Introduction
Osteoporosis (OP) is a chronic skeletal disorder that frequently occurs in elderly women and greatly impacts their quality of life [ 1 ]. Bone homeostasis can be influenced by bone resorption and bone formation, and dysregulated bone homeostasis is caused by reduced osteogenesis and increased bone loss, which eventually results in the development of OP [ 2 ]. There is a rise in incidence of OP with the rapid growth of geriatric population. Osteoporotic fracture remains a leading cause of deaths of elderly people [ 3 ]. Currently, several drugs have been approved to treat OP, such as bisphosphonates, calcitonin, and denosumab (RANKL monoclonal antibody). However, long-term usage of the above drugs may have side effects, including venous thromboembolism, tumorigenesis, stroke, and osteonecrosis [ 4 , 5 ]. Thus, it is urgent to identify more effective and safer agents to delay OP development.
Mounting evidence has proved that irisin can serve as an agent for alleviating OP [ 6 ]. It has been recognized that irisin, a cleavage peptide by fibronectin type III domain-containing protein 5 (FNDC5), is mainly secreted by skeletal muscles during exercise [ 7 ]. Although its promotive role in bone growth has been widely acknowledged, the protective mechanisms of irisin in OP remain unclear. A previous study has supported integrin αV/β5 as a downstream target of irisin [ 8 ], which provided clues for the investigation of potential mechanisms of irisin.
Euchromatic histone lysine methyltransferase 2 (EHMT2), also called G9a, is a methyltransferase that functions as a gene repressor via catalysis of the dimethylation of histone H3 lysine 9 (H3K9) [ 9 ].The regulation of EHMT2 in bone formation has been documented. For example, EHMT2 was involved in DNA methylation of PTEN during osteoblast differentiation of mesenchymal stem cells [ 10 ]. Another study reported that inhibition of EHMT2 attenuated bone loss during multiple myeloma progression [ 11 ]. Notably, EHMT2 was predicted to be modulated by S-phase kinase-associated protein 2 (SKP2)-mediated ubiquitination via bioinformatics analysis. In addition, Bond et al. demonstrated that SKP2 protein was stabilized by focal adhesion kinase (FAK) activation [ 12 ]. Furthermore, FAK could be activated by irisin’s downstream receptor integrin αV/β5 [ 13 ]. Therefore, we predicted that irisin might activate integrin αVβ5/FAK signaling to stabilize SKP2 protein and consequently caused ubiquitination of EHMT2.
Like ubiquitination, small ubiquitin-like modifier (SUMO) is another posttranslational modification. SUMO of proteins participates in the pathogenesis of multiple diseases, including OP [ 14 ]. Irisin has been documented to degrade HDAC4 by SUMO to relieve cardiac hypoxia/reoxygenation injury [ 15 ]. Interestingly, bioinformatics analysis predicted a potential modulation of EHMT2 by SUMO. In addition, EHMT2 was predicted to bind to distal-less homeobox 3 (DLX3) promoter and repressed DLX3 transcription via recruitment of H3K9me2.
Considering the above background, we speculated that EHMT2 protein might be degraded through regulation of its ubiquitination and SUMOylation by irisin, which subsequently transcriptionally activated DLX3, thereby enhancing osteogenic differentiation. This study aimed to validate this speculation, which shed light on a novel modulatory mechanism for osteogenesis.
Materials and methods
Animal experiment
Male SAMR1 and SAMP6 mice (16 weeks of age) were purchased from Vital River Laboratory Animal Technology (Beijing, China). The SAMP6 mice were randomly divided into five groups: SAMP6, SAMP6 + sh-NC, SAMP6 + sh-EHMT2, SAMP6 + oe-NC, SAMP6 + oe-FNDC5. AAV9 carrying sh-NC, sh-EHMT2, oe-NC, or oe-FNDC5 was injected into SAMP6 mice via tail vein. All mice were sacrificed after 6 weeks and the femurs were obtained. The animal experiments were approved by the Ethics Committee of University of South China (ethical number: 2024 − 281).
Isolation of mesenchymal stem cells (MSCs) and treatments
MSCs were isolated as previously reported [ 16 ]. Briefly, MSCs were isolated through bone marrow aspiration of SAMR1 and SAMP6 mice, followed by purification via density gradient centrifugation. The isolated MSCs were maintained with Dulbecco’s modified Eagle’s medium (DMEM, GIBCO, NY, USA) at 37 °C with 5% CO 2 with culture medium changed every 3 days. MSCs at passages 1–4 were used for subsequent experiments. MSCs were treated with 100 ng/mL irisin (MCE, NJ, USA) together with or without irisin inhibitor RGDS (100 nM, MCE) for 48 h.
Cell transfection
HEK293T cells were obtained from ATCC (Manassas, VA, USA), and were grown in DMEM containing 10% FBS (Gibco, Grand Island, NY, USA). MSCs were seeded into 6-well plates (5 × 10 5 cells per well) at 24 h before transfection. Lentivirus carrying short hairpin RNA (shRNA) targeting EHMT2 (sh-EHMT2), sh-SKP2, sh-protein inhibitor of activated STAT 4 (PIAS4), and negative control shRNA (sh-NC), oe-EHMT2, oe-SKP2, oe-PIAS4, or oe-NC were provided by GenePharma (Shanghai, China). Besides, MSCs were infected with lentivirus containing 8 µg/mL polybrene at a multiplicity of infection of 20, followed by selection of infected cells with 2.5 µg/mL puromycin after the infection for 2 days.
Osteogenic differentiation induction
MSCs were planted into 24-well plates (1 × 10 4 cells per well). After reaching 80% confluence, the culture medium was replaced with 1 mL osteogenic differentiation medium (Lonza, Switzerland) to induce osteogenic differentiation. The control cells were maintained in complete culture medium. The cells were cultured for 3, 6, 9, 12 days with medium changed every 3 days.
Histological analysis
For histological analysis, the femurs were received decalcification using 14% ethylenediamine tetra-acetic acid and embedding in paraffins, followed by slicing into 4 μm-sections. Hematoxylin and eosin (HE) staining was performed to evaluate pathological changes in femurs using a commercial HE Stain Kit (Solarbio, Beijing, China). In short, the sections were stained with hematoxylin for 5 min and rinsed with running tap water. The sections were incubated with eosin for 1 min, followed by dehydration in ethanol. Masson’s trichrome staining was performed using the Masson’s Trichrome Stain Kit (Solarbio), following the manufacturer’s instruction. Under a light microscope (Olympus, Tokyo, Japan), the stained sections were photographed.
Micro-CT
The femurs of mice were fixed with paraformaldehyde for 48 h. Subsequently, the femurs were scanned by the Micro-CT system (Siemens, Germany). Bone Volume/Total Volume (BV/TV) was analyzed by the Micro-CT analysis software.
Alkaline phosphatase (ALP) staining
ALP staining assay was conducted using the ALP staining kit (Wako, Japan). Briefly, MSCs were fixed with 4% paraformaldehyde for 30 min. The femur bone tissues were subjected to 4% paraformaldehyde fixation, demineralization, paraffin-embedding and cutting into 4 μm-sections. Afterward, the cells or tissue sections were stained with ALP staining solution for 1 h at 37 ℃, followed by rinsing with PBS. The images were captured under a light microscope.
Alizarin red S (ARS) staining
The 4% paraformaldehyde-fixed cells, or femur sections were stained with freshly prepared Alizarin Red S solution (Sigma-Aldrich, MO, USA) for 30 min at room temperature. After rinsing with PBS for three times, the staining was examined under a light microscope.
Immunohistochemical staining
After 4% paraformaldehyde fixation, the femurs were embedded in paraffins, cut into 4 μm-sections, and mounted onto gelatin-coated slides. After drying overnight, deparaffinization, and rehydration with gradient concentration of ethanol, the sections were blocked with 1% BSA to avoid nonspecific binding. Subsequently, the sections were probed with primary antibodies against EHMT2 (1:500, ab185050, Abcam, UK), SKP2 (1:50, ab183039, Abcam), PIAS4 (1:100, ab137500, Abcam), DLX3 (1:50, PA5-104482, Thermo Fisher) at 4 °C overnight. The sections were reacted with HRP-conjugated secondary antibody (1:1000, ab6721, Abcam) for 1 h. The images were taken using a light microscope.
Immunofluorescent staining
The pre-prepared femur sections were subjected to antigen-retrieval in boiling citrate buffer, permeabilized with 0.2% Triton X-100, and then blocked with 2% BSA for 1 h. Then, the sections were incubated with primary antibodies against EHMT2 (1:50, ab185050, Abcam), DLX3 (1:100, PA5-104482, Thermo Fisher), RUN2 (1:50, sc-101145, Santa Cruz, USA) overnight at 4 °C. After washing, the tissue sections were reacted with corresponding secondary antibodies for 1 h. After nuclear staining with DAPI, the images were photographed under a fluorescence microscope.
Real-time quantitative PCR (RT-qPCR)
The cell samples were added with TRIZOL reagent (Thermo Fisher, MA, USA) to extract total RNA following the procedures, followed by cDNA synthesis using the PrimeScript RT Master Mix (Cat No.: RR036A, Takara, Osaka, Japan). Next, qPCR was carried out using the SYBR green PCR kit (Cat No.: 4309155, Thermo Fisher), supplemented with forward and reverse primers. The relative gene levels normalized to GAPDH were quantified using the 2 −ΔΔCt method.
Western blotting
MSCs were treated with RIPA lysis buffer containing protease inhibitors (Thermo Fisher) to prepare protein samples. For protein isolation from bone tissues, the frozen femur tissues were pulverized in a liquid-nitrogen-cooled vessel by pestle and mortar. Bone tissues were lysed in a 50 mM Tris buffer containing 1% Triton X-100 (Sigma-Aldrich), 20 µg/mL leupeptin (Sigma-Aldrich), and 0.5 mM phenylmethylsulfonyl fluoride (Sigma-Aldrich). The lysates were cleared by centrifugation at 15,000 × g for 10 min. After determination of protein concentrations using BCA protein assay kit (Sangon, Shanghai, China), the protein solution was mixed with the sample buffer, denatured by boiling it in a water bath and stored at −20℃. Subsequently, the protein samples (30 µg) were separated using sodium dodecyl-sulfate polyacrylamide gel electrophoresis, and blotted onto polyvinylidene difluoride membranes. The nonspecific binding was blocked by incubation in 5% nonfat milk for 1 h. Then, the membranes were probed with primary antibodies against EHMT2 (1:1000, ab185050, Abcam), Runx2 (1:1000, ab236639, Abcam), OCN (1:1000, ab93876, Abcam), OPN (1:1000, ab283669, Abcam), Colla1 (1:1000, ab270993, Abcam), DLX3 (1:500, PA5-104482, Thermo Fisher), SKP2 (1:200, ab183039, Abcam), MIB1 (1:1000, ab124929, Abcam), RNF146 (1:500, PA5-78565, Thermo Fisher), RNF180 (1:500, ABIN1386111, Antibodies-online, PA, USA), MDM2 (1:1000, ab259265, Abcam), SMURF1 (1:1000, ab300408, Abcam), CBL (1:1000, ab32027, Abcam), FBXW7 (1:1000, ab192328, Abcam), SYVN1 (1:500, ab118483, Abcam), ITCH (1:1000, ab108515, Abcam), NEDD4 (1:1000, PA5-17463, Thermo Fisher), PIAS4 (1:1000, ab137500, Abcam), FAK (1:1000, ab40794, Abcam), p-FAK (1:1000, ab81298, Abcam), β-actin (1:2000, ab8227, Abcam) at 4℃ overnight. After reaction with HRP-conjugated secondary antibody (1:2000, ab6721, Abcam) for 1 h at room temperature, the blots were visualized using the ECL Substrate Kit (ab133406, Abcam).
Chromatin Immunoprecipitation (ChIP) and reChip
ChIP assay was conducted using the ChIP Assay Kit (Cat No.: 632012, Clontech, CA, USA) according to the manufacturer’s protocol. In short, MSCs (2 × 10 6 cells/mL) were crosslinked with 1% formaldehyde for 10 min and resuspended in SDS lysis buffer. After sonication for 6 min, the cell lysates were incubated with anti-EHMT2 (ab185050, Abcam), anti-H3K9me2 (ab176882, Abcam) antibody and protein G beads at 4 °C overnight. The immunoprecipitated DNAs were purified and their enrichments were measured by RT-qPCR.
For Re-ChIP assay, the immune complex was eluted by 20 mmol/L dithiothreitol and then diluted in ChIP dilution buffer. Then, re-ChIP reaction was conducted. After washing with Tris‐EDTA buffer containing 1% SDS, the immunoprecipitated DNAs were purified. The enriched DNAs were assessed by RT-qPCR.
Co-immunoprecipitation (Co-IP)
For the detection of endogenous interplay between SKP2 and EHMT2 proteins or the interaction between EHMT2 protein and MDM2 or SMURF1 protein, MSCs or femur tissues were lysed with IP lysis buffer to obtain lysates, followed by immunoprecipitation with anti-SKP2 (ABIN5701277, Antibodies-online, PA, USA), anti-EHMT2 (ab185050, Abcam), or anti-IgG (ab172730, Abcam) antibody conjugated with Protein A/G beads at 4 °C overnight. Finally, the samples were rinsed with IP lysis buffer and determined by Western blotting.
For the determination of protein ubiquitination, cells were infected with HA-Ubiquitin combined with or without treatment with 10 µM MG132 (Sigma-Aldrich). For SUMOylation and ubiquitination detection, similar procedures were performed to prepare cell lysates. Then, immunoprecipitation with protein A/G-coupled anti-EHMT2 (ab185050, Abcam), or anti-IgG (ab172730, Abcam) was performed at 4 °C overnight. Finally, the bound protein A/G beads were eluted and evaluated by Western blotting using anti-SUMO2/3 (ab81371, Abcam), anti‐EHMT2 (ab185050, Abcam) and anti-Ubiquitin (ab303664, Abcam), primary antibodies, respectively.
Ni 2+ -NTA agarose affinity pull-down assay
Ni 2+ -NTA agarose affinity pull-down assay was conducted according to a previous study [ 17 ]. Briefly, HEK293T cells were transfected with the indicated plasmids (HA-EHMT2, His-SUMO1, His-SIMO2, His-SUMO3 and Flag-Ubc9), and then lysed in 800 µL of Ni 2+ -NTA denaturing buffer [6 M Gu-HCl, 10 mM tris, 100 mM NaH 2 PO 4 (pH 8.0), and 20 mM N-Ethylmaleimide (NEM)]. After centrifugation at 12,000 g for 10 min, the collected supernatant was incubated with Ni 2+ -NTA agaroses (QIAGEN, Hilden, Germany) for 4 h. Finally, the agaroses were rinsed in successive.
washing steps in 1 mL of Ni 2+ -NTA washing buffer [8 M urea, 10 mM tris-HCl (pH 6.3), 100 mM NaH 2 PO 4 , and 0.1% Triton X-100] and lastly eluted by boiling the beads in 50 µL of 2× SDS-PAGE loading buffer containing 200 mM imidazole at 95 °C for 5 min. Bound proteins were measured by Western blotting.
Molecular docking
Docking simulations between irisin and EHMT2 were performed using the GRAMM Web Server database ( https://gramm.compbio.ku.edu/ ). Initial models of irisin and EHMT2 were retrieved from the Protein Data Bank (PDB ID: A0A6G1QYY8 for irisin and Q96KQ7 for EHMT2, https://www.rcsb.org/ ). Structural visualization was conducted using PyMOL.
Cellular thermal shift assay
The freshly collected MSCs were lysed, followed by centrifugation at 12,000 rpm for 15 min. Cell lysates were heated at different temperatures, 37, 40, 43, 46, 49, 52, 55, 58, 61, 64 °C for 3 min, respectively. Thereafter, the denatured protein was precipitated, and the protein level of EHMT2 in different treated assays were detected by Western blotting.
Statistical analysis
The experimental data from three biological repeats are presented as mean ± standard deviation (SD). Statistical analyses were performed by Student’s t test for two groups or one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test for multiple groups using GraphPad Prism 8.0 software. P < 0.05 was considered as statistically significant.
Results
Down-regulation of EHMT2 and up-regulation of DLX3 during osteogenic differentiation
First, we detected the osteogenic differentiation potential of MSCs at various time points. After maintenance in osteogenic differentiation medium, the osteogenic differentiation ability was assessed by ALP and ARS staining. We found that ALP and ARS staining was intensified with time extension, suggesting the enhanced osteogenic differentiation ability (Fig. S1A-B). Moreover, the levels of EHMT2, DLX3 and osteogenic markers including Runx2, OCN, OPN, and Colla1 were detected. The protein levels of EHMT2 were reduced, while DLX3 and osteogenesis-related markers were up-regulated with osteogenesis of MSCs in a time dependent manner (Fig. S1C&D). These findings implied that EHMT2 and DLX3 might participate in the regulation of osteogenic differentiation.
EHMT2 recruits H3K9me2 to cause transcription Inhibition of DLX3
EHMT2 has been indicated as a methyltransferase that can dimethylate the H3K9, which leads to the transcription repression of target genes [ 18 ]. We further evaluated the differential expression of EHMT2 and DLX3 in the bone tissues of osteoporotic SAMP6 mice and control SAMR1 mice. HE staining indicated a reduced bone mass in SAMP6 mice (Fig. 1 A). In addition, Masson’s trichrome staining showed that the generation of collagen fibers was decreased in in the bone tissues of SAMP6 mice (Fig. 1 A). Immunohistochemistry staining and Western blotting showed increased expression of EHMT2 and decreased expression of DLX3 in the bone sections of SAMP6 mice (Fig. 1 B &C ). Moreover, immunofluorescence double staining of EHMT2/DLX3 and RUNX2 was performed in the bone tissue sections. We found a co-staining of EHMT2/DLX3 and RUNX2 in the femurs, and EHMT2 level was higher, while DLX3 and RUNX2 levels were lower in the bone tissues of SAMP6 mice in comparison with SAMR1 mice (Fig. 1 D). We further explored whether EHMT2 could regulate expression of DLX3. Lentivirus carrying sh-EHMT2-1#, −2# were infected into MSCs isolated from SAMP6 mice to knock down EHMT2. Sh-EHMT2-1#, −2# strikingly decreased EHMT2 level, while increased DLX3 level in MSCs (Fig. 1 E). Furthermore, ChIPbase database predicted the potential binding of EHMT2 and H3K9me2 to DLX3 promoter regions (Fig. 1 F). ChIP assay showed that EHMT2 and H3K9me2 could directly bind to the promoter of DLX3 (Fig. 1 G). In addition, ChIP-reChIP indicated that EHMT2 deficiency evidently weakened the binding of EHMT2 and H3K9me2 to DLX3 promoter (Fig. 1 H). Together, the above data suggested that EHMT2 repressed DLX3 transcription and expression via recruitment of H3K9me2 in osteoporotic SAMP6 mice.
EHMT2 overexpression restrains osteogenic differentiation via down-regulating DLX3
To investigate whether EHMT2 affected osteogenic differentiation through regulation of DLX3, SAMP6 mice-derived MSCs were infected with lentivirus-carrying oe-EHMT2 or oe-NC. Western blotting validated that oe-EHMT2 infection evidently enhanced EHMT2 protein level, but reduced DLX3 protein level (Fig. 2 A). Next, we explored the regulation of EHMT2 in osteogenic differentiation. Both ALP and ARS staining indicated that EHMT2 overexpression weakened osteogenic differentiation of MSCs in osteogenic differentiation medium (Fig. 2 B). Accordingly, osteogenic differentiation medium-induced down-regulation of EHMT2, and up-regulation of DLX3, Runx2, OCN, OPN, and Colla1 in MSCs were partially reversed by oe-EHMT2 (Fig. 2 C). Our data demonstrated that EHMT2 overexpression reduced DLX3 expression to repress osteogenic differentiation.
Inhibition of EHMT2 promotes osteogenesis in osteoporotic SAMP6 mice via up-regulation of DLX3
To uncover the regulatory role of EHMT2 in osteogenesis in vivo, AAV9 carrying sh-EHMT2 was intravenously injected into SAMP6 mice. HE staining indicated that AAV9-mediated knockdown of EHMT2 effectively recovered the reduced bone mass of SAMP6 mice (Fig. 3 A). ARS staining demonstrated that the weakened ability of osteogenesis could be enhanced by EHMT2 deficiency (Fig. 3 B). Immunohistochemistry staining showed increased expression of EHMT2 and decreased expression of DLX3 in the bone sections of SAMP6 mice were partly reversed by EHMT2 silencing (Fig. 3 C). Immunofluorescence double staining showed a co-staining of EHMT2/DLX3 and RUNX2 in the femurs. EHMT2 silencing partly relieved the increased expression of EHMT2 and decreased expression of DLX3 and RUNX2 in the bone sections of SAMP6 mice (Fig. 3 D). Micro-CT showed that loss of EHMT2 improved the declined bone volume of SAMP6 mice (Fig. 3 E). In addition, up-regulation of EHMT2 and down-regulation of DLX3, Runx2, OCN, OPN, and Colla1 in the MSCs extracted from bone marrow of SAMP6 mice were partially counteracted by EHMT2 depletion (Fig. 3 F). To sum up, EHMT2 deficiency facilitated osteogenesis via enhancing DLX3 expression in vivo.
SKP2 contributes to ubiquitylation and degradation of EHMT2
To further elucidate the upstream regulatory mechanism for EHMT2, Ubibrowser database predicted 11 potential E3 ligases for EHMT2 (Fig. 4 A). Moreover, we detected the expression of 11 potential E3 ligases in the bone tissues of SAMP6 mice and control SAMR1 mice. As shown in Fig. 4 B, SKP2, MDM2, and SMURF1 were down-regulated in SAMP6 mice. Furthermore, Co-IP assay confirmed that EHMT2 protein could directly interacted with SKP2 protein, but not MDM2 or SMURF1 protein in the femurs of SAMP6 mice (Fig. 4 C). Thus, SKP2 was focused on in the subsequent experiments. Consistently, immunohistochemistry staining validated a lower expression of SKP2 in the bone tissues of SAMP6 mice (Fig. 4 D). SKP2 is an E3 ligase that mediates ubiquitylation and proteasome degradation of its substrates [ 19 ]. We tested whether SKP2 affected EHMT2 and DLX3 expression in MSCs. We discovered that oe-SKP2-mediated SKP2 overexpression strikingly lowered EHMT2 protein level, while raised DLX3 protein level (Fig. 4 E). Furthermore, the direct interaction between SKP2 and EHMT2 proteins was verified by Co-IP assay in MSCs from SAMP6 mice (Fig. 4 F). The reduced EHMT2 protein level in SKP2-overexpressed MSCs from SAMP6 mice was restored by proteasome inhibitor MG132 administration (Fig. 4 G). In addition, EHMT2 protein degradation upon CHX stimulation was accelerated by SKP2 overexpression (Fig. 4 H). Moreover, the ubiquitylation level of EHMT2 was evidently elevated after SKP2 overexpression (Fig. 4 I). Thus, SKP2 interacted with EHMT2 to facilitate its ubiquitylation and degradation.
SKP2 knockdown inhibits osteogenic differentiation via increasing EHMT2 expression
To determine the modulation of SKP2/EHMT2 axis in MSC osteogenic differentiation, MSCs were infected with lentivirus carrying sh-SKP2 or sh-EHMT2, or a combination of them. We found that sh-SKP2 strikingly reduced DLX3 expression, but raised EHMT2 expression in MSCs (Fig. 5 A). However, sh-EHMT2 led to opposite results and reversed sh-SKP2-mediated up-regulation of EHMT2 and down-regulation of DLX3, but not affected SKP2 expression (Fig. 5 A). Additionally, ALP and ARS staining suggested that osteogenic capacity was lowered by SKP2 knockdown, but enhanced by EHMT2 knockdown. The inhibitory effect of sh-SKP2 on osteogenic differentiation was abolished by EHMT2 down-regulation (Fig. 5 B&C). Moreover, SKP2 deficiency down-regulated SKP2, DLX3, Runx2, OCN, OPN, and Colla1, while up-regulated EHMT2 in MSCs cultured with osteogenic differentiation medium. Conversely, EHMT2 silencing exerted an opposite role and reversed sh-SKP2-mediated these changes (except for SKP2 expression) (Fig. 5 D). These findings suggested that SKP2 depletion suppressed osteogenic differentiation via up-regulation of EHMT2.
PIAS4 promotes sumoylation of EHMT2
PIAS4 is the main SUMO E3 ligase that contributes to SUMOylation of proteins [ 20 ]. We found a higher expression of PIAS4 in the femurs of SAMP6 mice than that in SAMR1 mice (Fig. 6 A&B). Consistently, PIAS4 was highly expressed in MSCs isolated from SAMP6 mice as compared with SAMR1 mice (Fig. 6 C). To further verify whether PIAS4 was involved in the modulation of EHMT2 protein, MSCs were infected with lentivirus carrying sh-PIAS4. Western blotting data revealed that EHMT2 was down-regulated, but DLX3 was up-regulated in PIAS4-knockdown cells (Fig. 6 D). Ni 2+ -NTA pull-down assay indicated that exogenous EHMT2 could bind to SUMO2/3, and SUMO2-mediated SUMOylation of EHMT2 was more apparent (Fig. 6 E). Similarly, the interaction between endogenous EHMT2 and SUMO2 was confirmed (Fig. 6 F). Furthermore, the EHMT2 SUMOylation level was higher MSCs from SAMP6 mice in comparison with SAMR1 mice (Fig. 6 G). Accordingly, we observed a decreased EHMT2 SUMOylation level in PIAS4-depleted HEK293T cells and MSCs (Fig. 6 H&I). To sum up, PIAS4 contributed to SUMOylation of EHMT2 to enhance EHMT2 protein expression.
PIAS4 inhibits osteogenic differentiation via SUMOylation of EHMT2
We further determined the involvement of PIAS4-mediated SUMOylation in the influence of osteogenic differentiation. We confirmed that PIAS4 overexpression enhanced EHMT2 protein abundance, while decreased DLX3 protein level (Fig. 7 A). In addition, EHMT2 SUMOylation was intensified by PIAS4 overexpression (Fig. 7 B). Moreover, enforced expression of PIAS4 reduced osteogenic capacity, which was abrogated by EHMT2 knockdown (Fig. 7 C). Similarly, we noticed that EHMT2 silencing reversed the elevated expression of EHMT2, and declined expression of DLX3, and osteogenic markers in PIAS4-overexpressed MSCs (Fig. 7 D). These data strongly proved that PIAS4 restrained osteogenic differentiation via promoting PIAS4-mediated SUMOylation of EHMT2.
Irisin down-regulates PIAS4 to promote osteogenic differentiation by repressing sumoylation of EHMT2
Irisin has been proved to suppress bone mass loss, thereby alleviating OP [ 21 ]. To test if irisin contributed to osteogenic differentiation through regulation of PIAS4, MSCs were infected with lentivirus-carrying oe-PIAS4, followed by treatment with irisin. Enforced expression of PIAS4 reduced irisin-induced improvement in osteogenic capacity (Fig. 8 A). Moreover, PIAS4 overexpression significantly abolished irisin-induced up-regulation of DLX3, Runx2, OCN, OPN, Colla1, and down-regulation of PIAS4 and EHMT2 (Fig. 8 B). It is worth noting that the decreased SUMOylation level in irisin-stimulated cells was elevated by PIAS4 overexpression (Fig. 8 C). Next, irisin-treated MSCs were co-treated with irisin inhibitor RGDS. As expected, irisin-mediated enhancement in osteogenic capacity was counteracted by RGDS co-treatment (Fig. 8 D). Besides, RGDS reversed irisin-induced increase in levels of p-FAK, DLX3, Runx2, OCN, OPN, Colla1, and reduction in levels of PIAS4 and EHMT2 (Fig. 8 E). Additionally, RGDS restored the declined SUMOylation level of EHMT2 induced by irisin (Fig. 8 F). Furthermore, irisin raised the ubiquitination level of EHMT2 protein, which was neutralized by RGDS (Fig. 8 G). To strengthen the correlation between irisin and EHMT2, molecular docking was performed. We found that irisin could bind to the Ser276, Arg474, Ala543, Ser525, Arg914, Thr1057, Arg1039, Arg1145, and Thr1146 of EHMT2 (Fig. S2A). In addition, the thermal stability of EHMT2 protein was reduced by irisin (Fig. S2B). Together, these observations revealed that irisin reduced EHMT2 expression via inhibiting its SUMOylation and promoting ubiquitination, thus promoting osteogenic differentiation.
FNDC5 overexpression promotes in vivo bone formation via EHMT2/DLX3 pathway
Given the findings in vitro, we aimed to validate whether irisin promoted bone formation via modulation of EHMT2/DLX3 pathway in mice. Given that irisin is a myokine produced by the cleavage of FNDC5. Thus, SAMP6 mice were intravenously injected with AAV9 carrying oe-FNDC5. Consistent with the results in vitro, FNDC5 overexpression facilitated new bone regeneration in the femurs of mice (Fig. 9 A&B). Next, we determined FNDC5, EHMT2, and DLX3 expression in the femur samples via immunohistochemical staining. We observed lower expression of FNDC5 and DLX3, but higher expression of EHMT2 in the femurs of SAMP6 mice, whereas these changes were partly reversed by FNDC5 overexpression (Fig. 9 C). Consistently, FNDC5 overexpression elevated bone volume of SAMP6 mice (Fig. 9 D). Besides, FNDC5, DLX3, Runx2, OCN, OPN, and Colla1 protein levels were up-regulated, while EHMT2 protein level was down-regulated by FNDC5 overexpression in SAMP6 mice (Fig. 9 E). Collectively, the in vivo bone regeneration was effectively improved by FNDC5 overexpression via regulation of EHMT2/DLX3 pathway.
Discussion
As a polypeptide hormone produced from skeletal muscles, the osteogenic functions of irisin have been confirmed [ 22 ]. Enhancing the exogenous or endogenous irisin level could protect against osteoblast apoptosis, thus promoting osteogenesis [ 23 , 24 ]. Consistent with these observations, our investigation demonstrated that EHMT2 expression was reduced by epigenetic modifications through irisin treatment, which subsequently led to activation of DLX3 transcription, and thereby contributing to osteogenic differentiation.
EHMT2, also known as G9a, has been recognized as a crucial modulator of bone homeostasis, which exerts a crucial function in the progression of OP [ 25 ]. Another work showed that EHMT2 contributed to osteoclastogenesis via catalyzing H3K27me1 in MMP9-dependent H3NT proteolysis [ 26 ]. In this work, a remarkably reduction in EHMT2 expression during osteogenic differentiation was observed, and restoration of EHMT2 expression evidently weakened the osteogenic activity. EHMT2 is a methyltransferase that suppresses gene expression via enrichment of H3K9me2 [ 27 ]. EHMT2 was documented to promote DNA hypermethylation of lncH19, thereby repressing its expression and causing bone loss in hind limb unloading rats [ 25 ]. DLX3, a member of the DLX family, has been verified to promote osteogenic differentiation of BMSCs [ 28 ]. In addition, DLX3 could restrain RAW264.7 cell osteoclastic differentiation [ 29 ]. Altogether, DLX3 may serve as an inhibitor of OP. Interestingly, our findings demonstrated that EHMT2 repressed transcription and expression of DLX3 through recruitment of H3K9me2. Therefore, down-regulation of EHMT2 triggered DLX3 transcription to facilitate osteoblastic differentiation.
SKP2 is an E3 ligase that governs multiple important biological processes via ubiquitination and degradation of target proteins. For example, SKP2-mediated ubiquitination of KDM5B was involved in the regulatory mechanism of circ_0087960 in osteogenic capacity of periodontal ligament stem cells [ 30 ]. Furthermore, carbamazepine alleviated OP via modulation of ubiquitination and expression of SKP2 [ 31 ]. These reports indicated that SKP2 represented as a regulator of osteogenic differentiation. Our study found that SKP2 targeted EHMT2 to facilitate its ubiquitination and degradation. Moreover, SKP2 knockdown blocked osteogenic induction medium (OM)-mediated osteogenesis via inhibiting EHMT2 ubiquitination. Besides, irisin could upregulate SKP2. This is the first report explained that irisin upregulated SKP2 to enhance osteogenic differentiation by promoting ubiquitination of EHMT2.
PIAS4 is a pivotal SUMO E3 ligase that mediates the SUMOylation of target proteins and consequently affects their stabilization [ 32 ]. Our data showed that down-regulation of PIAS4 strikingly reduced EHMT2 protein abundance, suggesting that PIAS4-mediated SUMOylation might participate in the modulation of EHMT2 protein stability. SUMO1, SUMO2 and SUMO3 are three well-known members of SUMO family, which directly bind to target proteins to form SUMO chain and then initiate the SUMO modification [ 33 ]. In this study, the direct binding of SUMO2/3 to EHMT2 protein was validated, which was attenuated by PIAS4 knockdown. Thus, PIAS4 was essential for SUMOylation of EHMT2 protein. As reported by Zhao et al., irisin could modulate its target protein HDAC4 through SUMOylation in hypoxia/reoxygenation-stimulated cardiomyocytes [ 15 ]. However, whether irisin-mediated EHMT2 protein degradation was associated with SUMOylation remains to be determined. In this research, irisin treatment decreased PIAS4 expression to restrain SUMO modification of EHMT2, thus lowering EHMT2 protein level.
There are several limitations in this study. Firstly, although we have evaluated the role of EHMT2/FNDC5 in OP in SAMP6 mice, our results need to be validated in other OP animal models. Secondly, the exact mechanisms by which irisin regulates the level of EHMT2 remain not fully understood. Thirdly, the downstream mechanism through which DLX3 promotes osteogenic differentiation needs to be further investigated. Hence, more investigations are needed in our future studies.
In summary, our findings uncovered that EHMT2 protein was destabilized via promoting ubiquitination but inhibiting SUMOylation of EHMT2 mediated by irisin. Furthermore, down-regulation of EHMT2 repressed methylation of DLX3 to enhance DLX3 transcription and expression, which intensified osteogenesis. Our results provide novel insight into the mechanisms of osteogenic differentiation.
Supplementary Information
Supplementary Material 1.
Abbreviations
Osteoporosis
Distal-less homeobox 3
Euchromatic histone lysine methyltransferase 2
Histone H3 lysine 9
S-phase kinase-associated protein 2
Focal adhesion kinase
Small ubiquitin-like modifier
Short hairpin RNA
Negative control shRNA
Protein inhibitor of activated STAT 4
Hematoxylin and eosin
Alkaline phosphatase
Alizarin red S
Real-time quantitative PCR
Enzyme-linked immunosorbent assay
Chromatin immunoprecipitation
Wild-type
Mutated
Co-immunoprecipitation
Standard deviation
Analysis of variance
Authors’ contributions
Mingxiang Zou, Wenkang Chen, Bangmin Luo and Ke Yin wrote the main manuscript text and Haowei Zhang, Ansong Liu, Jianhua Liu, Chenggong Wang and Da Zhong prepared figures. All authors reviewed the manuscript.
Funding
This work was supported by Interdisciplinary Research Program in Medicine and Engineering, The First Affiliated Hospital of University of South China (IRP-M&E-2025-02 to KY, IRP-M&E-2025-06 to MXZ); National Natural Science Foundation of China (82002364 to MXZ), Research Foundation of Education Bureau of Hunan Province (23B0412 to WKC).
Data availability
All data generated or analysed during this study are included in this published article.
Declarations
Ethics approval and consent to participate
The animal experiments were approved by the Ethics Committee of University of South China (ethical number: 2024 − 281).
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Mingxiang Zou, Wenkang Chen and Bangmin Luo contributed equally to this work.
References
1. Black DM, Rosen CJ. Clinical practice. Postmenopausal osteoporosis. N Engl J Med. 2016;374(3):254–62.
2. Tao H, Li W, Zhang W, Yang C, Zhang C, Liang X, et al. Urolithin A suppresses RANKL-induced osteoclastogenesis and postmenopausal osteoporosis by, suppresses inflammation and downstream NF-kappaB activated pyroptosis pathways. Pharmacol Res. 2021;174:105967.
3. Kanis JA, Harvey NC, McCloskey E, Bruyere O, Veronese N, Lorentzon M, et al. Algorithm for the management of patients at low, high and very high risk of osteoporotic fractures. Osteoporos Int. 2020;31(1):1–12.
4. Nayak S, Greenspan SL. A systematic review and meta-analysis of the effect of bisphosphonate drug holidays on bone mineral density and osteoporotic fracture risk. Osteoporos Int. 2019;30(4):705–20.
5. Kim SJ, Hwang YH, Mun SK, Hong SG, Kim KJ, Kang KY, et al. Protective Effects of 2,3,5,4’-Tetrahydroxystilbene-2-O-beta-d-glucoside on Ovariectomy Induced Osteoporosis Mouse Model. Int J Mol Sci. 2018;19(9):2554.
6. Hu X, Wang Z, Wang W, Cui P, Kong C, Chen X, et al. Irisin as an agent for protecting against osteoporosis: A review of the current mechanisms and pathways. J Adv Res. 2024;62:175–86.
7. Brenmoehl J, Albrecht E, Komolka K, Schering L, Langhammer M, Hoeflich A, et al. Irisin is elevated in skeletal muscle and serum of mice immediately after acute exercise. Int J Biol Sci. 2014;10(3):338–49.
8. Kim H, Wrann CD, Jedrychowski M, Vidoni S, Kitase Y, Nagano K, et al. Irisin mediates effects on bone and fat via AlphaV integrin receptors. Cell. 2019;178(2):507–8.
9. Casciello F, Al-Ejeh F, Kelly G, Brennan DJ, Ngiow SF, Young A, et al. G9a drives hypoxia-mediated gene repression for breast cancer cell survival and tumorigenesis. Proc Natl Acad Sci U S A. 2017;114(27):7077–82.
10. Shen WC, Lai YC, Li LH, Liao K, Lai HC, Kao SY, et al. Methylation and PTEN activation in dental pulp mesenchymal stem cells promotes osteogenesis and reduces oncogenesis. Nat Commun. 2019;10(1):2226.
11. Garcia-Gomez A, Li T, de la Calle-Fabregat C, Rodriguez-Ubreva J, Ciudad L, Catala-Moll F, et al. Targeting aberrant DNA methylation in mesenchymal stromal cells as a treatment for myeloma bone disease. Nat Commun. 2021;12(1):421.
12. Bond M, Sala-Newby GB, Newby AC. Focal adhesion kinase (FAK)-dependent regulation of S-phase kinase-associated protein-2 (Skp-2) stability. A novel mechanism regulating smooth muscle cell proliferation. J Biol Chem. 2004;279(36):37304–10.
13. Liang Z, Liu H, Zhang Y, Xiong L, Zeng Z, He X, et al. Cyr61 from adipose-derived stem cells promotes colorectal cancer metastasis and vasculogenic mimicry formation via integrin alpha(V) beta(5). Mol Oncol. 2021;15(12):3447–67.
14. Zhang Y, Yang K, Yang J, Lao Y, Deng L, Deng G, et al. SENP3 suppresses osteoclastogenesis by De-conjugating SUMO2/3 from IRF8 in bone Marrow-Derived monocytes. Cell Rep. 2020;30(6):1951–63. e4.
15. Zhao YT, Wang H, Zhang S, Du J, Zhuang S, Zhao TC. Irisin ameliorates hypoxia/reoxygenation-induced injury through modulation of histone deacetylase 4. PLoS ONE. 2016;11(11):e0166182.
16. Fan S, Li J, Zheng G, Ma Z, Peng X, Xie Z, et al. WAC facilitates Mitophagy-mediated MSC osteogenesis and new bone formation via protecting PINK1 from Ubiquitination-Dependent degradation. Adv Sci (Weinh). 2025;12(2):e2404107.
17. Yan L, Zhang T, Wang K, Chen Z, Yang Y, Shan B, et al. SENP1 prevents steatohepatitis by suppressing RIPK1-driven apoptosis and inflammation. Nat Commun. 2022;13(1):7153.
18. Jana A, Bellver-Sanchis A, Grinan-Ferre C, Banerjee DR. Repurposing of raltitrexed as an effective G9a/EHMT2 inhibitor and promising anti-Alzheimer’s agent. ACS Med Chem Lett. 2023;14(11):1531–6.
19. Wu T, Gu X, Cui H. Emerging roles of SKP2 in cancer drug resistance. Cells. 2021. 10.3390/cells10051147.
20. Galanty Y, Belotserkovskaya R, Coates J, Polo S, Miller KM, Jackson SP. Mammalian SUMO E3-ligases PIAS1 and PIAS4 promote responses to DNA double-strand breaks. Nature. 2009;462(7275):935–9.
21. Li H, Luo D, Xie W, Ye W, Chen J, Alberton P, et al. Irisin reduces senile osteoporosis by inducing osteocyte mitophagy through Ampk activation. iScience. 2024;27(11):111042.
22. Chen X, Sun K, Zhao S, Geng T, Fan X, Sun S, et al. Irisin promotes osteogenic differentiation of bone marrow mesenchymal stem cells by activating autophagy via the Wnt//beta-catenin signal pathway. Cytokine. 2020;136:155292.
23. Xue Y, Hu S, Chen C, He J, Sun J, Jin Y, et al. Myokine Irisin promotes osteogenesis by activating BMP/SMAD signaling via AlphaV integrin and regulates bone mass in mice. Int J Biol Sci. 2022;18(2):572–84.
24. Zhang J, Valverde P, Zhu X, Murray D, Wu Y, Yu L, et al. Exercise-induced irisin in bone and systemic irisin administration reveal new regulatory mechanisms of bone metabolism. Bone Res. 2017;5:16056.
25. Li B, Zhao J, Ma J, Chen W, Zhou C, Wei W, et al. Cross-talk between histone and DNA methylation mediates bone loss in hind limb unloading. J Bone Miner Res. 2021;36(5):956–67.
26. Kim K, Shin Y, Kim J, Ulmer TS, An W. H3K27me1 is essential for MMP-9-dependent H3N-terminal tail proteolysis during osteoclastogenesis. Epigenetics Chromatin. 2018;11(1):23.
27. Tachibana M, Matsumura Y, Fukuda M, Kimura H, Shinkai Y. G9a/GLP complexes independently mediate H3K9 and DNA methylation to silence transcription. EMBO J. 2008;27(20):2681–90.
28. Li D, Yuan Q, Xiong L, Li A, Xia Y. The miR-4739/DLX3 axis modulates bone marrow-derived mesenchymal stem cell (BMSC) osteogenesis affecting osteoporosis progression. Front Endocrinol (Lausanne). 2021;12:703167.
29. Zhao N, Han D, Liu Y, Li Y, Zeng L, Wang Y, et al. Dlx3 negatively regulates osteoclastic differentiation through microrna-124. Exp Cell Res. 2016;341(2):166–76.
30. Liu Y, Zhou Y. Circ_0087960 stabilizes KDM5B by reducing SKP2 mediated ubiquitination degradation and promotes osteogenic differentiation in periodontal ligament stem cells. Regen Ther. 2022;19:122–30.
31. Wang H, Liu Z, Niu D, Li H, Han Y, Peng J, et al. Carbamazepine regulates USP10 through miR-20a-5p to affect the deubiquitination of SKP2 and inhibit osteogenic differentiation. J Orthop Surg Res. 2023;18(1):820.
32. Liang M, Cai Z, Jiang Y, Huo H, Shen L, He B. Senp2 promotes VSMC phenotypic switching via myocardin de-SUMOylation. Int J Mol Sci. 2022. 10.3390/ijms232012637.
33. Vertegaal AC. Small ubiquitin-related modifiers in chains. Biochem Soc Trans. 2007;35(Pt 6):1422–3.
Associated Data
Supplementary Materials
Supplementary Material 1.
Data Availability Statement
All data generated or analysed during this study are included in this published article.