EZH2 Supports Osteoclast Differentiation and Bone Resorption Via Epigenetic and Cytoplasmic Targets

Abstract

Key osteoclast (OCL) regulatory gene promoters in bone marrow-derived monocytes harbor bivalent histone modifications that combine activating Histone 3 lysine 4 tri-methyl (H3K4me3) and repressive H3K27me3 marks, which upon RANKL stimulation resolve into repressive or activating architecture. Enhancer of zeste homologue 2 (EZH2) is the histone methyltransferase component of the polycomb repressive complex 2, which catalyzes H3K27me3 modifications. Immunofluorescence microscopy reveals that EZH2 localization during murine osteoclastogenesis is dynamically regulated. Using EZH2 knockdown and small molecule EZH2 inhibitor GSK126, we show that EZH2 plays a critical epigenetic role in OCL precursors (OCLp) during the first 24 hours of RANKL activation. RANKL triggers EZH2 translocation into the nucleus where it represses OCL-negative regulators MafB, Irf8, and Arg1. Consistent with its cytoplasmic localization in OCLp, EZH2 methyltransferase activity is required during early RANKL signaling for phosphorylation of AKT, resulting in downstream activation of the mTOR complex, which is essential for induction of OCL differentiation. Inhibition of RANKL-induced pmTOR-pS6RP signaling by GSK126 altered the translation ratio of the C/EBPβ-LAP and C/EBPβ-LIP isoforms and reduced nuclear translocation of the inhibitory C/EBPβ-LIP, which is necessary for transcriptional repression of the OCL negative-regulatory transcription factor MafB. EZH2 in multinucleated OCL is primarily cytoplasmic and mature OCL cultured on bone segments in the presence of GSK126 exhibit defective cytoskeletal architecture and reduced resorptive activity. Here we present new evidence that EZH2 plays epigenetic and cytoplasmic roles during OCL differentiation by suppressing MafB transcription and regulating early phases of PI3K-AKT-mTOR-mediated RANKL signaling, respectively. Consistent with its cytoplasmic localization, EZH2 is required for cytoskeletal dynamics during resorption by mature OCL. Thus, EZH2 exhibits complex roles in supporting osteoclast differentiation and function. © 2019 American Society for Bone and Mineral Research.

Keywords: CYTOSKELETAL DYNAMICS; EPIGENETICS; EZH2; OSTEOCLAST; PI3K-AKT-mTOR SIGNALING.

Figure 1.. Inhibition of EZH2 prevents repression of OCL negative factors and inhibits OCL formation.

BMM cells were cultured with MCSF for 3 days (d-3 to d0) to form OCLp before addition of RANKL/MCSF to generate OCL. (A) OCLp were cultured with RANKL/MCSF in the presence of the indicated concentrations of GSK126 for 5 days. TRAP+ OCL formed were analyzed by microscopy (representative images on left) and the counts for multinucleated (n≥3) TRAP+ OCL/well graphed (right). (B) GSK126 (10 μM) was added at the indicated time intervals during the processes of OCLp and OCL (d4) formation and representative images and counts for multinucleated (n≥3) TRAP+ OCL/well are shown. (C) OCLp (d0) or stimulated with RANKL/MCSF +/− GSK126 (10 μM; GSK) for the indicated times were analyzed for Ezh2 mRNA by qPCR. (D) OCLp (d0) or stimulated with RANKL/MCSF for the indicated times were analyzed for total cellular levels of EZH2, H3K27me3, total H3 and βactin by WB. (E) OCLp cultures (d0) or after stimulation with RANKL/MCSF +/− GSK126 (10 μM) for 1 day were analyzed for total cellular levels of EZH2, H3K27me3, H3K9ac, H3 and βactin by WB. (F) OCLp (d0) or after stimulation with RANKL/MCSF +/− GSK126 (10 μM) for 1 or 5 days were analyzed for the expression of OCL-specific genes NFATc1, RANK, CatK, DC-STAMP by qPCR. (G) OCLp (d0) or after stimulation with RANKL/MCSF +/− GSK126 (10 μM) for 1 day were analyzed for the expression of OCL-negative regulators MafB, Irf8, Arg1, and Bcl6 by qPCR. WB analyses show representative of 3 independent experiments and error bars in mRNA experiments represent SEM for 4-17 biological replicates. Statistical analysis used one-way ANOVA with Dunnett’s multiple comparisons test comparing to no GSK126 control (B), 0 time (C), or with Tukay’s multiple comparisons test for all means (G), and two-way ANOVA with Sidak multiple comparisons for day and treatment (F). Note that in C, d1 and d5 were not found to be significantly different from time 0.

Figure 2.. EZH2 is required for OCL differentiation.

BMM cells were infected with lentivirus (LV) during OCLp formation with MCSF. (A, B) OCLp infected with SCR or sh-Ezh2 (EZH2-KD) LV were cultured with RANK/MCSF for 1 day and analyzed for (A) Ezh2 and MafB expression by qPCR and (B) total cellular levels of EZH2, H3K27me3, MafB and βactin by WB. (C) SCR and EZH2-KD OCLp were stimulated with RANKL/MCSF for 4 days and analyzed for OCL formation. Representative images (left) and counts for multinucleated (n≥3) TRAP+ OCL/well are shown. (D, E) OCLp cells infected with control or EZH2-OE LV were stimulated with RANKL/MCSF for 1 day and analyzed for (D) Ezh2 and MafB expression by qPCR and (E) total cellular levels of EZH2, H3K27me3, H3 and βactin by WB. (F) Control and EZH2-OE OCLp were stimulated with RANKL/MCSF for 3 days and analyzed for OCL formation. Representative microscopy images (left) and counts for multinucleated (n≥3) TRAP+ OCL/well are shown. (G) Control and EZH2-OE OCLp were plated on bone slices in the presence of RANKL/MCSF for 8 days. Representative microscopy images for Toluidine blue stained resorption pits are shown and graphs (with SD) represent ImageJ quantitation of 5 distinct OCL resorption areas, counts for multinucleated (n≥3) TRAP+ OCL/bone area (μm²), and resorption per OCL. WB, TRAP assays and resorption experiments were conducted at least 2 independent times. Error bars in mRNA experiments represent SEM for 2-5 biological replicates. Statistical analyses used Student’s t test.

Figure 3.. EZH2 methyltransferase activity is critical within the first 24 hours OCLp differentiation.

OCLp (Figure 1) were stimulated with RANKL/MCSF as indicated and treated with GSK126 (10 μM) for the indicated time intervals. Counts of multinucleated (n≥3) TRAP+ OCL/well at day 4 and representative microscopy images are shown. (A) OCLp were stimulated with RANKL/MCSF +/− GSK126 (10 μM) at d0 only and GSK126 removed at the indicated times. (B) OCLp were stimulated with RANKL/MCSF +/− GSK126 (10 μM) at d0 and d2 and GSK126 removed at the indicated times. (A, B) Media from a second set of OCL cultures grown along-side the inhibitor experiments with identical RANKL/MCSF treatment was used to replace the GSK126 containing media. This prevented additional activation of OCL cultures that would interfere with the GSK126 treatment inhibition window. (C) OCLp were stimulated with RANKL/MCSF at d0 only and treated with GSK126 (10 μM) for the indicated time intervals. Statistical analysis used one-way ANOVA with Dunnett’s multiple comparisons test comparing to no GSK126 control. Linear trend was also analyzed (with d0-d3 order flipped in C) and found to be highly significant (p<0.0001) for A-C.

Figure 4.. GSK126 blocks RANKL-activated PI3K-AKT-mTOR signaling axis resulting in increased binding of C/EBPβ LAP to MafB.

OCLp were formed as in Figure 1. (A, B) OCLp (d0) were stimulated with RANKL/MCSF +/− GSK126 (10 μM) for 1 day were harvested and either (A) total cellular lysates were analyzed for cellular levels of C/EBPβ, MafB and βactin by WB or (B) cytoplasmic and nuclear extracts were analyzed for C/EBPβ, α-tubulin, H3K27me3, H3, and EZH2. Cytoplasmic EZH2 was normalized to α-tubulin and nuclear EZH2 to H3 and graphed. (C) SCR and EZH2-KD OCLp were stimulated with RANKL/MCSF for 1 day and harvested lysates analyzed were subjected to WB analysis for total C/EBPβ levels using two different antibodies where C/EBPβ-LIP+LAP Ab (α-C-term) detects both isoforms, while C/EBPβ-LAP Ab (α-N-term) only detects the LAP isoform. (D) ChIP-qPCR amplicons (depicted underneath the bar graphs) were used to detect EZH2 and H3K27me3 enrichment at the promoter (sites 0 and 1) and within the structural MafB gene (site 2) in OCLp (d0) or after RANKL/MCSF +/− GSK126 (10 μM) for 1 day. (E) C/EBPβ LAP and LIP isoform binding at the MafB promoter was determined by ChIP-qPCR using distinct C/EBPβ antibodies described in (C). (F) Lysates from OCLp (d0) or after stimulation with RANKL/MCSF +/− GSK126 (10 μM) for the indicated times were analyzed for pPI3K, pPDK1, pAKT, and mTOR pathway effectors by WB. The OCLp were pre-treated with GSK126 for 2 h prior to RANKL addition. WB analyses show representative of 3 independent experiments and error bars in ChIP results represent SEM for 3 biological replicates. (G) Schematic representation of GSK126 effects on the PI3K-AKT-mTOR signaling pathway during RANKL induced OCLp differentiation. Statistical analysis used one-way ANOVA with Dunnett’s multiple comparisons test comparing to “D1RL” (D, E) with trending data indicated (°; p=0.0681).

Figure 5.. EZH2 exhibits distinct localization pattern during differentiation of OCL.

(A) BMM induced to form OCLp as in Figure 1 were harvested and plated on glass bottom imaging chambers were then stimulated with RANKL/MCSF +/− GSK126 (10 μM) for 1 day or with RANKL/MCSF alone to form differentiated multinuclear OCL (d4). Fixed cells were immunofluorescently stained for EZH2 (red) and DAPI used to detect nuclear DNA (blue). (B) OCLp plated on glass bottom imaging chambers were stimulated for 1 day with MCSF +/− RANKL +/− GSK126, fixed, immunofluorescently stained for EZH2 using secondary antibody-Alexa Fluor 488 and DAPI. The nuclear fraction of the total EZH2 fluorescent intensity of each cell was determined in a population of 124-248 cells/group. Statistical analysis used one-way ANOVA with Dunnett’s multiple comparisons test comparing to MCSF+RANKL. (C) OCLp cultured on bone slices were stimulated with RANKL/MCSF +/− GSK126 (10 μM) for 1 day. Fixed cells were immunofluorescently stained for EZH2 (green) and DAPI (blue). (D) Multinuclear OCL differentiated with RANKL/MCSF on bone slices for 8 days (8d) were fixed and immunofluorescently stained for Calcitonin receptor (CTR) (red), EZH2 (green), and DAPI (blue). Since the cells were permeabilized, the CTR staining is visible both at the membrane and in the cytoplasm and allows visualization of the OCL membrane/cell shape. All images represent fluorescent microscopy confocal sections and scale bars indicate 20 μm.

Figure 6.. GSK126 treatment of mature OCL impairs cytoskeletal architecture and F-actin ring formation.

OCLp were plated onto bone slices and either maintained in MCSF or cultured with RANKL/MCSF for 8 days to generate mature OCL as indicated. (A) GSK126 (10 μM) was added to the OCL at day 6 (bottom panels) and the fixed cells were immunofluorescently stained for CTR (red), EZH2 (green), and DAPI (blue). (B) OCL were formed by RANKL/MCSF alone or with GSK126 (10 μM) added for the entire time (starting at d0) or only added to the OCL at day 6 as indicated. The fixed cells were fluorescently stained with Phalloidin rhodamine to detect actin (red) and DAPI (blue). All images represent fluorescent microscopy confocal sections and scale bars indicate 20 μm.

Figure 7.. Cytoplasmic EZH2 methyltransferase activity is required for OCL resorption.

OCLp were plated onto bone slices, and stimulated with MCSF alone or RANKL/MCSF for 9 days, except the sample on the far right which was harvested at day 7 as a control. GSK126 (10 μM) was applied to cultures at day 0 or day 7 as indicated. The fixed cells were stained for TRAP and also fluorescently stained with Phalloidin rhodamine (red) and DAPI (blue). (A) The TRAP stained OCL (Brightfield microscopy top panels) are shown using widefield fluorescent microscopy together with the fluorescent stains Phalloidin (red in middle panels) and DAPI (blue in bottom panels). The dark areas in the fluorescent images reveal the presence of the TRAP stain through interference. Scale bars represent 20 μm. (B) The cells were removed and Toluidine blue was applied to visualize OCL resorbed areas. (C) ImageJ was used to quantitate the resorption area. (D) Counts of TRAP+ multinuclear OCL/bone area (μm²) formed on the bone surface. Statistical analysis used two-way ANOVA with Tukay’s multiple comparisons test for all means.

Figure 8.. Summary for nuclear and cytoplasmic roles for EZH2 during osteoclastogenesis.

EZH2 is a RANKL inducible positive regulator of OCL differentiation required for silencing of MafB expression by both adding H3K27me3 as well as increasing C/EBPβ-LIP at the MafB gene promoter. Therefore, blocking EZH2 methyltransferase activity with GSK126 de-represses MafB both by decreased epigenetic silencing and by lack of proper RANKL-activation of the PI3K-pAKT-pmTOR signaling pathway. This results in increased C/EBPβ-LAP isoform levels with its enhanced binding/activation of the MafB gene. Since blockade of EZH2 methyltransferase activity in mature OCL results in condensed cytoskeletal architecture, impaired F-actin ring formation and reduced OCL resorption, we hypothesize that by modulating actin dynamics, EZH2 modulates cytoskeletal rearrangements in mature OCL.