategy for the treatment of osteoporosis.

Materials and Methods

Human Samples.

Sample collections were performed after written informed consent. To protect patient privacy, all personal information was de-identified. Institutional approval was obtained from the Shanghai General Hospital Institutional Review Board (Approval No. 2021094). The standards for sample collection are presented in SI Appendix.

Animals.

REGγ KO mice with a C57BL/6 genetic background were a friendly gift from Dr. John J. Monaco of the University of Cincinnati, and have been backcrossed in our facility for over 10 generations. REGγ^fl/fl, LSL-REGγ, and LysM-Cre mice were generated as described elsewhere (1, 2). To generate LysM-Cre/Traf6^fl/fl/REGγ KO (DKO) mice, REGγ KO mice were first crossed with Traf6^fl/fl mice and LysM-Cre mice separately to generate Traf6^fl/fl/REGγ KO mice and LysM-Cre/REGγ KO mice. Subsequently, these mice were further crossed together to bring out LysM-Cre/Traf6^fl/fl/REGγ KO genotype. Nip30^4A/4A mouse line was used in the current study. Nip30^4A/4A mice were designed and constructed in our laboratory. REGγ KO Nip30^4A/4A mice were generated by crossing Nip30^4A/4A mice with REGγ KO mice. All mice were housed in a specific pathogen-free (SPF) facility following standard humane animal husbandry protocols, approved by the animal care and use committee of the Institute of Microbiology (Chinese Academy of Sciences).

Human samples information, Plasmids information, micro-CT assays, immunohistochemistry assays, immunofluorescence assays, In vitro osteoclast differentiation assays, In vitro osteoblast differentiation assay, TRAP staining assay, pit formation assay, ALP staining, Alizarin Red S staining and Von Kossa staining, RNA extraction and Real-Time Quantitative PCR assays, western blotting assay, Co-immunoprecipitation assay, in vitro proteolytic analysis assay, and Enzyme-Linked Immunosorbent Assay, as well as reagents and resources, are listed in SI Appendix, Materials and Methods.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

The mass spectrometry data is being deposited to the ProteomeXchange Consortium via the iProX repository with the data identifier IPX0010088000 (ProteomeXchange dataset Identifier: PXD057237) (65). All other data are included in the article and/or SI Appendix.